Synthesis and evaluation of spermidine siderophores for iron chelation therapy and the role of cyclohexaamylose in catal...

MISSING IMAGE

Material Information

Title:
Synthesis and evaluation of spermidine siderophores for iron chelation therapy and the role of cyclohexaamylose in catalytic hydrolysis
Physical Description:
vi, 148 leaves : ill. ; 29 cm.
Language:
English
Creator:
Burton, Philip Seth, 1949-
Publication Date:

Subjects

Subjects / Keywords:
Shelating Agents   ( mesh )
Starch   ( mesh )
Chemistry, Physical   ( mesh )
Spermidine   ( mesh )
Iron Chelates   ( mesh )
Medicinal Chemistry thesis Ph.D   ( mesh )
Dissertations, Academic -- medicinal Chemistry -- UF   ( mesh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1981.
Bibliography:
Bibliography: leaves 142-147.
Statement of Responsibility:
by Philip Seth Burton.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000892083
oclc - 25097388
notis - AEK0553
System ID:
AA00009117:00001


This item is only available as the following downloads:


Full Text










SYNTHESIS AND EVALUATION OF SPERMIDINE SIDEROPHORES
FOR IRON CHELATION THERAPY AND THE ROLE
OF CYCLOHEXAAMYLOSE IN CATALYTIC HYDROLYSIS




BY

PHILIP SETH BURTON


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

UNIVERSITY OF FLORIDA


1981

































This work is dedicated to Pat and Eric for continued patience

and support through its completion.













ACKNOWLEDGEMENTS

I would like to thank my research advisor, Dr. Raymond J. Bergeron,

without whose help and guidance this work would not have been completed.

Dr. Richard R. Streiff of the V.A. Hospital, Gainesville, Florida, has

also been extremely helpful in the completion of the siderophore studies.

The computer program, PHFIT, used to analyze the potentiometric

titration data, was kindly supplied by Dr. Daniel Leussing of Ohio State

< University, Columbus, Ohio.

' The 220 MHz nmr chemical shift measurements of dodecakis-2,6-0-
Smethylcyclohexaamylose were provided by Dr. Michael A. Channing of the

National Institutes of Health, Bethesda, Maryland.
Q
4The 200 MHz spectra of Nl,NB-bis(2,3-dihydroxybenzoyl)spermidine

were provided by Mr. Leroy F. Johnson of Nicolet Technology, Mountain

View, California.
C Last, but certainly not least, I would like to thank Kathy McGovern
0
m for her many contributions to this work.

0

C>













TABLE OF CONTENTS

CHAPTER PAGE

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

ABSTRACT....................... ............. ............ v

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

II PREPARATION AND EVALUATION OF SPERMIDINE DERIVATIVES AS
IRON CHELATORS AND ANTINEOPLASTIC AGENTS................. 24

Experimental......................... ........ ..... ... 24
Results.............................................. 72
Discussion......................................... .. 87

III EVALUATION OF ROLE OF THE CYCLOHEXAAMYLOSE C-3 HYDROXYL
IN CATALYTIC HYDROLYSIS...................... ...... ..... 119

Experimental .......................................... 119
Results....................................... 122
Discussion........................................... 130

IV CONCLUSIONS.................. .................... ..... 139

REFERENCES............................................... 142

BIOGRAPHICAL SKETCH .................................... 148












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


SYNTHESIS AND EVALUATION OF SPERMIDINE SIDEROPHORES
FOR IRON CHELATION THERAPY AND THE ROLE OF
CYCLOHEXAAMYLOSE IN CATALYTIC HYDROLYSIS

By

Philip Seth Burton

August, 1981

Chairman: Raymond J. Bergeron
Major Department: Medicinal Chemistry

The first part of the dissertation describes the synthesis and evalua-

tion of spermidine siderophores related to two natural products isolated

from Paracoccus denitrificans. Employing N4-benzylspermidine as the key

starting material, the synthesis of the natural product, NI,N8-bis(2,3-

dihydroxybenzoyl)spermidine is described in three high yield steps.

A biomimetic approach to the synthesis of N4-acylated NI,N8-bis(2,3-

dihydroxybenzoyl)spermidine siderophores is also described. The key to

the sequence is the use of Cu(II) as a transitory protecting group for

the catechols of N1,N8-bis(2,3-dihydroxybenzoyl)spermidine in the conden-

sation of this synthon with the model acylating agent N-(2-hydroxybenzoyl)-

glycine.

Preliminary investigations aimed at characterizing the Cu(II) chelate

are also described. Proton nmr paramagnetic line broadening studies con-

firm the catechols of N1,N8-bis(2,3-dihydroxybenzoyl)spermidine to be the

chelating functionality. The stoichiometry of the complex is determined

v







by spectrophotometric methods while the cumulative formation constants

of the various protonated species are established by potentiometric tech-

niques.

The toxicity and iron binding characteristics of both NI,N -bis(2,3-

dihydroxybenzoyl)spermidine and N [N(2-hydroxybenzoyl)glycyl]Nl,N8-bis(2,3-

dihydroxybenzoyl)spermidine are evaluated establishing their potential

as clinical agents for the treatment of iron overload syndromes.

Also described is the coupling of NI,NS-bis(2,3-dihydroxybenzoyl)-

spermidine with 2,3-dihydro-1H-imidazo[1,2-b]pyrazole (IMPY). The fea-

sibility of this compound as a delivery device for antineoplastic agents

is discussed.

The second part of the dissertation examines both the static and dy-

namic aspects of the sodium 4-nitrophenolate and sodium 2,6-dimethyl-4-

nitrophenolate dodecakis 2,6-0-methylcyclohexaamylose complexes. The geo-

metries of these complexes are determined based on intermolecular homo-

nuclear nuclear Overhauser enhancements and chemical shift analyses while

the dynamic parameters of complexation i.e., rotational correlation times

and coupling constants, are calculated from 13C['H]T1 measurements. Com-

parison of these results with analogous measurements of the unmethylated

cyclohexaamylose inclusion complexes of these same substrates indicates

that the substrates are prevented from equivalent penetration of the methy-

lated derivative. This casts some doubt upon the premise that the cyclo-

hexaamylose 3-hydroxyls are inherently unreactive and therefore uninvolved

in cyclohexaamylose catalyzed hydrolyses.












CHAPTER I

INTRODUCTION
The first part of this dissertation is concerned with the synthesis

and evaluation of spermidine derivatives for use as therapeutic iron che-

lators. Iron is an essential element for life as we know it.1-3 Most,

if not all, cells have varying iron requirements for existence. In man,

iron is involved almost exclusively in processes related to cellular

respiration.4 Most of the iron in the body is either heme associated5 or

bound to iron transport and/or storage proteins.5 Iron metabolism is char-

acterized by efficient reutilization within an essentially closed system.

In normal individuals a daily absorption of about 1 mg of iron is balanced

by an equivalent loss.6 The total iron content of the body is about 4-5

grams,7 and is in dynamic equilibrium between the various iron "pools."

For instance, in the hemoglobin cycle, plasma iron is transferred to the

bone marrow, incorporated into red blood cells, and after the cells die,

returned to the plasma.8 The storage iron proteins, ferritin and hemo-

siderin, function as internal iron reserves to be used in the event of

sudden iron losses resulting from heavy bleeding.4

Since there is no mechanism for the excretion of large amounts of

iron from the body, its absorption from the intestine is very closely

controlled. While, in the ordinary diet, 10-20 mg of iron are ingested

daily, usually no more than 10% of this is actually absorbed.9 In the

event of increased iron absorption, the excess iron is concentrated in

the storage proteins i.e., hemosiderin in the spleen, liver and bone mar-

row, and ferritin in the parenchymal cells of the liver, muscle and other





2
organs.10 A large excess of iron in the body,accumulated over extended

periods of time, results in damage to internal organs and eventual death

for the individual.11'12 This process of iron accumulation is referred

to as iron overload.

Iron overload resulting from any of a number of defects is referred

to as hemochromatosis.10 These disorders are further divided into pri-

mary idiopathicc) and secondary hemochromatosis. Primary hemochromatosis

is the result of inherited deficiencies in iron metabolism which lead to

increased iron absorption."1 In advanced stages of this disease, frequent

symptoms are liver hepatomegaly, diabetes and hepatocellular carcinoma.

Furthermore, increased iron body stores of twenty to fifty times that of

normal are often accompanied by congestive cardiac failure.10

Treatment of idiopathic hemochromatosis involves the removal of the

excess iron and treatment of the related complications. The excess iron

is usually removed by phlebotomy.10 Approximately 500 ml of blood removed

weekly mobilizes up to 250 mg of iron. Usually, a regimen of two to three

years of weekly phlebotomy is required to return a heavily overloaded

patient to normal iron stores. However, these individuals still require

regular phlebotomy in order to maintain negative iron balance.

The other distinct, and clinically more difficult to manage, form of

iron overload is secondary hemochromatosis.10 The excess iron in this

instance often occurs as a result of treatment for anemia. In these

anemias errors in red blood cell metabolism frequently necessitate regu-

lar transfusions of blood as the only red blood cell source. This trans-

fused iron (250 mg/500 ml blood) is, in some cases, supplemented even fur-

ther by increased iron absorption due to ineffective erythropoiesis.11

Included within this spectrum of disorders are B-thalassemia, sickle cell

and other inherited haemolytic anemias. The distribution of iron deposits







and related damages in these cases is similar to that in primary hemo-

chromatosis. In addition, the incidence of hepatic cirrhosis and deposi-

tion of iron in the myocardium and endocrine organs is much higher in

this group.10

The clinical history of a typical case of B-thalassemia illustrates

the magnitude of the problem. Thalassemia is one of the most commonly

inherited disorders and is the most prevalent cause of hemochromatosis

in the world. Good health may be maintained in the first decade of life

with regular blood transfusions. However, growth usually slows in the

second decade and may be accompanied by disturbed endocrine-function.

