Citation
Hexahydropyrimidines as masked spermidine vectors in drug delivery and as reagents in the synthesis of h-shaped octacoordinate actinide ligands for human and environmental decontamination

Material Information

Title:
Hexahydropyrimidines as masked spermidine vectors in drug delivery and as reagents in the synthesis of h-shaped octacoordinate actinide ligands for human and environmental decontamination
Creator:
Seligsohn, Howard Wayne, 1960-
Publication Date:
Language:
English
Physical Description:
vii, 134 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Amines ( jstor )
Excretion ( jstor )
Intraperitoneal injections ( jstor )
Ligands ( jstor )
pH ( jstor )
Plutonium ( jstor )
Polyamines ( jstor )
Rats ( jstor )
Resins ( jstor )
Thorium ( jstor )
Actinoid Series Elements -- toxicity ( mesh )
Biological Transport, Active ( mesh )
Dissertations, Academic -- Medicinal Chemistry -- UF ( mesh )
Drug Carriers ( mesh )
Environmental Pollutants -- toxicity ( mesh )
Medicinal Chemistry thesis Ph.D ( mesh )
Polyamines -- metabolism ( mesh )
Spermidine -- metabolism ( mesh )
Vehicles ( mesh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1987.
Bibliography:
Includes bibliographical references (leaves 130-133).
Additional Physical Form:
Also available online.
General Note:
Typescript.
General Note:
Vita.
General Note:
Leaf vii (abstract) is lacking.
General Note:
Leaf vii (abstract)--is lacking.
Statement of Responsibility:
by Howard Wayne Seligsohn.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Howard Wayne Seligsohn. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
18458526 ( OCLC )
030281505 ( ALEPH )

Downloads

This item has the following downloads:


Full Text











HEXAHYDROPYRIMIDINES AS MASKED SPERMIDINE VECTORS IN DRUG DELIVERY AND AS REAGENTS IN THE SYNTHESIS OF
H-SHAPED OCTACOORDINATE ACTINIDE LIGANDS FOR HUMAN AND
ENVIRONMENTAL DECONTAMINATION








By



HOWARD WAYNE SELIGSOHN

















A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULLFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY


UNIVERSITY OF FLORIDA

1987



























Through their love,

patience,

understanding, inspiration, and constant support,





my mother and father have contributed

immeasureably to my career. In recognition of this, I proudly dedicate this dissertation to them.













ACKNOWLEDGEMENTS

I would like to extend my sincere gratitude to my major advisor, Dr. Raymond J. Bergeron, for all his support, patience and understanding throughout my studies. I also thank my committee, Drs. Richard Strieff, Kenneth Sloan, Merle Battiste, and Margaret James for their contribution to this work. To Dr. James Navratil and the Rocky Flats research group I extend my thanks for the plutonium data. Deserving special thanks are Drs. Nelson Scarborough and Steven Prudencio and Ms. Kady Crist for their surgical expertise, Diana Tukalo for her synthetic efforts, and Micheal Ingeno for all of his help with my in vitro experiments. Lastly, thanks to all of my good friends and family who have helped me through the difficult times. I could not have made it without your support.














TABLE OF CONTENTS


ACKNOWLEDGEMENTS............................................ iii

ABSTRACT......... ........................................... vi

CHAPTERS

I. INTRODUCTION................................... 1..... ..................1

II. H-SHAPED OCTACOORDINATE ACTINIDE LIGANDS FOR HUMAN AND
ENVIRONMENTAL DECONTAMINATION......................... ............ 13

Background.................................................................................................. 13
Synthesis...................................................................................................... 15
Catecholamide Ligands.......................................................................... 15
Parent catecholamides............................................................... 15
Nitro derivatives........................................ ....................................... 18
Resin-Bound Catecholamide Ligands................................ ........... 27
The catecholamide................................................ 27
The resin................................................................................... 27
Ligand Stoichiometry: Job's Plots...................................... .............. 35
Ligand Precipitation Techniques: Removal of Actinides from Water..................38
Insoluble Chelators: Removal of Actinides from Water and Blood Plasma........... 45
XAD-4-Sorbed Ligands.................................................................... 45
Covalently-Bound Ligands.............................. 49
IRP-64-bound catecholamides............................................................ 49
CH-Sepharose-4B-bound catecholamides.......................... ........ 49
Equilibrium Solution Chemistry: Thermodynamic Binding Constant
Measurements........................................................................... 52
Eriochrome Black T Competition........................................52
Competition Studies with Nitro Derivatives of Catecholamide Ligands........... 55
Pharmacology................................................................................................ 72
Vehicle Development.................................................. 72
Biological Toxicity............................................................................. 74
Clearance Studies.................................................................................... 74
Urinary clearance.......................................................................... 75
Fecal clearance............................................................................ 75
Biliary clearance......................................................................... . 78
Combination chelator therapy........................... .......................80








iv









III. ELUCIDATION OF THE SOLUTION STRUCTURE OF POLYAMINES IN
RELATION TO THE MECHANISM OF CELLULAR UPTAKE.............................87

Background................................................. ............ ........................ 87
Backbone Variations......................................................................... 87
N4-Substitution.................................... ................................................ 87
N1,N8-Bis Substitution ................................................................................90
The Role of Protonation State: Potentiometric Measurements......................... 90
The Role of Hydrogen Bonding............................................................... 98
The Reasoning...................................................... 98
The Evidence: Hexahydropyrmidines.................. ...... ................ 99

IV. EXPERIMENTAL DETAILS... ............................................................................ 107

Synthetic Procedures................................................. ............................. 107
Job's Plots............................................ .................................................... 122
Precipitation Techniques............................... 122
Eriochrome Black T Competition .............................. 123
Competition with Nitrocatecholamides...................... 124
Resin Experiments.............................. 125
Biological Evaluation............................... 126
Potentiometric Measurements............................. 127
Stability Studies.................................. 127
Inhibition of Spermrnidine Uptake ................................. 127
IC Measurements........................................................................................ 128

SUMMARY AND CONCLUSION... ....................... .................................................... 129

REFERENCES........................................................................................................ 130

BIOGRAPHICAL SKETCH ............... ............... . . ............................. 134






















v













Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fullfillment of the Requirements for the Degree of Doctor of Philosophy



HEXAHYDROPYRIMIDINES AS MASKED SPERMIDINE VECTORS IN
DRUG DELIVERY AND AS REAGENTS IN THE SYNTHESIS OF
H-SHAPED OCTACOORDINATE ACTINIDE LIGANDS FOR HUMAN AND ENVIRONMENTAL DECONTAMINATION




By

HOWARD WAYNE SELIGSOHN

August 1987

Chairman: Dr. Raymond J. Bergeron
Major Department: Medicinal Chemistry
The cellular uptake of several polyamines is evaluated in terms of their intermolecular hydrogen bonding and their charge at physiological pH. N-(4-Aminobutyl)hexahydropyrimidine and N-(3-aminopropyl)hexahydropyrimidine are shown to compete with spermidine for uptake by L1210 cells. This observation is in keeping with the idea that spermidine may adopt a hydrogen-bonded cyclic structure in the course of transport. Furthermore, the differences in the ability of spermidine, homospermidine, and norspermidine to utilize the spermidine uptake apparatus of L1210 cells is related to the protonation state of the amines. These states are calculated for each triamine from measured pKa data. The hexahydropyrimidines used in the polyamine uptake sudies were next used in the synthesis of octacoordinate catecholamide H-shaped ligands. These chelators were tested for their ability to remove actinides from aqueous solution, human plasma, and rats. H-Shaped ligands removed more than 99.9 percent of the thorium from a 0.32 pM aqueous solution. In addition, approximately 30 percent of the


vi















CHAPTER I

INTRODUCTION

In recent years, scientists have become increasingly aware that polyamines play an important role in cellular metabolic processes. Considerably increased levels of omithine decarboxylase, the rate-limiting enzyme in polyamine biosynthesis, is an apparent "priming" event for growth in a variety of systems (1). Many neoplasias have been shown to concentrate polyamines intracellularly (2, 3, 4). Polyamine biosynthesis and intracellular accumulation are intimately associated with rapid cell proliferation. The increase in intracellular spermidine has been shown to correlate well with the increase in intracellular RNA (5). Elevated polyamine levels in urine have been detected in tumor-bearing patients (6). Bergeron and Porter have demonstrated that polyamine-deficient L 210 cells divide much more slowly than control cells (7, 8, 9, 10). Thus it appears that an intracellular polyamine demand, polyamine biosynthesis and intracellular accumulation are intimately associated with rapid cell proliferation.

Since neoplastic tissues exhibit a high polyamine requirement when compared to normally differentiated tissues, the idea of using polyamines as vectors for delivering antineoplastics is an attractive one. If one were to design a polyamine derivative which possessed antineoplastic properties, and could also utilize the polyamine transport apparatus for cellular uptake, drug selectivity may be improved. Such a derivative could, of course, be taken up into normal tissue as well as rapidly dividing tissue. However, since neoplastic tissues exhibit high intracellular polyamine demand and accummulation relative to normal tissues, selective toxicity should be observed.

The design of polyamine antiproliferatives must be based upon a careful consideration of the polyamine transport apparatus. Implicit in this is an understanding of the solution conformations attained by substrates of the polyamine transport apparatus. With these

1








2
considerations in mind, several experiments were designed to ascertain the structural boundaries for substrates of the polyamine transport apparatus. More exactly, we wanted to know the degree to which a polyamine could be derivatized without significantly affecting uptake by the polyamine transport apparatus. Previous studies led us to postulate the importance of an intramolecularly hydrogen bonded structure (11, 12). To test this hypothesis, hexahydropyrimidine derivatives were prepared (13, 14) and found to be effective inhibitors of spermidine uptake.

These hexahydropyrimidines, as well as having interesting biological activity, tumed out to be very useful as synthetic reagents, leading to a variety of actinide chelating agents which are based on the structures of naturally-occurring siderophores possessing a polyamine backbone.

Actinide contamination in humans does not represent a major health hazard because of the federal and self-imposed safety precautions adopted by those in the nuclear industry. Still, the potential for contamination exists and incidents involving human exposure and contamination have occurred (15, 16, 17). Environmental contamination is more common than internal human contamination. In a laboratory situation, when contamination occurs, trained personnel are on hand to contain the hazard and treat the problem as is necessary. When contamination occurs outdoors, the problem is not easily contained and is, therefore, more serious. Not only can water supplies, soil, vegetation, and livestock become contaminated, but curious individuals--simply by their presence--may unknowingly contaminate themselves at the site of an accident as well. One can imagine how easily a localized contamination could spread, since the danger cannot be seen.

Environmental and extemal human decontamination ultimately generate a large volume of aqueous solution which contains very low levels of hazardous actinide. This presents a problem in terms of disposal. At the present time most of this water is buried in uninhabited regions of the country, but this is only a temporary solution. Eventually, the integrity of the storage containers may fail, leading to new contamination.







3
If, on the other hand, this water could be exposed to an actinide-specific ligand and the complex removed by some physical means, the water could be decontaminated. More specifically, if a ligand were bound to a polymer backbone, a resin with a high affinity for actinides would be the result. This could be used to concentrate the contamination onto a small amount of solid resin. In this way, the complex would be converted into a form from which the metal could be recovered--by pH manipulation, ashing, etc. A resin-bound ligand would be practical for many other aspects of decontamination as well. For instance, this type of ligand could be very useful as a tool for dialysis. An insoluble ligand could also be used for prevention of absorption of ingested actinides.

Environmental contamination can lead to both intemal and external human contamination (due to our irresistible urge to touch things) and intemal human contamination. Since actinides can invade the body via several routes, such as the lungs (18), GI tract (19) or open wounds (20), the degree to which a particular organ or body compartment is contaminated will depend on the site of entry (21, 22). For example, lung contamination is observed in cases of actinde inhalation. Once absorbed, actinide deposition initially occurs in many tissues of the body: liver, kidneys, muscle tissue, lungs, heart, testes, and bone (23). With time, the metal continually migrates to the liver and bone, while other soft tissue levels decrease. Once in these sites the metal is much more difficult to remove. Clearly, the sooner a contaminated individual can receive treatment, the easier the decontamination process.

No adequate therapeutic treatment exists for persons internally contaminated with actinides. Contamination of humans is generally external and most often the metal is removed long before absorption can occur, minimizing the seriousness of the problem. However internal contamination does occur and is more serious, owing to the limited therapeutic options.

The treatment of choice is currently diethylene triamine pentaacetic acid (DTPA) chelation therapy, figure 1, but DTPA has several problems associated with it. In addition to the fact that the drug removes an insufficient quantity of actinide (24, 25, 26), DTPA has toxicity problems,






4











KCOOH



HOOC ~ N N ~ COOH

HOOC ' COOH





FIGURE 1
DTPA (Diethylene Triamine Pentaacetic Acid).










most of which are associated with its ability to sequester and remove zinc from the body (27, 28). Although the latter difficulty has been overcome by administering the drug as its zinc complex

(29), the former problem remains--DTPA alone simply cannot remove an adequate amount of actinide. For example, when DTPA was administered to rats intravenously over a three hour period 24 hours after intramuscular injection of plutonium nitrate or citrate, a dose of 120 ipoles/kg was necessary to achieve excretion of only 20 percent of the injected dose, relative to control animals (25). It is unlikely that this problem can be solved even by modification of administrative routes; e.g., continuous infusion (30). Because of its highly polar zwitterionic nature, DTPA is poorly absorbed from the bloodstream and is rapidly excreted in the urine (N. L. Scarborough, private communication), implying that DTPA's inadequacy is a result of its volume of distribution (31, 32). To obviate the shortcomings of DTPA, more lipophilic catecholamide chelators were prepared (33, 34, 35).

Some of the most stable metal complexes known are formed between catecholamide ligands--siderophores--and iron(lll). For example, the parabactin-iron(lll) formation constant is 1048 M-1 (K. N. Raymond, private communication), parabactin being a hexacoordinate catecholamide ligand with a spermidine backbone. Plutonium(IV) and iron(Ill) are similar with respect to the chemical properties responsible for complex stability; i.e., high charge, small ionic size, and high acidity. For example, Pu(IV) has a charge-to-radius ratio of 444 e/lpm as compared to 460 e/pm for Fe(lll) (36). Consequently the coordination chemistry of the actinide metals and iron is very similar. This similarity extends to their biological properties as well. For example, both metals are bound tightly to transferrin (37, 38, 39) and ferritin (40) and associate with the trabecular bone. Due to plutonium's larger ionic radius, it forms more stable complexes with octadentate ligands, in contrast to Fe(lll), which forms very stable chelates with hexacoordinate ligands, figure 2.

Based on the above similarity, researchers have used their experience with iron-specific hexacoordinate chelators to design of octacoordinate actinide-specific chelators: the design of














OH OH H O OH O HO

Parabactin





OH

"o.$=o )j o
0 H 0 N



Vibriobactin



OH

0 HH 0

CH2 2







HO HO

Enterobactin


FIGURE 2
Naturally-Occurring Siderophores.








7
actinide chelators is based on the assumption that octacoordinate ligands represent the optimum situation for binding actinides, and that actinides will form the best chelate with the catechol moiety. From this reasoning emerged a variety of effective cyclic and linear octacoordinate catecholamide ligands, referred to as CYCAM (cyclic catecholamide)and LICAM (linear catecholamide), figure 3 (33, 34, 41). These compounds have been shown to stimulate actinide excretion ( 21, 25, 30, 32, 33, 34, 42, 43, 44). A derivative of LICAM in which each catechol has a 4-carboxyl group attatched, LICAMC, was the most promising of these, figure 4. It was found that LICAMC did access different body compartments than DTPA (42). For example, DTPA removed much less of the actinide associated with the skeleton than LICAMC. Unfortunately, LICAMC shows signs of nephrotoxicity, as indicated by a 300 percent increase in the amount of plutonium deposited in this organ, relative to controls.

Since no one therapeutic agent, such as DTPA or LICAMC, for example, is likely to be able to access all sites of actinide deposition, it is unreasonable to expect that a single drug can serve to remove all of a toxic metal from a contaminated individual. In addition, even if a chelator can remove all of a particular actinide, it may be ineffective at removing other actinides. For example, Mays et al. found that DTPA removed 70 and 60 percent of the deposited plutonium and americium, respectively, from beagles (30). At the same dosage, LICAMS and LICAMC were able to remove 86 and 88 percent of the deposited plutonium, respectively, but these same chelators were only able to remove 33 and 28 percent of the injected americium. Unfortunately, LICAMS also shows signs of toxicity in the form of renal hemmorhage, edema, and an abrupt rise in blood urea nitrogen (45). These observations suggest that the answer to chelation therapy lies in the application of several non-toxic chelators, each with a different volume of distribution, so that all actinides are accessable to at least one agent.

Volf came to the same conclusion, leading him to measure the ability of the combination of DTPA and LICAMC to stimulate plutonium excretion relative to stimulation by each chelator individually (43). He found that the combination did show increased removal, but LICAMC






8







=C N ( H2) NC=O

(CH2)m (CH2 )m
N N

HO�=:/ (CH2) C=
Ho.b 2, 1DOH


1,m,n-CYCAM





OH OH OH OH C=O C=O C=O C=O I I I I HN (CH2 N (CH2 )m (CH ) NH



n,m,n-LICAM




FIGURE 3
Synthetic Catecholamide Ligands.














COOH COOH COOH COOH


OH OH OH OH
C=O C=O c=O C=O


(CH2)4 (CH2)3




HOzS HO3H 0S H HHO S H O H H

OH OH OH OH
C=O C=O C=O C=O
I I I I
HN (CH 1 N (CH2) (CH NH






FIGURE 4
LICAMC (Linear Catecholamide, Carboxylated) and LICAMS (Linear Catecholamide, Sulfonated).








10



toxicity was still evident. Mays examined plutonium excretion by DTPA and/or LICAMS. His results indicate that the combination of chelators is as effective as DTPA alone at americium removal, and as effective as LICAMS alone at plutonium removal (30). Therefore combination chelator therapy represents a practical approach to the problem of total actinide decorporation in man.

In hopes of alleviating the problems stated above, a series of new octacoordinate catecholamide chelators for application to both biological and environmental actinide decontamination was designed, figure 5. We have developed a structurally novel system referred to as H-shaped ligands, in reference to the unusual structure of their backbone. Each point of the H is bound to a catechol moiety. Our goal was to synthesize a series of non-toxic ligands, each with its own unique backbone, which possessed a selective affinity for actinide metals, and could acess more or different body compartments than DTPA.

This series of ligands is not cartoxylated; the carboxyl groups are one potential source of LICAMC's observed toxicity. Additionally, H-shaped ligands are based upon a spermidine backbone rather than spermine. Based upon experience with N1,NS-bis(2,3-dihydroxybenzoyl)spermidine (Cpd IIl)--a probable product of metabolism, if metabolism occurs--the H-shaped ligand backbone should also be non-toxic (46).

High yield syntheses were designed and carried out to give the thirteen symmetric and assymmetric model catecholamide chelator systems in figure 5. The aim of the design was to synthesize a series of ligands whose geometries and lipophilic properties differed in order that the binding geometry of the ligand to a specific metal could be optimized, and in order that the ligands would exhibit differences in lipophilicity which would carry them into different body compartments.

By way of analogy, imagine that each metal and each ligand of interest are a sphere and hand, respectively, and that the optimum geometry of the chelate, or "grip", is achieved when










OH HO


OH HO

0 NH- (CH2)a-N- (CH2)b- N 0 C=O







NH (CH2)N- (CH2)d



"Mixed" or
Symmetric Ligands Asymmetric Ligands cpd n a b c d cpd n a b c d la 3 3 3 3 Ig 3 3 3 4 lb 2 4 44 4 lh 2 3 3 4 4 Ic 3 4 3 4 ii 3 4 4 4 Id 3 3 3 j 3 3 3 4 le 3 4444 k 333 4 4 If 3 4 3 4 11 3 4 4 4 Im 4 3 4 3 4



FIGURE 5
H-Shaped Ligands.








12


only the catechol moieties, or "fingertips," of a hand can comfortably hold a "sphere." If a hand is small in relation to the sphere of interest, the grip will be weak. Conversly, as the hand grows large in relation to the sphere, it becomes difficult to hold the sphere with the tip of every finger. Clearly, a ligand which binds, or grips, metals indiscriminately would quickly be rendered inactive, since other ions, or spheres, present would soon saturate the metal binding sites. It was demonstrated that H-shaped ligands do indeed exhibit a selective affinity for actinides in aqueous solution, human plasma, and rats.















CHAPTER II

H-SHAPED OCTACOORDINATE ACTINIDE LIGANDS FOR HUMAN AND ENVIRONMENTAL DECONTAMINATION


Background
High yield syntheses were designed and carried out on the thirteen symmetric and asymmetric model catecholamide chelator systems in figure 5. The aim of the synthetic program was to design and synthesize a series of ligands whose geometries and lipophilic properties differed. If successful, the binding geometry of the ligand to a specific metal could be optimized, and the ligands would exhibit differences in lipophilicity, carrying them into different body compartments. These two issues are relevent in the following ways.

It is clear that different actinide metals will have different optimum ligand binding geometries. For example a ligand which binds plutonium may not bind americium as effectively. However, whether the metals' chelation geometries are significantly different remains to be and must be established.

With regards to lipophilicity, because the distribution volumes of the actinide metals are not equivalent throughout the various tissues (47), designing ligands having different distribution volumes is essential if all actinide pools are to be acessed. This is critical in the development of a chemotherapeutic system for total actinide removal. In the case of LICAMS and DTPA, when only LICAMS was used, only americium was effectively excreted. Conversely, if only DTPA was used, only plutonium was excreted effectively. When administered in combination, however, both actinide metals were effectively excreted. This illustrates how a combination of chelators can effectively eliminate all of the actinide burden from an animal where single chelators could



13







14

not. Of course one hopes that a single drug will remove all of the toxic metals; however, this is unlikely.

During the synthesis of these compounds, it became evident that H-shaped ligands are only slightly soluble in neutral or acidic aqueous solutions. If, however, one could lower the pKa's of the catechol moieties then the pH range over which the ligand is soluble would expand. Raymond and his associates have made several attempts to do this involving carboxylation and sulfonation of the aromatic portions of his catecholamide ligands (33, 34). These derivatives did not exhibit a very impressive change in water solubility. In fact, the final products were isolated by recrystallization from water. With this same goal in mind, N,N-diethyl-(2,3-dihydroxy-5-nitro) benzamide was prepared for puposes of potentiometric titration, along with its parent compound N,N-diethyl-(2,3-dihydroxy)benzamide. If the pKa's of this nitrated monomeric derivative are substanially lowered, relative to those of the unsubstituted monomer, then the water solubility of the tetranitro H-shaped ligand should be greatly improved relative to the unsubstituted ligand. In addition to increasing water solubility, the visible absorption spectrum of the nitro monomer was found to be pH-dependent, implying that a nitro ligand could be a useful analytical tool for metal binding. It was not our intention to apply this ligand to biological decontamination since nitro compounds, in general, are notoriously toxic. This was observed with CYCAM-N02 (44). However, it may be useful for environmental decontamination, increasing the pH range over which the ligand is effective.

During the course of these studies, it became evident that an insoluble form of chelator would be very useful, possible applications being water decontamination, prevention of gastrointestinal actinide absorption, dialysis, and prevention of percutaneous absorption. Thus, a resin-bound catecholamide was prepared. If contaminated water was passed through a column containing the resin-linked ligand, the contamination would be concentrated onto the column material, effectively decontaminating the water. If inert to the body's digestive processes, a resin-bound ligand could sequester any ingested actinides and prevent their







15
absorption into the body. Although gastrointestinal absorption is not a problem in the case of PuO2, other forms of plutonium are absorbed and retained (48). The bound metal-ligand complex would then be excreted intact via the feces. This same type of system could be used to dialyze actinides without exposing the patient to systemic chelators. It should be pointed out that we are not trying to synthesize a single derivatized resin to perform all of these feats. We intend only to create a single such derivative in order to determine if matrix-linked ligands are a useful concept. If the concept works, different resin materials can be investigated to optimize the effectiveness of each application.

Synthesis

Catecholamide Liaands

Parent catecholamides

It should be pointed out that all syntheses were designed to facilitate modification of the molecule. This is critical in drug design, as structural changes are frequently required in order to improve on pharmacological behavior. Without such flexibility, one would be faced with starting from the beginning--i.e., developing a new synthetic sequence for each structural analog--rather than simply introducing a modified reagent into the present design.

Note that the thirteen analogs shown in figure 5 fall into two classes--symmetric (la-t and lm) and asymmetric or "mixed" ligands (lgj). Symmetric ligand precursors (~-1 and 2m) were synthesized by coupling an appropriate acid dichloride (succinyl, glutaryl, or adipoyl) with two equivalents of either N',N8-bis(2,3-dimethoxybenzoyl)spermidine, norspermidine, or homospermidine, figure 6.

In the case of an asymmetric ligand, one must selectively acylate each end of a diacid with a different spermidine derivative, necessitating a two-step process; i.e., the stepwise addition of each amine derivative. This is achieved by the aminolysis of succinic or glutaric anhydride (49) by the nor- or homospermidine derivative, figure 7. The "half-acids" thus produced (Qa-) can next be coupled to a different spermidine derivative, yielding methylated asymmetric ligand







16



OCH3 CH30


OCH3 CH30

NH- (CH2)a- NH- (CH2)b- NH


0 0 Cl (CH2)nCl / NEt3 OCH3 CH30


OCH3 CH3 0

NH- (CH2)a N- (CH2)t- NH 0

C=O


(CH2)n
OCH3 1= CH3 0

NH- (CH2)a-N- N-(CH NH 0

cpd a b n 2a 3 3 2 2b 4 4 2 2c 3 4 2 2d 3 3 3 2e 4 4 3 2f 3 4 3 2m 3 4 4


FIGURE 6
Synthesis for Symmetric Ligands.







17


NH- (CH2)a--NH- (CH2)b-NH

OCH3 o CH30

OCH (CH2)n CH3


COOH
OCH3 CHO
(CH2)n
OCH 1 CH3 i
3 C=O CH
NH- (CH2)a-N- (CH2)b-NH

3a 3 3 2 3b 4 4 2 3c 3 3 3 3d 4 4 3

CHN 4 OCH CH
OCH 2 CH3O-Yf NH-(CH2)-NH-,- N2)d- NH'

NH- (CH2)-N- (CH2)b-NHO


6 CH 3 ((CH2)n
OCH3 C=0 CH30
I
0 NH- (CH2) c-N- (CH2)d-NH

CH3 2313131412 CH

OCH 23 3 4 4 2 CH30
CH 2i 44342 3
2j 3 3 3 4 3 2k 3 3 414 3 21 4 4 3 4 3



FIGURE 7
Synthesis for Assymmetric Ligands.







18

precursors (2g1). The removal of the methyl protecting groups results from the action of boron tribromide, figure 8.

Nitro derivatives

Initially, synthesis of a tetra-nitro derivative of an H-shaped ligand was attempted via the tetra-BOC precursor 4d shown in figure 9. This method yielded the desired compound, but required many steps; thus the total synthesis was very time consuming. Also, one must acylate each amine with the same group; we ultimately hoped to be able to place a different acyl group on each amine so that we could "fine tune" the biological properties of the drug. To make this synthesis more cost and time efficient and more flexible, other synthetic methods were sought.

We decided to utilize the synthesis used for symmetric ligands, figure 10. This synthesis was expected to proceed without difficulty, based upon the experience of other similar molecules. The problems, we expected, would occur in the synthesis of the tetramethoxy precursor A. Until this time, literature methods (50, 51) were used to prepare tetramethoxy precursors, figures 11 and 12. This methodology is unsuitable for the synthesis in question. Although 2,3-dimethoxy-5-nitrobenzoyl chloride is easily prepared and smoothly acylates N4-benzyl norspermidine, the hydrogenolysis conditions necessary for cleavage of the N-benzyl protecting group would surely reduce the aromatic nitro groups to nitrosos and/or amines. Therefore another protecting group was sought. After several other groups were investigated, the tetramethoxy precursor A was synthesized using the formaldehyde adduct of norspermidine, N-(3-amino-l-propyl)hexahydropyrimidine, figure 13. To this hexahydropyrimidine--prepared by the method of McManis (13, 14)--was added two equivalents of 2,3-dimethoxy-5-nitrobenzoyl chloride, giving , The deprotection was performed by several different aldehyde trapping agents, but dimedone (5,5-dimethyl dihydroresorcinol) proved to be the most reproducible and versatile method, effectively deprotecting a wide variety of hexahydropyrimidine derivatives in high yield. Compound B was reacted with succinyl dichloride to yield the tetranitro precursor 5a, figure 10. It seems safe to assume that the asymmetric






19



OCH3 CH30 OCH3 CH30

NH- (CH2)a, N- (CH2) NH
C=O
OCH3 1 CH30
(CH2)n
OCH3 1= CH30

NH- (CH2)c- N- (CH2)d- NH 0
(2a-m)

OH BBr3 HO




NH- (CH2)a--N- (CH2)b NH
C=O
OH HO
(CH2)n
OH I HO
c=o
NH- (CH2 )c--N (CH2)d-NH

( a-m)



FIGURE 8
Removal of Protecting Groups.







20




0 00


Cl CI



I NEt3


0o NH N NH 0
0 r 0


0 0 0

1. CF3C00H

2. 02N-- H S OCH


02N OCH3 CH30 NO2
C H3 CH3
0 NH N NHAO
02N- OCH 0 CH3 0 NO2

CH3 0 CH 0
0 NH NN%.NHAO
5d



FIGURE 9
Synthesis for Tetra Nitro Ligand.






21

02N OCH CH 00 NO2
LN . OCHH CH 0- 7 CH
O N N N 0

o0
Cl"j O Cl NEt3
02N'S f< OCH3 CH3O N02




ON OCH CH30 SOCH CH 0
0 NH - HN NH 0 O




02N OCH CH r3 NO2
0 NH ~N NH 0



02N OH BBr HO NO2

0 NH NSNH 0

02N OH HO N02 OH o=0 HO
0 NHN 0NH 0




Figure 10
Alternate Synthesis for Tetra Nitro Ligand.







22






NH2











SYN~C=N NCN








N--NH2 N----NH2 N--- NH2

G NH2 NH2 C NH2 FIGURE 11
Synthesis for N4-Benzyl Triamines.







23






H2N~ N NH2

+
OCH3
OCH3



NEt3





OCH CH 0 NN NH H2/catalyst


CH CH30
I-OCH3 H CH30
qONH N NH 0




FIGURE 12
Synthesis for Tetramethoxy Precursors.







24

H2N' N N-" NH2


CH2=O


HN N NH2


02N OCH3
j O lN% OCH3 00 ClH

02N OCH3 CH 0 N02 CH CH3 0
0 N NN - NH 0


0,t0




02N OCH CH 0 NO2
CH CH3 0 y
0 NH N NH 0
H




FIGURE 13
Synthesis for Bis Nitro Tetramethoxy Precursor.









25
synthesis will also proceed smoothly with nitro derivatives. Thus, one should be able to synthesize a bis nitro H-shaped ligand precursor in which two catechols are nitro-derivatized and two are not. Unfortunately, reaction of boron tribromide with 5a yielded a product which was not iron positive (52), indicating that free catechols are not present in the product. Trimethyl silyl halides were no more effective at freeing the catechols in this compound. This was very surprising in light of the fact that N,N-diethyl(2,3-dimethoxy-5-nitro)-benzamide deprotected in the usual manner, and that 6 was deprotected with facility as well.

N-(3-amino-l-propyl)hexahydropyrimidine (APHHP) has proven to be a very useful reagent to us for organic synthesis. In the preceding synthesis, the fact that the central amine of norspermidine is now tertiary has been exploited. Methods were also developed in which the difference between the primary and secondary amine is used to our advantage. If one were to add a single equivalent of an acid chloride to APHHP, the products would include primaryacylated, secondary-acylated, and bis-acylated hexahydropyrimidine, an undesirable result. By using an N-hydroxysuccinimide ester rather than an acid chloride, one can selectively acylate only the primary amine in near quantitative yield. Although other researchers have used hindered acyl transfer reagents or protecting groups to achieve the same end, these reagents are not nearly as selective (53, 54, 55). Once isolated, this mono-acylated product can react with a different acyl chloride, followed by deprotection with dimedone, to yield an asymmetrically substituted polyamine. Alternatively, the monosubstituted hexahydropyrimidine can be deprotected with dimedone, followed by reaction with a different N-hydroxysuccinimide ester to yield the identical product.

In order to demonstrate the usefulness of this method, j1 was prepared, figure 14. APHHP was acylated with (2,3-dimethoxy)benzoyl-N-hydroxysuccinimide, acylation occurring strictly on the primary amine. This product (2) was either reacted with 2,3-dimethoxy-5-nitrobenzoyl chloride followed by dimedone, or dimedone followed by (2,3-dimethoxy-5-nitro)benzoyl-N-hydroxysuccinimide or (2,3-dimethoxy)benzoy-N-hydroxysuccinimide, to yield 10 or







26




HN-* NNH2
CH30 o CH30 o
Cl





HN 'NH 0 0 K NH 0 CH0
CH3 01 CH CH 3 0 HN N M NH 0





CH /


2 0 2.
CH NO
2. CH30 20




02N. OCH CH30"o
,OCH CH3O

o0 NH NH NH 0




FIGURE 14
Synthesis for Mono Nitro Tetramethoxy Precursor.









27
N1,N7-bis-(2,3-dimethoxybenzoyl)norspermidine, respectively. To show that asymmetric synthesis will work with nitro derivatives, JQ was successfully reacted with aa, figure 15, to yield the mono-nitro octamethoxy precursor. In this case reaction with boron tribromide proceeded with ease to yield la.