The massive deposition of iron in the heart and conduction fibers causes

arrhythmias and cardiac failure with death usually occurring between the

ages of 16 and 22 years.12 Significantly, the cause of death is not the

anemia, but rather the iron overload which is secondary to the treatment

of that anemia.

The problem of removing excess iron in such cases of secondary iron

overload is more difficult than in the case of primary hemochromatosis.

Obviously, phlebotomy is inappropriate in this instance. Clinically, the

problem is dealt with by employing an iron chelator to mobilize the excess

iron and facilitate its excretion. The most effective and least toxic

chelating agent used to date is desferrioxamine, a hydroxamic acid natural

product isolated from Streptomyces pilosus.13 Unfortunately, desferriox-

amine is not very satisfactory as it is poorly absorbed across the stomach

or intestinal walls and therefore cannot be administered orally.14,15

Furthermore, the short half-life of the compound in the body (76 minutes

IV) contributes to its inability to remove significant amounts of iron.16'17

Consequently, the accepted route of administration of desferrioxamine is

via subcutaneous or intravenous continuous infusion. Even under these








optimal conditions, daily infusions of 4-15 g (0.7-2.8 x 10-2 moles) of
desferrioxamine over periods of up to 12 hours result in iron excretion

of only 10-40 mg (1.8-7.2 x 10-4 moles).10 Furthermore, the ability to

benefit from this therapy is limited by the availability of infusion

apparatus. Clearly, more effective and more easily administered iron

chelators are desirable in order to overcome these serious problems.

In recent years, the search for new, orally effective iron chelators

has been undertaken by a number of investigators using a variety of

screening techniques.18-20 One recent study, for example, involved the

screening of over forty different compounds in iron overloaded rats.19

None of these compounds proved to be clinically useful. Unfortunately,

few, if any, promising compounds have been identified as a result of these

investigations.

However, recent work with two conjugates of 2,3-dihydroxybenzoic acid

isolated from bacteria cultures indicated that these compounds may, indeed,

be of potential therapeutic value. In 1975, Tait isolated these two novel

siderophores from cultures of Micrococcus denitrificans (now referred to as

Paracoccus denitrificans)21 when grown on iron deficient media.22 Sidero-

phores are low molecular weight compounds with extremely high affinities

for iron secreted by bacteria in order to sequester extracellular iron and

transport into the cell.23 These accomplish this by "masking" the ionic

charge on the metal, thereby lowering the insertion energy required to

move it across the cell membrane. These natural iron chelators also facil-

itate mobilization of iron by improving the solubility of Fe3+ (Ksp Fe(OH)3-
10-39)4 in relatively neutral solutions.

Tait's compounds, N',N8-bis(2,3-dihydroxybenzoyl)spermidine and N4[N-

2-hydroxybenzoyl)threonyl]Nl,N8-bis(2,3-dihydroxybenzoyl)spermidine (Fig-

ure 1) were novel in that spermidine (N[3-aminopropyl]l,4-diaminobutane)















HO:Q
HO


OH
OH,


CONHCH CH2CH 2NHCH2CH2CH2CH2NHCO
2" 2 2 2 ,


OH


HO CH3 HCH
HO NH HO CO 2C2CH N
SONHCH 2Ch2CH2NCH2CH2 CH2 CH2NHCO


HO NH

HO-'
CONH


Figure 1.


NHC O
CHH2CH2CH2NHCO
2 2 2


Siderophores isolated from Paracoccus denitrificans
a) NI,N1-bis(2,3-dihydroxybenzoy )spermidine; b) N4-
[N-(2-hydroxybenzoy1)threonyl]N1,N"-bis(2,3-dihydroxy-
benzoyl)spermidine; c) N-[3-(2,3-dihydroxybenzamido)-
propyl]-N-[4-(2,3-dihydroxybenzamido)butyl]-2-(2-
hydroxyphenyl)-5-methyloxazoline-4-carboxamide.








was employed by the bacteria as the backbone to which the catechol and

salicyloyl chelating functionalities were fixed. Since Tait's original

isolation was reported, it has been shown that the threonine containing

compound is actually present in the bacteria in the form of an oxazole

(Figure Ic).21 The structure reported by Tait is a hydrolysis product

generated during his purification procedure.

Many compounds, both natural and synthetic, have been identified as

powerful ligands for iron in vitro, only to be shown ineffective in re-

moving iron in vivo in the many screening programs to date. However,

Jacobs, White and Tait showed that these spermidine siderophores were able

to very effectively remove iron from transferring in vitro (Figure 2) and

prevent incorporation of iron from Chang cells.25 Furthermore, comparison

of these results with those obtained for desferrioxamine was very encour-

aging. Both NI,N8-bis(2,3-dihydroxybenzoyl)spermidine and N [N(2-hydroxy-

benzoyl)threonyl]N1,N8-bis(2,3-dihydroxybenzoyl)spermidine were far more

effective at removing iron from transferring and preventing incorporation

of iron into Chang cells than desferrioxamine (Table 1). Thus, these

compounds were extremely desirable candidates for therapeutic evaluation

in the treatment of iron overload. However, Jacobs, White and Tait cited

the very low biological yields of these compounds, and the tedious iso-

lation procedures as major obstacles in their utilization.25

The results of this preliminary investigation along with the appar-

ent structure simplicity of the compounds suggested an approach to their

total synthesis be devised. The approach decided on was to fix the ter-

minal, primary and amino acyl groups to spermidine, followed by function-
alizationof the secondary nitrogen with an appropriate synthon. This concept

is outlined for N4[N(2-hydroxybenzoyl)threonyl]N',NB-bis(2,3-dihydroxy-
benzoyl)spermidine in Figure 3. The first problem to be overcome was that

















50




40






4.--
E
30



s: 0



Lo
10-


0 0

A-0 0

0.2 0.4 0.6 0.8 !.0
concentration of chelator (mmol/z)

Figure 2. The removal of iron from transferring by increasing
concentrations of three chelating compounds. Des-
ferrioxamine 0 N',N8-bis(2,3-dihydroxybenzoyl)-
spermidine NI[N-(2-hydroxybenzoyl)threonyl]-N ,N8,
bis(2,3-dihydroxybenzoyl)spermidineA(from ref. 25).








Table 1. Iron Uptake and Ferritin Synthesis


N1,N8-bis(2,3-dihydroxybenzoyl)
spermidine


0.025 mM

0.1 mM

0.5 mM


N4-[N-(2-hydroxybenzoyl)threonyl]
NI,Nb-bis(2,3-dihydroxybenzoyl)-
spermidine


a


59Fe Incorporation (% control)a
Whole cell Ferritin


0.025 mM

0.1 mM

0.5 mM


Desferrioxamine


0.025 mM

0.1 mM

0.5 mM


from ref. 25











H2N N NH


CONH N NH


NH2



OH

y -OH
NHCO


C OH
CO


NH
I
CH3CHCH
IO
HO CO


CONH N


-OH

OH


, CO
NHCO


Preparation of N'-[N-(2-hydroxybenzoyl)threonyl]-N',N8-
bis(2,3-dihydroxybenzoyl)spermidine.


HO-

HO'


HO

HO Y


Figure 3.






of selective N',N8-acylation of spermidine in the presence of the secon-

dary N4-nitrogen. The solution has been described previously and has led

to the key synthon N4-benzylspermidine.26,27 The significance of this

reagent is that it is synthesized with the N4-position of spermidine pro-

tected. Therefore, selective acylation of the terminal nitrogens is eas-

ily accomplished with a variety of acylating agents. Then, the N4-pro-

tecting group may be removed to yield the free amine which may be func-

tionalized with a different acylating agent. Thus, when N4-benzylspermi-

dine was reacted with excess 2,3-diacetoxybenzoyl chloride, NI,N8-bis(2,3-

diacetoxybenzoyl)N4-benzylspermidine was obtained in 98% crude yield.

Hydrolysis of the acetoxy protecting groups with sodium methoxide in meth-

anol followed by removal of the N4-benzyl group under conditions of cata-

lytic hydrogenolysis afforded the natural product, N',N8-bis(2,3-dihydroxy-

benzoyl)spermidine, in excellent yield (Figure 4).28

Initial toxicity and iron clearing experiments with this compound

have shown it to be fairly nontoxic in mice and capable of removing

significantly more iron from iron overloaded rats than desferrioxamine.

Furthermore, in vitro experiments have shown it to be effectively absorbed

across rat intestines suggesting its potential for oral efficacy.28

Use of N1,N8-bis(2,3-dihydroxybenzoyl)spermidine as a synthon for

the preparation of the N4-substituted natural product was complicated by
the mixed functionality present. The problem of polyacylation of the

phenolic oxygens along with the nitrogen had to be dealt with. The solu-

tion involved the transitory protection of the phenolic oxygens by com-

plexation with copper in a sequence designed to mimic the biosynthetic

pathway to the N4-acylated siderophore.29 In his original investigation,

Tait demonstrated that Paracoccus denitrificans uses N1,N8-bis(2,3-dihy-

droxybenzoyl)spermidine in the synthesis of N-(3-(2,3-dihydroxybenzamido)-

















H2 .
N


H C020

H 3C 02/CO
C ONN .


H OCOC-3

NHCO


HO O

HO T O
HH 0 OH
C 0 N NHCO




HO' 0N
CONH- -,- NH ,,
'-*' *^' ^' r^ ^ nn


Figure 4. Preparation of NI,N8-bis(2,3-dihydroxybenzoyl)-
spermidine.


a
N^

c^







propyl)-N-(4-(2,3-dihydroxybenzamido)butyl)-2-(2-hydroxyphenyl-5-methyl-

oxazoline)-4-carboxamide.22 When 14C labeled compound Nl,N8-bis(2,3-

dihydroxybenzoyl)spermidine was fed to the microorganism, along with sali-

cyloyl threonine, the 14C labeled addition product was isolated. This
implied that, in vivo, the catechol hydroxyls are, in some way, protected

from acylation.

The decision to use copper for the transitory masking of the catechol

oxygens during the course of the acylation reaction was based upon a number

of considerations. Copper is well known to form strong tetracoordinate

complexes with soft anions, the strength of the chelate being dependent

upon the pH.30 At high pH, with the chelate in the form of the polyanion,

the copper complex should be extremely stable. Furthermore, cpk space

filling models revealed that, in such a complex, the secondary nitrogen

would be easily accessible to an acylating agent, certainly more acces-

sible than in the parent compound itself. However, upon protonation of

the catechol oxanions, the complex should dissociate, thus providing a

mechanism for removing the copper once having acylated the secondary

nitrogen.

In order to evaluate the feasibility of this approach, it was first

necessary to demonstrate that NI,N8-bis(2,3-dihydroxybenzoyl)spermidine

did, indeed, bind copper and that the association was strong enough to

effectively act as a protecting group. This was accomplished by means of

potentiometric titration of the compound in the presence of copper. The

stoichiometry of the complex was determined to be 1:1 by means of Job's

plots employing the method of continuous variations. Furthermore, the

catechol hydroxyls were identified as the site of chelation by means of

a proton nmr line broadening experiment with Cu(II) serving as a paramag-

netic probe.




13

Having thus determined that the approach was feasible, the model com-

pound N [N(2-hydroxybenzoyl)glycyl]N1,N8-bis(2,3-dihydroxybenzoyl)spermi-

dine (Figure 2) was synthesized. This conversion was effected by reacting

the preformed Cu(II)-NI,N8-bis(2,3-dihydroxybenzoyl)spermidine complex

with an activated form of N(2-hydroxybenzoyl)glycine in dimethylacetamide.

After an acidic work-up the product was isolated in 70% yield with no evi-

dence of 0-acylation. Structurally, this compound is identical to Tait's

hydrolysis product, N4[N(2-hydroxybenzoyl)threonyl]N ,N8-bis(2,3-dihydroxy-

benzoyl)spermidine, lacking only the threonine side chain. Since it seemed

reasonable that the side chain was not functionally involved in iron seques-

tration, this compound was also screened for iron clearing ability in the

iron overloaded rat, where it was shown to be about as effective at iron

mobilization as desferrioxamine.31

While providing the essential starting point for all of the synthe-

ses attempted thus far in ouriron chelator, the polyamines, of which sper-

midine is a member, possess other interesting properties. They appear to

be ubiquitous in nature and are involved in a variety of inter- and intra-

cellular functions.32 Recent work related to the cellular incorporation

of spermidine suggested that the ability to synthesize these compounds

might be useful in capacities other than simply providing a backbone for

potential iron chelators. In investigations dealing with the uptake of

polyamines into L-1210 leukemia cells, it was shown that spermidine and

spermidine derivatives are concentrated in these cells.33 Furthermore,

it has recently been demonstrated that N4-benzylspermidine is specifi-

cally incorporated into these same test cells.34 These observations sug-

gested that spermidine and the substituted spermidines could be useful as

delivery devices for antineoplastic agents. Other recent work with Erh-

lich tumor cells resulted in a second interesting observation with respect







to Nl,N8-bis(2,3-dihydroxybenzoyl)spermidine. These investigations demon-
strated the ability of this compound to potentiate the activity of 2,3-

dihydroxy-1H-imidazo(1,2-b)pyrazole (IMPY) (Figure 5a). IMPY is an anti-

neoplastic compound that works by rapidly inhibiting DNA synthesis35 and

is currently in phase II clinical trial. It has further been shown that

the inhibition by IMPY involves the enzyme ribonucleotide reductase.36

This is an iron containing protein and the inhibitory effect of IMPY is

readily reversible upon addition of iron to the system. Furthermore, the

inhibition could be potentiated by addition of EDTA with IMPY.37 These

results suggested that IMPY's activity involved interaction with the pro-

tein's iron. It has since been shown that other iron chelators, among

them desferrioxamine, possess the same ability to potentiate IMPY induced

inhibition of ribonucleotide reductase.38

In a preliminary investigation employing N',N8-bis(2,3-dihydroxyben-

zoyl)spermidine, Cory found the most dramatic potentiation of IMPY acti-

vity yet observed for an iron chelator.38 For this reason it seemed rea-

sonable that a compound in which both the iron chelating functionality and

the IMPY moiety were covalently attached might have even greater activity.

Therefore, the IMPY containing chelate, N[3-(2,3-dihydroxybenzamido)propyl]-

N[4-(2,3-dihydroxybenzamido)butyl]-N[4-(2,3-dihydro-l H-imidazo((1,2-b)-

pyrazolo)carboxamido]butyramide (Figure 5b) was synthesized. The synthetic

route is fairly general and offers the advantage of allowing the introduc-

tion of other types of functionality. Therefore, coupling other antineo-

plastics to the spermidine derivatives would be feasible as well.

The sequence begins with, again, N4-benzylspermidine. This reagent

is condensed with two equivalents of 2,3-dimethoxybenzoyl chloride in

methylene chloride to give the N1,N8-bis(2,3-dimethoxybenzoyl)N"-benzyl-

spermidine adduct. Removal of the benzyl group is conveniently accomplished





































I N

CO

() 23


fCal-O
C'i.


U,


C ONH H2H2NCH2C2C2CH2NHCO
CONHC2CH2 CH 2 NCH2CH2CH. 2 NHCO


Structures of a) 2,3-dihydro-1H-imidazo[1,2-b]-
pyrazole; b) N-[3-(2,3-dihydroxybenzamido)propyl]-N-
[4-(2,3-dihydroxybenzamido)butyl]-N-[4-N-23-dihydro-
1H-imidazol [,2-b]pyrazolo)carboxamido]butyramide.


HO

w Hn


Figure 5.


1 iV^







by hydrogenolysis in acetic acid over a palladium catalyst. IMPY was

N-acylated with glutaric anhydride to give the zwitterionic amino acid

product. This compound was then activated at the acid function by forma-

tion of the N-hydroxysuccinimide ester by reaction with dicyclohexylcarbo-

diimide and N-hydroxysuccinimide in dry pyridine. The resulting active

ester was then smoothly appended to NI,N8-bis(2,3-dimethoxybenzoyl)spermi-

dine.

Removal of the methyl ether protecting groups resulted in the cate-

chol product (Ficure 6). The overall yield for the sequence was good,

with all the intermediates readily purified. The final product was suc-

cessfully chromatographed on Sephadex LH-20. The synthesis is obviously

flexible enough to allow for incorporation of a number of groups to the

terminal nitrogens. Also, it is possible to append a variety of inter-

esting compounds to the N4-nitrogen as well as vary the length of the

connecting carbon fragment.

The second part of this dissertation deals with an evaluation of the

role of cyclohexaamyloses, C-3 hydroxyl in catalytic hydrolysis. The

cycloamyloses are a series of cyclic oligosaccharides containing from 6-

12 a-1,4 linked glucopyranose units.39,40 They are produced by the action

of Bacillus macerans and Bacillus megaterium when these bacteria are grown

on culture media rich in amylose.41 Cyclohexaamylose contains six glucose

residues (Figure 7) and has been the most extensively investigated of the

series. A high field 1H nmr study revealed precise solution conformational

information.42 Chemical shifts and coupling constants were determined from

computer simulation of the spectrum recorded at 220 MHZ. The results

demonstrated that, on the nmr time scale, all six glucose units were con-

formationally identical and the molecule possessed hexagonal symmetry.
Additional high field nmr investigations in DMSO indicated strong hydrogen














H

CN

CO(CH2)3CO2H
I

'\NN/


<



'CO NH
H3CO"

CO NH0


-N

CO
U 0
IL


23 "01H3

, NHCO 3


cN-N
Nj


CO
(2H2)3 OH
CO
1 OH
N N HC
"^^ t^,^HCO


Preparation of N-[3-(2,3-dihydroxybenzamido)propyl]-
N-[4-(2,3-dihydroxybenzamido)butyl]-N-[4-(N-2,3-
dihydro-1H-imidazol[1,2-b]pyrazolo)carboxamido]-
butyramide


Figure 6.


