Resin-Bound Catecholamide Liaands

The catecholamide

The intention was to synthesize a catecholamide ligand with a chemical handle with which the molecule could be covalently bound to an inert resin matrix. The design of such a molecule required the integration of several factors: (1) in order for selectivity to be maintained it is crucial that the matrix-linked ligand retains the conformational mobility of the parent ligand, (2) the chemical handle must be capable of covalent binding with a resin material, (3) the placement of the handle into the molecule must be done in such a way that the geometry of the metal-ligand complex is undisturbed; i.e., the handle should be on the exterior of the chelate, (4) if the ligand were attached directly to a polymer matrix, it is conceivable that the matrix would impose conformational restraints upon the ligand; i.e., the molecule would have restricted mobility, affecting its ability to sequester the metals of interest. Thus the handle should allow enough distance between the resin matrix and the chelator moiety so that the latter does not interfere with the mobility of the chelator and, more importantly, with metal chelation.

Compounds l1 and 11 with a chemical handle (13 and 0l) were synthesized by simply replacing glutaric acid with N-protected glutamic acid, figure 16. In this way a protected amine functionality is introduced into the backbone of the molecule, in a location which is far removed from the chelating sites. The t-butoxycarbonyl (BOC) group was chosen to protect the amine handle. This moiety is quickly and quantitatively removed by the action of triflouroacetic acid (TFA). A linear organic "string" in the form of N-BOC-8-aminocaprylic acid was attached to the amine handle to increase the distance between the ligand and the resin surface, figure 17. Once attached, a new amine functionality can be exposed by treatment with TFA to yield 15f.







28


02N~ OCH3 CH30
Z 3 OCH CH 30 1
O=C C=O
NH1 NH - NH



OCH 3N COOH CH3
OCH / CH 0
0=C NHN NH- C=0 5a


CH3" N/CH3



O2NQ(OCH3 CH \; 'CH 3 CH 0"
0=C C=0


OCH CH0 C 3o

y-OCH 3 o CH 30 Q
O=C C=O
\NHM NM NHi


FIGURE 15
Synthesis for Mono Nitro Octamethoxy Precursor.







29

CH3 CH3O
OCH CH3
oH






0 0


0 NHIN NH0
. OCH3


H3 BOC-NH- CH30

OCH CH 0
0 NHN NH 0

12f
Y OCH3

3 CC H
OCH CH


CH 2 0

H3 NH2 CH 0

NHN NH 0



FIGURE 16
Synthesis for an Amino-H-Shaped Ligand.








30








OCH NH 3CH 0





CHCH 0










OCHHN N
0 0








N H N CH
OCH3 CH30


,. O.NH-C7 H14-C-oH/DCC/


2. CF3 COOH


OCH3

0 0

CH 0
N 277 H, N4-C N CH 3







OCH3 CH30


FIGURE 17
Attatchment of the Spacer Molecule.








31
Our first attempts to attach such a string involved 4-aminobutyric acid. The condensation went well but when exposed to TFA the product did not stain with ninhydrin, a reagent which qualitatively detects primary and secondary amines and amine salts. A possible explanation is shown in figure 18. However, when 8-aminocaprylic acid was used cyclization did not occur, yielding the desired product. An alternative synthetic pathway was also developed, figure 19. In this synthesis, N-BOC-8-aminocaprylyI-N-hydroxysuccinimate was reacted with glutamic acid, yielding the diacid derivatized with the organic string. This intermediate was reacted with two equivalents of N1,N8-bis(2,3-dimethoxybenzoyl)spermidine in high yield. In the previous synthesis, the H-shaped ligand was assembled prior to attachment of the organic spacer arm. This latter pathway renders the synthesis more cost effective because the synthesis is convergent rather than linear. Therefore, smaller building blocks, or intermediates, are used in the final steps.

The resin

It was decided that an amide linkage would be appropriate for coupling to the resin. Thus a carboxylic acid resin was sought. Merrifield's resin (56) is a commercially available resin composed of polyvinyl benzene monomers with a small percentage of the benzene moieties converted to their chloromethyl derivative, figure 20. This resin can be reacted with bicarbonate and dimethylsulfoxide to yield the aldehyde (57). Oxidation by dichromate in acetic and sulfuric acids yields the benzoic acid derivative 16 (58). Altematively, reaction of the chloromethyl resin with malonic acid in the presence of pyridine and piperidine yields the cinnamic acid derivative 1Z. The advantage of having these two resins is that one can evaluate the effects of changing the length of the connecting string. Another resin, Amberlite IRP-64, is made from acrylic acid monomers, and already exists as a carboxylic acid (1). This resin was also used. These three resins were converted with thionyl chloride to their acid chlorides (59, 60) and reacted with the methylated amino-ligand 15[, yielding 19Jf 20 and 11 respectively. The degree to which the ligand coupled was determined by weighing the resin before and after exposure to the ligand.








32



OCH3 H3COO
C=O O=C
H HH H R2
NH 0 H2NH NH2 CF3 COOH o .

0 0
H H H HR2 C=O O=C


OC3 3COOCH3 H3CO I OCHo H CO where R= H C H3CO

NH 0




-H2 0





C=O O=C C=O O=C



NH N 0

0 0

C=0 0=C C=0 O=C
OCH3 H3CO >2 : OCH3 H3C0>
OCM H3CO OCH3 H3CO



FIGURE 18
Cyclization Mechanism in Triflouroacetic Acid.







33

0

tBOC-NH-0 0-No +0 0
0 o
HO OH
NH2


0 1 HO

o'N H NH < CH3. NCH3

OCH3 CHD N HO OCH3 CH3 0 O=C C N=CN

OCH3




0 0 CH0
OyNH -C, H4- -NH CHI3 0



Q NH N *\/ y.NH \

H CH50
OCH3 CH3 0

FIGURE 19
Alternate Synthesis for Amino-H-Shaped Ligand with Spacer Molecule Attatched.









34





COOH

Merrifield's resin Amberlits IRP-64


NaHCO3/DMSO 1
CI-S-Cl




C-H -Cl Na2Cr2 0

COOH cOH2S4




0

-OH -OH


0 0


II II
CI-S-Cl Cl-S-Cl




C- C1 1
-Cl- Cl




FIGURE 20
Synthesis for Acid Chloride Derivatives of Commercially Available Resins.








35

Reaction with boron tribromide gave the free catecholamide resins 22 23f, and 24L as indicated by a positive reaction with ferric chloride solution (52).

Lastly, CH-Sepharose-4B activated resin was obtained. This resin was ideal for ligand immobilization. The matrix is composed of a polysaccharide backbone which has been derivatized with hexanoic acid present as the N-hydroxysuccinimide ester. Since this ester will react selectively with amines in the presence of alchohols, the methyl protecting groups were removed prior to coupling of the ligand to the resin material. Therefore the resin need not be exposed to the harsh chemical reaction conditions necessary for demethylation. This was not the case with the acid chloride resins described. Exposing 13 to boron tribromide gave a high yield of the desired "catecholamine" 25. figure 21. When 25 was coupled to the resin, giving 26, a spacer of six carbon units is inherently inserted between the ligand and the polymer surface, allowing for free motion of the molecule. Coupling was performed in 50 percent pH 7.2 phosphate buffer: 50 percent ethanol to ensure dissolution of the catecholamide. In addition, the resin was washed with a large volume of ethanolic buffer after the reaction was quenched to ensure removal of any uncoupled ligand. A final wash was performed with 50 percent pH 4.0 acetate buffer: 50 percent ethanol to protonate all of the catechoyl functionalities. The last wash was concentrated and reacted with FeCI3. No color change was observed, indicating that no free catechol was present in this last wash. It was gravimetrically determined that ligand coupling was quantitative, by the weight difference between the control resin and the catecholamide resin. When a small amount of this resin (2) was placed in FeCI3 solution, a dark purple color developed, indicating the presence of the catechoyl functionality.

Liaand Stoichiometrv: Job's Plots

Before ligand binding constants could be calculated, it was necessary to measure the metal-ligand ratio for all metals and ligands involved in the determination. Thus, Job's plots (61) were performed in an ammonia buffer at pH 9.2 with copper for ligands la, 27. and Cpd II (N',N8-bis[2,3- dihydroxybenzoyl]spermidine), figure 22, at a total concentration of metal and








36

R OH HO R

H HO
O NHN- NH 0 H-HBr

27 R = NO2
Cpd II R = H

M Stoichlometry_(L:M)
Th4+ 1:1
Cu2+
Cu 1:1


HOH HO R

OH HO
0 NHONNH 0


OH HO
OH a HO

ONH NNH 0


11a R= NO2
la R=H


M Stoichiometrq_(L:M)
Th4+ 1:2 CU2+ 1:2



FIGURE 22
Results of Job's Plots with Catecholamide Ligands.








37

ligand of 33 pM. The results indicate that Compound II and 2Z form 1:1 metal-ligand complexes with copper and la forms a 2:1 copper complex. These results are consistent with McGovern's findings for Compound 11 (62, 63). In addition, it is reasonable to expect 2:1 stoichiometry for la and copper, since la is essentially two Compound II molecules linked together.

In the case of copper, spectral changes upon metal binding are small, but are large enough for this type of measurement. With thorium, however, the spectral changes could not be used for Job's plot measurements. Therefore, the mono nitro H-shaped ligand 1.a, figure 22, was used in place of ligand la. In order to show that using 11a would be a true reflection of what happens when la is used, a Job's plot was generated for 11a and copper. If the nitro group caused a significant change in the ability of the catechol to bind a metal, the Job's plot would not indicate 2:1 stoichiometry. For example, if the nitro-containing catechol binds copper more effectively, the first equivalent of copper would associate primarily with this group. This phenomenon would be reflected in the Job's plot as 1:1 stoiciometry, since the second equivalent of copper would not affect the nitro-containing catechol, which is already associated with the first equivalent of metal.

Conversely, if the nitro-containing catechol binds copper less effectively, the first equivalent of copper would not affect the visible spectrum since the metal would not associate with the nitro-containing catechol. Neither of these effects were observed, indicating that 11a does indeed accurately reflect the chelation chemistry of la.

Thus, compunds 2Z and 11a were used as a measure of the stoichiometry of the thorium complexes of Cpd II and 1J, respectively. For compound la another experiment was performed to insure the accuracy of using 11 in its place. A Job's plot was performed with la and thorium with eriochrome black T (EBT) present. It has already been shown that EBT does not displace thorium from H-shaped ligands under these conditions. Therefore, EBT can be used to complex any thorium which is not bound to 1. All of these experiments yielded the same information: Cpd II and 27Z form 1:1 metal-ligand complexes with thorium, and la and 11a form








38
2:1 complexes with thorium. This was somewhat surprising in light of the 1:1 stoichiometry observed with hexadentate catecholamide chelators and iron. This stoichiometry is most likely due to the ability of hydroxide to compete with la for the coordination sites of thorium at the experimental pH; i.e., ThO2 is the specie bound. These results are summarized in figure 22.

Lioand Precioitation Technioues: Removal of Actinides from Water

Several preliminary findings indicated that precipitation techniques would be a very effective means of decontaminating radioactive solutions. In earlier studies on hexacoordinate catecholamide chelators, it was observed that neither the chelator nor the iron chelate could be dialyzed. Both species were held up on the dialysis membrane (64), suggesting that these chelators could be bound by a filtration membrane. Additionally, when a millimolar aqueous solution of H-shaped ligand was prepared at pH 10.0 and the pH was lowered below 8.5, precipitation occurred. This observation recurred when an equivalent of thorium(IV) or plutonium(IV) was present, implying that complex formation does not interfere with the process of precipitation.

Several filtration techniques were developed to utilize the pH-dependent solubility properties of H-shaped ligands for decontamination. In the first of these, it was determined that a metal-ligand complex could be removed from solution by filtration. A solution of thorium(IV) chloride buffered at pH 9 was added to several tubes. To these were added one equivalent of an octacoordinate ligand in a small volume of methanol. Controls received only methanol. After neutralization, samples were filtered and the filtrates analyzed for thorium. In order to have a benchmark for the data, the ligand 3,4,3-LICAM was synthesized and evaluated along with H-shaped ligands. Arsenazo III (ARS, figure 23) was used to colorimetrically detect the thorium(IV) remaining in solution (65). Absorbance maxima for the thorium-ARS complex were observed at 615 nm and 666 nm. The 666 nm peak was used for the quantitative determination of thorium(IV) because there was less absorbance by free ARS at this wavelength than at 615 nm. It was found that 96.39-99.97 percent of the actinide could be removed in this








39


CI





HO

CI-PRN SH2 HO OH H2%fs o


N= N N=N





Arsenazo I I I



HO S-sH
OH

N=N NO2

Eriochrome Black T FIGURE 23 Colorimetric Reagents used for the Quantitative Detection of Metals.








40

way, first column of table 1. Note that 3,4,3-LICAM did not perform as well as H-shaped ligands in this experiment, removing 88.9 percent of the actinide. Controls without a ligand showed no removal of actinide.

In a series of experiments conducted in a joint effort between Rocky Flats and the University of Florida it was demonstrated that plutonium could be removed from aqueous solution by the same precipitation technique (J. D. Navratil, private communication), table 2. In practice, this means that if contaminated water is adjusted to high pH and a ligand is added, the contamination can be concentrated onto a filter by neutralization prior to filtration.

One might argue that a portion of the actinide, due to the nature of the experiment, was simply trapped by precipitation of the ligand in a non-specific way. If this were the case, the ligand would be able to bind more metal than the amount dictated by the ligand-metal complex. It was shown that this is not the case. When the ligand precipitation experiment was repeated with excess thorium, the ligand did not remove more than the stoichiometric amount of metal, indicating that non-specific precipitation did not occur.

The next experiments determined that H-shaped ligands can bind thorium(IV) when the metal is introduced into an acidic solution containing the precipitated ligand and, conversely, that the ligands can effectively precipitate thorium when introduced into an acidic solution containing the actinide. This renders it unnecessary to adjust the solution pH prior to addition of the chelate, an important financial consideration.

First, ligand was precipitated from a basic solution by acidification. One equivalent of thorium(IV) was then added. The suspension was stirred for several minutes, filtered, and the effluent assayed for thorium. Then, an acidic thorium solution was prepared. One equivalent of a ligand was added as a methanolic solution. The suspension was stirred for several minutes, filtered, and the effluent assayed for thorium. These methods proved very effective, removing 93.04-99.91 percent of the metal, table 3.








41




TABLE 1
Results of Thorium Precipitation Experiments With Catecholamide Ligands.


Percent Thorium Removala
Compound Competing Metals Competing Metals Percent Absent Prese nc Iifference

la 96.90 t 1.42 (5)b 85.86 � 4.19 (6) 11.39 lb 99.76 0.35 (6) 89.67 4.19 (5) 10.11 Ic 97.77 2.72 (6) 88.49 2.86 (6) 9.49 id 96.75 0.95 (6) 85.58 4.95 (6) 11.55 le 99.38 � 1.29 (6) 93.76 � 4.40 (6) 5.66 if 96.39� 1.06 (6) 84.32 � 4.33 (6) 12.52 lh 99.74 � 0.45 (6) 89.54� 2.31 (6) 10.23 lj 99.59 � 0.55 (4) 84.96 3.93 (6) 14.69 1k 99.97 � 0.05 (5) 85.38 � 5.87 (6) 14.59

LICAM 88.90 � 3.40 (5) 71.05 � 7.53 (6) 20.08 a [Th]intial=3.24 x 10-5 M, [ligand]initial=3.3 x 10-5 M. b The number in parentheses is the number of experiments performed. c Competing metals used were Fe(lll), Ca(ll), Mg(ll), Mn(ll), Zn(ll), and Hg(ll). One equivalent of each metal was used.









42







TABLE 2
Results of Plutonium Precipitation Experiments With Catecholamide Ligands.

% Pultonium Removal
Coround Withut Iron With iron


la 68�3 59�5 lb 90+6 86�6 id 63�11 58�6 le 55�1 47+4

If 92 10

11 88�9

1k 66 11 54�6


[Pu]initial =[ligandinitial =1 x 10-5 M.








43






TABLE 3
Results of Catecholamide Pre-Precipitation Experiments with Thorium.

Percent Thorium Removal
Compound Ligand Added to an Metal Added to an Acidic Metal Solution Acidic Lioand Susension la 99.91 � 0.16 (4)a 96.90 �3.41 (6) lb 98.21 � 3.15 (5) 99.16 �+0.82 (6) Ic 99.83 � 0.19 (6) 98.90 � 1.66 (6) id 99.54 � 0.79 (5) 93.04 : 6.98 (6) 1 e 95.60 � 6.21 (6) 99.08 � 1.44 (6) If 99.40 � 0.63 (6) 98.93 � 0.51 (5) 1h 99.82 � 0.31 (4) 99.37 � 0.46 (6) ij 94.58� 4.18 (5) 99.35� 0.57 (6) 1k 94.31 � 7.20 (5) 98.40 � 1.81 (5)

LICAM 97.64 � 2.00 (5) 97.95 � 1.02 (6) aThe number in parentheses is the number of experiments performed. [Thlinitial=3.24 x 10-5 M, [ligandlintial=3.3 x 10-5 M.








44
A property which is requisite for the H-shaped ligands in order to be practical for decontamination purposes is their ability to sequester actinides in the presence of other metal ions. Thus, experiments were designed to determine whether other polyvalent cations could compete with actinides for the ligand binding site. Cations were selected because of their +2 and +3 oxidation state, their relative abundance and importance in biological systems, and their known ability to compete with actinides for binding with other chelating agents. These experiments involved repeating the procedures outlined above with an equivalent of each competing metal present before introduction of the ligand. Control experiments without thorium(IV) were also performed to determine if the other cations could form a complex with ARS, causing a colorimetric reaction. Very little interference was detected, which suggests that the cations should not affect formation of the thorium-ARS complex to any significant extent. Although substantial thorium removal was still observed--93.76 to 84.32 percent--the competing metals did have some effect, and some ligands were affected to a greater extent than others, table 1. For example, le demonstrates the most selectivity while ji and 11 are the least selective H-shaped ligands. Once again, H-shaped ligands removed much more thorium than the 71.05 percent removed by 3,4,3-LICAM. In addition to removing less metal, 3,4,3-LICAM is the least selective catecholamide, demonstrating a 20.08 percent decrease in thorium removal.

At Rocky Flats, the ability of iron (III) to compete with plutonium for the ligand was measured, table 2. It was found that iron exerted a small effect on plutonium binding, a result consistent with our observations. Since iron forms a very stable chelate with the catechoyl moiety, it was somewhat surprising to us that iron did not compete more effectively for the octacoordinate ligand. These results indicate that it is reasonable to expect octacoordinate catecholamide ligands to experience little interference from biological iron when searching for absorbed actinides.








45

Insoluble Chelators: Removal of Actinides from Water and Blood Plasma

Although very effective as chelators, H-shaped ligands have some limitations for applicability. For example, a chelator taken by mouth to prevent GI absorbtion would be useless if absorbed from the gut. Additionally, the filtration technique is not applicable to basic solutions, where the ligands are very soluble. A soluble ligand would also be inappropriate for dialysis. Therefore we wanted to modify the ligand into an insoluble form, to overcome the types of problems just mentioned.

XAD-4-Sorbed Lioands

The Rocky Flats group demonstrated that when our H-shaped ligands are adsorbed onto XAD-4, a macroreticular resin, and contacted with aqueous plutonium-containing solutions, plutonium(IV) adsorbs very effectively onto the resin (66). In these early studies, excessive amounts of ligand were coated onto the resin, resulting in a situation where a great deal of the ligand was not available to sequester plutonium from solution; i.e., ligand exposure on the resin was minimal. Presumably the ligand molecules stacked on top of one another rather than forming a monolayer. In spite of this, remarkable decontamination factors (Df's) were observed, values in the hundreds. The Df is defined as the moles of metal bound per gram of ligand divided by the moles of metal in solution per milliliter of solution. This value, of course, is difficult to use for comparison of various ligands because the molecular weight of the ligand is incorporated into this term. Therefore, it is more useful to measure the Kd for the system, defined as the moles of metal bound per mole of ligand divided by the moles of metal in solution per milliliter of solution. Df's are still presented to facilitate comparison to earlier work. In later experiments where the ligand loading had been optimized, Df's on the order of 107 were attained (67). Furthermore, the Rocky Flats group demonstrated that a significant amount of plutonium(IV) was removed from human plasma (up to 36 percent) when 10 mL were exposed to 250 mg of resin at 370C overnight. This result indicates that resin-bound H-shaped ligands may be practical for dialysis. Their results are summarized in tables 4, 5, and 6.








46





TABLE 4
Kd's and Df's for Plutonium at a Ligand Loading of 10'6 moles ligand/0.25 g XAD-4.

Comrund Kd Df
la 2.69 x 106 1440 Ib 7.48 x 105 1140 Ic 9.57 x 105 1160 id 5.30 x 106 1560 le 2.25 x 106 1430 if 2.12 x 107 2300 1h 3.75 x 106 1501 1j 6.60 x 106 1810 1k 3.69 x 105 1110 11 4.40 x 106 1530 LICAM 4.81 x 106 1980


250 mg loaded resin was exposed to 10 mL solution containing 5 x 10-7M plutonium. Samples were mixed overnight prior to measurement.
Values have been corrected for plutonium removal by control resin.








47





TABLE 5
Kd's and Df's for Plutonium at a Ligand Loading of 10"7 moles ligand/0.25 g XAD-4.

Compund Kd Of
la 2.88 x 107 14,600 Ic 7.50 x 106 11,900 id 1.15 x 108 10,200 le 2.74 x 107 12,500 if 4.50 x 107 11,500 1i 5.31 x 106 11,500 1k 3.95 x 106 11,200 11 3.85 x 107 14,700 LICAM 2.03 x 107 16,100


250 mg loaded resin was exposed to 10 mL solution containing 5 x 10-5M plutonium. Samples were mixed overnight prior to measurement.
Values have been corrected for plutonium removal by control resin.









48






TABLE 6
Results of XAD-4 Experiments With Human Plasma.

Percent Plutonium Removala Compound A B

la 8+6b 3+19

lb 0�6 O+6

ic 21 8 26�10

Id 17t3 0�7

le 28 17 8�18

if 36�8 9�4 1h n.a. 3.4 Ij 30 13 6 6 1k 22 0 n.d.

11 21�8 0�17




A = 10-3 moles ligand/0.25g XAD-4, B = 10-6 moles ligand/0.25g XAD-4. a250 mg loaded resin was exposed to 10 mL plasma containing 4.18 x 10-7M plutonium. Samples were mixed ovemight priorto measurement. bValues have been corrected for plutonium removal by control resin.








49

Covalentlv-Bound Liaands

IRP-64-bound catecholamides

The precipitation experiments previously described for free ligand were repeated using resin Z4 in place of a free ligand, but without adjusting the pH to neutrality prior to filtration. Remember, pH adjustment was necessary to cause precipitation of the ligand-metal complex. In this case, the ligand is insoluble at every pH. Therefore, pH manipulation is unecessary. It was found that control resin--in this case the methylated catecholamide resin--removed as much thorium as the catecholamide resin. It was concluded that the polymer matrix was responsible for a large amount of metal binding. Unfortunately, many free carboxyl groups remain in the derivatized resin; only about ten percent reacted with a ligand. These free carboxyl groups are able to bind actinides, and therefore can interfere with attempts to measure metal binding due only to the bound ligand. Although these carboxyls appear to enhance thorium binding, this binding is non-specific. In order to measure the ability of the chelator, once bound to a polymer matrix, to bind thorium selectivly, binding due to the non-specific matrix must be either measured or prevented; we chose the latter approach.

Control resin was placed in pH 9.2 buffer with thorium and different concentrations of EDTA. it was found that an EDTA concentration of 150 mM was needed to prevent binding of thorium to the control resin, table 7. Unfortunately, this concentration is high enough to compete effectively under these conditions with the catecholamide for the metal, as indicated by reduced thorium removal by 24f at this concentration. It was concluded that a different resin should be used.

CH-Seoharose-4B-bound catecholamides

Next, resin 26 (figure 21) was tested. In this case 2.6 mM EDTA was enough to prevent the control resin from binding thorium when present at concentrations up to 0.64 irM, figure 24. In this case the control resin was reacted with ethanolamine, the reagent used to quench the excess reactive sites on the catecholamide resin. This concentration of EDTA did not affect








50







TABLE 7
Results of Experiments With IRP-64 Resin Derivative.

[EDTA] Resin % Removal Hours Kd

15 mM - 0 3 2.22 x 103
OMe 23.5 OH 71.1


150 mM - 0 6 2.11 x 103
OMe 8.1 OH 47.8


150 mM - 0 24 5.21 x 103
OMe 0 OH 65








51






100



90



a 80



70


[EDTAI [Th] 60 - 2.52 mM 0.64 pM A 2.29 mM 0.03 IpM S2.68 mM 0.076 IM S50- M 0.27 mM 0.14pM C, 2.65 mM 7.04 pM
-2.65 mM 60.7 pM
4 0 ' l ' I ' l ' I ' I ' l '
0 1 2 3 4 5 6 7 8 9 10 mg Control Resin Added



FIGURE 24
Determination of the Ability of EDTA to Prevent Adsorption of Thorium onto
CH-Sepharose-48 Control Resin at pH 9.2 as a Function of the Milligrams of Resin Used.








52

thorium binding by the catecholamide portion of 26, figure 25. For these experiments, 23thorium, a high energy a-emitter, was used, enabling very precise measurements of trace amounts of metal to be made.

By adding 230thorium to a solution of 232thorium, the concentration of metal can be increased while the number of counts added remains constant. When the concentration of thorium was increased in this way from 30 nM to 76 nM, it was found that control resin did bind some thorium, approximately six percent. When the thorium concentration was raised to 7.04 pM, more binding was observed; approximately 29 percent of the thorium in solution was non-specifically bound by control resin. In both cases, 26 was still able to remove more than 98 percent of the thorium in solution, figure 25. When competing metals were added, no significant difference in thorium removal by resin-bound catecholamides was observed, figure 26. These data indicate that matrix-bound ligands represent an effective form of H-shaped ligands which could be applied to many facets of decontamination.

Equilibrium Solution Chemistry: Thermodynamic Bindina Constant Measurements Eriochrome Black T Comnetition

Many techniques cannot be used to accurately determine very large metal-ligand binding constants due to the fact that the free metal in solution, or the free ligand in solution, is present at a concentration which is so low that it cannot be measured accurately. Therefore, it is essential to approximate this constant before designing an experiment for an accurate determination. A spectrophotometric technique was developed to approximate the magnitude of the thorium-ligand association constant. The colorimetric reagent eriochrome black T (EBT) was very well suited for this approximation, figure 23. EBT forms a very stable 1:1 complex with thorium(IV), the absolute binding constant being on the order of 1027 M-1. Its working pH ranges from 8 to 10, which is ideal for H-shaped ligand solubility (69). At its working pH EBT exhibits an absorbance maximun at 660nm while the thorium-EBT complex has an absorbance maximun at 570nm (70). A 10-5M solution of EBT can quantitatively detect thorium(IV) to








53







100

90

80

,70

60

S - EDTA] IThS 40- o 2.29 mM 0.03 IM Sa 2.68 mM 0.076pM
S0- 2.68 m 0.076pM 7 . 2.65 mM 7.04 p M Q 20 - 2.65 mM 7.04 IpM 13 2.65 mM 60.7 pM
10F- G 2.65 mM 60.7 pM



0 2 4 6 8 10 12 14 mg Resin Added




FIGURE 25
Determination of the Ability of 26 to Remove Thorium from Aqueous solution (pH 9.2).
Control Resin is Indicated by Squares and 2 is Indicated by Circles.








54





100-

90

E 80

- 70
ca 3-C Control resin, c 60
SControl resin, 50 present
-h Catecholamide resin, = 40 - thorium only
- "Catecholamide resin,
30 - --. competing metals ....present
Q 20
h.

10



0 1 2 3 4 5 6 7 8 9 10 mg Resin Added






FIGURE 26
Determination of the Ability of Other Metals to Interfere with Removal of Thorium from
Aqueous Solution (pH 9.2) by26. Metals Used were Fe(lll), Ca(ll), Mg(ll), Zn(ll), Mn(ll), and Hg(ll). Solid Lines Represent 7.04 uM Metals, Dotted Lines Represent 60.4 uM Metals.








55
concentrations as low as 10"7M (a 100:1 ratio of EBT to thorium) with a variance of one percent, figure 27.

To estimate the catecholamide-thorium(IV) association constant, a solution buffered at pH 9.2 was prepared, containing one equivalent of ThC14 and one equivalent of EBT. The solution showed a pink color, indicative of the thorium-EBT complex. Upon addition of one equivalent of a catecholamide, a color change, from pink to blue, was observed, indicating that the catechol displaced thorium from EBT, according to the equation


2 EBT-Th + Cat EBT + Cat-Th2 pink blue
The resultant absorbance spectrum was identical to that of the uncomplexed EBT dye, table 8. From these data it can be concluded that less than 0.1 percent of the total amount of thorium(IV) in the solution is associated with EBT. As a consequence of this "competition," the relative conditional binding constant of the catecholamide ligands relative to EBT must be on the order of 103 or greater.

Competition Studies with Nitro Derivatives of Catecholamide Liaands

In the previous experiment, the data could not be used for an accurate determination of the Jl-Th conditional metal binding constant because the fraction of thorium bound to EBT could not be measured. If one were to use ligands which shared the metal in such a way that the amount of metal associated with each ligand could be measured, accurate calculations of relative conditional binding constants could be made. If all proton stability constants, metal hydrolysis constants, metal-ligand complex stability constants, and metal-proton-ligand complex stability constants are known for one of the ligands, as well as the stoichiometry of the ligand-metal complex, then the conditional binding constant, defined by the equations








56








0.30 - 650nm
corr.-0.97



0.25




U 0.206.
n 700nm
S0.15- corr.-O.997



540nm
0.10 corr.=0.993




0.05 I I
0.0 0.5 1.0 1.5 2.0 2.5


[Thorium] (pM)





FIGURE 27 Standard Curve for ThCI4 with EBT (10 M).









57





TABLE 8
Color Reactions of Various Ligands with the Thorium-EBT Complex. Conn Coor Reaction Triethanolamine

EDTA

Salicylic acid

2,3-dihydroxybenzoic acid

3,4,3-LICAM + Compound la + Compound lb + Compound id + Compound le + Compound If + Compound 1k +


(+) indicates that the solution turned from pink to blue in 15 seconds or less upon addition of the ligand of interest. One equivalent each of EBT, thorium and ligand was used.








58
xM + nL MxLn

S (MxLn)
and K(L)n , where
(M)x(L)n



(M) the total concentration of metal which is not complexed to the ligand.
(L) = the total concentration ligand which is not complexed to the metal.
(LM) = the total concentration of metal-ligand complex.
x,n = the stoichiometric amount of metal and ligand, respectively, in the complex,

can be calculated for that ligand-metal complex (71). If the relative conditional binding constant, defined as

Kre= K2/ K1 , where

K.1 = the conditional binding constant for the first ligand-metal complex
K2 = the conditional binding constant for the second ligand-metal complex,

is measured, then the conditional binding constant for the second ligand can be calculated, since
K2 = Krel*'K,*.

By repeating this two-ligand-one-metal competition, replacing one ligand with another, one can eventually arrive at the conditional binding constant for a ligand of interest. In an analogous fashion one can perform a one-ligand-two- metal competition experiment to arrive at the conditional stability constant for a particular metal or metals of interest.

In this particular case, it was hypothesized that another catechol would be a suitable candidate for competition with la. The problem was that during the process of compound characterization it was found that the spectral qualities of the H-shaped ligands--ultraviolet, visible, flourescent, etc. under a variety of conditions--did not change significantly enough upon thorium binding to be a useful analytical tool. Also, the changes of one catecholamide ligand upon metal binding could not be distinguished from changes in the other catecholamide. To avoid this complication a series of relative conditional binding constants were measured, utilizing nitrocatecholamides.








59
Besides increasing water solubility, N,N-diethyl(2,3-dihydroxy-5-nitro)-benzamide exhibited another potentially very useful property. During the course of titration, it was noted that the color of the solution containing this compound was pH-dependent. At pH lower than 7, the solution was yellow. At pH higher than 10, the solution was purple. At intermediate pH, the color ranged from yellow to orange to red to purple. We hoped to be able to put this "indicator" property of the nitro derivative to good use.

Since the effect of metal binding upon the UV-visible spectra of a compound is the same as loss of protons, it was our belief that, upon metal binding at the proper pH, the tetranitro derivative of an H-shaped ligand would show the same type of color changes exhibited by the anions of N,N-diethyl (2,3-dihydroxy-5-nitro)benzamide. As explained earlier, our attempts to synthesize such a derivative were unsuccessful. We were able, however, to synthesize a bis nitro derivative of Compound II, and hoped this compound would exhibit the same type of color changes as those seen with the anions of N,N-diethyl-(2,3-dihydroxy-5-nitro)-benzamide Thus, 6 was deprotected to yield N,N-bis(2,3-dihydroxy-5-nitrobenzoyl)-norspermidine (2 ), the idea being that upon metal binding spectral changes will occur in the visible range; in this range the parent catecholamide does not interfere with the absorption of the nitrocatecholamide. Thus if two catecholamide ligands are allowed to compete for one metal in solution, when one of the ligands, e.g. 2Z, contains a nitrocatechol and the other, e.g. N,N-bis(2,3-dihydroxybenzoyl)spermidine (Compound II) or an H-shaped ligand, does not, the the amount of metal which is bound by the nitrocatecholamide can be quantitated by the changes in its absorption spectrum in the visible range.

Bergeron and McGovern (62, 63) have measured all of the equilibrium constants associated with formation of a metal complex with copper and Compound II. Since this type of data is not available for any thorium complex--due to the fact that the thorium hydrolysis constants are not known--we felt that we could use the data generated by McGovern, along with the copper-ammonia complex formation constants (72), to measure the copper-la conditional








60
complex formation constant and eventually the thorium-la conditional complex formation constant.

When copper was added to a constant concentration of the nitrocatecholamide 2Z, the absorbance at 400 nm was linear when plotted against the amount of copper added, figure 28. If we assume that all of the added copper is bound by the ligand, then a linear relationship exists between the analytical amount of copper and the concentration of ligand-copper complex, figure 29. If the same measurement is made with compound II or la present at the same concentration as 2Z. 2Z binds less copper than in the absence of the other catecholamide. By using the standard curve generated in the absence of a second catecholamide, the amount of copper bound to 2Z in the presence of compound II or la can be quantitated, and the percent of metal associated with the nitrocatecholamide can be calculated from the ratio of the slopes of the lines generated with and without a second catecholamide present, figure 29. If one knows the stoichiometric amount of copper in each complex of interest, the concentration of each metal complex can be calculated, leading to the relative conditional stability constant for the two ligands and copper.

In the cases of the Cpd II-copper complex and the 2-copper complex, x=n=1, and the relative conditional binding constant can be expanded to


2 1
. _ (M-L2)(L )
rl (M-L )(L)
where
M-L1 - the equilibrium concentration of the Cpd II-copper complex
M-L2 = the equilibrium concentration of the 2-copper complex
L1 = the equilibrium concentration of Cpd II
L2 = the equilibrium concentration of 2Z.
Since excess ligand is present in all cases, free copper can be ignored, and CM = (M-1) + (M-L2)
CL = (M-L1)+(L1)
CL2 = (M-L2) + (L2).
Since CL1 = CL2 = 33.3 pM, let CL1 = CL2 = CL. Then








61






0.75



0.70



o 0.650 0.60 0 0.55

Control
0.50 Cpd II



0.45
0 5 10 15 20 25 30 35 C(Cu) (pM)





FIGURE 28 Results of Two-Ligand-One-Metal Competition Experiments where 2Z is One Ligand and Copper is the Metal.









62







35


30


25
control slope=1.0 20


S 15Cpd 11
6 slope=.6909


10
la
slope=.0823 5


0
0 7 13 20 27 33


C(Cu) (JM)





FIGURE 29
Plot of the Copper-2Z Complex Versus the Amount of Copper Added. The Slope of the Line
Represents the Fraction of Copper Associated with 2Z at Equilibrium with an Equimolar
Amount of Competing Catecholamide Present.








63


CL = (L1) + (M-L1) CL= (L)+ (M-L).
Rearranging these equations,

(L1) = CL- (M-L1)
(L2) = C - (M-L2) , and (M-L2) [CL- (M-L)]
Kre (ML) "[CL- (M-L2)] If we let a = the fraction of the copper which is associated with L2 and f = the fraction of copper which is not bound to a ligand, then (M-L2) = a CM (M-L1) = (1-f-a) CM Since (1-a) >> f, (M-L') = (l-a) CM. If we let CM = b-CL, then


S a.[l1-(1-ea)b] K rel - (1-a)( 1 -ab)


where
a = fraction of copper bound to L2
b= ChMCL.
For this case, a = 0.6909. When 0.1 equivalents of metal are present; i.e., when b = 0.1,


K-reI = 2.327.


Using McGovem's data for the Cpd lI-copper(ll) complex along with the formation constants for copper(ll) and ammonia--the buffer used for these experiments--the conditional complex formation constant (K pd II-Cu) was calculated to be 1.20 x 1015 M1, and


S2.79 * 1015 M-.
K-Cu Cpd Il-Cu K1 = 2.79 x tM








64


For la and copper, x-2 and n-1. Therefore, the relative conditional binding constant expands to


K _ (M-L)(L ) rel (M-L )(L )(M)


where
(M-L1) = the concentration of 2Z-copper complex (M-L2) - the concentration of la-copper complex
(L1) = the concentration of ?Z (L ) the concentration of la
(M) = the concentration of uncomplexed copper(ll).

Since CL1 = CL2 = 33.3 IM, let CL1 = CL2 = CL. Then the mass balance equations can be written as

CM = (M + (M-L1) + 2 (M2-L2) CL= (L ) + (M-L CL - (L) + (M2-L ). Rearranging these equations into the forms ( - CM - (M-L1) - 2- (M2-L2)
(L) CL-(M-L1)
(L ) = CL - (M2-L2) and substituting,

(M2-L2) [CL- (M-L )] K'el (M-L).[CL(M2LZ) (M) If we let a = the fraction of the copper which is associated with L2 and f = the fraction of copper which is not bound to a ligand, then

(M) = fCM
(M-L') - aCM (M2-L2) = [(l-f-a)/2]CM. Since (I-a) >> f, (M2-L2)= [(l-a)/2]-CM. If we let CM = b'CL, then








65


(1-a)(1-ab)
K =l
rel a[2-( 1-)b](M)

where
a - fraction of copper bound to L2 b- C/CL.
For this case, a = 0.0823. When 0.1 equivalents of metal are present; i.e., when b = 0.1, K re = 5.795/(M). The concentration of free copper in solution can be calculated from the conditional complex formation constant for the 2Z-copper complex as follows


K * (27-Cu) (27-Cu) K - = - (Cu)=
2Z-cu (27)(Cu) (27)-K 27-cu


Since
(2Z-Cu) = a CM
(2 = C2- a-CM 10 CM - aCM K 2Z-Cu = 2.8 x 1015 M-1,


(Cu) = a"e , = 0.0823
[(10-a) K 27Cu (10 - 0.0823)-2.79 x 10 1M


= 2.96 x 10is M.
Therefore,
K'rei = 1.96x 1018 M-1 ,and


K * K* 33m2.
a- =K 2-Cu x re = 5.47 x 10 M-2


Once arriving at K for the la-Cu2 complex, it remains only to measure K" for the la-Th2 complex. In this a problem arises, again due to the fact that spectral changes upon metal binding are small. In addition, copper and thorium cause about the same degree of change. Therefore








66
we needed to distinguish between the two metal complexes or the two free metals. We were able to accomplish this using the complexometric reagent 1-[2-(5-chloropyridyl)azo]-2-naphthol (CI-PAN, figure 23). Synthsized by the method of Shibata et al. (72), Cl-PAN is practically insoluble in water but very soluble in a variety of organic solvents, including ethanol, acetone, chloroform, dioxane, carbon tetrachloride, diethyl ether, and dimethylformamide (73, 74). Depending upon the conditions, CI-PAN has been used to quantitatively detect a large number of metals. The free chromophore exhibits a yellow color, its Xmax occurring at 470 nm. Upon addition of thorium there is an almost imperceptible change to a yellow-orange color, indicating formation of the 4:1 CI-PAN-thorium complex. This change is insufficient for quantitation of thorium. However, when copper is added to this reagent a wine red color is observed ('max = 550 nm), indicating formation of the 2:1 complex (75) of CI-PAN with copper.

When copper was added to 70.9 I.M CI-PAN, the absorbance at 560 nm was linear when plotted versus the analytical concentration of copper, figure 30. When equimolar amounts of copper and thorium were added, the observed spectra are identical to those observed in the absence of thorium. Therefore, with CI-PAN one can quantitatively measure free copper when free thorium is present in the same solution. The thorium is bound by CI-PAN, but does not significantly alter the absorbance of the solution at 560 nm, as long as enough CI-PAN is present to fill all the coordination sites of the metal present. If one assumes that all free copper is bound by CI-PAN then the concentration of added copper is linearly related to the absorbance at 560 nm.

Since CI-PAN is water insoluble and la is insoluble in benzene, an equilibrium was established in which la remains in the basic buffer while CI-PAN remains in benzene. Any metal which is not bound by la will be extracted into the benzene layer by CI-PAN. In this way la will not interfere with the absorbance of CI-PAN, figure 31. To the biphasic system composed of pH 9.2 buffer and benzene was added Cl-PAN, copper, and varied amounts of la (0 to 0.5 equivalents relative to copper). If one plots the absorbance at 560 nm versus the analytical








67






0.35

vithout thorium
0.30 - vith thorium


0.25


%D 0.20
- corr.= 0.9987 ,
* 0.15O.Is0 < 0.10 S corr.= 0.9933
0.05
,'e



0 5 10 15 20 25 C(Cu) (JpM)








FIGURE 30 Standard Curve for Copper(ll)Acetate with CI-PAN (70.93 1IM) in the Absence and Presence
of Thorium(IV)Chloride at the Same Concentration as Copper.









68





red


CI =N yellow








benzene

water


:OH HO.C 0. ,-0

ONH^ N^ NH NH N NH 0, NH N~NH 0-NH ^^ NH
OH HO O 3utOOH HO








FIGURE 31
Equilibrium for CI-PAN Copper Extraction.








69

concentration of la, a linear relationship exists, where more la causes less (CI-PAN)2-Cu to form, resulting in a decrease in the absorbance at 560 nm, figure 32. When an equivalent of thorium is present along with copper, the change of absorbance is decreased, indicating formation of the la-Th2 complex, which results in more extracted copper and less decrease in the absorbance at 560 nm, figure 33. By treatment of the data in a fashion analogous to the two-ligand-one-metal competition experiment, x-2 and n=l, and the relative conditional binding constant expression expands to


a (Th2-la)-(Cu)2 (Cu2- I a) (Th)2 Since CTh - CC = 33.3 11M, let C-h = CCu, CM. Then the mass balance equations can be written as
CM = (Th) + 2 (Th2-la) CM = (Cu) + 2 (Cu2-a) and rearranged into the forms
(Th)= CM - 2 (Th2-la)
(Cu) = CM - 2 (Cu2-.a) If we let a - the fraction of la associated with copper, then (Cu2-la) - aCla (Th2-1a) -(1-a)-C. , and
(Cu) - CM- 2 a-Cl
(Th) - CM-2 (1-a)Cl . If we let ClA = b-CM, then (Cu2-la) -a b-CM (Th2-1a) - (1-a)- b'CM
(Cu) -CM - 2 a- bCM
(Th) - CM - 2 (1-a) bCM , and









70




1.6


1.4


1.2- +

1.0
corr.= -0.9936

0.8


0.6
co.or= -0.9994

0.4


0.2 - vith thorium
. standard curve,
without thorium
0.0 I I ' I I I
0 5 10 15 20 25 30 35 C(la) (JAM)






FIGURE 32
Results of la-Thorium-Copper Competition Experiments.








71



red g ellow-orange


/ - I',i
- "0e~ - 0 h.....
0--C --0 0o---.-T
/ I










CU2+ Th4
+ +


,-Th o a :Cu:
oC __o



NH ^ NH NH N^" NH
-:Th o C--u: o








FIGURE 33
Equilibrim for CI-PAN Copper/Thorium Extraction.








72

K* (1-a)(I-2ab)2

a.( 1 +2ab-2b)2


For this case, a - 0.7414. When 0.1 equivalents of la are present; i.e., when b = 0.1,



K rel = 0.2814, and


K = K c2x K = 1.54 x 1033 M-2.



These results are summarized in table 9.

Pharmacolov
Vehicle Development

It was observed that the catecholamide ligands are very soluble in alchohols and soluble at millimolar concentrations in ethyl acetate and nitrobenzene. Unfortunately, most organic solvents are not acceptible as vehicles for injection. As previously stated, ligand solubility in aqueous media at physiologic pH was slight. In order to solublize our ligands in physiologic buffer, Chremophor RH 40 (BASF) was used. This is a FDA-approved surfactant whose generic name is glycerol polyethylene glycol oxystearate. It is used commercially for solubilization of fat soluble vitamins, essential oils, and some hydrophobic pharmaceutical preparations. The LD50 for mice injected intraperitoneally is >6.4 g/kg.

Compounds were solubilized by maintaining a mixture of ligand and Chremophor at 550C with frequent ultrasonic mixing. Once the mixtures were homogeneous in appearance--after one to two hours of this treatment-phosphate buffered saline at pH 7.4 (PBS) was added until the final solution was 40 percent Chremophor/PBS. These solutions were rather viscous, but suitable for injection.








73






TABLE 9
Results of Conditional Binding Constant Calculations for an H-Shaped Ligand and ThC4.

Conditional Binding
Fraction Bound By Constant for the First Ligand(s) Metal(s) First Competitor Krela Competitor


Cpd ll Cu - - 1.2 x 1015 M-1 2Z/Cpd II Cu 0.6909 2.327 2.79 x 1015 M-1 a/2Z Cu 0.9177 1.96 x 1018 M-1 5.47 x 1033 M-2 la Th/Cu 0.2586 0.2814 1.54 x 1033 M-2 aCalculated for one equivalent of each competitor and 0.1 equivalent of the substance for which they are competing, in 1.00 F ammonia buffer, pH 9.2.








74

Biological Toxicity

Any therapeutic device must be nontoxic, a situation not realized with other octacoordinate actinide chelators. In keeping with this concern all octacoordinate H-shaped ligands were tested for acute toxicity. The compounds were prepared by dissolving them in a mixture of Chremophor and PBS as previously described. A/J male mice (five mice per dose) weighing 20-25 grams were given intraperitoneal injections of these samples up to and including doses as high as 1000 jg/kg. It should be noted that this concentration is equivalent to administering 70 grams of drug to the average human, much more than the anticipated therapeutic dosage. The compounds were administered to five mice at each concentration. All mice seemed healthy and active and remained that way until the end of the 72 hour experiment, at which time the animals were sacrificed. Vehicle-injected mice exhibited no symptoms of toxicity.

Clearance Studies

Whether our H-shaped catecholamide chelators can remove actinides from animals with internal contamination is a question of paramount importance. To effectively perform this role the compounds must be able to selectively sequester actinides from body tissues, proteins, and fluids, which may themselves bind the metals. It is no mean task for the chelator to selectively bind an actinide metal in a millieu replete with various cations and endogenous sequestering agents. Next, the chelate must be able to utilize an excretion route. This was a problem in the case of Raymond's CYCAM series, where the chelators bound plutonium but were not excreted

(44).

The following experiments represent preliminary attempts to ascertain the ability of catecholamide chelators to promote actinide excretion. No attempt has been made to optimize dosing. The intention was to test the concept of using H-shaped octacoordinate catecholamide chelators to remove actinides in the same way that hexacoordinate catecholamide chelators have been shown to eliminate iron.








75
Urinarv clearance

The ability of DTPA, Compound II, and H-shaped ligands to stimulate clearance of thorium in the urine was evaluated. Thorium (1 nmole in 0.5 mL in 0.09 percent citrate buffer) was first administered to rats by intraperitoneal injection. One hour later a chelating agent (200 pmoles/kg) was administered intraperitoneally in the 40 percent Cremophor/PBS vehicle. Animals were placed in metabolic cages with food and water ad libitum. Urine samples were collected at six hour intervals and assayed for thorium by scintillation counting of the metal's alpha radiation. The data indicate that DTPA is effective at stimulating thorium excretion in the urine, and that this process is very rapid, figure 34. The fact that most of the DTPA-stimulated clearance takes place within the first six hours is consistent with the short half-life (20 minutes) of this drug in the body. Although the catecholamide chelators tested in this way (lc- and M stimulated thorium output in the urine to a very small extent, if we make the assumption that additional methylenes increase lipophilicity, a structure-activity relationship exists between the lipophilicity of the ligand and the percent urinary thorium excretion. As one would expect, the analog presumed to be the least lipophilic, ld, stimulated the most urinary excretion while le was the least effective.

To determine if a connection exists between this result and the lipophilicity or degree of ionization of the ligand, the partition coefficient for two ligands-le and 1t--were measured in octanol/phosphate buffered saline, pH 7.4 (PBS). As expected, la was found to be more lipophilic than It, its partition coefficient being 1.63 times greater, table 10. Fecal clearance

The inability of our catecholamide actinide chelators to stimulate thorium excretion via the urine came as no great surprise. It has been shown that highly lipophilic chelators are excreted as their plutonium complexes almost exclusively in the feces (76). Our ligands, of course, fall into this lipophilic class. Furthermore, experience with hexacoordinate catecholamide iron chelators has shown that these lipophilic compounds are also excreted mainly in the feces (77).








76
10
Cpd 11 (3)
to n (4)
SId (5)
l (4)
8 - f (4)
im (4)
7 - CONTROL (9)





S


4





2

1



0
0 24 48 Time (hrs.)





FIGURE 34 Effects of Catecholamide Chelators on the Urinary Excretion of Thorium in Rats Over 48
Hours, at a Dosage of 200 pmoles/kg each. Chelators were Administered Intraperitoneally
One HourAfter Intraperitoneal Injection of 230Thorium. Numbers in Parentheses Represent the Number of Rats Tested.







77






TABLE 10
Results of Partition Coefficient Measurements for LUgands land f.
mamx c320 nmrn [ligad]octnoI
Compound octanol PBS octanol PBS [liOgandIps


.le 313nm 317nm 12,800 13,320 39.96 11 313nm 319nm 12,840 13,040 24.50








78
Considering the structural similarities of our catecholamide actinide chelators to these iron chelators, and their greater molecular weight, one would predict that these compounds would be excreted by the fecal route as well.

The potential health hazard of 230thorium, the isotope used for these studies, is well recognized; precautions and procedures for its safe use have been implemented to avoid contamination of individuals and working areas. Because of the necessary restrictions in handling thorium-containing samples, a safe technique for measuring the thorium content of feces has not been developed. Additionally, it is difficult to gather kinetic data from fecal samples, due to irregularity of sample excretion.

Biliarv clearance

To avoid the handling difficulties associated with measuring actinide output in feces, we have developed a series of experiments to collect and assay biliary tisues and fluids before and after chelation therapy.

Gall bladder excision. As an initial experiment, 36 mice were administered thorium (50,000 dpm) intraperitoneally, followed one hour later by an H-shaped ligand (200 irnoles/kg in 0.5 mL 40 percent Chremophor/PBS). At three hour intervals, 3 mice were sacrificed, and their gall bladders were removed and assayed for thorium content. Since the gall bladder acts only as a bile storage site, the amount of thorium in the gall bladder would depend on the diet composition, the amount consumed, and when it was consumed. This experiment will therefore only qualitatively indicate whether or not biliary excretion of thorium can be stimulated by H-shaped ligands. Mice that recieved an H-shaped ligand demonstrated approximately a three-fold increase in the thorium content of their gall bladders, figure 35. This data suggests that our chelators are excreted by the biliary route, and are effective actinide chelators.

Bile duct cannulation. In order to get quantitative excretion data and accurate kinetic data it was decided to collect bile for several days. This was accomplished by placing an indwelling cannula into the bile duct of adult male Sprague-Dawley rats, allowing for continuous collection








79



1000

900800E 700600= 500E 400a
: 300-
200
0]0 If-treated 100 - animals
SControl animals I I , I I I I
0 5 10 15 20 25 30 35 40 45 50 Time of Sacrifice (hours)





FIGURE 35 Effect of 11 on the Thorium Content of Mouse Gall Bladders Over 48 Hours, at a Dosage of
200 imoles/kg. Chelators Were Administered Intraperitoneally One Hour After Intraperitoneal Injection of 230Thorium (50,000 cpm).








80

of bile. By using the rat, one does not face the problems associated with the gall bladder since this organ is absent. The cannula emerges from the incision and, using a skin-tunneling needle, is directed under the skin to behind the rat's neck. The emerging cannula is directed through a torque-transmitting spring tether to a fluid swivel located above the metabolic cage. The animal is able to move about freely in the cage while urine and bile were collected. Urine was collected for 24 hour intervals while bile was collected for three hour intervals. Although more involved, this technique allows us to obtain very reproducible data for metal excretion. When the animal began to recover from anethesia and good bile flow had been established, thorium (1 nmole) was administered intraperitoneally. One hour later a chelator was given also intraperitoneally (100 or 200 iimoles/kg). Surgical techniques were performed by Dr. N. L. Scarborough, Dr. S. A. Prudencio, and Ms. Kady Crist. I assisted in sample preparation for scintillation counting.

To study the effect of altering the spermidine portion of H-shaped ligands, . la, and 1t were tested. It was found that the assymmetric SPD backbone is much more effective at stimulating thorium elimination via the bile than either symmetric backbone, figure 36. Since both addition to and removal from the spermidine backbone of a single methylene group causes a significant decrease in biliary thorium removal, it is reasonable to assume that any further alterations in a, b, c, and d (figure 5) will result in a further decrease in thorium excretion.

To examine the role of the connecting diacid, _i, If, and 1m were tested. The glutaryl moiety appears to be better than either the succinyl or adipoyl diacids. Therefore, for the same reasons, no other diacids were tested. Total thorium removal is shown in figure 37. Combination chelator theraov

The data show that H-shaped ligands stimulate thorium excretion almost exclusively in the bile whereas DTPA leads to urinary clearance of the actinide. Based on these findings we felt that perhaps if both drugs were administered in combination, the total actinide clearance would increase. To test this concept, the bile duct cannulation procedure was employed. In this experiment, rats were administered thorium intraperitoneally (1 nmole) followed by DTPA (100








81
40
SO0 Cpd 11 (3)

35- 1o (4)
* Id (5) A 1 (4)
30 - If (5)
M 1m (3)
* Control (8)
25


a 20


S 15
I










0 6 12 18 24 30 36 42 48 Time (hrs.)





FIGURE 36
Effects of Catecholamide Chelators on the Biliary Excretion of Thorium in Rats Over 48 Hours, at a Dosage of 200 rnoles/kg each. Chelators were Administered Intraperitoneally One Hour
After Intraperitoneal Injection of 230Thorium. Numbers in Parentheses Represent the Number of Rats Tested.








82
35
SCpd II (3)
-N Ic (4)
30- Id (5)
Sle (4)
I 1f (4)
25- Im (3)
0 Control (8)

20








I I




0 24 48
Time (hours)






FIGURE 37
Effects of Catecholamide Chelators on the Total Excretion of Thorium in Rats Over 48 Hours, at a Dosage of 200 rnoles/kg each. Chelators were Administered Intraperitoneally One Hour
After Intraperitoneal Injection of 230Thorium. Numbers in Parentheses Represent the Number of Rats Tested.








83

gpmoles per kg) and/or It (100 gmoles/kg), also intraperitoneally. Urine and bile were collected and assayed for thorium content, as previously described.

Urinary clearance. In the urine DTPA was able to remove 28.0 percent of the injected dose of thorium when administered alone at this lower dose, figure 38. When given in combination with ., only 22.3 percent of the injected dose was found in the urine, a 20.4 percent reduction.

Biliary clearance. In the bile a similar situation was observed, figure 39. If alone removed 23.6 percent of the injected dose of metal, but only 17.8 percent was removed in the bile by the combination of chelators, a 24.6 percent reduction.

Total clearance. The data may indicate that I does indeed access different body compartments than DTPA, as evidenced by increased thorium excretion. Also, it appears that there are some compartments that are frequented by both drugs, as reflected by the fact that the combination of drugs removes less than an additive amount of metal. If DTPA and were independent, then the metal removed by DTPA in the urine when administered by itself added to the metal removed by It in the bile when administered by itself would be approximately equal to the total output when the drugs are administered in combination. Since the actual output of metal is 21.6 percent less than the additive amount, there is some overlap in the volumes of distribution for the two chelators. That is, there are some body compartments which are accessible to both chelators, some which are accessible to only one chelator, and some which are not accessible to either chelator. When administered in combination, in the compartments accessible to both chelators a portion of the available thorium is bound by each chelator. for example, some of the thorium which would have been bound by DTPA and eliminated in the urine is bound by 1, causing an apparent decrease in the effectiveness of DTPA. Although combination therapy decreased the appparent effectiveness of each chelator, the total excretion did increase, figure 40.







84







30
[ DTPA
1f
25- F DTPA and If

a

at 200 15



S 10



L) 5




0 24 48 Time (hrs.)




FIGURE 38 Effects of It or DTPA on the Urinary Excretion of Thorium in Rats Over 48 Hours, at a Dosage
of 100 nmoles/kg each. Chelators were Administered Intraperitoneally One Hour After Intraperitoneal Injection of 230Thorium. Numbers in Parentheses Represent the Number of Rats Tested.








85





25
-- DTPA


M DTPA and If
20








, 10



1 5





0 6 12 18 24 30 36 42 48 Time (hrs.)




FIGURE 39
Effects of If or DTPA on the Biliary Excretion of Thorium in Rats Over 48 Hours, at a Dosage of
100 Wpmoles/kg each. Chelators were Administered Intraperitoneally One Hour After Intraperitoneal Injection of 23Thorium. Numbers in Parentheses Represent the Number of Rats Tested.








86



60
[ DTPA
1] if
! 50 - DTPAand If
experimental
* DTPA and If S40 theoretical
S40
L

g. 30










01 I

0 24 48 Time (hrs.)




FIGURE 40 Effects of 11 or DTPA on the Total Excretion of Thorium in Rats Over 48 Hours, at a Dosage of
100 .noles/kg each. Chelators were Administered Intraperitoneally One Hour After Intraperitoneal Injection of 230Thorium. Numbers in Parentheses Represent the Number of Rats Tested.















CHAPTER III

ELUCIDATION OF THE SOLUTION STRUCTURE OF POLYAMINES
IN RELATION TO THE MECHANISM OF CELLULAR UPTAKE Background

Bergeron, Porter and Stolowich have provided a great deal of information regarding the structural requirements, preferences, and limitations associated with substrates of the polyamine transport apparatus (7, 8, 9, 10, 78, 79). Backbone Variations

Eleven structural analogs of spermidine and putrescine were synthesized and studied both for their ability to inhibit [3H-spermidine uptake by the polyamine transport apparatus of L1210 cells and/or their ability to stimulate growth in vitro in polyamine-depleted L1 210 cancer cells, table 11. One should note that the analogs used in this study differ from the naturallyoccurring polyamines only in the number of methylene units separating the various amine functionalities. Uptake inhibition was found to be competitive in all cases, homospermidine being the most competitive polyamine. Any further alterations in the number of methylene units in the polyamine backbone resulted in diminished competitive inhibition. Replacement of the secondary nitrogen of spermidine by a methylene unit produced a molecule which was an effective inhibitor of [3H]-spermidine uptake. In growth stimulation experiments, however, little effect was caused by these analogs. This study suggests that the secondary amine is essential for polyamine functions while its importance is less pronounced for uptake. N4-Substitution

Alkylation. N4-alkylated spermidine derivatives were synthesized to determine both how N4-alkylation would affect uptake characteristics and how large of an alkyl substituent would be tolerated by the polyamine transport apparatus, table 12. It was found that the spermidine

87








88




TABLE 11
Inhibition of [3H-Spermidine Uptake into L1210 Cells by Polyamines or Their Homologs.

.3HI-Soermidine Untake Inhibitionb
Homolog Kia Picomoles per Percent of luM 10'icemrn Contrdl
none - 56.1 100 DA3C >500 54.0 96 DA4(putrescine) 171.3 44.6 80 DA5 459.0 54.0 96 DA6 63.2 40.1 71 DA7 18.2 23.0 41 DA8 22.1 25.2 45 3TA3d 8.4 16.1 29 4TA4 3.5 7.3 13 3TA5 12.3 19.8 35 3TA6 13.1 19.6 35 3TA7 13.0 20.0 36 3TA8 7.8 13.5 24 Spermine 9.1 17.1 30

Reference 79.
aPrewarmed L1210 cells(5x106) were incubated for 20 min in 1 mL of
RPMI-1640 media containing 2 percent Hepes-Mops and 0.2, 0.5, 1.0, 2.0, 5.0, or 10.0 pM [3H]-spermidine and 100 pM homolog. Uptake data were fitted by computer for competitive inhibition; the Michaelis constant for spermidine uptake was 2.0 pM, and the maximum velocity of the reaction was 117
pmole/min per milligram of protein.
bCells were incubated for 20 minutes at 370C with 10 grM [3H]-spermidine
plusl00 gM putrescine or spermidine homolog.
CThe abbreviation for putrescine homologs having the general structure
NH2(CH2)nNH2 is DAn (for diamine) where n=3 to 8.
dThe abbreviation for spermidine homologs having the general structure
NH2(CH2)nNH(CH2)n,NH2 is nTA n, (for triamine) where n is 3 or 4 and n' is 3 to
8.








89





TABLE 12
Effects of N4-Alkylated Spermidines on [3H]-Spermidine Transport into Ascites L1210 Leukemia Cells.

Substituent Relaive Utakea K fuvb

-H 100

-CH3 17 3.4 -CH2-CH3 14 3.1 -(CH2)5-CH3 69 34 -CH2-C6H5 67 39

Reference 79.
aCells were incubated for 20 minutes at 370C with 10 IM [3H]-spermidine plus100 rM analog.
bPrewarmed L1210 cells(5x106) were incubated for 20 min in 1 mL of
RPMI-1640 media containing 2 percent Hepes-Mops and 0.2, 0.5, 1.0, 2.0,
5.0, or 10.0 uM [3H]-spermidine and 10 or 100 uM analog.








90
molecule could be alkylated rather extensively at the N4-position and still participate effectively in experiments related to uptake, including inhibition of [3H]-spermidine uptake, prevention of MGBG-induced cytotoxicity, and Intracellular detection by HPLC. MGBG, methylglyoxal-bisguanylhydrazone, is an inhibitor of polyamine biosynthesis.

AGylation. N4-acylation had more pronounced effects upon uptake than did alkylation. A series of N4-acyl spermidine derivatives and their alkyl counterparts--for example, ethyl and acetyl--were prepared and evaluated for their ability to compete with [3H]-spermidine for cellular uptake, table 13. It was found that both series of derivatives could behave as polyamine transport apparatus substrates, but there was an obvious preference for the N4-alkylated derivatives. In addition, the secondary nitrgen could be extensively modified without a significant decrease in inhibitory effect, as evidenced by the uptake of derivatives such as N4-hexyl and N4-benzyl spermidine. This is consistent with the conclusions of the previous study, where the primary amines appeared to be more critical for uptake. N1.N8-Bis Substitution

A series of N1,NS-bis acyl derivatives were prepared and tested in a similar manner, table 14. Of the derivatives tested--N1,N8-bis-methyl and N1,NB-bis-formyl spermidine were not tested--N1,N8-bis ethyl spermidine was the only molecule capable of competing effectively with spermidine for uptake; any modification to this derivative rendered the molecule incapable of behaving as a polyamine transport apparatus substrate.


The Role of Protonation State: Potentiometric Measurements
In an attempt to more accurately define the role of charge in polyamine uptake, the protonation state of N4-benzylspermidine, as well as the nor- and homospermidine analogs, was studied. To achieve this the pKa of each nitrogen of the three analogs, and homospermidine itself was determined potentiometrically. The benzyl analogs were chosen because considerable uptake data has been accumulated on these compounds.








91





TABLE 13
Comparison Between the Effects of N4-Alkylated and N4-Acylated Spermidines on
[3H]-Spermidine Transport into Ascites L1210 Leukemia Cells.

Substituent Relative Uakea 1K uMb

-H 100 -CH2-CH3 14 3.1 -(C=O)-CH3 81 115 -CH2-(CH2)4-CH3 69 34

-(C=O)-(CH2)4-CH3 88 135

-CH2-C6H5 67 39

-(C=O)-C6H5 92 500

Reference 79.
aCells were incubated for 20 minutes at 370C with 10 gM [3H]-spermidine plus100 gM analog.
bPrewarmed L1210 cells(5x106) were incubated for 20 min in 1 mL of RPMI-1640 media containing 2 percent Hepes-Mops and 0.2, 0.5, 1.0, 2.0,
5.0, or 10.0 gM [3H]-spermidine and 10 or100 uM analog.








92







TABLE 14
Effects of Terminally Modified Spermidines on [3H]-Spermidine Uptake into L1210 Leukemia Cells.

Sustltuent RelsMve ULakeU Ki uh

-H 100

-(C=O)-CH3 91 508

-CH2-CH3 69 62

-(C=O)-CH2CH3 92 550 -CH2-CH2CH3 80 117 -tBOC 91 521 -2,3-Dimethoxybenzoyl 91 256 Reference 79.
aCells were incubated for 20 minutes at 3700C with 10 gM [3H]-spermidine plus100 RM analog.
bPrewarmed L1210 cells(5x106) were incubated for 20 min in 1 mL of RPMI-1640 media containing 2 percent Hepes-Mops and 0.2, 0.5, 1.0, 2.0,
5.0, or 10.0 .M (3HI-spermidine and 10 or100 pM analog.








93


When an aqueous titration was completed, a set of data consisting of pH measurements versus volume of titrant added was generated. The computer program PHFIT (D. Leussing, private communication) was used to analyze this data. PHFIT handles as many as four independent species and fifteen associated species. The program first calculates, by way of a standard Newton-Raphson iteration, the distribution of species at each data point based on the total concentration of species at that point and the estimated pKa provided for each specie. Thus, an initial theoretical titration curve is obtained; assuming all significant equilibria have been considered, the difference between this calculated curve and the observed data is then minimized by further refinement of the initial estimates. To calculate the acid dissociation constants, one must provide estimates for these constants, the analytical concentrations of all independent species involved in the chemical equilibria, the autodissociation constant (Kw) for water under the experimental conditions, and the activity coefficient (y) of H+ under the experimental conditions. Also, certain assumptions are made--that the titrations performed under conditions of constant temperature and ionic strength.

By using a water-jacketed cell in conjunction with a constant temperature water bath, the condition of constant temperature (250C) is satisfied. To insure the condition of constant ionic strength, an excess of a strong electrolyte, potassium chloride, is added to both the solution of interest and the titrant, so that changes in the ionic state of the acid and titrant are insignificant in comparison to the total ion population. In practice this means that the ligand concentration (approximately 2 x 103 M) is much less than the ionic strength of the solution (0.1M KCI). For an aqueous solution at 250C with an ionic strength of 0.1M, Kw and yare 10-13.787 and 0.78 respectively (80, 81). These two terms are related by the expression



K,,=[H+][OH-]= a 2








94
where aH and aOH are the activity of H+ and OH" respectively. The results of the pKa studies are presented in table 15 along with other selected polyamine pKa's.