GC10H


HO


CH2 H


Figure 7. Cyclohexaamylose







bonding between the C-2 and C-3 hydroxyls on contiguous glucose units.43

Overall, then, the molecule assumes the conformation of a torus shaped

cavity with the C-2 and C-3 hydroxyls situated at the large face and the

C-6 primary hydroxyls at the narrow face. The strong hydrogen bonding

between the C-2 and C-3 hydroxyls imparts rigidity to the structure, al-

lowing little rotational freedom about the glycosidic bond.44

One of the more interesting properties of the cycloamyloses is their

ability to form complexes by inclusion of a suitable substrate within the

cavity. Over one hundred different substrates have been identified, rang-

ing from the noble gases to fatty acyl coenzyme A compounds.45

The cycloamyloses also exhibit the ability to catalyze certain types

of reactions with suitably functionalized molecules while in the bound

state. For instance, in the case of the aqueous hydrolysis of esters,

inclusion of the substrate followed by nucleophilic attack on the ester

by either the C-2 or C-3 hydroxyl oxygen results in an acyl cycloamylose

intermediate.40,46 Finally, hydroxide ion from the solvent cleaves the

acyl intermediate to produce free cycloamylose and acid.46 This reaction

sequence is very similar to reactions catalyzed by enzymes. Furthermore,

the cycloamyloses also exhibit many other enzyme characteristics such as

stereoselective binding, competitive binding and Michaelis-Menten type

kinetics.47 Accordingly, they have been widely employed as enzyme active

site models.48,49

In all of the cycloamylose hydrolyses of guest molecules, it has been

assumed that the cycloamylose C-2 hydroxyls are active and not the C-3

hydroxyls.46'47 This assumption is based on two lines of evidence: 1) un-

der basic conditions, the C-2 hydroxyls are substantially more reactive
with electrophilic reagents than the C-3 hydroxyls,50 suggesting they have







a lower pKa, and 2) when cycloheptaamylose and cyclohexaamylose are ex-

haustively 2,6-0-methylated,51 they lose their catalytic activity under

conditions in which they were previously active.46,52

With respect to the first line of evidence, early x-ray studies53,54

suggested the C-3 hydroxyl hydrogens were hydrogen bonded to the C-2 hy-

droxyl oxygens. This was later substantiated by 1H nmr studies in DMSO.43

Such a hydrogen bonding scheme is certainly consistent with the idea that

the C-2 hydroxyl has the more acidic proton. However, it must be empha-

sized that the forces responsible for the crystal structure of the cyclo-

amyloses are not the same as those in solution and, furthermore, that the

hydrogen bonding network of the cycloamyloses in DMSO might well be dif-

ferent in water. Thus, conclusions regarding the relative values of the

C-2 and C-3 hydroxyl pKa's based on these measurements are not necessar-

ily valid in aqueous solution.55

However, the fact that the 2,6-per-O-methyl cycloamyloses lose their

hydrolytic activity still seemed strong evidence in support of the fact

that the C-2 hydroxyls were active in hydrolysis. It seemed reasonable

that an investigation of the structure of dodecakis-2,6-0-methylcyclo-

hexaamylose substrate complexes in aqueous solution might shed some light

on this apparent anomaly.

Cyclohexxamylose, because of its sixfold symmetry, provides a par-

ticularly simple nmr spectrum and is well suited to nmr investigations.

By measuring the chemical shift effects observed upon complexation (espe-

cially the cycloamyloses' interior H-3 and H-5 protons) it is possible

to determine the direction and depth of penetration of the substrate into

the cavity.56,57 Similarly, observations of any substrate-experienced

homonuclear Overhauser enhancements upon decoupling of the cycloamylose







interior H-3 and H-5 protons provide another means of determining the

depth of penetration.57,58 Finally, changes in the overall molecular

rotational correlation times obtained from '3C['H]TI measurements of both

host and guest upon complexation provide a means of probing the dynamic

aspects of the binding process.58'59 A dynamic coupling coefficient, e,

the ratio of substrate correlation times to that of host, may be calcu-

lated which describes the "tightness" of the complexation.

In an earlier investigation addressed to the mechanism of the cyclo-

hexaamylose accelerated hydrolysis of para- and meta-nitrophenylacetates,

the nmr techniques described above proved to be very useful.58 60 Employ-

ing sodium 4-nitrophenylate and sodium 2,6-dimethyl 4-nitrophenylate as

models for the corresponding esters (Figure 8), both the position and

dynamic coupling of the substrates upon binding in cyclohexaamylose were

determined. The results of these experiments indicated that 2,6-dimethyl-

4-nitrophenolate coupled with the cyclohexaamylose cavity twice as effec-

tively as the 4-nitrophenolate anion. Furthermore, rotation of the methyl

groups of 2,6-dimethyl-4-nitrophenolate about the methyl aromatic carbon-

carbon single bond was slowed by cyclohexaamylose complexation. This re-

duced rotation of the methyl groups was attributed to contact between the

methyl groups and the "rim" of the cycloamylose cavity.58 This proposed

interaction is in keeping with the tighter coupling observed between the

2,6-dimethyl substrate and the cycloamylose cavity.

The success of these measurements suggested that these techniques

could be employed to probe the effect of methylation of the cyclohexaamy-

lose C-6 and C-2 hydroxyls on the disposition and coupling of these sub-

strates. If sodium 4-nitrophenolate and 2,6-dimethyl-4-nitrophenolate

penetrated the 2,6-0-methylated cyclohexaamylose to the same degree as in












0


0-


Binding geometries of 3- and 4-nitrophenylacetate (a and b,
respectively) versus sodium 4-nitrophenolate and 2,6-dimethyl-
4-nitrophenolate (c and d, respectively).


Figure 8.







the unmethylated case, they were expected to couple more tightly to this

derivative than the parent oligosaccharide due to increased "rim" inter-

actions. In analogy to the previous case, it was further expected that

the 2,6-dimethyl-4-nitrophenolate couples more tightly to the methylated

oligosaccharide than the 4-nitrophenolate.

Earlier studies on the rate of complexation45 and the initial dynamic

coupling experiments indicated that these substrates should rotate many

times before leaving the cavity. For example, the off rate of the sodium

4-nitrophenolate cyclohexaamylose complex is 3.1 x 104 sec-15 while the
58
rotation rate of that substrate in the cavity is 1.7 x 10- sec-1.

Thus, the guest molecule can rotate about 106 times in the cavity before

the complex dissociates. Clearly, the substrate does not just move in

and out of the cavity but resides sufficiently long to "sense" steric

inhibition, i.e., to fall into the slots provided by the methyl groups.

However, it was found that neither sodium 4-nitrophenolate nor 2,6-

dimethyl-4-nitrophenolate penetrates the dodecakis-2,6-0-methylcyclohexa-

amylose (DMCD) cavity as deeply as they do the cyclohexaamylose cavity.

This finding strongly suggests that arguments for the singular involve-

ment of the C-2 hydroxyls in the hydrolysis of the nitrophenylacetates

based on the lack of catalytic activity of dodecakis-2,6-0-methylcyclo-

hexaamylose are invalid.













CHAPTER II

PREPARATION AND EVALUATION OF SPERMIDINE
DERIVATIVES AS IRON CHELATORS
AND ANTINEOPLASTIC AGENTS

Experimental

Materials and Methods

For the synthetic sequences, all reagents were purchased from Aldrich.

The 2,3-dihydro-1H-imidazo[l,2-b] pyrazole (IMPY) was obtained from the

Division of Cancer Treatment, National Cancer Institute, Bethesda, Mary-

land. Methylene chloride and N,N-dimethylacetamide (DMA) were distilled
o o
and stored over 3A and 4A molecular sieves. Unless specified otherwise,

Na2S04 was used as a drying agent. Preparative thin layer chromatography

was done on 20 x 20 cm silica gel plates obtained from Analtech. Sepha-

dex LH-20 was purchased from Pharmacia Fine Chemicals. Melting points

were taken on a Thomas-Hoover apparatus and are uncorrected. Samples for

1H nmr were prepared in DCC13 with chemical shifts given in parts per mil-

lion relative to an internal Me4Si standard unless stated otherwise.

These spectra were recorded on a Varian T-60 and/or a Joel FX-100 spec-

trometer. Samples for ir spectra were prepared in KBr unless stated other-

wise, the spectra being recorded on a Beckman ir 4210 spectrophotometer.

Results are given in cm-1. Elemental analyses were performed by Galbraith

Laboratories, Knoxville, Tennessee, or MicAnal, Tucson, Arizona.

For the physical measurements, UV spectra were recorded on a Beckman

model 25 UV/VIS spectrophotometer or Cary 219. PH measurements were made

on either a Radiometer pHM-84 or pHM-62 digital pH meter. The stock iron
solution used was an atomic absorption standard purchased from Aldrich.

24







N4Benzyl-N1,NB-bis(benzoyl)spermidine
A flask was charged with a solution of N4-benzylspermidine (2.50 g,

0.01 mol) in 50 ml dry, degassed pyridine under N2. The flask was pro-

tected from light, cooled to -780C and the benzoyl chloride (3.00 g,

0.021 mol) in 32 ml dry, degassed methylene chloride added over a period

of 1.5 hrs. The reaction mixture was stirred under N2 an additional

thirty hours,allowing it to warm slowly to room temperature. It was then

poured into 125 ml H20 and extracted 1 x 15, 1 x 25 ml benzene. The

combined benzene extracts were washed 3 x 15 ml aqueous 15% Na2CO3 (w/v),

5 x 25 ml H20, dried, filtered and reduced in vacuo to yield 3.37 g of

the desired compound. By adjusting the pH of the original aqueous solu-

tion to 11.0 with KOH, extracting 2 x 20 ml HCC13 and combining, drying

and evaporating the organic extracts as before, an additional 0.78 g

product was isolated for a total yield of 4.15 g (95% yield) of the de-

sired compound: mp (EtOAc/pet ether) 100.5-102.5 C; nmr: 1.63 (m, 6H),

2.47 (m, 4H), 3.47 (m, 6H), 6.43 (broad s, 2H), 6.93-7.83 (m, 15H); ir:

3340 (s), 1630 (s), 1525 (s), 690 (s).

Anal. calcd. for C28H33N302: C, 75.81; H, 7.50; N, 9.47. Found:
C, 75.73; H, 7.46; N, 9.37.