Knowing the pKa's, it is possible to calculate the concentration of the mono-, di-, and trication of a polyamine in solution at any given pH, using the equations O= KK2K3
[H+]3+ Kt[H+]2+ KiK2[H+]+ KIK2K3


K1K2[H+]
(X(I)= Kf1<[H+]
[H+]3+ K1[H*]2+ KIK2[H ]+ KtK2K3


KI [H+]2
[H ]3+ KI[H*]2+ KtK2[H*]+ K1K2K3


((3)= [H3
[H]3.+ KIlH*]2+ K1K2 [H*]+ KIK2K3 where a(i) is the fraction of polyamine existing as the i+ cation. Recall that the order of uptake inhibition is homospermidine > spermidine > norspermidine. It is clear that there are substantial differences in the relative concentrations of polycations at a particular pH. For example, at pH 7.4 only 67.4 percent of norspermidine is in the form of the trication while 89.9 percent of spermidine is in this form and 97.1 percent of homospermidine is triprotonated, figure 41. Similar trends are seen with the benzyl compounds, wherel 11.4 percent, 34.3 percent, and 69.1 percent, respectively, exist as the trication.

The values for a(3) seem to correlate well with C. W. Porter's in vitro measurements of inhibition constants (7, 8). By multiplying the actual concentration of polyamine by the fraction of the polyamine which exists as the trication at pH 7.4 one can determine the "effective" concentration of polyamine. When one recalculates the Ki's of various polyamines based on these corrected concentrations, the large differences in affinity for the transport apparatus are greatly diminished, table 16. However, the role of a 3+ cation is unclear in view of the fact that




Full Text
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 EIJBV554F_ETGS09 INGEST_TIME 2015-03-27T18:42:32Z PACKAGE AA00029781_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

HEXAHYDROPYRIMIDINES AS MASKED SPERMIDINE VECTORS IN DRUG DELIVERY AND AS REAGENTS IN THE SYNTHESIS OF H-SHAPED OCTACOORDINATE ACTINIDE LIGANDS FOR HUMAN AND ENVIRONMENTAL DECONTAMINATION By HOWARD WAYNE SELIGSOHN iU A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULLFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1987

PAGE 2

Through their love, patience, understanding, inspiration, ; t " j s"'' constant support. my mother and father have contributed immeasureabiy to my career. In recognition of this, I proudly dedicate this dissertation to them.

PAGE 3

ACKNOWLEDGEMENTS I would like to extend my sincere gratitude to my major advisor, Dr. Raymond J. Bergeron, for all his support, patience and understanding ttirougtiout my studies. I also thank my committee, Drs. Richard Strieff, Kenneth Sloan, Merle Battiste, and Margaret James for their contribution to this work. To Dr. James Navratil and the Rocky Fiats research group I extend my thanks for the plutonium data. Desen/ing special thanks are Drs. Nelson Scarborough and Steven Poidencio and Ms. Kady Crist for their surgical expertise, Diana Tukalo for her synthetic efforts, and Micheal Ingeno for all of his help with my in vitro experiments. Lastly, thanks to all of my good friends and family who have helped me through the difficult times. I could not have made it without your support.

PAGE 4

TABLE OF CONTENTS ACKNOWLEDGEMENTS nj ABSTRACT. vi CHAPTERS I, INTRODUCTION 1 II. H-SHAPED OCTACOORDINATE ACTINIDE LIGANDS FOR HUMAN AND ENVIRONMENTAL DECONTAMINATION 13 Background 13 Synthesis 15 Catecholamide Uganda 15 Parent catecholamides 15 Nitro derivatives 18 Resin-Bound Catectiolamide Ligands 27 The catechDlamide 27 The resin 27 Ligand Stoichiometcy: Job's Plots 35 Ligand Precipitation Techniques: Removal of Actinides from Water 38 Insoluble Chelators: Removal of Actinides from Water and Blood Plasma 45 XAD-4-Sorted Ligands 45 Covalently-Bound Ligands 49 IRP-64-bound catecholamides 49 CH-Sepharose-4B-bound catecholamides 49 Equilibrium Solution Chemistry: Thermodynamic Binding Constant Measurements 52 Eriochra me Black T Competition 52 Competition Studies with Nitro Derivatives of Catecholamide Ligands 55 Pharmacology 72 Vehicle Development 72 Biological Toxicity 74 Clearance Studies 74 Urinary clearance 75 Fecal clearance 75 Biliaiy clearance 78 Combination chelator therapy 80

PAGE 5

III. ELUCIDATION OF THE SOLUTION STRUCTURE OF POLYAMINES IN RELATION TO THE MECHANISM OF CELLULAR UPTAKE 87 Background 87 Backbone Variatbns 87 N*-Substitutlon 87 N^N^-Bis Substitution 90 The Role of Protonation State: Potentiometric Measurements 90 The Role of Hydnogen Bonding 98 The Reasoning 98 The Evkjence: Hexahydropyrimidines 99 IV. EXPERIMENTAL DETAILS 107 Synthetic Procedures 107 Job's Pkjts 122 Precipitation Techniques 122 Eriochrome Black T Competition 123 Competition with Nitrocatecholamides 124 Resin Experiments 125 Biological Evaluation 126 Potentiometric Measurements 127 Stability Studies 127 Inhibition of Spermidine Uptake 127 ICgg Measurennents 128 SUMMARY AND CONCLUSION 129 REFERENCES 130 BIOGRAPHICAL SKETCH I34

PAGE 6

r ' Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fullfillment of the Requirements for the Degree of Doctor of Philosophy HEXAHYDROPYRIMIDINES AS MASKED SPERMIDINE VECTORS IN DRUG DELIVERY AND AS REAGENTS IN THE SYNTHESIS OF H-SHAPED OCTACOORDINATE ACTINIDE LIGANDS FOR HUMAN AND ENVIRONMENTAL DECONTAMINATION By HOWARD WAYNE SELIGSOHN August 1987 Chairman: Dr. Raymond J. Bergeron Major Department: Medicinal Chemistry The cellular uptake of several polyamines is evaluated in terms of their intermolecular hydrogen bonding and their charge at physiological pH. N-(4-Aminobutyl)hexahydropyrimidine and N-{3-aminopropyl)hexahydropyrimidine are shown to compete with spermidine for uptake by L1210 cells. This observation Is in keeping with the idea that spermidine may adopt a hydrogen-bonded cyclic structure in the course of transport. Furthemiore, the differences in the ability of spemiidine, homospermidine, and norspermidine to utilize the spermidine uptake apparatus of L1210 cells is related to the protonation state of the amines. These states are calculated for each triamine from measured pK^ data. The hexahydropyrimidines used in the polyamine uptake sudies were next used in the synthesis of octacoordinate catecholamide H-shaped ligands. These chelators were tested for their ability to remove actinides from aqueous solution, human plasma, and rats. H-Shaped ligands removed more than 99.9 percent of the thorium from a 0.32 pM aqueous solution. In addition, approximately 30 percent of the

PAGE 7

': . CHAPTER! INTRODUCTION In recent years, scientists have become Increasingly aware that polyamines play an Important role in cellular metabolic processes. Considerably increased levels of ornithine decarboxylase, the rate-limiting enzyme in polyamine biosynthesis, is an apparent "priming" event for growth in a variety of systems (1). Many neoplasias have been shown to concentrate polyamines intracellularly (2, 3, 4). Polyamine biosynthesis and intracellular accumulation are intimately associated with rapid cell proliferation. The increase in Intracellular spermidine has been shown to correlate well with the increase in intracellular RNA (5). Elevated polyamine levels in urine have been detected in tumor-bearing patients (6). Bergeron and Porter have demonstrated that polyamine-deficient L1210 cells divide much more slowly than control cells (7, 8, 9, 10). Thus it appears that an intracellular polyamine demand, polyamine biosynthesis and intracellular accumulation are intimately associated with rapid cell proliferation. Since neoplastic tissues exhibit a high polyamine requirement when compared to normally differentiated tissues, the idea of using polyamines as vectors for delivering antineoplastics is an attractive one. If one were to design a polyamine derivative which possessed antineoplastic properties, and could also utilize the polyamine transport apparatus for cellular uptake, drug selectivity may be improved. Such a derivative could, of course, be taken up into normal tissue as well as rapidly dividing tissue. However, since neoplastic tissues exhibit high intracellular polyamine demand and aocummulation relative to normal tissues, selective toxicity should be observed. The design of polyamine antiproliferatives must be based upon a careful consideration of the polyamine transport apparatus. Implicit in this is an understanding of the solution conformations attained by substrates of the polyamine transport apparatus. With these 1

PAGE 8

2 considerations in mind, several experiments were designed to ascertain the structural txjundaries for substrates of the polyamine transport apparatus. More exactly, we wanted to know the degree to which a polyamine could be derivatized without significantly affecting uptake by the polyamine transport apparatus. Previous studies led us to postulate the importance of an intramolecularly hydrogen bonded structure (11, 12). To test this hypothesis, hexahydropyrimidine derivatives were prepared (13, 14) and found to be effective inhibitors of spermidine uptake. These hexahydropyrimidines, as well as having interesting biological activity, turned out to be very useful as synthetic reagents, leading to a variety of actinide chelating agents which are based on the structures of naturally-occurring siderophores possessing a polyamine backbone. Actinide contamination in humans does not represent a major health hazard because of the federal and self-imposed safety precautions adopted by those in the nuclear industry. Still, the potential for contamination exists and incidents involving human exposure and contamination have occurred (15, 16, 17). Environmental contamination is more common than Internal human contamination. In a laboratory situation, when contamination occurs, trained personnel are on hand to contain the hazard and treat the problem as is necessary. When contamination occurs outdoors, the problem is not easily contained and is, therefore, more serious. Not only can water supplies, soil, vegetation, and livestock become contaminated, but curious individuals-simply by their presence-may unknowingly contaminate themselves at the site of an accident as well. One can imagine how easily a localized contamination could spread, since the danger cannot be seen. Environmental and external human decontamination ultimately generate a large volume of aqueous solution which contains very low levels of hazardous actinide. This presents a problem in terms of disposal. At the present time most of this water is buried in uninhabited regions of the country, but this is only a temporary solution. Eventually, the integrity of the storage containers may fail, leading to new contamination.

PAGE 9

3 If, on the other hand, this water could be exposed to an actinide-specific ligand and the complex removed by some physical means, the water could be decontaminated. More specifically, if a ligand were bound to a polymer backlxsne, a resin with a high affinity for actinides would be the result. This could be used to concentrate the contamination onto a small amount of solid resin. In this way, the complex would be converted into a form from which the metal could be recovered-by pH manipulation, ashing, etc. A resin-bound ligand would be practical for many other aspects of decontamination as well. For instance, this type of ligand could be very useful as a tool for dialysis. An insoluble ligand could also be used for prevention of absorption of ingested actinides. Environmental contamination can lead to both internal and external human contamination (due to our irresistible urge to touch things) and internal human contamination. Since actinides can invade the body via several routes, such as the lungs (18), Gl tract (19) or open wounds (20), the degree to which a particular organ or body compartment is contaminated will depend on the site of entry (21, 22). For example, lung contamination is observed in cases of actinde inhalation. Once absorbed, actinide deposition initially occurs in many tissues of the body: liver, kidneys, muscle tissue, lungs, heart, testes, and bone (23). With time, the metal continually migrates to the liver and bone, while other soft tissue levels decrease. Once in these sites the metal is much more difficult to remove. Clearly, the sooner a contaminated individual can receive treatment, the easier the decontamination process. No adequate therapeutic treatment exists for persons internally contaminated with actinides. Contamination of humans is generally external and most often the metal is removed long before absorption can occur, minimizing the seriousness of the problem. However internal contamination does occur and is more serious, owing to the limited therapeutic options. The treatment of choice is currently diethylene triamine pentaacetic acid (DTPA) chelation therapy, figure 1 , but DTPA has several problems associated with it. In addition to the fact that the drug removes an insufficient quantity of actinide (24, 25, 26), DTPA has toxicity problems.

PAGE 10

HOOC^ N' HOOC' IN r COOH 'N' ^COOH 'COOH FIGURE 1 DTPA (Diethylene Triamine Pentaacetic Acid).

PAGE 11

most of which are associated with its ability to sequester and remove zinc from the Ixjdy (27, 28). Although the latter difficulty has been overcome by administering the drug as its zinc complex (29), the former problem remains-DTPA alone simply cannot remove an adequate amount of actinide. For example, when DTPA was administered to rats intravenously over a three hour period 24 hours after intramuscular injection of plutonium nitrate or citrate, a dose of 120 (imoles/kg was necessary to achieve excretion of only 20 percent of the injected dose, relative to control animals (25). It is unlikely that this problem can be solved even by modification of administrative routes; e.g., continuous infusion (30). Because of its highly polar zwitterionic nature, DTPA is poorly absorbed from the bloodstream and is rapidly excreted in the urine (N. L. Scartiorough, private communication), implying that DTPA's inadequacy is a result of its volume of distribution (31, 32). To obviate the shortcomings of DTPA, more lipophilic catecholamide chelators were prepared (33, 34, 35). Some of the most stable metal complexes known are formed between catecholamide ligands-siderophores--and iron(lll). For example, the parabacfin-iron(lll) formation constant is 10 M'"^ (K. N. Raymond, private communication), parabactin being a hexacoordinate catecholamide ligand with a spermidine backbone. Plutonium(IV) and iron(lll) are similar with respect to the chemical properties responsible for complex stability; i.e., high charge, small ionic size, and high acidity. For example, Pu(IV) has a charge-to-radius ratio of 444 e/^m as compared to 460 e/jim for Fe(lll) (36). Consequently the coordination chemistry of the actinide metals and iron is very similar. This similarity extends to their biological properties as well. For example, both metals are bound tightly to transferrin (37, 38, 39) and ferritin (40) and associate with the trabecular bone. Due to Plutonium's larger ionic radius, it forms more stable complexes with octadentate iigands, in contrast to Fe(lil), which forms very stable chelates with hexacoordinate ligands, figure 2. Based on the above similarity, researchers have used their experience with iron-specific hexacoordinate chelators to design of octacoordinate actlnide-speciflo chelators: the design of

PAGE 12

OH OH NH-. a. -OH HO Parabactin OH OH OH OH OH OH ^. ^o 1 ' I ' \, i^'J » -v Vibrlobactin OH OH NH CH< \h2 O oh m ""i ho ho Enterobactin FIGURE 2 Naturally-Occurring Sideropho res.

PAGE 13

7 actinide chelators is based on the assumption that octacoordinate iigands represent the optimum situation for binding actinides, and that actinides will form the best chelate with the catechol moiety. From this reasoning emerged a variety of effective cyclic and linear octacoordinate catecholamide Iigands, referred to as CYCAM (cyclic £atecholamide)and LICAM (Inearcatechoiamide), figure 3 (33, 34, 41), These compounds have been shown to stimulate actinide excretion ( 21, 25, 30, 32, 33, 34, 42, 43, 44). A derivative of LICAM in which each catechol has a 4-cart)oxyl group attatched, LICAMC, was the most promising of these, figure 4. It was found that UCAIutC did access different body compartments than DTPA (42). For example, DTPA removed much less of the actinide associated with the sl
PAGE 14

""^wr HO HO X? 0=C ^^(CHj) OH OH (CH2) m rr^ C=0 NT (CHj) m 0=C HO HO *(CH2)f l.m.n-CYCAM c=o OH OH ^:: cijc:: 9;: p:: C=0 I HN C=0 I ,N C=0 I N, C = I ,NH (CHj);;^ ^iCW^)^ ^iZH^)^ n,m,n-LICAIi FIGURES Synthetic Catectiolamide Ligands.

PAGE 15

COOH COOH 'OH ^'^^^OH OH COOH COOH 'OH ^^^^OH OH C=0 I HN^ OH C=0 I C=0 I OH C=0 I .NH (CH2)f ^(CH2)^ ^(CH2)f ^v^.^^OH "V^i^^^^OH ^^^Y^OH ^Y^^ OH OH C=0 I HN C=0 I C=0 I C=0 I .NH (CH2)f ^(CH2)^ ^(CH2)^ FIGURE 4 LICAMC (Linear Catecholamide, Carboxylated) and LICAMS (Linear Catecholamide, Sulfonated).

PAGE 16

"'.w»W»-. 10 toxicity was still evident. Mays examined plutonium excretion by DTPA and/or LICAMS. His results indicate tliat the combination of chelators is as effective as DTPA alone at americium removal, and as effective as LICAI^S alone at plutonium removal (30). Therefore combination chelator therapy represents a practical approach to the problem of total actinide decorporation in man. In hopes of alleviating the problems stated above, a series of new octacoordlnate catecholamide chelators for application to both biological and environmental actinide decontamination was designed, figure 5. We have developed a structurally novel system referred to as H-shaped ligands, in reference to the unusual stmcture of their backbone. Each point of the H is bound to a catechol moiety. Our goal was to synthesize a series of non-toxic ligands, each with its own unique backtjone, which possessed a selective affinity for actinide metals, and could acess more or different body compartments than DTPA. This series of ligands is not carboxylated; the carboxyl groups are one potential source of LICAI^C's observed toxicity. Additionally, H-shaped ligands are based upon a spermidine backbone rather than spermine. Based upon experience with N\N^-bis(2,3-dihydroxybenzoyl)spermidine (Cpd ll)-a probable product of metabolism, if metabolism occurs-the H-shaped ligand backbone should also be non-toxic (46). High yield syntheses were designed and carried out to give the thirteen symmetric and assymmetrio model catecholamide chelator systems in figure 5. The aim of the design was to synthesize a series of ligands whose geometries and lipophilic properties differed in order that the binding geometry of the ligand to a specific metal could be optimized, and in order that the ligands would exhibit differences in lipophilicity which would carry them into different body compartments. By way of analogy, imagine that each metal and each ligand of interest are a sphere and hand, respectively, and that the optimum geometry of the chelate, or "grip", is achieved when

PAGE 17

11 NH (CH2)a(CH2)t,C=0 (CH2)„ c=o (C»2h " ^^^T^d NH sy mmetri : Ligands As "Mixed" or ymmetric L igands cpd n a b c d cpd n a b c d la 2 3 3 3 3 ig 2 3 3 3 4 lb 4 4 4 4 Ih 3 3 4 4 Ic 3 4 3 4 Ij 3 4 4 4 Id 3 3 3 3 3 Ij 3 3 3 3 4 1e 4 4 4 4 Ik 3 3 4 4 If 3 4 3 4 11 3 4 4 4 Im 4 3 4 3 4 FIGURES H-Shaped Ligands.

PAGE 18

12 only the catechol moieties, or "fingertips," of a hand can comfortably hold a "sphere." If a hand is small in relation to the sphere of interest, the grip will be weak. Conversly, as the hand grows large in relation to the sphere, it becomes difficult to hold the sphere with the tip of every finger. Clearly, a ligand which binds, or grips, metals indiscriminately would quickly be rendered inactive, since other ions, or spheres, present would soon saturate the metal binding sites. It was demonstrated that H-shaped ligands do indeed exhibit a selective affinity for actinides in aqueous solution, human plasma, and rats. 1 ' i ' i ' . ! i ' . ^ •'' * . r :: . f-

PAGE 19

~^*7 CHAPTER II H-SHAPED OCTACOORDINATE ACTINIDE LIGANDS FOR HUMAN AND ENVIRONMENTAL DECONTAMINATION Background High yield syntheses were designed and carried out on the thirteen symmetric and asymmetric model catecholamide chelator systems in figure 5. The aim of the synthetic program was to design and synthesize a series of iigands whose geometries and lipophilic properties differed. If successful, the binding geometry of the ligand to a specific metal could be optimized, and the Iigands would exhibit differences in lipophilicity, carrying them into different body compartments. These two issues are reievent in the following ways. It is clear that different actinide metals will have different optimum ligand binding geometries. For example a ligand which binds plutonlum may not bind amerlcium as effectively. However, whether the metals' chelation geometries are significantly different remains to be and must be established. With regards to lipophilicity, because the distribution volumes of the actinide metals are not equivalent throughout the various tissues (47), designing Iigands having different distribution volumes is essential if all actinide pools are to be acessed. This is critical in the development of a chemotherapeutic system for total actinide removal. In the case of LICAMS and DTPA, when only LICAMS was used, only americium was effectively excreted. Conversely, if only DTPA was used, only plutonium was excreted effectively. When administered in combination, however, both actinide metals were effectively excreted. This illustrates how a combination of chelators can effectively eliminate all of the actinide burden from an animal where single chelators could 13

PAGE 20

14 not. Of course one hopes that a single drug will remove all of the toxic metals; however, this is unlikely. During the synthesis of these compounds, it became evident that H-shaped iigands are only slightly soluble in neutral or acidic aqueous solutions. If, however, one could lower the pKg's of the catechol moieties then the pH range over which the ligand is soluble would expand. Raymond and his associates have made several attempts to do this involving carboxylation and suifonation of the aromatic ponions of his catecholamide Iigands (33, 34). These derivatives did not exhibit a very impressive change in water solubility. In fact, the final products were isolated by recrystallization from water. With this same goal in mind, N,N-diethyl-{2,3-dihydroxy-5-nitro) benzamide was prepared for puposes of potentiometric titration, along with its parent compound N,N-diethyl-(2,3-dihydroxy)benzamide. If the pK^'s of this nitrated monomeric derivative are substanially lowered, relative to those of the unsubstituted monomer, then the water solubility of the tetranitro H-shaped ligand should be greatly improved relative to the unsubstituted ligand. In addition to increasing water solubility, the visible absorption spectrum of the nitro monomer was found to be pH-dependent, implying that a nitro ligand could be a useful analytical tool for metal binding. It was not our intention to apply this ligand to biological decontamination since nitro compounds, in general, are notoriously toxic. This was observed with CYCAM-NO, (44). However, it may be useful for environmental decontamination, increasing the pH range over which the ligand is effective. During the course of these studies, it became evident that an insoluble form of chelator would be very useful, possible applications being water decontamination, prevention of gastrointestinal actinide absorption, dialysis, and prevention of percutaneous absorption. Thus, a resin-bound catecholamide was prepared. If contaminated water was passed through a column containing the resin-linl
PAGE 21

15 absorption into the body. Although gastrointestinal absorption is not a problem in the case of PuOj, other forms of plutonium are absorbed and retained (48). The bound metal-ligand complex would then be excreted intact via the feces. This same type of system could be used to dialyze actinides without exposing the patient to systemic chelators. It should be pointed out that we are not trying to synthesize a single derivatized resin to perform all of these feats. We intend only to create a single such derivative in order to determine if matrix-linked ligands are a useful concept. If the concept works, different resin materials can be investigated to optimize the effectiveness of each application. Synthesis Catecholamide Ligands Parent catecholamides It should be pointed out that all syntheses were designed to facilitate modification of the molecule. This is critical in drug design, as stnjctural changes are frequently required in order to improve on phannacological behavior. Without such flexibility, one would be faced with starting from the beginning-i.e., developing a new synthetic sequence for each stnjcfural analog-rather than simply introducing a modified reagent into the present design. Note that the thirteen analogs shown in figure 5 fall into two classes-symmetric (lad and Im) and asymmetric or "mixed" ligands flg-IV Symmetric ligand precursors (23:1 and 2m) were synthesized by coupling an appropriate acid dichloride (succinyl, glutaryl, or adipoyi) with two equivalents of either N^N^-bis(2,3-dimethoxybe^zoyl)spermidine, norspermidine, or homospermidine, figure 6. In the case of an asymmetric ligand, one must selectively acylate each end of a diacid with a different spermidine derivative, necessitating a two-step process; i.e., the stepwise addition of each amine derivative. This is achieved by the aminolysis of succinic or glutaric anhydride (49) by the noror homospermidine derivative, figure 7. The "half-acids" thus produced (3a:d) can next be coupled to a different spermidine derivative, yielding methylated asymmetric ligand

PAGE 22

16 OCH, OCH, (CH2)a NH (CH2)b CH3O A.C1/I Cl--'^(CH2);f^Cl / NEtj (CH2), (CH2), — N I C=0 I (CH2)„ 1 c=o I — N (CHo) 2'b CH3O (CH2)t, NH cpd a b n ,' ' "1 2a 3 3 2 ' 2b 4 4 2 2c 3 4 2 2d 3 3 3 2e 4 4 3 2f 3 4 3 2m 3 4 4 FIGURE 6 Synthesis for Symmetric Ligands.

PAGE 23

17 NH — (CH2)a NH iCH^)^ OCH, OCH, ^'^^2^n CH3O >^^ OCH, COOH 1 CH3O (CH2)n C=0 CH3O NH — (CH2)a N (CHj)^ — NH a b n 3a 3 3 2 3b 4 4 2 3c 3 3 3 3d 4 4 3 C"3^H^'="3 H=C=Hv a /cr-'Xi o=c CHjO :a? c=o "NH— (CH2)-NH-(CH2)j— HH' 0=f^NH — (CH2)a N (CH2)i, NH ' C=0 Cs^OCH, -^OCH,
PAGE 24

18 precursors {2jtDThe removal of the methyl protecting groups results from the action of boron tribromide, figure 8. Nitro derivatives , ' . 3 .; . _ Initially, synthesis of a tetra-nitro derivative of an H-shaped ligand was attempted via the tetra-BOC precursor id shown in figure 9. This method yielded the desired compound, but required many steps; thus the total synthesis was very time consuming. Also, one must acylate each amine with the same group; we ultimately hoped to be able to place a different acyl group on each amine so that we could "fine tune" the biological properties of the dnjg. To mal
PAGE 25

19 OCH, OCH, NH (CH2)a N (-CH^)^ NH C=0 CH3O OCH, OCH^ (CH2)„ c=o NH (CH2)c N (CH2), 2'd (2a-m) BBr, NH (CH2)a N (CH2)[j NH C=0 I (CH2)„ (CH2)j NH (la-m) FIGURES Removal of Protecting Groups.

PAGE 26

20 O'^NH 0,N >^^V OCH, -OCH CH,0 NO, FIGURES Synthesis for Tetra Nitro Ligand.

PAGE 27

21 O2N CH3O CH. ,X--T'" 0,N 0,N XT" 7 Figure 10 Alternate Synttiesis for Tetra Nitro Ligand.

PAGE 28

22 -C=N C=N 'NH 2 ^N' 'NH, -NH, NH, -N^^^^NH, NH, FIGURE 11 Synthesis for N^-Benzyl Triamines.

PAGE 29

23 H2N'! l^ FIGURE 12 Synthesis for Tetramethoxy Precursors.

PAGE 30

24 H2NH »NH, CH2=0 02N-v^>S/0CH3 NH-^O 0==*^NH FIGURE 13 Synthesis for Bis Nitro Tetrametlioxy Precursor.

PAGE 31

-w^^y r p 25 synthesis will also proceed smoothly with nitro derivatives. Thus, one should be able to synthesize a bis nitro H-shaped ligand precursor in which two catechols are nitro-derivatized and two are not. Unfortunately, reaction of t)oron tribromide with 5a yielded a product which was not iron positive (52), indicating that free catechols are not present in the product. Trimethyl silyl haiides were no more effective at freeing the catechols in this compound. This was very surprising in light of the fact that N,N-diethyl(2,3-dimethoxy-5-nitro)-benzamide deprotected in the usual manner, and that S was deprotected with facility as well. N-(3-amino-1-propyl)hexahydropyrimidine (APHHP) has proven to be a very useful reagent to us for organic synthesis. In the preceding synthesis, the fact that the central amine of norspermidine is now tertiary has been exploited. Methods were also developed in which the difference between the primary and secondary amine is used to our advantage. If one were to add a single equivalent of an acid chloride to APHHP, the products would include primaryacylated, secondary-acylated, and bis-acylated hexahydropyrimidine, an undesirable result. By using an N-hydnjxysuccinimide ester rather than an acid chloride, one can selectively acylate only the primary amine in near quantitative yield. Although other researchers have used hindered acyl transfer reagents or protecting groups to achieve the same end, these reagents are not nearly as selective {53, 54, 55). Once isolated, this mono-acylated product can react with a different acyl chloride, followed by deprotection with dimedone, to yield an asymmetrically substituted polyamine. Alternatively, the monosubstituted hexahydropyrimidine can be deprotected with dimedone, followed by reaction with a different N-hydroxysuccinimide ester to yield the identical product. In order to demonstrate the usefulness of this method, IQ. was prepared, figure 14. APHHP was acylated with (2,3-dimethoxy)benzoyl-N-hydroxysuccinimide, acylation occurring strictly on the primary amine. This product (2) was either reacted with 2,3-dimethoxy-5-nitrobenzoyl chloride followed by dimedone, or dimedone followed by (2,3-dimethoxy-5-nitro)benzoyl-N-hydroxysuccinimide or (2,3-dimethoxy)ben2oyl-N-hydroxysuccinimide, to yield Ifl or

PAGE 32

A? 26 CH3O.J »NH, o'^ci CH3O CH3O CH. 0*^0-H I CH3O CH3O 9 u Ojjv-'^O 2. •="3°inr °-p O'^O-Mj 0,N ^T^^' OCH, OCH, CH3O CH3O 'NH FIGURE 14 Synthesis for Mono Nitro Tetramethoxy Precursor.

PAGE 33

27 NVN'-bis-(2,3-diniethoxyben2oyl)norspermidine, respectively. To show that asymmetric synthesis will work with nitro derivatives, 111 was successfully reacted with 2a, figure 1 5, to yield the mono-nitro octamethoxy precursor. In this case reaction with boron tribromide proceeded with ease to yield 11a . Resin-Boun d Catecholamirie Liganris The catecholamide The intention was to synthesize a catecholamide ligand with a chemical handle with which the molecule could be covalently bound to an inert resin matrix. The design of such a molecule required the integration of several factors: (1) in order for selectivity to be maintained it is crucial that the matrix-linked ligand retains the conformational mobility of the parent ligand, (2) the chemical handle must be capable of covalent binding with a resin material, (3) the placement of the handle into the molecule must be done in such a way that the geometry of the metal-ligand complex is undisturbed; i.e., the handle should be on the exterior of the chelate, (4) if the ligand were attatched directly to a polymer matrix, it is conceivable that the matrix would impose conformational restraints upon the ligand; i.e., the molecule would have restricted mobility, affecting its ability to sequester the metals of interest. Thus the handle should allow enough distance between the resin matrix and the chelator moiety so that the latter does not interfere with the mobility of the chelator and, more importantly, with metal chelation. Compounds la and H with a chemical handle f13e and i3D were synthesized by simply replacing glutaric acid with N-protected glutamic acid, figure 16. In this way a protected amine functionality is Introduced into the backbone of the molecule, in a location which is far removed from the chelating sites. The t-butoxycarbonyl (BOC) group was chosen to protect the amine handle. This moiety Is quickly and quantitatively removed by the action of triflouroacetic acid (TFA). A linear organic "string" in the form of N-BOC-8-aminocaprylic acid was attached to the amine handle to Increase the distance between the ligand and the resin surface, figure 17. Once attached, a new amine functionality can be exposed by treatment with TFA to yield I5f.

PAGE 34

28 OjN 'XX OCH, OCH, 0=C. «NH' «NH' + «NHOCH, OCH, COOH ''"s CH1)9 -NH' 5a »NH' CHj^^^CHj 6 O2N XX OCH, ,N=C=Nv OCH, 0=C :. V'. . ^NH' cr — r] ^^^:>Nj-^0CH3 >^0CH3 CH, "N-" — ^»NH '° CH3O CH3O .C=0 o=c. .c=o 'NH' «NH' FIGURE 15 Synthesis for Mono Nitro Octamethoxy Precursor.

PAGE 35

29 OCH, OCH, CHj^^^CHj H=C=H> 6 /Cf'"--"io CH30 CH30 NH-BOC "Ovy'k,^^^"" FIGURE 16 Synthesis for an Amino-H-Shaped Ligand.

PAGE 36

cor. 30 ^^y^OCHj 1. 2. OCH CH,-v„^CH, o'^nh-C7H,4-c-oh/dcc/ ^ CF3COOH OCH, FIGURE 17 Attatchment of the Spacer Molecule.

PAGE 37

31 Our first attempts to attach such a string involved 4-amlnobutyric acid. The condensation went weii but when exposed to TFA the product did not stain with ninhydrin, a reagent which qualitatively detects primary and secondary amines and amine salts. A possible explanation is shown in figure 18. However, when 8-aminocaprylic acid was used cyclization did not occur, yielding the desired product. An alternative synthetic pathway was also developed, figure 19. In this synthesis, N-BOC-8-aminocaprylyl-N-hydroxysuccinimate was reacted with glutamic acid, yielding the diacid derivatized with the organic string. This intermediate was reacted with two equivalents of N^N^-bis(2,3-dimethoxybenzoyl)spermidine in high yield. In the previous synthesis, the H-shaped ligand was assembled prior to attachment of the organic spacer arm. This latter pathway renders the synthesis more cost effective because the synthesis is convergent rather than linear. Therefore, smaller building blocks, or intermediates, are used in the final steps. The resin It was decided that an amide linl
PAGE 38

•if:.-. 32 owhere R = NRH,CO''^V'^ 'HH^^O ^T-v^OCHj H3CO ^V^OCH, H,CO C=0 ' ^ 0=C C=0 0=C 3 "3"C=0 0=C -OCHj H3CO FIGURE 18 Cyclization Mechanism in Triflouroacetic Acid.

PAGE 39

33 tBOC-NH CH3O FIGURE 19 Alternate Synthesis (or Amino-H-Shaped Ligand with Spacer Molecule Attatched.

PAGE 40

34 W // -CI Herri field '3 resin Q-COOH Amberlite IRP-64 FIGURE 20 Synthesis for Acid Chloride Derivatives of Commercially Available Resins.

PAGE 41

;. 35 Reaction with boron tribromide gave the free catecholamide resins 22L 221, and 24L as indicated by a positive reaction with ferric chioride solution (52). Lastiy, CH-Sepharose-4B activated resin was obtained. This resin was ideai for ligand immobilization. The matrix is composed of a polysaccharide backbone which has been derivatized with hexanoic acid present as the N-hydroxysuccinimide ester. Since this ester will react selectively with amines in the presence of alchohols, the methyl protecting groups were removed prior to coupling of the ligand to the resin material. Therefore the resin need not be exposed to the harsh chemical reaction conditions necessary for demethylation. This was not the case with the acid chloride resins described. Exposing 121 to tioron tribromide gave a high yield of the desired "catecholamine" 25, figure 21. When 2£ was coupled to the resin, giving 2S, a spacer of six carbon units is inherently inserted between the ligand and the polymer surface, allowing for free motion of the molecule. Coupling was performed in 50 percent pH 7.2 phosphate buffer: 50 percent ethanol to ensure dissolution of the catecholamide. In addition, the resin was washed with a large volume of ethanolic buffer after the reaction was quenched to ensure removal of any uncoupled ligand. A final wash was performed with 50 percent pH 4.0 acetate buffer: 50 percent ethanol to protonate all of the catechoyi functionalities. The last wash was concentrated and reacted with FeClj. No color change was observed, indicating that no free catechol was present in this last wash. It was gravimelrically determined that ligand coupling was quantitative, by the weight difference between the control resin and the catecholamide resin. When a small amount of this resin (22) was placed in FeClg solution, a dark purple color developed, indicating the presence of the catechoyi functionality. Lioand Stnichinm Rlrv: Job's Pint.-; Before ligand binding constants could be calculated, it was necessary to measure the metal-ligand ratio for all metals and ligands involved in the determination. Thus, Job's plots (61) were performed in an ammonia buffer at pH 9.2 with copper for ligands la, ZL and Cpd II (N^N^-bis[2,3dihydroxybenzoyljspermidine), figure 22, at a total concentration of metal and

PAGE 42

36 27 Cpd II R R H n Stolchlometr y (L:M) Th'** 1 : 1 Cu^* 1:1 1 la NO^ la R = H Th Cu H Stoichlometr u ( L:t1 ) 4+ 1:2 2+ 1:2 FIGURE 22 Results of Job's Plots with Catectiolamide Ligands.

PAGE 43

37 ligand of 33 nM. The results indicate that Compound li and 2Z form 1 :1 metal-ligand compiexes with copper and la forms a 2:1 copper complex. These results are consistent with N/lcGovern's findings for Compound II (62, 63). In addition, it is reasonable to expect 2:1 stoichiometry for la and copper, since la is essentially two Compound II molecules linked together. In the case of copper, spectral changes upon metal binding are small, but are large enough for this type of measurement. With thorium, however, the spectral changes could not be used for Job's plot measurements. Therefore, the mono nitro H-shaped ligand Ha, figure 22, was used in place of ligand la. In order to show that using Ha would be a taie reflection of what happens when la is used, a Job's plot was generated for Ha and copper. If the nitro group caused a significant change in the ability of the catechol to bind a metal, the Job's plot would not indicate 2:1 stoichiometry. For example, if the nitro-containing catechol binds copper more effectively, the first equivalent of copper would associate primarily with this group. This phenomenon would be reflected in the Job's plot as 1:1 stoiciometry, since the second equivalent of copper would not affect the nitro-containing catechol, which is already associated with the first equivalent of metal. Conversely, if the nitro-containing catechol binds copper less effectively, the first equivalent of copper would not affect the visible spectnjm since the metal would not associate with the nitro-containing catechol. Neither of these effects were observed, indicating that Ha does indeed accurately reflect the chelation chemistry of la. Thus, compunds 21 and Ha were used as a measure of the stoichiometry of the thorium complexes of Cpd II and la, respectively. For compound la another experiment was performed to insure the accuracy of using Ha in its place. A Jobs plot was performed with la and thorium with eriochrome black T (EBT) present. It has already been shown that EBT does not displace thorium from H-shaped ligands under these conditions. Therefore, EBT can be used to complex any thorium which is not bound to la. AM of these experiments yielded the same information: Cpd II and 21 form 1 :1 metal-ligand complexes with thorium, and la and Ha form

PAGE 44

^ 38 2;1 complexes with thorium. This was somewhat surprising in light of the 1 :1 stoichiometry observed with hexadentate catecholamide chelators and iron. This stoichiometry is most iil
PAGE 45

39 CI-PHN Eriochrome Black T FIGURE 23 Colorimetric Reagents used for the Quantitative Detection of Metals.

PAGE 46

40 way, first column of table 1 . Note that 3,4,3-LICAM did not perform as well as H-shaped llgands In tfiis experiment, removing 88.9 percent of the actinide. Controls without a ligand showed no removal of actinide. in a series of experiments conducted in a joint effort between Rocky Flats and the University of Florida it was demonstrated that plutonlum could be removed from aqueous solution by the same precipitation technique (J. D. Navratil, private communication), table 2. In practice, this means that if contaminated water is adjusted to high pH and a ligand Is added, the contamination can be concentrated onto a filter by neutralization prior to filtration. One might argue that a portion of the actinide, due to the nature of the experiment, was simply trapped by precipitation of the ligand in a non-specific way. If this were the case, the ligand would be able to bind more metal than the amount dictated by the llgand-metal complex. It was shown that this Is not the case. When the ligand precipitation experiment was repeated with excess thorium, the ligand did not remove more than the stoichiometric amount of metal, indicating that non-specific precipitation did not occur. The next experiments detemiined that H-shaped llgands can bind thorlum(IV) when the metal is Introduced Into an acidic solution containing the precipitated ligand and, conversely, that the llgands can effectively precipitate thorium when introduced into an acidic solution containing the actinide. This renders it unnecessary to adjust the solution pH prior to addition of the chelate, an important financial consideration. First, ligand was precipitated from a basic solution by acidification. One equivalent of thorlum(IV) was then added. The suspension was stirred for several minutes, filtered, and the effluent assayed for thorium. Then, an acidic thorium solution was prepared. One equivalent of a ligand was added as a methanollc solution. The suspension was stirred for several minutes, filtered, and the effluent assayed for thorium. These methods proved very effective, removing 93.04-99.91 percent of the metal, table 3.

PAGE 47

41 TABLE 1 Results of Thorium Precipitation Experiments With Catechoiamide Ligands. Percent Thorium RRmnval^ Compound Competing Metals Competing Metals Percent fitiSSDl Present^ Diffenencfi la 96.90 ±1.42 (5)b lb 99.76 ±0.35 (6) ic 97.77 ±2.72 (6) Id 96.75 ±0.95 (6) 1« 99.38 ±1.29 (6) If 96.39 ±1.06 (6) ih 99.74 ±0.45 (6) 1i 99.59 ±0.55 (4) Ik 99.97 ±0.05 (5) LICAM 88.90 ±3.40 (5) 85.86 ±4.19 (6) 89.67 ±4.19 (5) 88.49 ±2.86 (6) 85.58 ±4.95 (6) 93.76 ±4.40 (6) 84.32 ±4.33 (6) 89.54 ±2.31 (6) 84.96 ±3.93 (6) 85.38 ±5.87 (6) 71 .05 ±7.53 (6) 11.39 10.11 9.49 11.55 5.66 12.52 10.23 14.69 14.59 20.08 ^ n'hlinitial=3.24 x 10-5 M, [ligand]|niiiai=3.3 x 10-5 M. ^ The number in parentheses is the number of experiments performed. = Competing metals used were Fe(lll), Ca(ll), Mg(ll), Mn(ll), Zn(ll), and Hg(ll). One equivalent of each metal was used. iv^ i,;-

PAGE 48

42 TABLE 2 Results of Plutonium Precipitation Experiments WAh Catecliolamide Ligands. % Plutoniiim Ramnval Oornixmi WitiioiJtImn Witiiimn 13 68±3 59±5 lb 90 ±6 86 ±6 1" 63±11 58±6 "8 55±1 47±4 H ^ 92 ±10 1) 88 ±9 "« 66±11 54±6 [Pulinitial =[ligand]initiai =1 x 10-5 m.

PAGE 49

43 TABLE 3 Results of Catecholamide Pre-Precipitation Experiments with Thorium. Compound Percent Thorium Removal Ligand Added to an Metal Added to an Acidic Metal Solution Acidic Liaand Si isnfinsbn la 99.91 ±0.16 (4)3 96.90 ±3.41 (6) lb 98.21 ±3.15 (5) 99.1 6 ±0.82 (6) 1c 99.83 ±0.19 (6) 98.90 ±1.66 (6) Id 99.54 ±0.79 (5) 93.04 ± 6.98 (6) le 95.60 ± 6.21 (6) 99.08 ±1.44 (6) If 99.40 ± 0.63 (6) 98.93 ±0.51 (5) ih 99.82 ± 0.31 (4) 99.37 ±0.46 (6) If 94.58 ±4.18 (5) 99.35 ± 0.57 (6) Ik 94.31 ±7.20 (5) 98.40 ±1.81 (5) LICAM 97.64 ±2.00 (5) 97.95 ± 1 .02 (6) ^The number in parentheses is the number of experiments performed. n"h]initial=3-24 x lO'S M, [ligand]initja|=3.3 x lO'S M. v';5

PAGE 50

44 A property which is requisite for the H-shaped ligands in order to be practical for decontamination purposes is their ability to sequester actinides in the presence of other metal ions. Thus, experiments were designed to determine whether other polyvalent cations could compete with actinides for the ligand binding site. Cations were selected because of their +2 and +3 oxidation state, their relative abundance and importance in biological systems, and their l
PAGE 51

45 Insoluble Chelators: Removal of Artinidss from Watsr and Blood Plasma Although very effective as chelators, H-shaped ligands have some limitations for applicability. For example, a chelator taken by mouth to prevent Gl absorbtion would be useless if absorbed from the gut. Additionally, the filtration technique is not applicable to basic solutions, where the ligands are very soluble. A soluble ligand would also be inappropriate for dialysis. Therefore we wanted to modify the ligand into an insoluble form, to overcome the types of problems just mentioned. XAD-4-SnrfiPH I l^janrK; The Rocky Flats group denxjnstrated that when our H-shaped ligands are adsorbed onto XAD-4, a macroreticular resin, and contacted with aqueous plutonium-containing solutions, plutonium(IV) adsorbs very effectively onto the resin (66). In these early studies, excessive amounts of ligand were coated onto the resin, resulting in a situation where a great deal of the ligand was not available to sequester plutonium from solution; i.e., ligand exposure on the resin was minimal. Presumably the ligand molecules stacked on top of one another rather than forming a monolayer. In spite of this, remarkable decontamination factors (Df's) were observed, values in the hundreds. The Df is defined as the moles of metal bound per gram of ligand divided by the moles of metal in solution per milliliter of solution. This value, of course, is difficult to use for comparison of various ligands because the molecular weight of the ligand is incorporated into this term. Therefore, it is more useful to measure the Kd for the system, defined as the moles of metal bound per mole of ligand divided by the moles of metal in solution per milliliter of solution. Df's are still presented to facilitate comparison to earlier work. In later experiments where the ligand loading had been optimized, Df's on the order of 10'' were attained (67). Furthermore, the Rocky Flats group demonstrated that a significant amount of plutonium(IV) was removed from human plasma (up to 36 percent) when 10 mL were exposed to 250 mg of resin at 370c overnight. This result indicates that resin-bound H-shaped ligands may be practical for dialysis. Their results are summarized in tables 4, 5, and 6.

PAGE 52

46 TABLE 4 Ktfs and Dfs for Plutonium at a Ligand Loading of 10"^ moies iigand/0.25 g XAD-4. Compoiind \ 7.48 xtO^ 1140 10 9.57 xlO^ 1160 Id 5.30 xlO^ 1560 te 2.25x10^ 1430 If 2.12x10^ 2300 th 3.75x10^ ;t501 1j 6.60x10^ t»10 tk 3.69x10^ 1110 11 4.40x10® 1530 LICAM 4.81 X 10® 1980 250 mg ioaded resin was exposed to 10 mL solution containing 5 x lO'^M Plutonium. Samples were mixed overnigtit prior to measurement. Values have been corrected for plutonium removal by control resin. ,....*iv. i; u.: % >w^; , , '..C ' V , tr ' z.l.

PAGE 53

47 TABLE 5 Kd's and Df's for Plutonium at a Ligand Loading of 10''' moles ligand/0.25 g XAD-4. ComcQUPd la Df 1a 2.88x10^ 14,600 1 1 ,900 10,200 1« 2.74x10'' 12,500 1* 4.50x10'' 11,500 11,500 1c 7.50x10^ 1d 1.15x10^ 1J 5.31x10^ 1k 3.95 xlO^ tl 3.85x10^ LICAM 2.03x10^ 11,200 14,700 16,100 250 mg loaded resin was exposed to 10 mL solution containing 5 x 10"=M Plutonium. Samples were mixed overnight prior to measurement. Values have been corrected for plutonium removal by control resin. i ' ^

PAGE 54

48 TABLE 6 Results of XAD-4 Experiments With Human Plasma. .; i <' r Compainri 1a lb 10 Id 1e 1f 1h ii Ik II PerCRnt Plu tonium Rfimnval ^ 8±6° 0±6 21 ±8 17i:3 28 ±17 36±8 n.a. 30 ±13 22±0 21 ±8 3±19 0±6 26±10 0±7 8±18 9±4 3±4 6±6 n.d. 0±17 A = 10-3 moles ligand/0.25g XAD-4, B = 10-6 moles ligand/0.25g XAD-4. ^250 mg loaded resin was exposed to 10 mL plasma containing 4.18 x 10-'M Plutonium. Samples were mixed overnigtit prior to measurement. "Values have been corrected for plutonium removal by control resin.

PAGE 55

49 Covalentlv-Bound Ligands IRP-64-bou nd catecholamides The precipitation experiments previously described lor free ligand were repeated using resin 2M in place of a free iigand, but without adjusting the pH to neutraiity prior to filtration. Remember, pH adjustment was necessary to cause precipitation of the ligand-metal complex, in this case, the ligand Is insoluble at every pH. Therefore, pH manipulation is unecessary. It was found that control resin--in this case the methylated catecholamide resin-removed as much thorium as the catecholamide resin. It was concluded that the polymer matrix was responsible for a large amount of metal binding. Unfortunately, many free carboxyl groups remain in the derivatlzed resin; only about ten percent reacted with a ligand. These free carboxyl groups are able to bind actinides, and therefore can interfere with attempts to measure metal binding due only to the bound ligand. Although these carboxyls appear to enhance thorium binding, this binding is non-specific. In order to measure the ability of the chelator, once bound to a polymer matrix, to bind thorium selectivly, binding due to the non-specific matrix must be either measured or prevented; we chose the latter approach. Control resin was placed in pH 9.2 buffer with thorium and different concentrations of EDTA. It was found that an EDTA concentration of 150 mM was needed to prevent binding of thorium to the control resin, table 7. Unfortunately, this concentration is high enough to compete effectively under these conditions with the catecholamide for the metal, as indicated by reduced thorium removal by 241 at this concentration. It was concluded that a different resin should be used. CH-Sepharose-4R-hniinrin alecholamides Next, resin 22 (figure 21) was tested. In this case 2.6 mM EDTA was enough to prevent the control resin from binding thorium when present at concentrations up to 0.64 jiM, figure 24. In this case the control resin was reacted with ethanolamine, the reagent used to quench the excess reactive sites on the catecholamide resin. This concentration of EDTA did not affect

PAGE 56

TABLE 7 Results of Experiments With IRP-64 Resin Derivative. 150 mM OMe 8.1 OH 47.8 150 mM — OMe OH 65 6 2.11x10-' 24 5.21 X 10^ 50 [EDTA] Resin % Removal Hours Kd 15 mM — 3 2.22x10^ OMe 23.5 OH 71.1

PAGE 57

51 100 [EDTA] [Th] [] 2.52 mM 0.64 |iM H 2.29 mM 0.03 |iM [•] 2.68 mM 0.076 |iM H 0.27 mM 0.14|iM \°] 2.65 mM 7.04 |iM [] 2.65 mM 60.7 |iM I'M 7 8 ~1 10 mg Control Resin Added FIGURE 24 Determination of the Ability of EDTA to Prevent Adsorption of Thorium onto CH-Sepharose-4B Control Resin at pH 9.2 as a Function of the Milligrams of Resin Used ^•1 > L./

PAGE 58

52 thorium binding by the catecholamide portion of 2fi, figure 25. For these experiments, ^^"thorium, a high energy a-emitter, was used, enabling very precise measurements of trace amounts of metal to be made. By adding 230^f,Qf|yn., ,q g solution of 232,f|(jrium, the concentration of metal can be increased while the number of counts added remains constant. When the concentration of thorium was increased in this way from 30 nM to 76 nM, it was found that control resin did bind some thorium, approximately six percent. When the thorium concentration was raised to 7.04 \iM, more binding was observed; approximately 29 percent of the thorium in solution was non-specifically bound by control resin. In both cases, 2S was still able to remove more than 98 percent of the thorium in solution, figure 25. When competing metals were added, no significant difference in thorium removal by resin-bound catecholamides was observed, figure 26. These data indicate that matrix-bound ligands represent an effective form of H-shaped ligands which could be applied to many facets of decontamination. Equilibrium Solution Chemistry: Thermndvna mic Binding Constant Msasiirfimpnts Eriochrome Black T Cnmnatitinn Many techniques cannot be used to accurately determine very large metal-ligand binding constants due to the fact that the free metal in solution, or the free ligand in solution, is present at a concentration which is so low that it cannot be measured accurately. Therefore, it is essential to approximate this constant before designing an experiment for an accurate determination. A spectrophotometric technique was developed to approximate the magnitude of the thorium-ligand association constant. The colorimetric reagent eriochrome black T (EBT) was very well suited for this approximation, figure 23. EBT forms a very stable 1 :1 complex with thorium(IV), the absolute binding constant being on the order of 10^^ M'''. Its working pH ranges from 8 to 10, which is ideal for H-shaped ligand solubility (69). At its working pH EBT exhibits an absorbance maximun at 660nm while the thorium-EBT complex has an absorbance maximun at 570nm (70). A 10"^M solution of EBT can quantitatively detect thorium(IV) to ^^^ -j,'.,^^: \ C^'" r ^'L, . .: :}'it t'i

PAGE 59

53 100 c o o V) c c c '5 E Oi GC 0) a. 6 8 10 mg Resin Hdded 12 14 .V *.• FIGURE 25 Determinatton of the Ability of 2S to Remove Thorium from Aqueous solution (pH 9.2). Control Resin is Indicated by Squares and 2S is Indicated by Circles.

PAGE 60

54 c c 'S £ E p o JZ a. D Control resin, thorium only Control resin, competing metals present O I Cateoholamide resin, thorium only • Cateoholamide resin, competing metals present 3 4 5 6 7 mg Resin Added FIGURE 26 Determination of the Ability of Other Metals to Interfere with Removal of Thorium from Mn,m"\^!!i'i°" *o^-^' "^^^ ^^'^" Used were Fe(lll). Ca{II), Mg(ll), Zn(ll), Mn{ll), and Hg(ll). Solid Lines Represent 7.04 nM Metals, Dotted Lines Represent 60 4 nM Metals

PAGE 61

. '%•-. 55 concentrations as low as lO'^M (a 100:1 ratio of EBT to thorium) with a variance of one percent, figure 27. . ,. . ^ To estimate the catecholamide-thorium(IV) association constant, a solution buffered at pH 9.2 was prepared, containing one equivalent of ThCl4 and one equivalent of EBT. The solution showed a pinl< color, indicative of the thorium-EBT complex. Upon addition of one equivalent of a cateoholamide, a color change, from pink to blue, was observed, indicating that the catechol displaced thorium from EBT, according to the equation 2 EBT-Th + Cat EBT + Cat-Thj pi nk bl ue The resultant absorbance spectmm was identical to that of the uncomplexed EBT dye, table 8. From these data it can be concluded that less than 0.1 percent of the total amount of thorium(IV) in the solution is associated with EBT. As a consequence of this "competition," the relative conditional binding constant of the cateoholamide ligands relative to EBT must be on the order of lO"' or greater. Competition Studies with Nitro Dfiri vatives of Catechnlamirie Ligands In the previous experiment, the data could not be used for an accurate determination of the ll-Th conditional metal binding constant because the fraction of thorium bound to EBT could not be measured. If one were to use ligands which shared the metal in such a way that the amount of metal associated with each ligand could be measured, accurate calculations of relative conditional binding constants could be made. If all proton stability constants, metal hydrolysis constants, metal-ligand complex stability constants, and metal-proton-ligand complex stability constants are known for one of the ligands, as well as the stoichiometry of the ligand-metal complex, then the conditional binding constant, defined by the equations

PAGE 62

56 o u c a t o in a < 0.30 -, 0.250.200.150.100.05 650nin corr.=-0.997 2.0 2.5 [Thorluml ()iM) FIGURE 27 Standard Curve for ThCl4 with EBT (1 nM).

PAGE 63

57 TABLE 8 Color Reactions of Various Ligands with the Thorium-EBT Complex. ComCQUnd Color Rsartinn ' Triethanoiamine EDTA Salicylic acid 2,3-dihydroxybenzoic acid 3,4,3-LICAM + Compound la + Compound lb " + Compound Id + Compound 1e ' ' + Compound If + Compound Ik {+) indicates that the solution turned from pink to blue in 15 seconds or less upon addition of the ligand of interest. One equivalent each of EBT, thorium and ligand was used.

PAGE 64

xM + nL . MxL„ and K = — — , where 58 (Mr(L)" (M). the total concentration of metal which is not complexed to the ligand. (L) = the total concentration ligand which is not complexed to the metal. (LM) the total concentration of metai-ligand complex. x,n = the stoichiometric amount of metal and ligand, respectively, in the complex, can be calculated for that ligand-metal complex (71). If the relative conditional binding constant, defined as Kfgi = Kj / K, , where onstant for the first liganc ;onstant for the second 11; is measured, then the conditional binding constant for the second ligand can be calculated, K , _ = the conditional binding constant for the first ligand-metal complex Kg = the conditional binding constant for the second ligand-metal complex, By repeating this two-ligand-one-metal competition, replacing one ligand with another, one can eventually arrive at the conditional binding constant for a ligand of interest. In an analogous fashion one can perform a one-ligand-twometal competition experiment to arrive at the conditional stability constant for a particular metal or metals of interest. In this particular case, it was hypothesized that another catechol would be a suitable candidate for competition with la. The problem was that during the process of compound characterization it was found that the spectral qualities of the H-shaped ligands-ultraviolet, visible, flourescent, etc. under a variety of conditions-did not change significantly enough upon thorium binding to be a useful analytical tool. Also, the changes of one catecholamide ligand upon metal binding could not be distinguished from changes in the other catecholamide. To avoid this complication a series of relative conditional binding constants were measured, utilizing nitrocatecholamides.

PAGE 65

59 Besides increasing water solubility, N,N-diethyl(2,3-dihydroxy-5-nitro)-benzamide exhibited another potentially very useful property. During the course of titration, it was noted that the color of the solution containing this compound was pH-dependent. At pH lower than 7, the solution was yellow. At pH higher than 10, the solution was purple. At intermediate pH, the color ranged from yellow to orange to red to purple. We hoped to be able to put this "indicator property of the nitro derivative to good use. Since the effect of metal binding upon the UV-visible spectra of a compound is the same as loss of protons, it was our belief that, upon metal binding at the proper pH, the tetranitro derivative of an H-shaped ligand would show the same type of color changes exhibited by the anions of Isl.N-diethyl (2,3-dihydroxy-5-nitro)benzamide. As explained earlier, our attempts to synthesize such a derivative were unsuccessful. We were able, however, to synthesize a bis nitro derivative of Compound II, and hoped this compound would exhibit the same type of color changes as those seen with the anions of N,N-diethyl-(2,3-dihydroxy-5-nitro)-benzamide Thus, £ was deprotected to yield N,N-bis(2,3-dihydroxy-5-nitrobenzoyl)-norspermidine (2Z), the idea being that upon metal binding spectral changes will occur in the visible range; in this range the parent catecholamide does not interfere with the absorption of the nitrocatecholamide. Thus if two catecholamide ligands are allowed to compete for one metal in solution, when one of the ligands, e.g. ZL contains a nitrocatechol and the other, e.g. N,N-bis(2,3-dihydroxybenzoyl)spermidine (Compound II) or an H-shaped ligand, does not, the the amount of metal which is bound by the nitnscatecholamide can be quantitated by the changes in its absorption spectmm in the visible range. Bergeron and McGovern (62, 63) have measured all of the equilibrium constants associated with fomnation of a metal complex with copper and Compound II. Since this type of data is not available for any thorium complex-due to the fact that the thorium hydrolysis constants are not known--we felt that we could use the data generated by McGovern, along with the copper-ammonia complex formation constants (72), to measure the copper-la conditional

PAGE 66

60 complex formation constant and eventually the thorlum-la conditional complex formation constant. When copper was added to a constant concentration of the nitrocatecholamide 22, the absortjance at 400 nm was linear when plotted against the amount of copper added, figure 28. If we assume that all of the added copper is taound by the ligand, then a linear relationship exists between the analytical amount of copper and the concentration of ligand-copper complex, figure 29. If the same measurement is made with compound II or la present at the same concentration as 22, 22 binds less copper than In the absence of the other catecholamide. By using the standard cun/e generated in the absence of a second catecholamide, the amount of copper bound to 22 in the presence of compound II or la can be quantitated, and the percent of metal associated with the nitrocatecholamide can be calculated from the ratio of the slopes of the lines generated with and without a second catecholamide present, figure 29. If one knows the stoichiometric amount of copper in each complex of interest, the concentration of each metal complex can be calculated, leading to the relative conditional stability constant for the two ligands and copper. In the cases of the Cpd ll-copper complex and the 22-copper complex, x=n=1, and the relative conditional binding constant can be expanded to ,.* _ (m-l^)(l' ) "' ' (m-l')(l2) where M-L^ = the equilibrium concentration of the Cpd ll-copper complex M-L^ = the equilibrium concentration of the 22-copper complex = the equilibrium concentration of Cpd II L'= the equilibrium concentration of 22Since excess ligand is present in all cases, free copper can be ignored, and Cm =(M-J) + (M-l2) Cl1 = (M-L') + (J) Cl2 = (M-l2) + (l2). Since Cl1 = Cl2 = 33.3 \iM. let Cl1 = Cl2 = C|_. Then

PAGE 67

W " o o 0) u c CD n L. o in n < 61 u./a — ;."> .: 1 0.70 -J >f^ — _ _ \,»^\^^ , , f 0.65\v* V ^^ 0.60\« ^%v^^ 0.55\^ • Control \ 0.50Cpd II la • \ \ 0.45 — N. ' 1 ' 1 1 ' 1 ' 1 ' 1 30 35 C(CU) (jiM) FIGURE 28 Results of Two-Llgand-One-Metal Competition Experiments wtiere 2Z is One Ligand and Copper is the Metal.

PAGE 68

62 35-, C(Cu) (,it1) FIGURE 29 Plot Of the Copper-2Z Complex Versus ttie Amount of Copper Added. The Slope of the Line Represents the Fraction of Copper Associated with 2Z at Equilibrium with an Equimolar Amount of Competing Catecholamide Present.

PAGE 69

63 Cl=(l2) + (M-l2). Rearranging these equations, (L^) =Cl(M-L^) , and ^rt^ Cl(M-L^) (m-l^)[c^^-(m-l')] (M-J)[C|_-(M-L^)1 If we let a = the fraction of the copper which is associated with L^ and f = the fraction of copper which is not bound to a ligand, then Since(i-a)»f, If we let Cy = bCL, then (M-l2) = aCy (M-O) = (l-f-a)Cj^. (M-J)= (1-a)C|. K » _ a[1-(l-a)b] '••1 "^ (1-8)(I-ab) where a = fraction of copper bound to L^ • ' b= Cm/Cl, For this case, a = 0.6909. When 0.1 equivalents of metal are present; i.e., when b = 0.1, K„, = 2.327. Using McGovem's data for the Cpd ll-copper(ll) complex along with the formation constants for copper(ll) and ammonia--the buffer used for these experiments--the conditional complex fomation constant (k'q^ ^_qJ was calculated to be 1 .20 x 1 0^* M"'' , and K 27-Cu = K cpd l,-cux'<,,l = 2.79 xlO'^M-'.

PAGE 70

64 For la and copper, x=2 and n=1. Therefore, the relative conditional binding constant expands to ^* . (m-l^xl') "' ' (M-l')(L^)(M) where (M-L^) = the concentration of 22-copper complex (H/l-L^) = the concentration of la-copper complex (L^ ) = the concentration of 2Z (L^ ) = the concentration of la (M) = the concentration of uncomplexed copper(ll). Since C^i = Cl2 = 33.3 |iM, let Cl1 = C|_2 = C|_. Then the mass talance equations can be written as Cm (M) + (M-L^ ) + 2(Mj-L^) ;l= (l2)4.(M2-l2). Rearranging these equations into the forms (M] .Cm-{M-LJ)-2-(M2-L2) {L;)-Cl-(M-l1) (L^) = Cl (Mj-L^) and substituting, , _ (M2-L^)[C^-(ri-L')l K '" ' (M-l')[Cl-(M2-L^)1(M) If we let a = the fraction of the copper which is associated with L^ and f = the fraction of copper which is not bound to a ligand, then Since (1 -a) »f. If we let C., (M) = (-Cm (M2-L2) = [(1-f-a)/21CM. (M2-L2)= [(1-a)/2]Cn,

PAGE 71

65 _ (l-a)(l-ab) '"' " a[2-(l-a)b](M) where a = fraction of copper bound to L^ b= C^/Cl. For tfiis case, a = 0.0823. Wfien 0.1 equivalents of metal are present; i.e., when b 0.1 , K\g| = 5.795/(M). The concentration of free copper in solution can be calculated from the conditional complex formation constant for the 2Z-copper complex as follows (27-Cu) , (27-Cu) K ,7-r„ = =^ (CU) 2^-C"' (27)(Cu) ^ (27)K*v.cu Since (2Z-Cu)=aCM {2Z)=C2z_-a-CM= iOCM-aCM K*2Z-Cu =2.8x10^5^-1, (Qu) = ^'V _ 0.0823 [( 1 0-a)1S^]-K'^.(.^ (10-0.0823)-2.79x10'^M'' Therefore, = 2.96x10'1^lu1 K I = l.sexlO^^M"'' ,and Ku-Cu2 =K2I-CuxK,„ = 5.47x 10"M-2. Once arriving at K for the la-Cug complex, it remains only to measure k' for the la-Th, complex. In this a problem arises, again due to the fact that spectral changes upon metal binding are small. In addition, copper and thorium cause about the same degree of change. Therefore

PAGE 72

.'. , .;;;'• •!«. J ' ' >',-.» 66 we needed to distinguish between the two metal complexes or the two free metals. We were able to accomplish this using the complexometrio reagent 1-[2-(5-chloropyridyl)azo]-2-naphthol (CI-PAN, figure 23). Synthsized by the method of Shibata et al. (72), CI-PAN is practically insoluble in water but very soluble in a variety of organic solvents, including ethanol, acetone, chloroform, dioxane, carbon tetrachloride, diethyl ether, and dimethylformamide (73, 74). Depending upon the conditions, CI-PAN has been used to quantitatively detect a large number of metals. The free chromophore exhibits a yellow color, its X^^ occurring at 470 nm. Upon addition of thorium there is an almost imperceptible change to a yellow-orange color. Indicating formation of the 4:1 CI-PAN-thorium complex. This change is insufficient for quantitation of thorium. However, when copper is added to this reagent a wine red color Is observed (X^gx " 550 nm), indicating formation of the 2:1 complex (75) of CI-PAN with copper. When copper was added to 70.9 |xM CI-PAN, the absorbance at 560 nm was linear when plotted versus the analytical concentration of copper, figure 30. When equimolar amounts of copper and thorium were added, the observed spectra are identical to those observed in the absence of thorium. Therefore, with CI-PAN one can quantitatively measure free copper when free thorium is present in the same solution. The thorium is bound by CI-PAN, but does not significantly alter the absorbance of the solution at 560 nm, as long as enough CI-PAN is present to fill all the coordination sites of the metal present. If one assumes that all free copper is bound by CI-PAN then the concentration of added copper is linearly related to the absorbance at 560 nm. Since CI-PAN is water insoluble and la is insoluble in benzene, an equilibrium was established in which la remains in the basic buffer while CI-PAN remains in benzene. Any metal which is not bound by la will be extracted into the benzene layer by CI-PAN. In this way la will not interfere with the absorbance of CI-PAN, figure 31 . To the biphasic system composed of pH 9.2 buffer and benzene was added CI-PAN, copper, and varied amounts of la (0 to 0.5 equivalents relative to copper). If one plots the absorbance at 560 nm versus the analytical

PAGE 73

; ^. . <., J t .* u ' 67 0.35 -, 0.050.00 -11 0.30|«-"I without thorium ! ! vith thorium 0.25 -, e O 0.20y"' u u e « o n < 0.150.10corr.= 0.9987^ / , ^corr.= 0.9933 10 C(Cu) (jiM) T" 15 20 25 FIGURE 30 Standard Curve for Copper(ll)Acetate with CI-PAN (70.93 (iM) in the Absence and Presence of Thorium(IV)Chloride at the Same Concentration as Copper.

PAGE 74

68 red CI yellov HoXr =0 HO-^ — ^^ HH ^>^ N """^^ NH ^ FIGURE 31 Equilibrium for CI-PAN Copper Extraction.