NI,N8-Bis(benzoyl)spermidine

A solution of N4-benzyl-NI,N8-bis(benzoyl)spermidine (0.277 g, 0.62

mmol) in 2.5 ml glacial acetic acid was added to a suspension of PdO
(0.043 g, 0.3 mmol) in 1.25 ml glacial acetic acid. The hydrogenolysis

was allowed to proceed until no more hydrogen was taken up. The reaction

mixture was then filtered, the acetic acid evaporated and the residue

taken up in 10 ml anhydrous methanol. The resulting alcoholic solution

was adjusted to pH 11 with NaOMe and reduced in vacuo. The crude product







was taken up in 10 ml HCC13 and washed 3 x 5 ml cold H20. The organic
phase was then dried, filtered and evaporated to give 0.209 g (95% yield)

of the desired white crystalline product: mp 130.5-133.00 C; nmr: 1.16

(m, 7H), 2.63 (m, 4H), 3.43 (m, 4H), 6.53 (broad s, 2H), 7.13-7.73 (m,

10H); ir: 3310 (s), 2920 (m), 2860 (m), 1615 (s), 1515 (s).

Anal. calc. for Cz1H27N302: C, 71.36; H, 7.70; N, 11.89. Found: C,

71.18; H, 7.60; N, 11.73.

2-(2-tri fl uoroacetoxyphenyl)-5-oxazol one

Trifluoroacetic anhydride (0.252 g, 1.2 mmol) was added to a suspen-

sion of N-(2-hydroxybenzoyl)glycine (0.195 g, 1.0 mnol) in 10 ml dry

CH2C12. The reaction was allowed to proceed with stirring under N2 until

a clear solution was obtained.

After two hours, the solution was cooled to 00C and evaporated to

dryness by means of high vacuum. The residue was redissolved in 10 ml

dry CH2C12 and cooled to 00C. Immediately, 1,8-bis(dimethylamino)naph-

thalene (324 mg, 1.5 mmol) in 5 ml CH2CI2 was added. The resulting orange-

red solution was allowed to warm to room temperature with continued stir-

ring.

After sixteen hours the reaction mixture was washed 3 x 5 ml 3% (w/v)

ice cold HCI and 1 x 10 ml ice water. The organic phase was dried over

MgSO4, filtered and solvent removed to yield 250 mg (91% crude yield) of

product. Recrystallization from CH2C1, yielded clean, crystalline com-

pound.

The structure of the oxazolone was interpreted as involving three

equilibrating tautomers (Figure 9). The nmr spectra of these tautomers

have been assigned as follows: I corresponds to the singlet at 6 4.28,
II corresponds to the two singlets at 6 4.57 and 4.8 respectively and

III corresponds to the singlets at 6 5.08 and 6 2.80. The 2.80 resonance




















0 I











()
0
0
I





LL. l
O













I


.)



id


0O

(*cti~


--'c
S 0
0
N
Q.


en
0



I



0
4-c

0
S-

0

r"-






0

S.-
E
0
ro











0-
4--
CM







0
S-
'-->

S-0
L
0l
0-




s-







was assigned as belonging to the hydroxyl proton as it is readily exchanged

by addition of D20. Attempts to follow the incorporation of the deuterium

to all the tautomerswere frustrated by the hydrolysis of the oxazolone over

time. It appears, then, that tautomer II is the predominant form in solu-

tion. This is consistent with earlier work with isoxazolone systems.61,62

Mp (CH2C1,) 150C (d); nmr: (CD3COCD3) 2.80 (s), 4.28 (s), 4.57 (s),

4.80 (s) total 2H, 7.0-8.0 (m, 4H); ir: (kBr) 3270 (W), 1855 (s), 1705 (s),

1190 (s).

Anal. calcd. for CzIH6N 04F3: C, 48.37; H, 2.21; N, 5.13. Found:

C. 48.11; H, 2.08; N, 4.90.

N-(2-hydroxybenzoyl)diethylamidoglycine

To a suspension of N-(2-hydroxybenzoyl)glycine (195 mg, 1.2 mmol) in

10 ml dry CHC1 was added trifluoroacetic anhydride (252 mg, 1.2 mmol)

and the resulting solution allowed to stir at room temperature under N2.

After two hours reaction time the mixture was cooled to 00C and

evaporated to dryness under high vacuum. The residue was redissolved in

10 ml dry CH2C12, cooled to 00C, and 1,8-bis(dimethylamino)naphthalene

(430 mg, 2.0 mmol) dissolved in 5 ml dry CH2C12 was added dropwise. The

resulting solution was allowed to warm to room temperature with continued

stirring under N2.

The reaction was stopped after eighteen hours. The solution was

washed 3 x 5 ml 3% (w/v) cold HC1, and 1 x 10 ml cold H20, dried over MgS04

and solvent removed under vacuum.

In order to effect complete hydrolysis of the trifluoroacetoxy groups,

the crude product was dissolvedin 10 ml methanol and sodium methoxide

(54 mg, 1.0 mmol) added with stirring under N,. After three hours, meth-

anolic HC1 was added to achieve a pH of 3 (hydrion pH paper). The precipi-

tated NaCl was filtered off and solvent removed under vacuum. Chromatography







of the crude product on silica gel eluting with 5% MeOH in CHC13 (v/v)

afforded 210 mg (84%) of the desired compound: mp (MeOH/CHC13) 85-870C;

nmr: 1.0-1.5 (m, 6H), 3.42 (m, 4H), 4.27 (d, 2H), 6.63-8.0 (m, 5H), 12.18

(s, 1H); ir: 3260 (s), 1625 (s), 1590 (s), 1250 (s).

Anal. calcd. for C130H1N203: C, 62.38; H, 7.25; N, 11.19. Found:

C, 62.35; H, 7.23; N, 11.19.

N'-[N-(2-Hydroxybenzoyl)glycyl]-N1,N8-bis(benzoyl)spermidine

Trifluoroacetic anhydride (0.446 g, 2.1 mmol) was added with stirring

to a suspension of N-(2-hydroxybenzoyl)glycine (0.165 g, 0.185 mmol) in

20 ml CH2C12. The resulting mixture was refluxed at 500C for two hours.

The solution was evaporated to dryness under high vacuum and the N-(2-

trifluoroacetoxybenzoyl)glycyl trifluoroacetic anhydride residue redis-

solved in 15 ml CH2C12. After cooling to -780C and addition of 1,8-bis-

(dimethylamino)naphthalene (0.398 g, 1.86 mmol) in 5 ml CH2C12, the N',N8-

bis(benzoyl)spermidine acetate (0.257 g, 0.62 mmol) in 10 ml CH2Cl2 was

added dropwise. The reaction mixture was allowed to warm to room tempera-

ture with continued stirring under N2. After 30 hrs, the solution was

washed with 3 x 5 ml ice cold 1.1% aqueous HC1 (w/v), 3 x 5 ml ice water,

dried, filtered and evaporated. The residue was dissolved in 25 ml degassed

methanol and NaOMe added to a pH of approximately 9. After stirring under

N2 for thirty minutes, the pH was adjusted to approximately 2 by addition

of HCl gas dissolved in methanol. The solution was evaporated; the resi-

due was dissolved in 50 ml CHC12, filtered and evaporated to yield 294mg
(89% crude yield) of the product.

An analytical sample was purified on a silica gel plate eluting
with 5% MeOH in EtOAc: nmr: 1.17-2.33 (broad m, 6H), 3.33 (broad m, 8H),
4.13 (broad s, 2H), 6.50-8.00 (broad m, 17H), 12.07 (s, 1H); ir: 3260

(m), 2915 (m), 1628 (s), 1525 (m), 1299 (m).







Anal. calcd. for C3uH34N405: C, 67.91; H, 6.46; N, 10.56. Found:
C, 67.72; H, 6.55; N, 10.44.

N4-Benzyl-N1,N8-bis(2,3-diacetoxybenzoyl)spermidine Hydrochloride

A solution of N4-benzylspermidine (2.80 g, 11.9 mmol) and 1,8-bis-
(dimethylamino)naphthalene (5.00 g, 23.3 mmol) in 400 ml CH2C12 was cooled

to 00C under N2. The 2,3-diacetoxybenzoyl chloride (6.00 g, 23.4 mmol)

in 250 ml CH2Cl2 was added over a seven hour period. After addition was

completed, the reaction vessel was allowed to warm slowly to room temper-

ature with continued stirring under N2. After 20 hours, the reaction

mixture was again cooled to 00C, washed In 3x3O0ml ice cold 1.1% aqueous

HC1 (w/v), 3 x 30 ml ice cold H20, dried, filtered and evaporated to

give 5.62 g (66% crude yield) of the desired product, a white semicrys-

talline solid. Yields as high as 98% were obtained when a 20% excess of

the acid chloride was used.

An analytical sample was purified by high pressure liquid chromato-

graphy on a 100 A u Styragel column eluted with THF. This purification
demonstrated the crude product was in excess of 95% pure: mp 121-1230C;

1H nmr: 1.50 (m, 6H), 2.23 (2, 12H), 2.93 (m, 4H), 3.27 (m, 4H), 3.83

(s, 2H), 7.27 (m, 14H); ir: 3220 (m), 1765 (s), 1563 (m), 1374 (m), 1201
(s) cm-1.

Anal. calcd. for C36H,2N3010C1: C, 60.71; H, 5.94; N, 5.90; C1, 4.98.
Found: C, 60.59; H, 5.95; N, 5.71; C1, 4.80.

N4-Benzyl-N1,N8-bis(2,3-dihydroxybenzoyl)spermidine Hydrochloride

A solution of N4-benzyl-NI,N8-bis(2,3-diacetoxybenzoyl)spermidine
hydrochloride (1.0 g, 1.4 mmol) and sodium methoxide (0.3 g, 5.6 mmol)
in 20 ml dry degassed methanol was allowed to stir under N2. After six

hours the pH of the solution was adjusted to approximately 2 by addition





31

of HC1 saturated methanol. The resulting suspension was filtered through

sintered glass and solvent removed to yield 750 mg (98% crude yield) of

the desired compound.

An analytical sample was purified by chromatography on Sephadex-LH-

20 eluting with absolute ethanol. This purification demonstrated that

the crude product was in excess of 98% pure: mp (EtOH) 1100C (d); IH nmr

(trifluoroacetic acid, chemical shifts calculated relative to internal

CH2C 2, 5.28 ppm: 2.05, 2.45 (two broad overlapping m, 6H), 3.48, 3.81
(two overlapping m, 8H), 4.51 (m, 2H), 7.25 (m, 9H), 8.58 (broad m, 3H);

ir: 3300 (s), 16.48 (m), 1550 (s), 1275 (s) cm-1. Anal. calcd. for

C28H34N306CI: C, 61.82; H, 6.80; H, 7.72. Found: C, 61.59; H, 6.17;
N, 7.48.

N',N8-bis(2,3-dihydroxybenzoyl)spermidine Hydrochloride

The PdC12 (0.06 g, 0.34 mmol) was added to a solution of N4-benzyl-N1,N8-

bis(2,3-dihydroxybenzoyl)spermidine hydrochloride (0.4 g, 0.74 mmol) in

20 ml trifluoroacetic acid and the resulting suspension stirred under a

hydrogen atmosphere. After nine hours total reaction time, the catalyst

was filtered off and solvent removed in vacuo. The resulting crude resi-

due was then twice suspended in 50 mL 20% (w/v) HC1 and sonicated, the

solvent being removed under high vacuum. The resulting white solid was

chromatographed on Sephadex LH-20 (1.2 x 35 cm) eluting with anhydrous

ethanol. Fractions containing the pure compound were pooled, the sol-

vent removed and residue triturated with deionized water and freeze-

dried to yield 370 mg (95%) of the desired product, a white crystalline

solid: mp (EtOH), 170C (d); 1H nmr: (trifluoroacetic acid, chemical

shifts downfield from TMS, calculated relative to internal CH2C12, 5.28

ppm): 2.13 (broad m, 8H), 3.50, 3.86 (two broad overlapping m, 8H), 7.30







(broad m, 8H); ir: 3100 (broad m), 1638 (m), 1585 (s), 1537 (s). UV
spectra of this compound were identical with Tait's data.

Anal. calcd. for C21H28N306C1-H20: C, 53.45; H, 6.40; N, 8.90.
Found: C, 53.53; H, 6.64; N, 8.76.

N4-[N-(2-hydroxybenzoyl)glycyl]-N ,N8-bis(2,3-dihydroxybenzoyl)spermidine

Trifluoroacetic anhydride (0.300 g, 1.5 mmol) was added with stir-

ring to a suspension of N-(2-hydroxybenzoyl)glycine (0.130 g, 0.66 mmol)

in 15 mL dry CH2C12. The resulting mixture was allowed to stir at room

temperature under N2for 0.5 hour. The solution was then evaporated to

dryness under high vacuum and the N-(2-trifluoroacetoxybenzoyl)glycyl

trifluoroacetic anhydride residue redissolved in 25 mL of DMA. The re-

sulting solution was cooled to -100C, and 1,8-bis(dimethylamino)naphtha-

lene (0.240 g, 1.12 mmol) in 5 mL DMA added. Immediately, a solution of
N1,N8-bis(2,3-dihydroxybenzoyl)spermidine hydrochloride (0.252 g, 0.55

mmol), CuS04 (0.260 g, 1.65 mmol) and 1,8-bis(dimethylamino)naphthalene
(0.470 g, 2.2 mmol) in 10 mL DMA was added dropwise. The reaction mix-

ture was allowed to warm to room temperature with continued stirring

under N2. After thirty hours, the reaction was stopped by addition of

20 ml of pH 5 phosphate buffer. The resulting yellow solution was ex-

tracted 10 x 50 ml CH2 C2. The combined extracts were washed 5 x 25

mL pH 5 phosphate buffer, 12 x 25 mL H20, dried, passed through 2 x 1.5

cm column of silica gel and solvent removed in vacuo. The resulting

brown gum was dissolved in 25 mL degassed MeOH and NaOCH3 added to a pH

of approximately 9. After stirring 0.5 hour, the pH was adjusted to 2

by addition of methanolic HC1 and the solvent removed in vacuo. The

gummy solid remaining was absorbed onto Sephadex LH-20, applied to a 3

x 10 cm column of Sephadex LH-20 and eluted with 15% ethanol/benzene


































































































o0






Li
LL_






34

















I-





qw


OJ



"O

-2 EE


U,


0-
z0
r-
U,







oC


/E
4-c
.0








O
a







E
L .
0r
z oW
t S-O






















r "A







r
-o




C

0


0












a.
r--


- N












a1

U
-P







o0 o












E
S-
c--
U

0




0






















S.-
~r-






































U
0}







ci)
r

Ci)
>,










x
.0
O












'-C')
"a
*(-
E






O










4-1)
a)
0
N
C

















a)
S0
X
O






1 0
I



E
.[ (1




r
I




O






Q.



E
C
n:




r-


r-




cn
iZ





















r-

E
a)

o L



0
>)



N
C




-o
*r-


30
z






S--
0









S--
u







o




c i
N


















0.
O












-0
I
t I
















r

3:
































Va






Ci
E



(C








N




0



r-
I





-0




o























0
i?

























4-o
- "


-r













En













r--
ou

























C
-5
'r
E

L






4-







N
-o










S-
I




a )
r-



:+







*r-
0



E0
0

















Lr-
-U























1r





03





-5
C^)

























0I
S-
,O


4-









S-
--




0
(3)
E









N











I



vr
.Q











0










5-0
N















Ic





(U
EC -)


a)


SU


LU
E r-
C-































01
I-
w

S-





LI.-
E-r-
c +^























0


>)

-o


C-



.r-
.0
o


















I
S0
















L--
*r-
I-








E,










0





L4 0)
So
0














r-


E -0



O
- 0N
0 t-
U-5



*r--
C:N
Cv
le-a)










































()


*r-
E

a)



i N








N




*r--
Sa
-o









4--
0

E
i O










*r-


-)
A-













C'i

























Io_ t

;r:
c
1 c_ -- ---------:----; ot
r-




--------~- .- -~--

i=--
r


I,; ~ r.
i
--- 7
--
P


- -


__->g




-r K


-~w





i E


0L
QI
1/


---








E '1C o


__ ~ _~ ~_

a



25
-L
f C-.

-~ s-.


--






i -~----cc
i


r


"; S i


























I --










o
r .
o





F- N -- ,
0






I;
____________ ___ --J a



__________ 0




<:-- ------------------- -,




K .





C
!- -




____ -___ --- 0--


____------ I-







45


































C

O




-r

E




r-
0a)

0




N

o
0




4-
0






I



O


E











s.-







C)
r-
LL
Q-
t/l











sI-

























--
L -I


..


IH




H -





F-


L~ -I





r 2













-





-



Z

z


aC

c-
*r-

*r-
E


0
(U1
Q-




C:
N
r
a)

*-

.0


















O
N

a)
o





























tn
C"

S-







0
C'
0


S-0






















C"
a-



c,
a,
VI
Ct:












r-
ILl




































































































_ _
-i


47



















ca
r-


E


O.
a)



C
N


-o







C
I




-o


0
N






a-)




"i


*-



a-












.,
4-,




s-4





*+--
Lr










s.-



LL












NOISSIWSNV&L IN3D3d


~z:~-; -~-


- 4 -- -


"-7
-~ A-


^5





I




0 2
-zi








c^2


u


- -- -.-. -
*-


NOISSIWSNY.a IN ai)do


Bi





I


I


-----1



ci
- -- -o -


E


g ,--

- .





>1



S -
,








4 0
0 N









,--
=-:: -





---: o5






0-
(A
-1
1 4-



:---1 0



S^ J Q:
:j t t-


IL-'/--


r--i------
~--:-- =t~-~Lf~lTTt_'


Ml


-~--C
-+-- --


t=
---

r
-~ --C--L--i=


~f=















NOISSIWSNVai lN3D3!d


~., 0
0


gI-0



0


C I


S-~ --



- --- -


E
E


I--v- z -


2E'Ns ^-Ac3=


~ ZtTE

-rf -=T
>t z2K


~~Et~-I Tt~-Fiz-t I:


4 >A

C 9 0 C C 0 0 C C


o
0 0


- = 1- I i t-=


NOISSIWSNY~I IN3Ml3d


-:- --~---t-1-~-i------~i-~-1Z~~i~-~;;~-:~


8 .




a .)







.0
I N






4
o

SC"
0-
vi 0



ao N






r-
I





Ur-
.0






-o








( 0


c )
4.-

0


v,
l/l






r -''
LL


-- ~ ~"


''


-=gg-=


a







50









I



o

r-c-
N




x



r-









.0





C,
0
- I































LJ
cu
























.f.
















E*
0 *o



























I
a z









rl
c"J
o,

































CM
I
Zf





=5 (








(v/v). The elution was followed by analytical thin layer chromatography

and the solvent removed yielding 228 mg (70% yield) of the desired com-

pound: mp (ethanol/benzene) 117-120C; nmr (CD2Cl2) 1.32-3.28 (m, 6H);

3.04-3.72 (m, 8H); 4.22 (s, 2H); 6.48-8.12 (m, 15H); 12.10 (s, 1H);

12.74 (s, 1H); 13.09 (s, 1H); ir (KBr); 3370 (s), 1635 (s), 1535 (s),

1270 (s), 790 (s).

Anal. calcd. for C3oH34N409.2H20: C, 57.14; H, 6.07; N, 8.88.

Found: C, 57.20; H, 5.67; N, 8.77.