PAGE 75

69 concentration of la, a linear relationship exists, wtiere more la causes less (CI-PANjj-Cu to form, resulting in a decrease In the absorbanca at 560 nm, figure 32. When an equivalent of thorium Is present along with copper, the change of absorbance is decreased. Indicating formation of the la-Thj complex, which results in more extracted copper and less decrease in the absorbance at 560 nm, figure 33. By treatment of the data in a fashion analogous to the two-ligand-one-metal competition experiment, x=2 and n=1, and the relative conditional binding constant expression expands to • (Th2-la)(Cuf ^ r»l T (Cu2-1a)(Th) Since Cji, = Cqu = 33.3 |iiVI, let Cj^^ = C^jj = C^^. Then the mass balance equations can be written as Cm =(Th) + 2(Th2-ia) Cm= (Cu) + 2-(Cu2-ia) and rearranged into the forms (Th) = Cm 2(Thg-la) (Cu) = Cm 2(Cuj-la) If we let a = the fraction of la associated with copper, then (Cu2-la) a-C^ (Thg-la) =(l-a)C^ ,and (Cu)=CM-2aC^ (Th) =CM-2{1-a)C^. If we let C^ = bC^ , then (Cu2-la) =abCM (Thg-la) -(1-a)bCM (Cu) Cm 2abCM (Th) =CM-2(1-a)bCM,and

PAGE 76

1.6-, 70 a 1.4 1.2 1.00.80.60.40.2corr.= -0.9936 0.0vith thorium • standard curve, vithout thorium 35 FIGURE 32 Results of la-Thorium-Copper Competition Experiments.

PAGE 77

71 '^^ ' -. < I . ! • + -5o z^ ge11av-orang« -:r|^0-..Th-..oJ^ ,:cuc II "t FIGURE 33 Equilibrim for CI-PAN Copper/Thorium Extraction.

PAGE 78

K » _ (l-a)(l-2ab) "'" a(l+2ab-2b)^ 72 2 Forthis case, a = 0.7414. When 0.1 equivalents of la are present; i.e., when b = 0.1. K ,g| = 0.2814, and K U-V ^ U-Cu/^el ='-54x1 0^3 M-2. These results are summarized in table 9. Pharmacology Vehicle Developmsnt It was observed that the catecholamide ligands are very soluble in alchohols and soluble at millimolar concentrations in ethyl acetate and nitrobenzene. Unfortunately, most organic solvents are not acceptible as vehicles for Injection. As previously stated, llgand solubility in aqueous media at physiologic pH was slight. In order to solublize our ligands in physiologic buffer, Chremophor RH 40 (BASF) was used. This is a FDA-approved surfactant whose generic name is glycerol polyethylene glycol oxystearate. It is used commercially for solubilization of fat soluble vitamins, essential oils, and some hydrophobic pharmaceutical preparations. The LDgg for mice injected intraperitoneally is >6.4 g/kg. Compounds were solubilized by maintaining a mixture of llgand and Chremophor at 55°C with frequent ultrasonic mixing. Once the mixtures were homogeneous in appearance-after one to two hours of this treatment-phosphate buffered saline at pH 7.4 (PBS) was added until the final solution was 40 percent Chremophor/PBS. These solutions were rather viscous, but suitable for injection.

PAGE 79

73 TABLE 9 Results of Conditional Binding Constant Calculations for an H-Sfiaped Ligand and ThCl4. Ligand(s) Metal(sj Fraction Bound By First Competitor K'rel^ Conditional Binding Constant for the First Competitor CpdII Cu — — 1.2 x1015m-1 2Z/Cpd II Cu 0.6909 2.327 2.79x10^5 1^-1 12/21 Cu 0.9177 1 .96 X 1018 m1 5.47x1033m-2 la Th/Cu 0.2586 0.2814 1.54x1033M-2 ^Calculated for one equivalent of each competitor and 0.1 equivalent of the substance for which they are competing, in 1.00 F ammonia buffer, pH 9.2. .V -i • :.} '-jr.

PAGE 80

74 BiolOQical Toxiritv Any therapeutic device must be nontoxic, a situation not realized with other octacoordinafe actinide cheiators. In l
PAGE 81

75 Urinary clearance The ability of DTPA, Compound II, and H-shaped llgands to stimulate clearance of thorium in the urine was evaluated. Thorium (1 nmole in 0.5 mL in 0.09 percent citrate buffer) was first administered to rats by intraperitoneal injection. One hour later a chelating agent (200 nmoles/kg) was administered intraperitoneally in the 40 percent Cremophor/PBS vehicle. Animals were placed in metabolic cages with food and water ad libitum. Urine samples were collected at six hour intervals and assayed for thorium by scintillation counting of the metal's alpha radiation. The data indicate that DTPA is effective at stimulating thorium excretion in the urine, and that this process is very rapid, figure 34. The fact that most of the DTPA-stimulated clearance takes place within the first six hours is consistent with the short half-life {20 minutes) of this dnjg in the body. Although the catecholamide chelators tested in this way ( ic-f and Im) stimulated thorium output in the urine to a very small extent, if we make the assumption that additional methylenes increase lipophilicity, a staicture-activity relationship exists between the lipophilicity of the ligand and the percent urinary thorium excretion. As one would expect, the analog presumed to be the least lipophilic, li stimulated the most urinary excretion while l£ was the least effective. To determine if a connection exists between this result and the lipophilicity or degree of ionization of the ligand, the partition coefficient for two ligands-la and H-were measured in octanol/phosphate buffered saline, pH 7.4 (PBS). As expected, l^was found to be more lipophilic than IL its partition coefficient being 1 .63 times greater, table 1 0. Fecal clBarancfl The inability of our catecholamide actinide chelators to stimulate thorium excretion via the urine came as no great surprise. It has been shown that highly lipophilic chelators are excreted as their plutonium complexes almost exclusively in the feces (76). Our llgands, of course, fall into this lipophilic class. Furthermore, experience with hexacoordinate catecholamide iron chelators has shown that these lipophilic compounds are also excreted mainly in the feces (77).

PAGE 82

10 98 — 776 : 643 2 1 Cpd II (3) 1c (4) Id (5) 1e(4) If (4) \.:/-V 1m (4) / f" ,' CONTROL (9) FIGURE 34 Effects of Catecholamide Chelators on the Urinary Excretion of Thorium in Rats Over 48 Hours, at a Dosage of 200 nmoles/l
PAGE 83

77 '.'' \ C TABLE 10 Results o( Partition Coefficient Measurements (or Uganda le and J(,. ''320nm [liflandJoetanol Compound octanol PBS octanol PBS [ligandlpgs \e 3I3nm 3I7nm 12,800 13,320 39,96 » 313nm 319nm 12,840 13,040 24.50

PAGE 84

78 Considering the structural similarities of our catecholamide actinide chelators to these iron chelators, and their greater molecular weight, one would predict that these compounds would be excreted by the fecal route as well. The potential health hazard of ^•'"thorium, the isotope used for these studies, is well recognized; precautions and procedures for its safe use have been implemented to avoid contamination of individuals and working areas. Because of the necessary restrictions in handling thorium-containing samples, a safe technique for measuring the thorium content of feces has not been developed. Additionally, it is difficult to gather kinetic data from fecal samples, due to irregularity of sample excretion. Biliary clearance To avoid the handling difficulties associated with measuring actinide output in feces, we have developed a series of experiments to collect and assay biliary tisues and fluids before and after chelation therapy. Gall bladder excision. As an initial experiment, 36 mice were administered thorium (50,000 dpm) intraperitoneally, followed one hour later by an H-shaped ligand (200 jimoles/kg in 0.5 mL 40 percent Chremophor/PBS). At three hour intervals, 3 mice were sacrificed, and their gall bladders were removed and assayed for thorium content. Since the gall bladder acts only as a bile storage site, the amount of thorium in the gall bladder would depend on the diet composition, the amount consumed, and when it was consumed. This experiment will therefore only qualitatively indicate whether or not biliary excretion of thorium can be stimulated by H-shaped ligands. Mice that recieved an H-shaped ligand demonstrated approximately a three-fold increase in the thorium content of their gall bladders, figure 35. This data suggests that our chelators are excreted by the biliary route, and are effective actinide chelators. Bile duct canniilatinn In order to get quantitative excretion data and accurate kinetic data it was decided to collect bile for several days. This was accomplished by placing an indwelling cannula into the bile duct of adult male Sprague-Dawley rats, allowing for continuous collection

PAGE 85

79 1000-. 900800E a. u 700e a) c 600500u E 3 400w 13002001000^ Htreated animals [I Control animals 1^ 10 15 20 T T 25 30 35 Time of Sacrifice (hours) I ' I 40 45 50 FIGURE 35 Effect of U on the Tfiorium Content of IVIouse Gall Bladders Over 48 Hours, at a Dosage of 200 iimoles/kg. Chelators Were Administered Intraperitoneally One Hour After Intraperitoneal Injection of ^•'"Thorium (50,000 cpm).

PAGE 86

80 Of bile. By using the rat, one does not face the problems associated with the gall bladder since this organ is absent. The cannula emerges from the incision and, using a skin-tunneling needle, is directed under the skin to behind the rat's neck. The emerging cannula is directed through a torque-transmitting spring tether to a fluid swivel located above the metabolic cage. The animal is able to move about freely in the cage while urine and bile were collected. Urine was collected for 24 hour intervals while bile was collected for three hour intervals. Although more involved, this technique allows us to obtain very reproducible data for metal excretion. When the animal began to recover from anethesia and good bile flow had been established, thorium (l nmole) was administered intraperitoneally. One hour later a chelator was given also intraperitoneally ' (100 or 200 nmoles/kg). Surgical techniques were performed by Dr. N. L. Scarborough, Dr. S. • A. Prudencio, and Ms. Kady Crist. I assisted in sample preparation for scintillation counting. ( To study the effect of altering the spermidine portion of H-shaped ligands, M la, and if were tested. It was found that the assymmetric SPD backbone is much more effective at stimulating thorium elimination via the bile than either symmetric backbone, figure 36. Since both addition to and removal from the spermidine backbone of a single methylene group causes a significant decrease in biliary thorium removal, it is reasonable to assume that any further alterations in a, b, c, and d (figure 5) will result in a further decrease in thorium excretion. To examine the role of the connecting diacid, l£, U and Im were tested. The glutaryl moiety appears to be better than either the succinyl or adipoyi diacids. Therefore, for the same reasons, no other diacids were tested. Total thorium removal Is shown in figure 37. Combination chelator thsranv The data show that H-shaped ligands stimulate thorium excretion almost exclusively in the bile whereas DTPA leads to urinary clearance of the actinide. Based on these findings we felt that perhaps if both dnjgs were administered in combination, the total actinide clearance would increase. To test this concept, the bile duct cannulation procedure was employed. In this experiment, rats were administered thorium intraperitoneally (1 nmole) followed by DTPA (100

PAGE 87

40-, 81 FIGURE 36 Effects of Catecholamide Chelators on the Biliary Excretion of Thorium in Rats Over 48 Hours, at a Dosage of 200 |imoles/kg each. Chelators were Administered Intraperitoneally One Hour After Intraperitoneal Injection of ^soxhonum Numbers in Parentheses Represent the Number of Rats Tested.

PAGE 88

35-1,—, 30L^ 252082 E £ U 1510Time (hours) FIGURE 37 Effects of Catecholamide Chelators on the Total Excretion of Thorium in Rats Over 48 Hours, at a Dosage of 200 jimoles/kg each. Chelators were Administered Intraperitoneally One Hour After Intraperitoneal Injection of 230jhorium, Numbers in Parentheses Represent the Number of Rats Tested.

PAGE 89

83 ixmoles per kg) and/or 1[ (1 00 iimoles/kg), also intraperitoneally. Urine and bile were collected and assayed for thorium content, as previously described. Urinary clearance. In the urine DTPA was able to renrrave 28.0 percent of the injected dose of thorium when administered alone at this lower dose, figure 38. When given in combination with 11, only 22.3 percent of the injected dose was found in the urine, a 20.4 percent reduction. Biliary clearance . In the bile a similar situation was observed, figure 39. U alone removed 23.6 percent of the injected dose of metal, but only 17.8 percent was removed in the bile by the combination of chelators, a 24.6 percent reduction. Total clearance . The data may indicate that 11 does indeed access different body compartments than DTPA, as evidenced by increased thorium excretion. Also, it appears that there are some compartments that are frequented by both dnjgs, as reflected by the fact that the combination of drugs removes less than an additive amount of metal. If DTPA and Jl were independent, then the metal removed by DTPA in the urine when administered by Itself added to the metal removed by H In the bile when administered by itself would be approximately equal to the total output when the daigs are administered in combination. Since the actual output of metal is 21.6 percent less than the additive amount, there is some overiap in the volumes of distribution for the two chelators. That is, there are some body compartments which are accessible to both chelators, some which are accessible to only one chelator, and some which are not accessible to either chelator. When administered in combination, in the compartments accessible to txith chelators a portion of the available thorium is bound by each chelator, tor example, some of the thorium which would have been bound by DTPA and eliminated in the urine is bound by IL causing an apparent decrease in the effectiveness of DTPA. Although combination therapy decreased the appparent effectiveness of each chelator, the total excretion did increase, figure 40.

PAGE 90

^^ 84 a > e S o » Q. > 3 E £ 3 u 30-, 25201510\m] DTPA DTPA and If FIGURE 38 Effects of 11 or DTPA on ttie Urinary Excretion of Thorium in Rats Over 48 Hours, at a Dosage of 100 iimoies/kg eacfi. Clielators were Administered Intraperitoneally One Hour After Intraperitoneal Injection of 230Thorium. Numbers in Parentlieses Represent the Number of Rats Tested.

PAGE 91

85 a > e S 0) oc 0) u L. 0) Q. 0) > 25-1 2015: 105[] DTPA DTPA and If 6 12 18 24 30 36 Time (hrs.) 42 48 FIGURE 39 Effects of II or DTPA on the Biliary Excretion of Thorium in Rats Over 48 Hours, at a Dosage of 100 p.moles/kg each. Chelators were Administered Intraperitoneally One Hour After Intraperitoneal Injection of 230Thorium. Numbers in Parentheses Represent the Number of Rats Tested.

PAGE 92

'^ 'i 86 60-, > e S e u u a> Q. a> > u [] DTPA DTPA and If experimental [o] DTPA and If theoretical ! . 24 Time (hrs.) FIGURE 40 Effects of 1! or DTPA on the Total Excretion of Tfiorium in Rats Over 48 Hours, at a Dosage of 100 nmoles/kg each. Chelators were Administered Intraperitoneally One Hour After Intraperitoneal Injection of ^^^Thorium. Numbers in Parentheses Represent the Number of Rats Tested.

PAGE 93

CHAPTER III ELUCIDATION OF THE SOLUTION STRUCTURE OF POLYAMINES IN RELATION TO THE MECHANISM OF CELLULAR UPTAKE Backnrniinrl Bergeron, Porter and Stolowich have provided a great deal of information regarding the structural requirements, preferences, and limitations associated with substrates of the polyamine transport apparatus (7, 8, 9, 10, 78, 79). BackhnnR VariatinnR Eleven stnjcfural analogs of spermidine and putresoine were synthesized and studied both for their ability to inhibit [^Hl-spermidine uptake by the polyamine transport apparatus of LI 21 cells and/or their ability to stimulate grovirth in vitro in polyamine-depleted LI 21 cancer cells, table 11. One should note that the analogs used in this study differ from the naturallyoccurring polyamines only in the number of methylene units separating the various amine functionalities. Uptake inhibition was found to be competitive in ail oases, homospermidine being the most competitive polyamine. Any further alterations in the number of methylene units in the polyamine backbone resulted in diminished competitive inhibition. Replacement of the secondary nitrogen of spermidine by a methylene unit produced a molecule which was an effective inhibitor of [^HJ-spermidine uptake, in growth stimulation experiments, however, little effect was caused by these analogs. This study suggests that the secondary amine is essential for polyamine functions while Its importance is less pronounced for uptake. N^-Substitiitinn ; ' '"^' '^'''y'a'iPH N''-alkylated spermidine derivatives were synthesized to determine both how N-t-alkylation would affect uptake characteristics and how large of an alkyl substituent would be tolerated by the polyamine transport apparatus, table 12. It was found that the spermidine 87

PAGE 94

88 TABLE 11 Inhibition of [^HJ-Spermidine Uptake into L1210 Cells by Polyamines or Ttieir Homoiogs. Homo log r^Hl-Soenniriine 1 Iptake Intiihitinnb Picomoles per Percent of m^cells-min Cnntml none 56.1 100 DAgO >500 54.0 96 DA4(putresclne) 171.3 44.6 80 t»5 459.0 54.0 96 °^i 63.2 40.1 71 DA7 18.2 23.0 41 DAg 22.1 25.2 45 sTAg^ 8.4 16.1 29 4TA4 3.5 7.3 13 3TAS 12.3 19.8 35 3TA6 13.1 19.6 35 3TA7 13.0 20.0 36 3TA8 . ' 7.8 13.5 24 Spermine 9.1 17.1 30 Reference 79. ^Prewarmed Li 210 cells{5xl06) were incubated for 20 min in 1 mL of RPMI-1640 media containing 2 percent Hepes-Mops and 0.2, 0.5, 1.0, 2.0, 5.0, or 10.0 nM [3H]-spermidine and 100 nM fiomolog. Uptake data were fitted by computer for competitive inhibition; the Michaelis constant for spermidine uptake was 2.0 nM, and the maximum velocity of the reaction was 117 pnxDie/min per milligram of protein. ''Cells were incubated for 20 minutes at 370C with 10 ^M [^HJ-spermidine pluslOO nM putrescine or spermidine homolog. ''The abbreviation for putrescine homoiogs having the general structure NH2(CH2)nNH2 is DAp (for diamine) where n=3 to 8. <^The abbreviation for spermidine homoiogs having the general structure NH2(CH2)nNH(CH2)n.NH2 is „TA „ (for triamine) where n is 3 or 4 and n' is 3 to 8.

PAGE 95

89 TABLE 12 Effects of N'*-Alkylated Spermidines on [^HJ-Spermidine Transport into Ascites LI 210 Leukemia Cells. Suhstiliient ReiaivB LIptakRa Ji^uMP-H -CH3 -CHp'CHq -(CH2)5-CH3 -CHj-CgHg 100 17 14 69 67 3.4 3.1 34 39 Reference 79. acells were incubated for 20 minutes at 37°C with 10 nM [^HJ-spermidine pluslOOjiM analog. Prewarmed L1210 cells(5x10S) were incubated for 20 min in 1 mL of RPMI-1640 media containing 2 percent Hepes-Mops and 0.2, 0.5, 1.0, 2.0, 5.0, or 10.0 nM pHJ-spermidine and 10 or 100 jiM analog.

PAGE 96

90 molecule could be alkylated rather extensively at ttie N*-position and still participate effectively in experiments related to uptake, including inhibition of [^H]-spermidine uptake, prevention of MGBG-induced cytotoxicity, and intracellular detection by HPLC. K/1GBG, methylglyoxal-bisguanylhydrazone, is an inhibitor of polyamine biosynthesis. Acvlation . N^-acylation had more pronounced effects upon uptake than did alkylation. A series of N'*-acyl spermidine derivatives and their alkyl counterparts--for example, ethyl and acetyl-were prepared and evaluated for their ability to compete w/ith [^H]-spermidine for cellular uptake, table 13. It was found that both series of derivatives could behave as polyamine transport apparatus substrates, but there was an obvious preference for the N'*-alkylated derivatives. In addition, the secondary nitrgen could be extensively modified without a significant decrease in inhibitory effect, as evidenced by the uptake of derivatives such as N'*-hexyl and N^-benzyl spermidine. This is consistent with the conclusions of the previous study, where the primary amines appeared to be more critical for uptake. N^N^ -Bis Substitution A series of N'' ,N^-bis acyl derivatives were prepared and tested in a similar manner, table 14. Of the derivatives tested--N',N^-bis-methyi and N\N^-bis-formyl spermidine were not tested-N' ,N°-bis ethyl spermidine was the only molecule capable of competing effectively with spermidine for uptake; any modifteation to this derivative rendered the molecule incapable of behaving as a polyamine transport apparatus substrate. TTie Role of Protonafion State: Pot entiometrin Measurements In an attempt to more accurately define the role of charge in polyamine uptake, the protonation state of N'*-benzylspermidine, as well as the norand homospermidine analogs, was studied. To achieve this the pK^ of each nitrogen of the three analogs, and homospermidine itself was determined potentiometrically. The benzyl analogs were chosen because considerable uptake data has been accumulated on these compounds.

PAGE 97

91 i..' TABLE 13 Comparison Between the Effects of N^-Alkylated and N*-Acylated Spermidines on pH]-Spermidine Transport into Ascites LI 210 Leul
PAGE 98

92 TABLE 14 Effects of Terminally Modified Spermidines on [^HJ-Spennldlne Uptake Into L1210 Leukemia Cells. Siibstitijent Relative Uptake^ K|(iiM^ -H 100 -{C=0)-CH3 91 508 -CH2-CH3 69 62 -(C=0)-CH2CH3 92 550 -CH2"CH2CH2 80 117 -tBOC 91 521 -2,3-Dimethoxybenzoyl 91 256 Reference 79. ^ells were incubated for 20 minutes at 37°C with 10 iM [^HJ-spermidine pluslOOnlul analog. ''Prewarmed LI 210 cells(5x10S) were incubated for 20 min in 1 mL of RPtull-1640 media containing 2 percent Hepes-Mops and 0.2, 0.5, 1.0, 2.0, 5.0, or 10.0 nM [3H]-spermidine and 10 orlOO nM analog. *-...-• /•,

PAGE 99

93 When an aqueous titration was completed, a set of data consisting of pH measurements versus volume of titrant added was generated. The computer program PHFIT (D. Leussing, private communication) was used to analyze this data. PHFIT handles as many as four independent species and fifteen associated species. The program first calculates, by way of a standard Newton-Raphson iteration, the distribution of species at each data point based on the total concentration of species at that point and the estimated pK^ provided for each specie. Thus, an Initial theoretical titration curve is obtained; assuming all significant equilibria have been considered, the difference between this calculated curve and the observed data is then minimized by further refinement of the initial estimates. To calculate the acid dissociation constants, one must provide estimates for these constants, the analytical concentrations of all Independent species involved in the chemical equilibria, the autodissociation constant (K^) for water under the experimental conditions, and the activity coefficient (7) of H'*' under the experimental conditions. Also, certain assumptions are made-that the titrations performed under conditions of constant temperature and ionic strength. By using a water-jacketed cell in conjunction with a constant temperature water bath, the condition of constant temperature (25°C) is satisfied. To insure the condition of constant ionic strength, an excess of a strong electrolyte, potassium chbride, is added to both the solution of interest and the titrant, so that changes in the ionic state of the acid and titrant are insignificant in comparison to the total ion population. In practice this means that the ligand concentration (approximately 2x10"^ M) is much less than the ionic strength of the solution (0.1 M KCI). For an aqueous solution at 25°C with an ionic strength of 0.1 M, K^ and vare lO"''^-^^^ and 0.78 respectively (80, 81). These two terms are related by the expression K^=[H*][0H-1= -^Sii^

PAGE 100

-Htpf-^. 94 Where a^, and aQH are the activity of H* and OH' respectively. The results of the pK^ studies are presented in table 15 along with other selected polyamine pK^'s. Knowing the pK^'s, it is possible to calculate the concentration of the mono-, di-, and trication of a polyamine in solution at any given pH, using the equations , , Ki Ko i^T [H*]*+ K,1H*]^+ K,K2[H*1+ KiKjKj a(i)= — = ^i^^i^!^^ [H*]^+ K,[H*l2+ K,K2[H*]+ K1K2K3 CX(2)= ^ [H*]'+ K,[H*]2+ K,K2[H*]+ K,K2K3 a(3)=^ Pl [H*l^+ KitH*]^* K,K2[H*]+ K,K2K3 where a(i) is the fraction of polyamine existing as the 1+ cation. Recall that the order of uptake inhibition Is homospermidine > spermidine > norspermidine. It is clear that there are substantial differences In the relative concentrations of polycations at a particular pH. For example, at pH 7.4 only 67.4 percent of norspermidine is in the form of the trication while 89.9 percent of spermidine is in this form and 97.1 percent of homospermidine is triprotonated, figure 41. Similar trends are seen with the benzyl compounds, wherel 1 .4 percent, 34.3 percent, and 69.1 percent, respectively, exist as the trication. The values for a(3) seem to correlate well with C. W. Porter's in vitro measurements of inhibition constants (7, 8). By multiplying the actual concentration of polyamine by the fraction of the polyamine which exists as the trication at pH 7.4 one can determine the "effective" concentration of polyamine. When one recalculates the Kj's of various polyamines based on these con-ected concentrations, the large differences in affinity for the transport apparatus are greatly diminished, table 16. However, the role of a 3+ cation is unclear in view of the fact that

PAGE 101

95 Q TABLE 15 'f''* Results of Acid-base Titrations of fvlH2-(CH2)a-N-{CH2)b-NH2 a b X pKi pK2 pKg # of Titrations^ 3 3 Hj 10.579 (56)'> 9.793 (78) 6.507 (66) 5 3 4 H2 10.647 (29) 9.836 (66) 7.118 (100) 6 4 4 Hj 10.835 (32) 10.111 (63) 7.752 (135) 6 3 3-0 9.979 (190) 7.763 (350) — 5 3 4 =0 10.511 (64) 8.970 (76) — 6 ^Performed at 250C in 0.1 M KCI, [polyamine] = 0.001 M. ''Numbers last digit. "Numbers in parenttieses represent the standard deviation associated witfi tfie

PAGE 102

96 NSPD SPD HSPD l^iH BNSPD BSPD 1 il liliiiSiiiiiiiijIiliiiliiiiiiiiiiiiiiiiiiiiiiiiiiii^ BHSPD iiiii i=2 §1=3 ' 1 ' 1 ' 1 ' 1 ' 1 ' 1 I'M 1 ' 1 Percent Existing as the 1+ Cation FIGURE 41 Distribution of Protonated Species at pH 7.4 for Several Triamines.

PAGE 103

97 TABLE 16 Correlation Between the Polyamine Inhibition Constant for Spermidine Uptake and the Fraction of Polyamine Existing as the Trication at Physiological pH. Ki* ai3p K|Xo(31 BNSPD<= 135 mM 0.114 15.4 mM BSPD 36 0.343 12.3 BHSPD 14 0.691 9.7 NSPO 8.4 0.674 5.7 H8PD 3.5 0.971 3.4 ^Reference 8. ''Calculated from measured pKa's, see text. "^Abbreviations used are: BNSPD N''-benzyl-norspermidine; BSPD N^-benzylspermidine; BHSPD N^-benzyl-homospermidine; NSPD norspermidine; HSPD homospermidine. .'..'• 1 *' r * •

PAGE 104

98 long chain diamines, which would be 2+ cations, compete so effectively for uptake. FurthernHsre, In looking at the concentration of the 2+ cation of triamlnes and competitton with spermidine for uptake an Inverse relationship seems to exist. Finally, there Is so little 1+ cation and free base that their roles would seem insignificant, it may be that the role of the trication is to bring the polyamine to the surface of the negatively charged cell membrane. Once associated with the membrane, other factors may come into play. Clearly, charge is not the sole explanation for the selectivity In polyamine uptake. The Role of Hydrogen Bonding The RRasnnini? Both the N'^-benzoyl and the N^-acetyl compound have restricted rotation about the amide C-N bond while the corresponding reduced compounds have more freedom of rotation. This implies that although the position in which the acetyl group Is held does not substantially hinder uptake, the benzoyl group does. A consideration of how charges on the terminal nitrogens of spermidine might be utilized to hold the polyamine in a particular conformation and the possible conformations which might allow a benzoyl versus a benzyl functionality-or, to a lesser extent, an acetyl versus an ethyl functionality-to interfere with this binding led us to postulate the importance of a hydrogen-txinded cyclic conformation for polyamine transport. It is clear that a trication cannot form intramolecular hydrogen bonds and that spemiidine exists mainly as a trication in an aqueous environment at pH 7.4. it may be that the trication is energetically unfavorable at the ceil membrane and that less charged species exist. It has been shown that the strength of a hydrogen bond is maximal when the bond is colinear; i.e., when the X-H-X angle formed by the nuclei is 180 degrees. In the case of intramolecular hydrogen bonds the optimal angle cannot always be achieved, resulting In a weaker bond. For example, hydrogen bonding of the terminal hydroxyls of compounds of the general stnjcture HO-ICHjln-OH have been measured by Infrared spectroscopic analysis (82). The data indicates that 1 ,4-butanediol (n=4) forms a stronger intramolecular hydrogen bond

PAGE 105

99 than does 1,3-propanediol (n.3), table 17. For terminal diamines H2N-(CH2)n-NH2 acid-base titrimetry has provided similar data; i.e., 1,4-diaminobutane is more basic than 1 ,3-diaminopropane (83). This may be due to increased electrostatic repulsion in the case of 1 ,3-diaminopropane and/or due to the presence of a stronger intramolecular hydrogen bond in the case of 1,4-diaminobutane; a seven-membered intramoieculariy hydrogen bonded ring can nmre closely achieve a conformation which optimizes the N-H~N bond angle. In previous uptake experiments, homospermidine inhibited (^H]-spermidine uptake better than norspermidine, implying that the aminobutyl sidechain is more easily recognized by the polyamine transport apparatus than the aminopropyl sidechain. Since the aminobutyl sidechain can form a more stable intramolecular hydrogen bond, a greater percentage of homospermidine may exist as the cyclic hydrogen-bonded conformer on the cell surface, leading to better recognition and inhibition. Although in the case of N'' ,N^-diaminoheptane and N^N^-diaminooctane no central nitrogen with which to form an intramolecular hydrogen bond exists, a psuedo-cyclic conformation may still be achieved by folding of these molecules. In addition, N''-benzoyl spermidine cannot compete effectively for uptake. This may reflect the inability of this molecule to achieve a cyclic conformation for steric reasons. However, the con-esponding benzyl compound can compete very well. Although this analog will exert steric restrictions similar to the benzoyl analog, the lone electron pair associated with the central amine may assist formation of a cyclic confomier by fonming an intramolecular hydrogen bond. Also, because of Increased freedom of rotation which exists about the bonds unlike that for the corresponding amide, the amine may be able to more easily adopt the required comformation. The Evidence: Hflxahv droovrimidinBS ''h-NMR were considered in order to show that the polyamines do indeed form cyclic conformers; however, it was clear this wouk) not definitively indicate whether or not the LI 210 cell polyamine transport apparatus transports sixand/or seven-membered hydrogen-bonded cyclic stnjctures; thus, the importance of cyclic confomiers in uptake had to be evaluated by

PAGE 106

100 TABLE 17 Infrared Data for Terminal Diols of the Stnjcture H0-(CH2)n-0H Vfrao Vbound ^v 3612 32 3558 78 3478 156 153 2 3644 3 3636 4 3634 S 3638 6 3638 Reference 84. mo148

PAGE 107

101 another method. It was decided to prepare cyclic methylene-bridged analogs of intramolecularly hydrogen Iwnded norspermidine and spermidine-hexahydropyrimidlne analogs, figure 42--and determine if these analogs would be transported by LI 210 cells. A hexahydropyrimidlne analog is easily synthesized from the appropriate amine and formaldehyde (13, 14). The product still possesses the polyamine backbone intact, but this backbone is "locked in" to a cyclic conformation by the methylene bridge. Since all three nitrogens remain amines, they can still bear positive charge. If cyclic conformers are involved in the uptake mechanism, these analogs should be readily recognized and transported by the polyamine transport apparatus. Two questions were adressed relative to these compounds, (1) Are they stable in solution, and (2) Do they compete for the polyamine transport apparatus? The stability was questioned simply because of the potentially reversible nature of hexahydropyrimidlne formation. The question was answered utilizing ^H-NMR. A deuterated version of the culture medium employed in the polyamine uptake studies (RPMI-1640) was prepared. N-(4-aminobutyl)-hexahydropyrimidine was added to the medium and a 300MHz ''h-NMR spectmn recorded both immediately and four hours later. A singlet characteristic of the hexahydropyrimidlne methylene bridge exhibits a chemical shift of 3.4 ppm downfield from TA/1S. The two spectra indicated that the methylene bridge of ABHHP was stable during the course of the biological studies. Since the mechanism of hydrolysis is the same for ABHHP and N-(3-aminopropyl)hexahydropyrimidine (APHHP), it was deemed unnecessary to repeat the experiment for the latter. Once having established the stability of hexahydropyrimidines under experimental conditions, spermidine competition studies were initiated. Briefly, LI 210 cells in log growth phase were exposed to ^^C-labelled spermidine at concentrations varying from 0.2 to 10 nM in the absence or presence of 10 |xl^ hexahydropyrimidlne analog. ^^C-spermidine cellular uptake was measured using scintillation procedures. The data were plotted according to the Hofstee method, figure 43, and the K|'s calculated from the equation

PAGE 108

102 nBHHP flPHHP FIGURE 42 Cyclic Spermidine Analogs.

PAGE 109

103 r a a. in 1086420-2-4-6-810p^ Control P^ APHHP P^ ABHHP [sPD]=v„„.(i5£5L)_,. I ' n ' I ' i ' I — I — 1 — I — I — I 1 100 200 300 400 500 600 700 800 [SPD] /Velocity (jimoles/liter-cpm) FIGURE 43 Hofstee Plot for Competitive Inhibition of '''c-Spermidine Uptake by Cyclic Polyamine Analogs. Cells Were Incubated for 20 Minutes at 37 Degrees with 0.0, 2.0, 4.0, 6.0, 8.0, or 10.0 nM [itCl-Spermidine and 10.0 (iM Analog.