Potentiometric measurements of N1,N8-bis(2,3-dihydroxybenzoyl)spermidine
pKa's

All solutions were prepared using doubly distilled, degassed water

and were stored under N2, which had been passed through a basic solution

of pyrogallol. The NaOH solutions were standardized by titration of

potassium hydrogen phthalate to a phenolphthalein endpoint.

Potentiometric measurements were made in a glass jacketed titration

vessel connected to a constant temperature circulating water bath. The

temperature was maintained at 25.0 0.50C. Typically, 20 mL solutions

were prepared by appropriate dilutions of a stock solution of NI,N8-bis-

(2,3-dihydroxybenzoyl)spermidine trifluoroacetate over a concentration

range of 4-8 x 10-4M. Because of the susceptibility of the compound to

oxidation when in basic solution, all titrations were conducted under an

N2 atmosphere, the N2 having again been passed through a basic pyrogallol

solution to remove all traces of oxygen.

All pH measurements were taken with a Radiometer PHM-84 digital pH

meter equipped with a Radiometer combination electrode. Meter calibra-

tions were checked immediately after each titration. A minimum of four-

teen data points was obtained for each titration, readings being taken

after the stirring motor had been turned off and the solutions allowed to
equilibrate.







All potentiometric data were evaluated using the computer program

PHFIT.63 As written, this program handles as many as four fundamental

and fifteen complex species. Initial input includes the composition of

all complexes postulated to be of significance, actual values of their

cumulative formation constants in logarithmic form when available, and

initial estimates of those to be defined as well as the initial concen-

tration of each species. The activity coefficient of H+ under the ex-

perimental conditions employed is also required; this quantity was assigned

a value of 0.78 throughout.

The program first calculates the distribution of species at each

data point based on the total concentration of species at that point and

the given formation constants for each adduct. Thus, an initial theor-

etical titration curve is obtained; assuming all significant equilibria

have been considered, the difference between this calculated curve and

that actually observed is then minimized by further refinement of the

estimated formation constants. Refinement is continued until either a

given number of cycles have been completed or the fit is improved by less

than 0.5%; the final values of these refined overall formation constants

along with estimates of their uncertainly are then output.

For convenience, stepwise rather than cumulative formation constants

were considered in evaluation of pKa titration data; thus, the output

uncertainty in this instance refers to the stepwise protonation constants.

This interconversion of stepwise and cumulative formation constants was

easily handled by means of a simple subroutine.63

The data set generated by each titration was evaluated independent-

ly of all others. The values obtained for each formation constant were

averaged; the errors in these averages were estimated as the square root







of the sum of the squared estimated standard deviations associated with

each constant divided by the number of constants averaged.

Potentiometric Measurements of Binding of Cu++ with N',N8-bis(2,3-dihy-
droxybenzoyl)spermidine

Stock copper solutions were prepared using Cu(N03)2-6H20 and the

resulting metal concentrations determined by flame photometry on a Perkin-

Elmer model 290B atomic absorption photometer.

Solutions were prepared to give final concentrations of NI,N8-bis-

(2,3-dihydroxybenzoyl)spermidine trifluoroacetate (8 x 10-4M) and cop-

per concentrations of 1 x 10-4 M. Thus, the ligand to metal ratio was

kept constant at about 8:1. The ionic strength of each solution was

adjusted to 0.1 by addition of KN03.

Finally, appropriate dilutions of the stock were made and titrated

as for the pKa determinations. For these titrations, considerably more

time was required for the solutions to equilibrate after addition of

NaOH.

200 MHz 'H NMR Spectra of N1,N8-bis(2,3-dihydroxybenzoyl)spermidine

A sample of N1,N8-bis(2,3-dihydroxybenzoyl)spermidine trifluoroace-

tate was prepared in CD30D. High field 'H nmr were recorded at 200 MHz

on a Nicolet superconducting spectrometer. Standard proton decoupling

experiments were also performed on this instrument to aid in spectrum

identification.

Copper Induced 'H Line Broadening of Nl,NB-bis(2,3-dihydroxybenzoyl)-
spermidine

A solution of 0.05 M in N1,Ns-bis(2,3-dihydroxybenzoyl)spermidine

trifluoroacetate and 0.2 M in sodium methoxide was prepared in CD30D

and its spectrum recorded. Aliquots of a 5 x 10-3 M solution of Cu(N03)

in CD30D were added to the nmr tube via a microliter syringe such that







the ligand to metal ratio varied from 500 to 36. The spectra were re-
corded on a JOEL FX-100 spectrometer at 270C.

Spectrophotometric Determination of the Stoichiometry of Binding of Cu+
with N1,N8-bis(2,3-dihydroxybenzoyl)spermidine

Stock solutions of Cu(N03)2 (9.44 x 10-s M) and N',Ns-bis(2,3-dihy-

droxybenzoyl)spermidine trifluoroacetate (1.05 x 10-4 M) were prepared

in 0.1 M KN03 using doubly distilled, degassed water. The pH of each

was adjusted to 7.6 by addition of concentrated NaOH and a series of

solutions prepared which increased in ligand while decreasing in metal

but holding the total number of moles constant. A second series of solu-

tions containing ligand alone was prepared similarly. The absorbance of

each solution was read at 337 nm.

Potentiometric Measurements of the Binding of Fe3+ with NI,N8-bis(2,3-
dihydroxybenzoyl)spermidine
Due to the extreme water insolubility of ferric hydroxide stock

0.01 M Fe(N03)2 was prepared in 0.1 N HC104.

Solutions were prepared 8 x 10-4 M in NI,N8-bis(2,3-dihydroxyben-

zoyl)spermidine trifluoroacetate and 1 x 10-4 M in Fe3+. In another

titration the concentration of ligand was 2 x 10-4 M and 2 x 10-5 M metal.

The ionic strengths were adjusted to 0.1 by addition of NaC104. Twenty milli-

liter samples weretitrated with 0.2 N NaOH as described previously.

In a slightly different titration, a stock solution was prepared
8 x 10-4 M in ligand, 1 x 10-5 M in Fe3+ and 3.2 x 10-3 M in NaOH. Again,

the ionic strength was adjusted to 0.1 with NaC104. This system was ti-

trated with 0.2 N HC104.

Spectrophotometric Determination of Stoichiometry of N1,N8-bis(2,3-dihy-
droxybenzoyl)spermidine Fe(III) Complex at pH = 7.4
Typically, a stock solution of N',N8-bis(2,3-dihydroxybenzoyl)sper-

midine (~ 8 x 10-5 M) was prepared in phosphate buffer I = 0.1, pH = 7.4.







The stock Fe(III) was 1.79 x 10-5 M in 2% HN03. Solutions were prepared

with increasing ligand concentration and decreasing Fe(III) concentration,

holding the total number of moles constant. The absorbance of each solu-

tion was measured at 49 nm.

Spectrophotometric Measurement of Fe(III) Induced Changes in the Spec-
trum of NI,N8-bis(2,3-dihydroxybenzoyl)spermidine

A series of solutions was prepared in which the concentration of

N1,NB-bis(2,3-dihydroxybenzoyl)spermidine trifluoroacetate was held con-

stant at about 4 x 10-5 M. The concentration of iron increased from

1 x 10-5 M to 8 x 10-5 M spanning a range of ligand to metal ratios of

4 to 0.5. The solutions were prepared in phosphate buffer I = 0.1,

pH = 7.4. The spectrum of each solution was recorded from 850 nm to 425

nm.

Potentiometric Measurement of N4-[N-(2-hydroxybenzoyl)glycyl]-NI,N-bis-
(2,3-dihydroxybenzoyl)spermidine pKa's

Stock solutions were prepared which were 5 x 10-5 M in N4-[N-(2-

hydroxybenzoyl)glycyl]-Nl,N8-bis(2,3-dihydroxybenzoyl)spermidine and

25 x 10-5 M in NaOH in doubly distilled, degassed water. The ionic

strength of the solutions was adjusted to 0.1 by addition of KNO3.

The measurements were performed on 20 mL samples titrating with
0.1N HN03 in the apparatus previously described.

Spectrophotometric Measurement of pH Induced Changes in the Spectrum of
N -[N-(2-hydroxybenzoyl1glycy 1-N ,N -bis(2,3-dihydroxybenzoyl)spermidine
Fe(III) Complex

A stock solution which was 4.5 x 10-5 M in N4-[N-(2-hydroxybenzoyl)-

glycyl]-NI,N8-bis(2,3-dihydroxybenzoyl)spermidine and 4.5 x 10-5 M in

Fe(III) was prepared in 99% 0.1 M KN03 and 1% ethanol. The iron was

introduced as a dilution of 1.79 x 10-2 M Fe(N03)3 in 2% HNO3. The pH

was adjusted to 11 by addition of concentrated NaOH. The absorbance
spectrum was recorded from 750 nm to 350 nm.







Titration of the stock with concentrated HN03 yielded a series of
spectra spanning the pH range 11 to 5.7. All manipulations were per-

formed under an N2 atmosphere which had been passed through basic pyro-

gallol.

Spectrophotometric Determination of Stoichiometry of N4-[N-(2-hydroxy-
benzoyl )glycyl -N ,N -bis(2,3-dihydroxybenzoyl )spermidine Fe(III) Complex

Typically, a stock solution of N4-N-(2-hydroxybenzoyl)glycyl-Nl,N8-
bis(2,3-dihydroxybenzoyl)spermidine was prepared in 5% EtOH, 95% phos-

phate buffer I = 0.1, pH = 7.4 to give a final concentration of about

8 x 10-5 M. The buffer was thoroughly degassed by bubbling N2 prior to

use. The Fe(III) stock was an Aldrich AA standard which was 1.79 x 10-2M

Fe(III) in 2% HNO3. A series of solutions was prepared in which the

ligand concentration decreased while the Fe(III) concentration increased

but holding the total number of moles constant.