PAGE 110

104 Utilizing a K^ of 0.77 [iM for spermidine, table 18. The magnitude of K,^ for spermidine and K, for norspermidine are somewhat different than those reported by C. W. Porter et al. (8). This difference can be attributed to the fact that our experiments utilized L1210 cells from tissue culture whereas previous work employed Ascites cells. However, in order to verify that the variance could not be assigned to the fact that we employed different concentrations of polyamines in our studies, a Dixon plot was generated, using norspermidine as a model. The data was plotted using a rearrangement of the Michealis-Menton equation for competitive inhibition 1 K„ [11 + V " _SV„K^_ The linearity of this plot, figure 44, suggests that the mode of inhibition is the same over the concentration range examined; i.e., K| is constant. The ability of the hexahydropyrimidines to prevent the growth of LI 210 cells was also measured. These analogs were found to be very effective at preventing cell division, as evidenced by their IC^^ values, table 18. This may be due to interference with or disruption of polyamine metabolism or function. Whatever the mode of action, it is clear from these data that the possibility of polyamines adopting cyclic conformations during transport is quite reasonable.

PAGE 111

105 TABLE 18 PerRftntofContmP SPD iiptakR inhihitinn ComDaind KfytMp IC5o/pMl= none 100 NSPD 29 4.67 ±0.33 0.7 , , . APHHP • -M^ '; 3.91 ±2.57 0.5 SPD ;n,,' J^ : 0.77 ± 0.31 (K J 1000.0"^ ABHHP 1.41 ±1.15 40.0 acells were incubated for 20 minutes at 37°C with 1 jiM (3H)-spermicline plusi 00 hM analog. ''Prewarmed L1210 C8lls(5x106) were incubated for 20 min in 1 mL of RPMI-1640 media containing 2 percent Hepes-Mops and 0.2, 0.5, 1.0, 2.0, 5.0, or 10.0 ^M pH]-spermidine and 10 orlOO nl^ analog. ^3x10'' cells/mL were incubated for 48 hours at 370C with to 1 00 ^H/l polyamine. "Reference 8.

PAGE 112

106 z.^t-<»f ,2.2E-41 V = 1.6E-4/ / *% 1.4E1.2E-4-4/ / / 3l / r >pe = 1 16E-6 ' / / corr. = 0.9975 I.OE-4> / ^ y-int. =8.56E-5 8.0E5^ ' ' 1 1 1 ' 1 ' 1 — '— [Norspermidlne] (jiM) FIGURE 44 Dixon Plot for Norspermidlne, a Representative Competitive Inhibitor of '•'C-Spermlcllne Uptake. Cells Were Incubated for 20 Minutes at 37 Degrees with 10 nlvl ['"CJ-Spermidine and 10, 40, 70, 100, or 130 nl^ Norspermidlne.

PAGE 113

CHAPTER IV EXPERIMENTAL DETAILS Synthetic Prnrsriiiras N'li'' -BiS(2.3-dimfilh0XYhen70vll-N ' *-succinvl norsngrmidina da) A solution of N\N^-bis(2,3-cllmethoxybenzoyl) norspermidine (4.95 g, 10.8 mmoles), triethylamine (5.4 g, 53 mmoles), and succinic anhydride (1.23 g, 13 mmoles) in 200 mL of CHjClj was refluxed for four hours prior to addition of 3-(dimethyiamino)propylamine (0.31 g, 3.0 mnrales). The resulting mixture was refluxed overnight. After cooling the reaction mixture was washed with 1M HCI (5 x 20 mL) and the combined aqueous washings were back-extracted with CHjClj (25 mL). The combined organic layers were dried with Na2S04, filtered, and evaporated to afford 5.37 g (88 percent) cmde product, which separated on silica gel (400 g, 10 percent MeOH:CHCl3) to yield 4.03 g (65 percent) pure product. 'H-NMR (CDCI3) 5 1.6-2.1 (m, 4H), 2.65 (s, 4H), 3.2-3.6 (m, 8H), 3.9 (s,12H),6.9-8.4 (m,8H), 10.6 (br, 1H). An analytical sample was prepared by loading lOOmg of product onto silica gel (15g, 12-25 percent MeOH:EtOAc). Analysis calculated for CggHgyNaOgHjO, Calc. 58.22 %C, 6.81 %H, 7.27 %N, found 58.44 %C, 6.73 %H, 7.24 %N. N'j^^-Bi5(2.3-d imsmoxvhsn70Vl)-N ^ -succinvlhnmnspermidinfif.^h> 2b was prepared and purified in the same manner as aa (65 percent). '' H-NMR (CDCU) 8 1.4-1.9 (m, 8H), 2.55 (S, 4H), 3.2-3.7 (m, 8H), 3.95 (S,12H), 6.9-8.2 (m,8H), 10.6 (br, 1H). Analysis calculated for CgoH^^NjOgHjO, Calc. 59.49 %C, 7.16 %H, 6.94 %N, found 59.30 %C, 7.13 %H, 6.92 %N. 107

PAGE 114

108 N^i:l^BiS(2,3-dimethQXvhen70vn-N' * -Qlutarvl norspfirmidinfi (^r^ 2c was prepared and purified in the same manner as 22 (65 percent). ''H-NMR (CDCI3) 5 1.6-2.2 (m, 6H), 2.2-2,6 {m, 4H), 3.2-3.8 (m, 8H), 3.95 (s,12H),7.0-8.3 {m,8H), 10.6 (br, 1H). Analysis calculated for CggHggNgOgHjO, Calc. 58.87 %C, 6.99 %H, 7.10 %N, found 58.78 %C, 7.05 %H, 7.05 %N. N^N3-Bisf2.3-riimethn>vhPn7nvh-N|5 .Qlutarvl homnsnflrmiriinp IM\ 2d was prepared and purified in the same manner as 2a (65 percent). '' H-NMR (CDCU) 5 1.4-1.9 (m, 8H), 2.3-2.6 (m, 6H), 3.2-3.7 (m, 8H), 3.95 (s,12H), 6.9-8.2 (m,8H), 10.5 (br, 1H). Analysis calculated for Cj^H^gNjOgHjO, Calc. 60.08 %C, 7.32 %H, 6.78 %N, found 60.05 %C, 7.31 %H, 6.78 %N. Bisi fN. N-bis (2.3-dimethoxvhRn7nvl-.^aminonrnnvh; succinamida f2al A solution of N'',N^-bis(2,3-dimethoxybenzoyl) norspermidine (2.81 g, 5.96 mmoles) and triethylamine (0.9 g, 8.9 mmoles) in 150 mL CHjClj was stirred at 0°C under a nitrogen atmosphere. Dropwise addition of succinyl dichloride (0.45 g, 2.90 mmoles) in 50 mL CHpCl, was completed before allowing the reaction mixture to warm slowly to room temperature and stirring for 24 hours. After cooling to 0°C, 3N HCI (75 mL) was added. Thirty minutes later the organic phase was washed with 3N HCI (3 x 75 mL), dried over Na2S04, filtered, and evaporated to yield 2.88 g (97 percent) crude product, which was chromatographed on silica gel (150 g, 10:45:45 percent MeOH:EtOAc:CHCl3) to yield 1.98 g (85 percent) of pure product. An analytical sample was prepared by HPLC on silica using the same solvent system. ^ H-NMR (CDCI3) 8 1.5-2.2 (m, 8H), 2.2-2.8 (m, 4H), 3.0-3.7 (m, 16H), 3.8 (s, 24H), 6.7-8.3 (m,16H). Analysis calculated for C52Hg3NgO,4, Calc. 62.39 %C, 6.85 %H, 8.39 %N, found 62.13 %C, 6.89 %H, 8.34 %N. BiSlN. N-biS(2.3-dimethoxvbenzovl-4-a minohiitvni succinamids IPh\ 2I2 was prepared and purified in the same manner as 2a (85 percent). '' H-NMR (CDCU) 8 1.3-1.8 (m, 16H), 2.1-2.8 (m, 4H), 3.0-3.7 (m. 16H), 3.8 (s, 24H), 6.8-8.1 (m, 16H). Analysis

PAGE 115

109 calculated for C55H7gNgOi4, Calc. 63.62 %C, 7.25 %H, 7.95 %N, found 63.66 %C, 7.29 %H, 7.95 %N. BiS[N-(2.3-dimelhQxvbenzovl-4-amlnobutvn-N-f2 3-riim s thoxvbenzovl-3-aminonmnvn] sucr.inamidg(?r.> 2C was prepared and purified In the same manner as 2a (85 percent). • H-NMR (CDCy 5 1.2-2.1 (m, 12H), 2.1-2.9 (m, 4H), 2.9-3.6 (m, 16H), 3.7 (s, 24H), 6.6-8.1 (m, 16H). Analysis calculated for Cs^Hj2NsO^^HjO, Calc. 61.94 %C, 7.12 %H, 8.03 %N, found 61.74 %C, 7.15 %H, 7.92 %N. BisfN. N-biS (2.3-dimethoxvhBn70vl-3-a minoDroDvh] ^lutaramiriB i7!d\ 2d was prepared and purified in the same manner as 2a (85 percent). ^ H-NMR (CDCI3) 5 1.5-2.1 (m, 8H), 2.1-2.6 (m, 6H), 3.0-3.7 (m, 16H), 3.8 (s, 24H), 6.7-8.3 {m, 16H). Analysis calculated for CsgH^gNgOi 4, Calc. 62.71 %C, 6.95 %H, 8.28 %N, found 62.51 %C, 7.03 %H, 8.17 %N. ' •' BiSf [N. N-biS (2.3-dimethnxvbenzovl-4-aminohiitvni fflMt a ramide f?fi> 2a was prepared and purified in the same manner as 2a {85 percent). '' H-NMR (CDCI,) 8 1.3-1.8 (m, 16H), 2.1-2.6 (m, 6H), 3.8 (s, 24H), 6.8-8.1 (m, 16H). Analysis calculated for *^57^^78'^60l4'2H20, Calc. 61.83%C, 7.46 %H, 7.59 %N, found 61.92 %C, 7.21 %H, 7.59 %N. Bi5rN-(2 , 3-ClimethgXVbfln70Vl-4-aminobiitvh-N-f2.3-riin i ethoxvhfin7nvl-3-aminnnrnpvH] alutaramirig (?n 21 was prepared and purified in the same manner as 2a (85 percent). '^ H-NMR (CDCL) 5 1.2-2.1 (m, 12H), 2.1-2.6 (m, 6H), 2.9-3.6 (m, 16H), 3.7 (s, 24H). 6.6-8.1 (m, 16H). Analysis calculatedforC55H74NgO,4-2H20, Calc. 61.21 %C, 7.28 %H, 7.79 %N, found 61.11 %C, 6.90 %H, 7.66 %N. N-fBi5f2.3-ClinielhOXVhen70vl-3-aminnnr n DVlll-N'-rN-r? 3-dimethnxvhenzovl4-aminnhi ityh-\|(2.3-dimBlhOXvhRn7nvl-3-aminnprn pvlllsiir(;inamiriP^Pnl A soluton of ac (1.0 g, 1.74 mmoles), dicyclohexylcarbodiimide (0.38 g, 1.85 mmoles), 4-(dimethylamino)pyridine (20 mg), and N\N8-bis(2,3dimethoxybenzoyOspermidine (1.02 g,

PAGE 116

110 2.09 mmoles) in CHgClj (150 mL) was stirred for 24 hours. The reaction mixture was filtered and evaporated to afford 1.97g cnjde oil which was purified b silica gel chromatography (20 g, eluted with 0:50:50 to 10:45:45 EtOH: EtOAc:CHCl3) to yield 1.68 g of pure product (85 percent). An analytical sample was prepared by HPLC on silica using the same solvent system. ''h-NMR (CDCy 8 1.5-2.2 (m, 10H),2.65 (s, 4H), 3.1-3.7 (m, 16H), 3.8 (s, 24H),6.9-7.7 (m,12H). Analysis calculated for C53H7(,N50,4H20, Calc. 61.61 %C, 7.02 %H, 8.13 %N, found 61.54 %C, 7.19 %H,7.84%N. Cn 1 ' ' N-[Bi5(2 , 3-dimethPXVben7QVl-3-aminoDroDvni-N'-rhisf ? .3-dimethn)fvhen7nvl-4-aminnhMtvn] succinamidfl(2h> 2tl was prepared and purified In the same manner as 2fl (85 percent). '•h-NMR (CDCU 8 1.5-2.2 (m, 12H), 2.65 (s, 4H), 3.1-3.7 (m, 16H), 3.85 (s, 24H), 6.9-7.7 (m, 12H). Analysis calculated for C54H72N60,4H20, Calc. 61.94 %C, 7.12%H, 8.03 %N, found 61.83 %C, 7,14 %H, 8.03 %N. ' : N-rBis(2.3-dimethoxvhflnzovl-4-aminn hiitvni-N'-[N-(2.3-dimfithnyvbenzovl-4-aminobiitvn-N(2.3-dimethoxvben70vl-3-aminnp roDvnisi]minamiriprPH 2i was prepared and purified in the same manner as 2a (85 percent). '' H-NMR (CDCI,) 5 1.5-2.2 (m, 14H), 2.65 (s, 4H), 3.1-3.7 (m, 16H), 3.85 (s, 24H), 6.9-7.7 (m, 12H). Analysis calculated for C55H74N5O14, Calc. 63.32 %C, 7.15 %H, 8.06 %N, found 63.36 %C, 7.32 %H, 7.89 %N. N-rBi5(2 . 3-ilirnelh0XVhfin7OVl-3-aminopronvm-N'-rN-fpr!d imfithnxvhpn7ovl-4-aminnhMtvlUN(2.3-dimethOXVhsn7nvl-3-aminnnrnp vl^lQlutaramide^PH 2i was prepared and purified in the same manner as 2a (85 percent). • H-NMR (CDCI,) 8 1.2-2.1 (m, 10H), 2.1-2.7 (m, 6H), 3.1-3.8 (m, 16H), 3.85 (s, 24H), 6.8-7,8 (m, 12H). Analysis calculated for C54H72N50,4, Calc. 63.02 %C, 7.05 %H, 8.17 %N, found 62.94 %C, 6.93 %H, 8,25 %N,

PAGE 117

,""« 111 N-;Bisf2.3-dimethoxvbenzovl-3-aminop ropvn]-N'-fbis(2.3-dimethoxvbenzovl-4-aminobutyl)l tilularamide(2kl 21s was prepared and purified in the same manner as 2a (85 percent). ^H-NMR (CDCI3) 8 1.4-2.1 (m, 12H), 2.2-2.5 (m, 6H), 3.1-3.7 (m, 16H), 3.9 (s, 24H), 6.9-7.7 (m, 12H). Analysis calculated for C55H74N50i4H20,Calc. 62.25 %C, 7.22 %H, 7.92 %N, found 62.36 %C, 7.21 %H, 7.88 %N. N-rBis(2.3-dimethoxvbenzovl-4-aminobutvlll-N'-IN-<2.3-dimethoxvbenzovl-4-aminohijtvll-N(2.3-dimethoxvbenzovl-3-aminopro pyniglutaramide(2n 21 was prepared and purified in the same manner as 2a (85 percent). ^H-NMR (CDCyS 1.4-2.1 (m, 14H), 2.2-2.5 (m, 6H), 3.1-3.7 (m, 16H), 3.9 (s, 24H), 6.9-7.7 (m, 12H). Analysis calculated for C5gH76NgOi4, Calc. 63.62 %C, 7.25 %H, 7.95 %N, found 63.71 %C, 7.12 %H, 7.81 %N. N-IBis(2.3-dimethoxvbenzovl-3-amino propvll]-N'-;bisf2.3-dimelhoxvbenzovl-4-aminobutvh] adipamidBf2m> 2inwas prepared and purified in the same manner as 2a (85 percent). ''H-NMR (CDCI3) 5 1.7 (br, 14H), 2.45 (br, 4H), 3.5 (t, 16H), 3.9 (d, 24H), 7.1 (q, 8H), 7.75 (m, 4H), 8.1 (br, 3H), 8.4 (t, 1H). Analysis calculated forC55H7gN50,4H20, Calc. 62.55 %C, 7.31 %H, 7.82 %N, found 62.84 %C, 7.31 %H, 7.81 %N. N.N'-Bisfbis(2.3-dihvdroxvbenzovl-3-a minQoropvl^1succinamidena^ A solution of 2a (1.2 g, 1.16 mmoles) in CHjClj (100 mL) was stirred under nitrogen at 0°C. A soluton of BBr3 (25 mmoles) in CHjClj (50 mL) was addded dropwise. After stining at room temperature for 16 hours, the reacton vessel was cooled to 0°C and ice cold HjO (20 mL) was added slowly with vigorous stirring. After stirring for 2 hours, the crude material was filtered and washed altematingly with HjO and CHjClj. The resultant solid was dissolved in MeOH and evaporated several times to afford 1 .3 g of crude product. This material was chromatographed on Sephadex LH-20 (120 g, eluted with 32 percent EtOH.benzene) to yield 0.98 g product (90 percent). An analytical sample was prepared by HPLC, eluting with 25 percent 0.4F phosphate buffer, pH 3.0 : 75 percent acelonitrile. The desired fractions were placed under a stream of

PAGE 118

112 nitrogen gas to remove the acetonitrile, extracted with EtOAc (3 x 3 mL per traction), dried over Na2S04, and evaporated. 'h-NMR (CD3OD) 8 1.7 (m, 8H), 2.7 (s, 4H), 3.4 (m, 16H), 6.6-7.4 (m, 12H). Anaiysis calculated for C44H52NgO,4, Caic. 59.45 %C, 5.90 %H, 9.45 %N, found 59.42 %C, 5.93 %H, 9.41 %N. N,N'-BlsrbiSf2.3-dihvdroxvhen7nvl-4-ami n obutvl^l arirrinamidfi MW lijwas prepared and purified in the same manner as la. ''h-NMR (CD3OD) 8 1.7 (m, 16H), 2.7 (s, 4H), 3.4 (m, 16H), 6.6-7.4 (m, 12H). Analysis calculated for C4gH5oNgOi4H20, Caic. 59.86 %C, 6.49 %H, 8.73 %N, found 59.90 %C, 6.31 %H, 8.60 %N. N , N'-BiSfbiS(2.3-dihvdroxvhenzovl-4-amin obiitvl^lsrirrinamirifinr^ l£ was prepared and purified in the same manner as la. '' H-NI«R (CD3OD) 8 1 .7 (m, 1 2H), 2.7 (s, 4H), 3.4 (m, 16H), 6.6-7.4 (m, 12H). Analysis calculated for C45H5gNgO,4, Caic. 60.25 %C, 6.1 6 %H, 9.1 6 %N, found 60.08 %C, 6.24 %H, 9.08 %N. N , N'-Bisrt)iS (2.3-dihvdroxvben7nvl-.?-aWii n oprnnvm nlntaramidfi rirti Id was prepared and purified in the same manner as la. ^ H-NH/IR (CD3OD) 5 1 .7 (m, 8H), 2.0-2.6 (m, 6H), 3.4 (m, 16H), 6.6-7.4 (m, 12H). Analysis calculated for C45H56N60,4-H20, Caic. 58.69 %C, 6.13 %H, 9.13 %N, found 58.50 %C, 6.24 %H, 8.89 %N. N . N'-BISfhiS (?.3-dihvdroxvhfin7nvl-4-aminn fa »tvhl nlirtammirl» (1^] l£ was prepared and purified in the same manner as la. • H-NMR (CD3OD) 8 1 .7 (m, 16H), 2.0-2.6 (m, 6H), 3.4 (m, 16H), 6.6-7.4 (m, 12H). Analysis calculated for C4gHg2N60i4, Caic. 61 .37 %C, 6.52 %H, 8.76 %N, found 61.33 %C, 6.53 %H, 8.74 %N. N , N'-B'SrN-f2 , 3-dlhVdr0XVbfinrQVl-4-aminohlJtvh-N'-f2 3-riihvdr o xvhfin7nvl-.?-aminonmnyni Qlutaramiflfi Mn 11 was prepared and purified in the same manner as I3. ^ H-NlvlR (CD3OD) S 1 .7 (m, 1 2H), 2.0-2.6 (m, 6H), 3.4 (m, 16H), 6.6-7.4 (m, 12H). Analysis calculated for C47H5gNgOi4H20, Caic. 59.48 %C, 6.37 %H, 8.86 %N, found 59.39 %C, 6.39 %H, 8.80 %N.

PAGE 119

113 N-rBis(2.3-dihvdroxvbenzovl-3-aminopr oDvhVN'-fN-f2.3-drhvdrnxvhfinzovl-4-aminnhi]tvh-Nf2.3-dihvdroxvbenzovl-3-aminnprn pvM1si]ndnamirifiMq > Ifl was prepared and purified in tlie same manner as la. ^ H-NMR (CD3OD) 5 1 .7 (m, 10H), 2.7 (s, 4H), 3.4 (m, 16H), 6.6-7.4 (m, 12H). Analysis calculated for C45H54NgO,4, Calc. 59.86 %C, 6.03 %H, 9.31 %N, found 59.62 %C, 6.24 %H, 9.38 %N. N-rBiS(2.3-dihvdrQxvbenzovl-3-aminoDroDvh1-N'-rhisf2 .3-ditivdroxvhsnzovl-4-aminnhi]tvl)l Ih was prepared and purified in ttie same manner as la. ^ H-NMR (CD3OD) 8 1 .7 (m, 12H), 2.7 (s, 4H), 3.4 (m, 16H), 6.6-7.4 (m, 12H). Analysis calculated for C46H5gNgO,4H20, Calc. 59.09 %C, 6.25 %H, 8.99 %N, found 59.21 %C, 6.48 %H, 9.21 %N. N-tBis(2.3-dihvdroxvbenzovl-4-aminohijt vll1-N'-rN-/2.3-dihvriroxvhRnzovl-4-aminohiitvh-N(2.3-dihvriroxvhsn7nv l-3-aminonropvm succinamirisM i> li was prepared and purified in the same manner as la. ^ H-NMR (CD3OD) 8 1 .7 (m, 14H), 2.7 (s, 4H), 3.4 (m, 16H), 6.6-7.4 (m, ^12H). Analysis calculated for C47H58N6O14H2O, Calc. 59.48 %C, 6.37 %H, 8.86 %N, found 59.62 %C, 6.48 %H, 8.69 %N. N-fBiSf2.3-diflvdroxvbenzovl-3-aminnp roDvlll-N'-fN-f2.3-dihvdrnxvben70vl-4-aminobijtvl-N(2.3-dihvrirnxvhsn7nvl-3-aminnnrnnv hlalutaramidsni^ li was prepared and purified in the same manner as la. ^ H-NMR (CD3OD) 8 1 .7 (m, 1 0H), 2.0-2.6 (m, 6H), 3.4 (m, 16H), 6.6-7.4 (m, 12H). Analysis calculated for C46H56NgO,4H20, Calc. 59.09 %C, 6.25 %H, 8.99 %N, found 58.97 %C, 6.09 %H, 9.08 %N. N-fBiS(2.3-dihvdroxvbenzovl-3-aminonrnt]vhl-N'-fhisf? 3 -dihvriroxvhsnzovl-4-aminnhiitvni alutaramidffMkl Us was prepared and purified in the same manner as la. '' H-NMR (CD3OD) 8 1 .7 (m, 1 2H), 2.0-2.6 (m, 6H), 3.4 (m, 16H), 6.6-7.4 (m, 12H). Analysis calculated for C47H5gNgOi4-2H20, Calc. 58.38 %C, 6.46 %H, 8.69 %N, found 58.29 %C, 6.41 %H, 8.46 %N.

PAGE 120

(2,3-dlhYClroxvhRn7nvl-3-aminnnrnpvi nali]taramidPMh 11 was prepared and purified in the same manner as la. 1 H-NH/IR (CD3OD) 8 1 .7 (m, 1 4H), 2.0-2.6 (m, 6H), 3.4 (m, 16H), 6.6-7.4 (m, 12H). Analysis calculated for C^gHgoNgOi^-HjO, Calc. 59.86 %C, 6.49 %H, 8.73 %N, found 59.93 %C, 6.50 %H, 8.73 %N. N-rBi5(2 . 3-flihYt)roxYbf?P?'OYl-3-aminonrnnvni-N'-rhisfp:^-dihvrtr o xvhfin7nvi-4-aminnh„)Ym adipamidfiMm^ '"" Imwas prepared and purified in tfie same manner as la.. ''h-NIYIR (CD3OD) 5 1.60{br, 12H), 1.8 (br, 4H), 2.35 (br, 4H), 3.4 (m, 16H), 6.6-7.4 (m, 12H). Analysis calculated for ^48^^6o'^60u"2°' Calc. 59.86 %C, 6.49 %H, 8.73 %N, found 59.87 %C, 6.52 %H, 8.48 %N. N.N'-BisfhisfN-RnC-n-aminnnmpvni nl ntaramiriR Mri^ To a solution of Ni,N''-bis(butyloxycarbonyl)norspermidine hydrocfiloride (1.95 g, 4.94 mmoies), triethylamine (1.46 g, 14.4 mmoles) and N,N-(dimetfiyl) aminopyridine (60 mg) in CH2CI2 (150 mL) was added glutaryl dichlorlde (.38 g, 2.25 mmoles). This solution was stirred at room temperature for 12 hours before transferring to a separatory funnel with CHgClg (300 mL) and washing with ^M HCI (200 mL) and 10 percent NaHCOj (200 mL) to yield 1.97 g cmde product (87 percent) ^H-NMR (CDCI3) 5 1 .45 (S, 36H), 2.0 (t, 4H), 2.4 (m, 8H), 2.75 (m, 2H), 3.2 (m, 16H). Analysis calculated for C37H70N6O10, Calc. 58.55 %C, 9.30 %H, 11.07 %N, found 58.61 %C, 9.28 %H, 11.22 %N. This compound was exposed to TFA at 0°C for 20 min. to yield the corresponding tetraamine which was used immediately. 2.3-Dimefhoxv-S-nitrnhfl nzovlf:hlnririfl To 2,3-dimethoxy-5-nrtro benzoic acid (1.92 g, 8.45 mmoles)~synthesized by the method of Cain and Simonsen (84)~was added thionyl chloride (50 mL). The solution was refluxed for 5 hours before cooling to room temperature. Excess thionyl chloride was evaporated and benzene (200 mL) was added. This solution was washed with 13 percent KgCOg (2 x 50 mL), dried over Na2S04, and evaporated to yield 1 .98 g product (92 percent). 1 H-NMR (CDCI3) 5 3.8 (s, 6H), 7.0 (m, 2H).

PAGE 121

115 N-hYdroxvsuccinimvlf2.3-rii methnxv-!;-nitrn1hBn7natq To a solution of 2,3-dimethoxy-5-nitro benzoic acid (.68 g, 3.0 mmoles) and N-hydroxysuccinimide (.40 g, 3.5 mmoles) in THF (25 mL) was added DCC (.74 g, 3.6 mmoles). The reaction was stirred 20 hours before filtration and evaporation to yield 1.05 g product (100 percent), which was used without further purification. N.N-dielhvU2.3-riimpthn» v-5-nitrn>hfin7amiria To a solution of diethylamine (.37 g, 5.05 mmoles) in CH2CI2 (25 mL) at 0°C was added dropwise (2,3-dimethoxy-5-nitro)ben2oyl chloride (.50 g, 2.03 mmoles) in CHjClj (25 mL). This solution was stirred at room temperature for 24 hours before washing with 3 M HCI (2 x 50 mL) to yield 0.58 g cnjde material (97 percent), which was chromatographed (15 g silica eluted with CHCI3) to yield the pure material (80 percent). ' H-NMR (acetone-dg) 5 1 .1 (t, 3H), 1 .2 (t, 3H), 3.1 (q, 2H), 3.5 (q, 2H), 4.0 (d, 6H), 7.6 (d, 1H), 7.8 (d, 1H). Analysis calculated for CijHigNjOg, CalC. 55.31 %C, 6.43 %H, 9.92 %N, found 55.48 %C, 6.39 %H, 9.90 %N. N.N-dielhvlf2.3-riihvrtrnvY-'^-nitrn^hPn7amirtB To a solution of BBrg (12 mL, 12 mmoies) in CHgClj (20 mL) at 0°C was added dropwise N,N-diethyl (2,3-dimethoxy-5-nitro)benzamide (.18 g, .67 mmoles) in CHjClj (40 mL). The suspension was stin-ed at room temperature before adding ice/water (20 mL) and stirring an additional 3 hours. The mixture was then evaporated to an aqueous solution which was extracted with ethyl acetate (6 x 25 mL). This extract was extracted with 10 percent NaOH (3 x 50 mL), acidified with concentrated HCI, and extracted again with ethyl acetate (3 x 75 mL) to yield the cnjde material, which was chromatographed (2 g silica eiuted with 10 percent MeOHiCHCIg) to yield .12 g product (76 percent). ^H-NMR ( acetone-dg) 5 1.1 (t, 6H), 3.5 (q, 2H), 4.1 (q, 2H), 7.79 (s, 2H). Analysis calculated for C, ^ H14N2OS, calc. 51 .97 %C, 5.55 %H, 1 1 .02 %N, found 52.14 %C, 5.69 %H, 11.22 %N.

PAGE 122

1 , "6 N^-(2.3-dimelhoxv-5-nitrob9nzovn-N3-r3-f2.3-dimfithnKv-fi .nitro-N-benzovh amino-1-propvl)hflx ahvdropvrimi(1inflfSl To a solution of APHHP (.39 g, 2.7 mmoles) and triethylamine (0.6 g, 5.93 mmoles) in CHjClj (40 mL) at 0°C was added 2,3-dimethoxy-5-nitrobenzoyl chloride (1.33 g, 5.41 mmoles). The solution was stirred 1 hour and evaporated. Ethyl acetate (25 mL) was added, causing an impurity to precipitate. After filtration, the mother liquor was chromatographed (85 g silica eluted with EtOAc) to yield 1 .24 g product (82 percent). '' H-NMR ( CDCL3) 5 1 .5-2.1 (m, 4H), 2.8 (m, 4H), 3.5 (m, 4H), 4.1 (s,12H), 4.5 (s, 2H), 7.8 (m, 2H), 8.5 (dd, 2H). Analysis calculated for CjsHgQNgO^o, Calc. 53.57 %C, 5.39 %H, 12.49 %N, found 53.55 %C, 5.41 %H, 12.44 %N. N-[3-f2.3-dimethoxv-N-benzovnamino-1-Dronvllhe xahvriroDvrimiriinfl(9^ » To a solution of APHHP (132 mg, .92 mmoles) in CH2CI2 (25 mL) at C°C was added N-hydroxysuccinimyl(2,3-dim8thoxy)benzoate (250 mg, .89 mmoles) in CH2CI2 (50 mL) over a 2-hour period before stining at room temperature for 24 hours. The solution was extracted with 1 M HCI (3 X 25 mL), made basic with NaOH (s), and extracted with CH2CI2 (4 x 25 mL). This crude material was chromatographed (20 g silica eluted with 1 percent triethylamine:MeOH) to yield 1 00 mg product (45 percent). • H-NlwlR (CDCI3) 5 1 .55 (quint, 2H), 1 .75 ( quint, 2H), 2.3 (d, 2H), 2.7 (m, 4H), 3.4 (s, 2H), 3.5 (t, 2H), 3.9 (S, 6H), 7.1 (m, 1H), 7.65 (dd, 1H), 8.3 (t,1H). Analysis calculated for C,gH25N303-2H20, calc. 55.96 %C, 8.51 %H, 12.24 %N, found 55.88 %C, 8.65%H, 12.44 %N. N^-f2.3-riimfithoxvhsn; rovhnorsnermidinfi A solution of 2 (100 mg, .29 mmoles), piperidine (2 drops), and dimedone (210 mg, 1.5 mmoles) in ethanol (20 mL) was refluxed for 18 hours. After evaporation, ethyl ether (50 mL) and 1 M HCI (20 mL) were added, and the organic layer was extracted with 1 M HCI (2 x 20 mL). The combined aqueous layers were saturated with NaCI(s) and made basic with NaOH(s) before extracting with CH2CI2 (6 x 20 mL) to yield the crude product, which was chromatographed

PAGE 123

117 (5.5 g silica eluted with 2 percent to 10 percent NH^OHiMeOH) to yield 100 mg product (100 percent). Spectral data were identical to those already reported (84). N^ -(2 . 3-dlmSthOXVhsn7nvn-N^-f2.3-dimRthoxv-.';-nitrnhfln zovl>nnrsnprmiriinpMn) 112 was prepared and purified in the same manner as 2''H-NMR (CDCy 8 1 .3 (t, 4H), 2.2 (m,2H), 2.65 (m, 2H), 3.1 (m, 2H), 3.65 (m, 4H), 3.9-4.1 (m, 12H), 7-8.5 (m, 7H). Analysis calculated for C24H32N40a-H20, calc. 55.16 %C, 6.56 %H, 10.72 %N, found 55.15 %C, 6.56 %H, 10.73 %N. N^N^-(2.3-dimftthnyv-'^nitrohsn7nvnnnrsnRrmiriina^fi) A solution of a (51 mg, .091 mmoles), piperidine (1 drop), and dimedone (140 mg, 1 mmoie) in 50 percent aqueous ethanol (60 mL) was refluxed for 24 hours, evaporated to an aqueous solution, and extracted with CHjClj (3 x 30 mL). This crude material was chromatographed (10 g silica eluted with percent to 5 percent NH40H:l^eOH) to yield 40 mg product (80 percent). This compound was found to be very unstabre and was used as soon as isolated. ''H-NMR (CDCy 8 1.9 (quint, 4H), 2.6 (s, 1H), 2.8 (t, 4H), 4.0 (d, 12H), 7.8 (d, 2H), 8.2 (t, 2H), 8.45 (d, 2H). BiSfN , N-lilS [ (2 . 3-dim(;thnXYbenzQVl-5-nitrnl-3-aminnnmnvl H nlMtaramiripf';ri> Method 1 . To a solution of S (220 mg, .4 mmoles) and triethylamine (90 mg, .89 mmoles) in CHjClj (10 mL) at 0°C was added dropwise glutaryl dichloride (32 mg, .19 mmoles) in CHjClj (15mL). The reaction was stin-ed 30 mins. before washing with 1 IWI HCI (3 x 10 mL) and 1 M NaOH (3x15 mL). Preparative chromatography (20 cm x 20 cm silica plate developed with 1 :2:7 methanol:chloroform:ethyl acetate and extracted with 50 percent methanol:ethyi acetate) yielded 150 mg product (66 percent). 1h-NMR (CDCI3) 8 1.9 (m, 8H), 2.4 (t, 8H), 3.5 (t, 16H), 4.1 (t, 24H), 7.9 (m, 4H), 8.1 (t, 2H), 8.45 (t, 2H), 8.5 (m, 4H). Analysis calculated for ''53^^6Nlo°22' Calc. 53.35 %C, 5.41 %H, 11.74 %N, found 53.21 %C, 5.62 %H, 1 1.62 %N. Meltmii. To a solution of M. (210 mg, .26 mmoles) and KjCOg (.44 g, 3.I8 mmoles) In dH20 (15 mL) at 4OC was added dropwise 2,3-dlmethoxy-5-nitro-benzoyl chloride (.34 g, 1.36

PAGE 124

118 mmoles) in benzene (25 mL) with vigorous stirring. Ttie reaction was stirred 4 hours before adding CHjClj (12 mL) and dHjO (5 mL). After 30 mins. the mixture was transferred to a separatory funnel with 5 percent cold NaOH (50 mL) and CHjClj (50 mL). The aqueous layer was extracted with MeClg (2 x 50 mL) to yield 300 mg product (98 percent). Spectral data were the same as those obtained by method l . N-[BiS(2.3-dimethoxvbenzovl-3-aminnp roDvni-N'-[N-f2.3-dimfithnxv-5-nitrohfin7nvl-.laminQPropvll-N-f2.3-dimethoxvhsnzovl-.3-amin nDroDvns[iccinamidfi N-[Bis(2,3-dimethoxybenzoyl-3-aminopropyl)]-N'-[N-(2,3-dimethoxy-5-nitro-benzoyl-3aminopropyl)-N-(2,3-dimethoxybenzoyl-3-aminopropyl)lsuccinamide was prepared and purified in the same manner as2fl. ''H-NMR (CDCI3) 5 1.9 (br, 6H), 2.3 (br, 2H), 2.7 (s, 4H), 3.5 (t, 16H), 7.0-8.6 (m, 15H). Analysis calculated for CsgHg^N^O^g calc. 59.70 %C, 6.46 %H, 9.37 %N, found 59.90 %C, 6.49 %H, 9.40 %N. N-rBis(2.3-dihvdroxvbenzovl-3-amtnonrnnvhl-N'-IN-(?3 dihvdroxv-5-nitrohfin7nvl-.?amlnOPrQPVl)-N-(2.3-rilhvdroxvbenzovl-3-aminnnr nDvnisiif:cinamidpnia> lia was prepared and purified in the same manner as 2a. ^ H-NMR (CDCI3) 5 1 .5 (m, 2H), 2.1 (m, 6H), 2.9 (s, 4H), 3.5 (m, 16H), 6.7-8.4 (m, 11H). Analysis calculated forC44H5,N70ig, Calc. 56.69 %C, 5.50 %H, 10.50 %N, found 56.41 %C, 5.63 %H, 10.47 %N. ilt£:£i£lM'li^-biSf2.3-dimelhoxvbenzovnsDermidinff|-N-R OC-alMtamamidflMPf) To a solution of N^N^-bis(2,3-dimethoxybenzoyl) homospermidine (1.16 g, 2.39 mmoles), dicylohexylcartjodiimide (0.50 g, 2.41 mmoles), and N-BOCglutamic acid (0.28 g, 1.15 mmoles) in CHjClj (30 mL) was added 4-(dimethylamino)pyridine (20 mg). The solution was stin-ed for 12 hours prior to filtraton and evaporation of the filtrate to yield 1 .45 g crude oil. This oil was purified by silica gel chromatography (75 g, eluted with 1:8:8 EtOAc:acetone: benzene) to yield 1.05 g product (60 percent). ^I H-NMR (CDCy 8 1.3 (s, 9H), 1.5-1.8 (m, 16H), 2.3-2.5 (m, 5H), 3.1-3.6 (m, 16H), 3.9 (s, 24H), 4.6 (br, 1H), 5.4 (br, 1H), 6.9-8.2 (m,16H). Analysis calculated for CggHggN^Oig, Calc. 62.21 %C, 7.22 %H, 8.46 %N, found 62.00 %C, 7.58 %H, 8.32 %N.

PAGE 125

119 N.N'-Bi5[N^ .N^-biS(2.3-dimfithnxvhBn 7ovnsnfirmiriinelnlijtamamidpTFAM?jf) Triflouroacetic acid (10 mL) was cooled to 0°C, added to 121 (0.6 g, 0.51 mmoles) and stirred at 0°C for 60 minutes before evaporating. Benzene (2 x 20 mL) was added and evaporated to yieid 0.75 g crude product, which was purified by silica gel chromatography (15 g, packed and loaded with 1:1 EtOAc:CHCl3 and eluted with 15;42.5;42.5 to 20:40:40 EtOH:EtOAc:CHCl3) to yield 0.7 g product (100 percent). 'h-NMR (CDCI3) 8 1.5 (br, 18H), 2.4-2.7 (m, 4H), 3.1-3.5 (m, 16H), 3.9 (s, 24H), 6.9-8.2 (m,16H). Analysis calculated for ^57"76'^70l6'^3' ^alC. 58.40 %C, 6.53 %H, 8.36 %N, found 58.56 %C, 6.79 %H, 8.19 %N. BiSlM^ .N^-bisf2.3-dihvdroxvhen7n vnsDermidine]Qlutamamirie HRr(PR^ 25 was prepared and purified in the same manner as compound la. ''h-NMR (CD3OD) 5 1.7 (m, 12H), 2.0-2.65 (m, 4H), 3.45 (m, 16H), 6.6-7.3 (m, 12H). Analysis calculated for ^47^eo^7°U^'' oalo54.97 %C, 5.89 %H, 9.55 %N, found 55.09 %C, 5.93 %H, 9.56 %N. N-BOC-8-aminnranrvlin arid BOC-ON (1.59 g, 6.46 mmoles) was added to a solution of 8-aminocaprylic acid (1.00 g, 6.28 mmoles) and triethylamine (1.81 g, 17.9 mmoles) in tetrahydrofuran (15 mL) and dHjO (15 mL). This solution was stirred 6 hours before evaporating the acetone. The aqueous layer plus 30 mL HjO was washed with ethyl ether (3 x 25 mL), acidified with concentrated HCI, and extracted with EtOAc (3 x 25 mL). The combined organic layers were dried with NajSO^ and evaporated to afford 1.65 g of oil (100 percent). ^H-NMR (CDCI3) 5 1.1-1.8 (m, 10H), 1.45 (s, 9H), 2.3 (t, 2H), 3.1 (m, 2H), 4.85 (br,1H), 10.6 (s. 1H). Analysis calculated forCigHgsNO^, Calc. 59.97 %C, 10.07 %H, 5.38 %N, found 59.93 %C, 10.10 %H, 5.33 %N. N-((8-N-BQC-am i n0CaprvlVl)hiSlN ^N3 -bi5f2.3-riimPthnwhpnTnyl lsDfirmiriinPHnliitamamidPM4f^ A solution of 121 (0.49 g, 0.41 mmoles), N-BOC-8-aminocaprylic acid (0.13 g, 0.50 mmoles), dicyclohexylcarbodiimide (0.11 g, 0.53 mmoles) and 4-(dimethylamino)pyridine (10 mg) in ethyl ether (20 mL) and CHjClg (2 mL) was stirred 24 hours, filtered, and evaporated to yield 0.65 g of crude product, which was purified on silica gel (60 g, eluted with 10:45:45

PAGE 126

.'f-s^ 120 EtOHrEtOAciCHCy fo yield 0.37 g of Oil (69 percent). This oil was dissolved in CHjCij (50 ml), wasiied with saturated NaHCOg (3 x 25 mL), dried over Na2S04, and evaporated to yield 0.29 g product (55 percent). ^H-NMR (CDCI3) 8 0.8-2.6 (m, 26H), 1.45 (s, 9H), 2.6-3.8 (m, 18H), 3.9 (s, 24H), 4.8 (br, 2H), 7.0-8.2 (m, 12H). Analysis calculated for CggHggNgOiy, Calc. 62.85 %C, 7.60 %H, 8.62 %N, found 62.80 %C, 7.67 %H, 8.78 %N. N-(8-amin0CaprYlvl)biS[N^ .N°-bis(2.3-dimethoxvhen7 nvnsDermidinelQlutamamids TFAn5n 141 (0.29 g ,0.22 mmoles) was stirred for 20 minutes at 0°C in TFA (1 mL) before evaporating, the remaining oil was dissolved in saturated aqueous NaCI (20 mL), made basic with NaHC03(s) and extracted with chloroform (5 x 25 mL) to yield 0.22 g product (82 percent). 'H-Nt^R (CDCI3) 8 1.3 (s, 9H), 1.5 (m, 22H), 2.3 (m, 3H), 3.5 (m, 18H), 3.9 (s, 24H), 6.5-8.3 (m, 17H). Analysis calculated for Cg3HggNgOi5, Calc. 63.09 %C, 7.56 %H, 9.34 %N, found 63.22 %C, 7.60 %H, 9.30 %N. IRP-64 and c hloride resinMSl Resin preparation. To IRP-64 (5.8 g) was added thionyl chloride (25 mL) and DMF (1 .0 mL) and the mixture was refluxed for 3 hours. After cooling, the resin was washed with dry benzene (100 mL) and dry THF (100 mL) and placed under high vacuum. Determination of eguivalency. To ^'*C-glycine ethyl ester (33.83 mg, 250 mmoles) was added CH2CI2 (15 mL) and triethylamine (2.5 mL). To each of two flasks was added IS (1 .65 mg and 7.72 mg, respectively) and the prepared solution (2.0 mL). To determine the degree of non-specific binding, to two additional flasks was added IRP-64 (15.04 mg and 12.12 mg, respectively) and the prepared solution (5.0 mL). The samples were mixed overnight and the radioactivity remaining in solution determined by scintillation counting. It was calculated that the equivalency of acid chloride groups was approximately 1.2 meq/gm of resin, about 11 percent conversion from the acid.

PAGE 127

121 IRP-64 msthvlatecl rasin derivativflfsm To CHjClj (10 mL) and triethylamine (1 mL) was added 1^ (50 mg) and Ifi (185 mg). The suspension was stirred 6 hours and refluxed an additional 12 hours before adding methanol (1 mL) and refluxing 3 hours. When cool, the suspension was filtered and washed well with CHjCL to yield 240 mg product. The equivalency was calculated gravimetrically to be 0.24 meq/gm resin. IRP-64 catechol resin (24f^ To 211 (77 mg) and CHjClj (6 mL) was added BBrj (2 mL of a 1.0 M solution in CHgClj). The mixture was stirred 6 hours. To the cooled flask was added dHjO (2 mL) with vigorous stirring for 2 hours. The resulting mixture was filtered and washed with CH2CI2 (50 mL), dHjO (50 mL) and methanol (30 mL) and placed under high vacuum to yield 45 mg of iron-positive product. CH-SeDharose-4B catfi colamide derivative f2fi> Phosphatg hi i ffsr . NaCI (2.9 g) and NajHPO^ (1.76 g) were added to dHjO (200 mL) and the pH adjusted to 7.2 with H3PO4 before diluting with dHgO to 500 mL. This solution was diluted to 1 .0 liter with ethanol. Acetate bufferNaCI (14.5 g) and sodium acetate (4.1 g) were added to dHjO (450 mL), the pH adjusted to 4 with AcOH, and diluted with dHgO to 500 mL. This solution was passed through Chelex resin and an equal voume of ethanol was added. Resin prepflrmion. CH-Sepharose 4B (2.09 g) was stirred in 1 mtwl HCI (30 mL) for 5 minutes. After filtering and washing with more HCI (40 mL) the resin remained in HCI for an additional 15 minutes, at which time it was filtered and washed with phosphate buffer (100 mL). Resin derivqii/atinn. The prepared resin was divided between two tubes. To one tube was added phosphate buffer (7 mL) followed by 2£ (76.6 mg) in methanol (.5 mL) and phosphate buffer to 10 mL. To the other tube was added phosphate buffer (10 mL). Both tubes were gently mixed for 24 hours before adding ethanolamine (25 mL) to each tube and

PAGE 128

122 mixing an additionai 6 hours. The resins were individually washed with phosphate buffer (50 mL), acetate buffer (50 mL), and again with phosphate buffer (50 mL). Job's Plots All ligands and metals were prepared as millimolar solutions. The same buffer prepared for competition studies was used (pH 9.2 ammonia). To each of 11 tubes was added buffer (100 mL), dHjO (2.8 mL), metal (0, 10, 20, 30, 40, 50, 60, 70, 80, 90, or 100 nL), and ligand (100, 90, 80, 70, 60, 50, 40, 30, 20, 10, or ^L). The UV-visible absorbance spectra were then recorded. The total molarity of these solutions remained constant at 33.3 iimoles/liter. Precinitatinn Terihniqiifis 1. To 2.50 mL of dHjO was added 7.2F NH3/NH4CI buffer, pH 9.2 (100 jiL), 0.9073 mM ThCl4 (100 nD, and 1.0 equivalent of a catecholamide ligand (100 jiL of methanolic solution). After 1 or 30 minutes the pH was adjusted to neutrality with concentrated HCI (30 jiL) and the suspension filtered through a 0.45 urn filter. Next, more HCI (150 nL) and ARS (2.0 mL) were added and the absorbance at 666 nm recorded. 2A 3.362 X 10"" 1^ solution of ligand buffered at pH 9.0 (2.7 mL) was acidified (pH 6) and stirred before one equivalent of thorium (100 nL) was added. The suspension was stirred for several minutes, filtered, and concentrated HCI (150 ixL) and ARS (2.0 mL) were added and the absorbance at 666 nm recorded. a. To 2.50 mL of dHjO was added 7.2F NH3/NH4CI buffer, pH 9.2 (100 |iL), 0.9073 mM ThCl4 (100 nD, and concentrated HCI (30 jiL). One equivalent of a catecholamide ligand (100 HL of methanolic solution) was added and this acidic solution mixed. The suspension was filtered, and more HCI (150 hL) and ARS (2.0 mL) were added and the absorbance at 666 nm recorded.

PAGE 129

123 Eriochrome Black T Competition EBT Soliition For accuracy and reproducibility of data, the commercially available indicator grade EBT was purified by a modified method of Robinson and Mills (85). A buffer solution was prepared by adding 150 mL of glacial acetic acid and 250 mL dHgO to a 500 mL volumetric flask. The pH was adjusted to 4.0 with NaOH prior to diluting to mark with dHgO. The impure EBT (1 g) was stirred with dHjO (20 mL) and heated to boiling with constant stin-ing. The above acetate buffer (20 mL) was added and the mixture allowed to cool. The solution was suction filtered and washed with additional buffer. This process was repeated six times. The resultant solid was boiled with absolute ethanol (40 mL), suction filtered, and washed with hot ethanol (60 mL). This procedure was repeated six times. The solid was dried in the presence of PjOg under high vaccuum at 65°C to yield the pure dye. A purity of 98.7 percent was determined by acid-base titration. An EBT stock solution was prepared by dissolving 56.77 mg EBT in methanol and diluting to a final volume of 250 mL with methanol (final concentration = 4.922 x 10''' M), A working solution was prepared by diluting 40 mL stock EBT to 1 00 mL. (1 .969 x 1 0"* M). Hvdroxvlami ne Solution • Hydroxylamine hydrochloride (1.75 g) was weighed into a 250 mL volumetric flask, dissolved in and diluted to mark with dHjO (final concentration = 0.100 F). Masking Agent Into 500 mL dHjO in a 1 .0 liter volumetric flask was dissolved triethanolamine (1 0.05 g) and disodium EDTA (37.24 g). The solution was diluted to mark, achieving final concentrations of 0.1009 M TEA and 0.1000 M EDTA. Thoriumnvi Solutinns ^^^ThCI^ (62.43 mg) was added to a 10 mL volumetric flask and diluted to mark with dHjO (concentration = 1.67 x 10"^ M). 3.00 mL of this solution were Iransfered to a 10 mL volumetric flask and diluted to mark with dHjO (final concentration = 5.01 x 1 0"^ M). A working solution was

PAGE 130

124 prepared by diluting 2.00 mL to 100 mL (1 .00 x 1 0"^ M). The concentration of the stock solution was determined by EDTA titration, using Xylene! Orange as the indicator (86). Buffer To a 1 .0 liter volumetric flask was added NH4CI (60 g), dHgO (200 mL), and NH^OH (400 mL). This solution was adjusted to pH 9.2 with concentrated HCI solution before diluting to mark with dHgO. Standard Ciirvs To each of nine 100 mL volumetric flasks was added hydroxylamine solution (1 .00 mL), ThCl4 solution (0.00, 2.50, 5.00, 7.50, 10.00, 15.00, 20.00, and 25.00 mL, respectively), masking agent (1.00 mL), and dHjO (25 mL). After adding EBT solution (5.00 mL) and buffer (3.00 mL), the solutions were diluted to mark with dHgO and allowed to stand for at least 10 minutes before reading their visible spectra. Experimental Samples A solution was prepared as described, with 25 mL of thorium. One equivalent of a ligand was added (100 \iL of methanolic solution), and the color was observed for 15 seconds. Development of a blue color during this time was recorded as positive (+). Compelilion With Nitrncat scholamidfis Stock Solutions Buffer . 3.0 F NH3 pH 7.2 in dHjO. Catecholamidfi stnt^k snlntinn-; 1 n mM in moihanni ovf-opi fnr 97 ..,hi>-h ,.,-|o pr^pcircd in dHaO. Thorium stock sglution. 9.073 x 10"'' M ^SOThC^ in dHjO, as determined by EDTA titration, using Xylenol Orange as the indicator (88).

PAGE 131

125 Copper SlOCK solution6. 1 9 x 1 0-3 M Cu(0Ac)2 m dH20. a working solution was prepared at a concentration of 1.00 mM, and used for subsequent experiments. CI-PAN stork sniMtinn 3.00 mM in DMF. One-Metal -Two-Liganri Comnetitinn 4.5 mL polypropylene tubes with snap caps were used. To each tube was added buffer (300 |iL),2Z {100 M.L), a catecholamide or methanol (100 nL), copper (0, 20, 40, 60, 80, or 100 jiL plus 100, 80, 60, 40, 20, or nL dHjO), and dHgO (2.6 mL) under a Nj atmosphere. Tubes were sealed and mixed 6 hours before recording their visible absorption spectra. Each experiment was performed three times. Two-Melals-One-Linand ComPBtitinn To each tube was added buffer (300 nL), la (0, 18, 36, 54, 72, or 90 \iL plus 90, 72, 54, 36, 18, or iiL methanol), 232,horium (100 \iL), copper (90 nL), CI-PAN (85 jiL), dHjO (2.65 mL), and benzene (1.3 mL). Tubes were sealed and mixed 24 hours. After sitting for 30-60 minutes, an aliquot of each benzene layer (900 tiL) was added to DMF (500 nL), shaken, and the visible absorption spectra recorded. Each experiment was performed three times. Resin Exnerimsnts Plasma Experiments Human plasma (30 mL) was filtered (0.2 nm) and combined with ^^Othorium (45,000 cpm, 0.45 ^moie] and mixed for 10 minutes. To each of six tubes was transferred 4.5 mL of spiked plasma. Two tubes were empty prior to the addition, two contained control resin (2-6 mg), and two tubes contained catecholamide resin (2-6 mg). These samples were sealed and mixed for 24 hours at 37°C. Each sample was filtered (0.45 nm) and counted by liquid scintillation (500 nL sample plus 1 mL Aquasoi 2).

PAGE 132

126 Biolonical Fvaliiatinn Urinary ClRaranm 220Thorium (100,000 cpm in 0.09 M citrate) was administered rntraperitoneally. One hour later, chelators (100 or 200nmoles/kg in 40 percent Chremophor/PBS) were injected intraperitoneally. Animals were placed in metabolic cages with food and water ad libitum. Urine samples were collected for 6 hour inten/als. Gall Bladder Fxnisinn Thirty-six mice (A/J) weighing 28-32 grams were given ^SOfhorium (50,000 cpm in 0.09 M citrate) by intraperitoneal injection. One hour later, the mice received vehicle (50 percent Chremophor/PBS) or JI (200 nmoles/kg) intraperitoneally. Three mice from each group were sacrificed after 3, 6, 12, 24, 36, and 48 hours, and their gall bladders were removed and digested in tissue solubilizer (1 mL Soluene 350, Packard). After 24 hours, hydrogen peroxide (0.5 mL, 30 percent) and ethanol (0.5 mL) were added. Scintillation fluid (5 mL Aquasol 2) was added to each sample, and the thorium content measured by liquid scintillation counting. Bile Duct Canniilatinn Chronic cannulae were placed in the bile ducts of adult male Sprague-Dawley rats, allowing for continuous collection of bile. The cannula emerges from the incision and, using a skin-tunneling needle, is directed under the skin to behind the rat's neck. The emerging cannula is directed through a torque-transmitting spring tether to a fluid swivel located above the metabolic cage. The animal is able to move about freely in the cage while urine and bile were collected. After completing the cannulation procedure, rats were given 23o,horium (50,000 cpm in 0.5 mL 0.09 M citrate) by intraperitoneal injection. One hour later, chelators (100 or 200 timoles/kg in 40 percent Chremophor/PBS) were administered intraperitoneally. Bile was collected for 3 hour intervals for 48 hours.

PAGE 133

127 Potentiometrin Ms asurements A 20.00 mL sample which contained 0.1 M KCI and approximately 0.001 M polyamine was placed in a water-jacl
PAGE 134

128 HL of 1.0 N NaOH at 60°C for 60 minutes and acidified with 700 nL 1.0 N HCI. 800 nL was transferred to a scintillation viai for counting. Percent nptakP L1210 celis were exposed to 10 nlW ^''C-spermidine and 100 nN/I spermidine analog, and the inhibited uptai
PAGE 135

SUMMARY AND CONCLUSIONS This work has described the diverse applications of hexahydropyrimidines, both as biologically active molecules and as a reagent for the selective acylation of triamines, leading to the preparation of a wide variety of organic molecules. Hexahydropyrimidines have been shown to compete very effectively for uptake by the polyamine transport apparatus, supporting the idea that a cyclic conformation, assisted by intramolecular hydrogen bonding, is an important part of substrate recognition. These hexahydropyrimidines were also shown to be effective inhibitors of LI 21 cell proliferation in vitro . In addition, hexahydropyrimidines sen/ed as intermediates in the synthesis of H-shaped octacoordlnate catecholamide chelators. These chelators were shown to bind actinides selectively in solution, and to selectively precipitate actinides from solution, effectively decontaminating the solution. A resin-bound octacoordlnate chelator was synthesized and shown to be able to decontaminate aqueous solutions as effectively as the soluble chelator. This resin-bound chelator was able to chelate the protien-lxjund actinides present in human plasma, as well. A nitro derivative of a catecholamide chelator was synthesized. This derivative was utilized to measure the conditional formation constant for the complex formed between an H-shaped ligand and thorium(iV), and the stoichiometry of these complexes. H-shaped ligands were shownto be non-toxic in mice and rats, and were able to stimulate excretion of a large portion of an injected actinide burden, mainly via the feces. 129

PAGE 136

REFERENCES 1. D. H. Russell and S. H. Snyder, Procedings of the National Academy of Science fifl 1420 (1968). 2. H. G. Williams-Ashman, G. L Coppoc and G. Weber, Cancer Research 32 1924(1972). 3. D. H. Russell and C. C. Levy, Cancer Research 21 248 (1 971 ). 4. W. J. P. Neish and L. Key, international Journal of Cancer2 69 (1967). 5. D. H. Russell and J. B. Lombardini, Biochimie Biophysica Acta 24Q 273 (1971 ). 6. D. H. Russell, C. C. Levy, S. C. Schimpff and I. A. Hawk, Cancer Research 211555(1971). 7. C. W. Porter and R. J. Bergeron, Science 2ia 1083 (1983). 8. C. W. Porter, R. J. Bergeron and N. J. Stolowich, Cancer Research 42 4072 (1982). 9. R. A. Casero, Jr., R. J. Bergeron and C. W. Porter, Journal of Cell Physioloav 121476(1984). "' 10. C. W. Porter, P. F. Cavanaugh, Jr., N. J. Stolow/ich, B. Ganis, E. Kelly and R. J. Bergeron Cancer Research 45 2050 (1985). 11. M. M. Kimberly and J. H. Goldstein, Analytical Chemistry 52 789 (1 981 ). 12. M. Delfini, A. L. Segre, F. Conti, R. Barbucci, V. Barone and P. Ferruti, Journal of the Chemical Society, Pertains Transactions II 900 (1980). 13. J. S. McManis, Doctoral Dissertation, Cornell University (1980). 14. K. Chantrapromma, J. S. McManis, and 8. Ganem, Tetrahedron Letters 21 2475 (1980). 15. B. D. Breitenstein, Health Physics 45 855 (1 983). 16. I.e. Nelson, K. R. Held, P. A. Fuqua and T. D. Mahony, Health Physics 22 925 (1 972). 17. C. R. Lagerquist, S. E. Hammond and D. B. Hylton, Health Physics 22 921 (1 972). 18. C. L. Sanders, Health Physics 22 607 (1972). 1 9. D. W. Baxter and M. F. Sullivan, Health Physics 22 785 (1 972). 20. R. W. Bistline, R. L. Wallers, and J. L. Lebel, Health Physics 22 829 (1 972). 130

PAGE 137

131 21 . P. W. Durbin, Health Physics 22 495 (1975). 22. O. Matsuoka, M. Kashima, H. Joshima and Y. Noda, Health Physics 22 713 (1972). 23. P. W. Durtiin, M. W. Horowitz and E. R. Close, Health Physics 22 731 (1972). 24. A. Seidel and V. Volf, Health Physics 22 779 (1972). 25. V. Volf, Health Physics 22273(1974). ' ". . 26. R. 0. Lloyd, S. S. McFarland, G. N. Taylor, J. L. Williams and C. W. Mays, Radiation Research £297 (1975). 27. D. R. Kalkwarf, et al.. Health Physics 45 937 (1983). 28. G. N. Taylor and C. W. Mays, Health Physics 35 858 (1979). 29. V. Volf and A. Seidel, Radiation Research 52 638 (1974). 30. C. W. Mays, G. N. Taylor, R. D. Lloyd,and M. E. Wrenn, Actinides in Man and Animals, Los Alamos, California, R and D Press, 1981. 31 . vol of dist 32. vol of dist 33. P. W. Durbin, E. S. Jones, K. N. Raymond and F. L. Welti, Radiation Research 21170(1980). 34. F. L. Welti and K. N. Raymond, Journal of the American Chemical Society 1122 2289 (1980). 35. F. L. Weill, K. N. Raymond and P. W. Durbin, Journal of Medicinal Chemistry 2i203 1981. 36. R. D. Shannon, Acta Crystallographica A 22151 751 (1 976). 37. B. J. Stover, F. W. Baienger and W. Stevens, Radiation Research 22 381 (1968). 38. G. A. Turner and D. M. Taylor, Radiation Research 22 22 (1968). 39. C. W. Mays and T. F. Dougherty, Health Physics 22 793 (1972). 40. G. Boocock, C. J. Dampure, D. S. Popplewell, and D. M. Taylor, Radiation Research 42 381 (1970). 41. F. L. Welti, K. N. Raymond, W. L. Smith, and T. R. Howard, Journal of the American Chemical Society IM 1170 (1978). 42. Z. Szot, M. Rochalska, M. Wojewodzka, A. Chimiak, and W. Pryzchodzen, Radiation and Environmental Biophysics 25 31 (1986). 43. V. Volf, International Journal of Radiation Biology, 42 449 (1 985).

PAGE 138

3C^ •fP^X132 44. P. W. Durbin, E. S. Jones, K. N. Raymond, and F. L. Weill, Radiation Researcli 21170(1980). 45. R. D. Lloyd, F. W. Bruenger, D. R. Atherton, C. W. Jones, F. L Welti, P. W. Durbin, K. N. Raymond, G. N. Taylor, W. Stevens, and C. W. Mays, Research! in Radiobiology, University of Utah, COO-1 19-254, 151 (1979). 46. R. J. Bergeron, P. S. Burton, K. A. McGovern, and E. J. St. Onge, Journal of ttie American Chemical Society, 22 1130 (1980). 47. M. F. Sullivan, Health Physics 22 159 (1980). 48. D.W. Baxter and M.F.Sullivan, Health Physics 22 785 (1972). 49. D. Pressman, J. H. Bryden and L. Pauling, Journal of the American Chemical Society Zai352 (1970). 50. R. J. Bergeron, S. J. Kline, N. J. Stolowich, K. A. McGovern and P. S. Burton, Journal of Organic Chemistry IS 4524 (1981). 51. R. J. Bergeronand J. R.Garlich, Synthesis 782 (1984). 52. J. M. Bobbitt, Thin Layer Chromatography, London, Reinhold Publishing Corporation, 1963. 53. C. M.TiceandB. Ganem, Journal of Organic Chemistry 42 2106 (1983). 54. A. G. M. Barrett and J. C. A. Lana, Journal of the Chemical Society, Chemical : Communications 471 (1978). 55. A. G. M. Ban-ett, J. C. A. Lana and S. Tograie, Journal of the Chemical Society, Chemical Communications 300 (1980). 56. R. B. Merrifield, Biochemistry 2 1385 (1964). T \ 57. J. T.Ayres and C.K.Mann, Polymer Letters 2 505 (1965). 58. J. M. Frechet and C. Schuerch, JACS 22 492(1 971). 59. J. Cavadore, J. Derancourt and A. Previero, FEBS Letters £S 155 (1975). 60. T. Kusama and H. Hayatsu, Chem. Pharm. Bulletin 18(2> 319 (1970). 61. P. Job, Annales de Chimie (10) 2113 (1928). 62. R. J. Bergeron, P. S. Burton, S. J. Kline, and K. A. McGovern, Journal of Organic Chemistry 42 3712 (1981). 63. K. A. McGovern, Doctoral Dissertation, University of Florida (1983). 64. R. J. Bergeron, in preparation.

PAGE 139

* 133 65. S.B.Sawin,Talanta 111 (1964). 66. R. J. Bergeron, S. J. Kline and J. D. Navratil, Radiochimica Acta 25 47 (1984). 67. C. J. King and J. D. Navratil, Ctiemica! Separations Volume II. Applications, Denver, Colorado, Litarvan Literature, 1986. 68. P. F. Lott, K. L. Cheng and B. C. H. Kwan, Analytical Chemistry 32 1 702 (1 960). 69. S. B. Dabhade and S. P. Sangal, Microchemical Journal 13. 1 93 (1 968). 70. F. J. C. Rossotti and H. Rossctti, The Determination of Stability Constants, New York, New York, McGraw-Hill, 1961. 71 . B. J. Hathaway and A. A. G. Tomlinson, Coordination Chemistry Review 5 1 (1970). 72. S. Shibata, M. Funjkawa, E. Kamata and K. Goto, Analytica Chimica Acta 5Q 439 (1970). 73. R. G. Anderson and G. Nickless, Analyst 22 207 (1 967). 74. S. Shibata, Analytica Chimica Acta 25 348 (1 961 ). 75. A. I. Busev and V. M. Ivanov, Zhurnal Analiticheskoi Khimii 12 1238 (1964). 76. Dr. R. J. Bergeron, private communication. 77. E. St. Onge, Master's Thesis, University of Florida (1982). 78. C. W. Porter, J. Miller and R. J. Bergeron, Cancer Research M 126 (1984). 79. N. J. Stolowlch, Doctoral Dissertation, University of Florida (1983). 80. F. H.Sweeton, etal.. Journal of Solution Chemistry a 191 (1974). 81. H. S. Harned and J. M. Harris, Jr., Journal of the American Chemical Society 52 2633 (1928). 82. L. P. Kuhn, Journal of the American Chemical Society 24 2492 (1951). 83. R. Barbucci, P. PaolettI and A. Vacca, Journal of the Chemical Society (A) 2202 (1970). 84. J. C. Cain and J. L Simonsen.Joumal of the Chemical Society 125 156 (1914). 85. C. Robinson and H. A. T. Mills, Proceedings of the Royal Society (London) Aiai 576 (1931). 86. J. Korbl and R. Pribil, Chemist-Analyst 45 102 (1956).

PAGE 140

BIOGRAPHICAL SKETCH Howard Wayne Seligsohn was born in Philadelphia, Pennsylvania on February 9, 1960. He moved to New Port Richey, Florida in 1973 and graduated from Hudson High School in 1977. The author enrolled in Eckerd College, located in St. Petersburg, Florida for two years before transferring to the University of Florida in Gainesville, Florida, where he earned the B. Sc. In mathematics in 1 980 and the B. Sc. in chemistry in 1 981 . He then entered graduate school at the University of Florida. Working under the guidance of Dr. Raymond J. Bergeron, he earned his Ph. D. In medicinal chemistry In 1987. 134

PAGE 141

I certify that I have read this study and that in my opinion it conforms to acceptable standards of schoiarly presentation and is fuiiy adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Raymond J.-Bergeron, Chairman Professor, 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. /. Richard R. Streitf /^ Professor, Medicinal Chemjgfry 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. x >" /'— Kenneth B. Sloan Associate Professor, 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. Merle A. Battiste Professor, 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. Margaret O. James Associate Professor, Medicinal Chemistry

PAGE 142

This dissertation was submitted to tfie Graduate l:aculty of the Coiiege of Pharmacy and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. '-' August 1987 / Dean, College of-Pharmacy Dean, College of-Pharmacy
PAGE 143

1 :t' < f UNIVERSITY OF FLORIDA 3 1262 08554 8328 ...yj :i