The absorbance of each solution was read at 500 nm, a wavelength

where neither the ligand or metal alone absorb.

Similar stock solutions and complex solutions were prepared as for

the determination at pH 7.4 except phosphate buffer I = 0.1 pH 9.0 was

used. The absorbance was again measured at 500 nm.

Preliminary Evaluations of the Toxicity of N1,N8-bis(2,3-dihydroxyben-
zoyl )spermidine

For the initial toxicity experiments, suspensions of NI,N8-bis(2,3-

dihydroxybenzoyl)spermidine in mineral oil were prepared. In another

series of tests, the compound was suspended in sterile isotonic saline.

Dilutions of the stock were made such that the total injection volume

was 2 mL. Swiss white mice (20 g) were injected intraperitoneally at

concentrations ranging from 1500 mg/kg down to 500 mg/kg. Control ani-

mals were injected with 2 mL of either mineral oil or saline.







N4-(4-carboxybutyryl )-N1 ,N8-bis(2,3-dimethoxybenzoyl)spermidine

A solution of N',N8-bis(2,3-dimethoxybenzoyl)spermidine (470 mg,

1.0 mmol) in 10 mL dry CH2C12 was cooled to 00C and 1 mL dry pyridine

added. Glutaric anhydride (125 mg, 1.1 mmol) dissolved in 5 mL CH2C12

was added dropwise. The resulting mixture was allowed to come to room

temperature and stir under a CaC12 drying tube.

After forty hours total reaction time the solvents were removed

under vacuum. The residue was dissolved in 100 mL EtOAc and washed

3 x 15 mL 2% (w/v) HC1 and 1 x 10 mL H20. The organic phase was dried

over sodium sulfate and solvent removed. Chromatography of the resulting

oil on silica gel eluting with 10% MeOH in CHC13 yielded 480 mg (82%)

of an extremely hydroscopic white solid: nmr (CD2C1,) 1.33-2.17 (broad m,

8H), 2.37 (m, 4H), 3.4 (m, 8H), 3.87 (s, 12H), 6.83-8.5 (m, 9H); ir:

3290 (s), 1720 (s), 1680 (s), 960 (s).

Anal. calcd. for C3uH41N309.H20: C, 59.50; H, 7.11; N, 6.94. Found:

C, 59.26; H, 7.11; N, 6.94.

3-(4-carboxybutyramido)quinoline

To a solution of 3-aminoquinoline (0.56 g, 3.89 mmol) in 25 mL CH2-

Cl2 at 00C was added glutaric anhydride (0.53 g, 4.65 mmol). The result-
ing solution was allowed to warm to room temperature with stirring under

a CaCl2 drying tube.

As the reaction proceeded, the product, a white solid, precipitated

from solution. After eighteen hours, the reaction was cooled to 00C and

the crude product filtered off to yield 900 mg (90%). Crystallization

from methanol yielded clean compound: mp (CH30H) 210C (d); nmr (CDO0D);

1.72-2.75 (m, 6H); 7.38-9.08 (m, 6H); ir: 3280 (s), 1680 (s), 1540 (s),

900 (s).







Anal. calcd. for C14H14N003: C, 65.11; H, 5.46; N, 10.85. Found:
C, 65.04; H, 5.53; N, 10.62.

N4-[4-(3-quinalinocarboxamido)butyryll-N1,N8-bis(2,3-dimethoxybenzoyl)-
spermidine

To a solution of 3-(4-carboxybutyramido)quinoline (0.15 g, 0.58

rmmol) and N-hydroxysuccinimide (0.1 g, 0.87 mmol) in dry 10 mL pyridine

was added dicyclohexylcarbodiimide (0.18 g, 0.87 mmol) in 2 mL pyridine.

The resulting solution was allowed to stir under N2.

After sixteen hours, the reaction mixture was cooled to 0C and

filtered to remove the precipitated dicyclohexylurea. After again cool-

ing to 00C, a solution of N1,N8-bis(2,3-dimethoxybenzoyl)spermidine

(0.33 g, 0.70 mmol) in 5 mL dry CHC13 was added dropwise. The reaction

mixture was allowed to warm to room temperature with continued stirring

under N2.

The reaction was stopped after twenty-four hours and the solvents

removed under high vacuum. The residue was dissolved in 25 mL EtOAc and

washed 5 x 3 mL H20, dried over MgS04 and evaporated. The resulting yel-

low oil was chromatographed on silica gel eluting with 5% ethanol in CH-

C13 yielded 370 mg (86%) of the desired product, a fluffy white, highly
hydroscopic solid: nmr 1.3-2.3 (broad m, 8H), 2.5 (m, 4H), 3.4 (m, 8H),

3.83 (s, 12H), 6.5-9.0 (broad m, 15H); ir (KBr): 2920 (s), 1630 (s),

1260 (s), 795 (s).

Anal. calcd. for C39H,7N,08.2H20: C, 62.71; H, 6.64; N, 9.16.

Found: C, 62.48; H, 6.85; N, 9.34.

N-(4-carboxybutyryl)-2,3-dihydro-1H-imidazo[1,2-b]pyrazole

To a solution of 2,3-dihydro-1H-imidazo[1,2-b]pyrazole (0.1 g, 1.83

mmol) in 10 mL dry CHC13 was added glutaric anhydride (0.13 g, 2.2 mmol)

in 3 mL CHC13. The resulting mixture was allowed to stir at room







temperature under an N2 atmosphere. As the reaction proceeds,.the product
precipitates from solution.

After stirring sixteen hours, the product was filtered off, washed

with cold CHC13 and dried under vacuum to yield 160 mg (76%) of the de-

sired compound. This material was analytically pure and used without

further purification: mp > 2500C nmr (trifluoroacetic acid, chemical

shifts relative to internal CH2C12 at 5.25 ppm) 1.83-3.17 (m, 6H), 4.8

(s, 4H), 6.7 (broad s, 1H), 8 (d, 1H); ir (KBr); 2910 (m), 1725 (s),

1670 (s), 1555 (s).

Anal. calcd. for C1UH13N303: C, 53.81; H, 5.87; N, 18.82. Found:

C, 53.68; H, 5.97; N, 18.72.

N-[3-(2,3-dimethoxybenzamido)propyl ]-N-(4-(2,3-dimethoxybenzami do)butyl]-
N-L4-(N-2,3-di hydro-lH-imidazoe ,2-bjpyrazolo )carboxamido]butyrami de

To a solution of N-(4-carboxybutyryl)-2,3-dihydro-1H-imidazo[l,2-b]-

pyrazole (0.1 g, 0.45 mmol) in 10 mL dry pyridine was added N-hydroxy-

succinimide (0.062 g, 0.54 mmol) and dicyclohexylcarbodiomide (0.115 g,

0.54 mmol) and the resulting solution allowed to stir under N2. After

sixteen hours, the precipitated dicyclohexylurea was filtered off and a

solution of NI,N8-bis(2,3-dimethoxybenzoyl)spermidine (0.25 g, 0.54 mmol)

in 5 mL CHC13 added.

After stirring an additional forty-eight hours under N2, the pyri-

dine was removed under high vacuum. The residue was dissolved in 50 mL

methylene chloride and washed 2 x 5 mL ice cold 3% (w/v) HC1 and 1 x 5 mL

cold H20, dried and solvent evaporated. The crude product was chromato-

graphed on silica gel eluting with 10% isopropylalcohol in methylene

chloride to yield 200 mg (66% yield) of a white, hydroscopic solid: nmr:

1.73 (broad m, 8H), 2.57 (m, 4H), 3.5 (m, 8H), 3.93 (s, 12H), 4.3 (m, 4H),

5.67 (broad s, 1H), 6.67-7.67 (m, 9H); ir (KBr) 3320 (m), 2920 (s),1620(s),
1260 (s).
























0
N




x
-o
0




















o
.-,



















u
4n
+.c



"U























S0
I





o












*-
f,







0
I-














4J

(D
o





EE
U












Cd.


0 -
s-
*r-




1.1.





















;i


0

0r
CJ
I




z
0
I
Z





03


-4


0
+





a,


a,

r-
4C-

o
C
.-




o
00



o
-0




C



o



4-


E
S-

,-J



E -










r-
; ^






" L
0 s_
nr
4-


























-o.





*r-














-o
0








E








*-










*r-
o








E

3 0





S-N
x
















u0
I-
ro
























E
0









C,,
r
- -

I -5
1-
E?






O

Lca


















(1

0
N





-o




s.-.J
0
N


r--
E




0
*-
1




r-
I







-o









o -








U
I-
























0
OS-
0


~ r-'
S-
- r4-
UC

iO



C *r
0 -



l<- (
0L






















4I-
o o

-4r-
SE
Eea
"- X
-0 0O








N
I -
C Uo






S-
la 0
I N












E ro
c E
I

.---

4-
0 0





W -0









S:- E r
-0LI


r N
EC
N r-
CE




o i




L .I
O r-








"0 10
2I-



01













W-" 0"
*4-
I..1_


o I


CI




.-i
U-I



O.0 *





E (0s
CN






CM
SO



*r-












I
N r-
c E


X-I
0 "-
IS.0

S A S-
- -o o
*r--








I-
-. r-
E











N0


0I
I roL
N --
L




















Sr- -
X i











I m --
0' I --
*oo

















4- O











o = o ,..4-' S(U
r-- -
MOU

NO










-r E i






-


a)
ES-z









(0.0-



1.n







66









a,
.--

*r-
E


U)


0
N
C
.a




r-o

-I-









r
a,
E


m


U

-o
I














r-
z





o



E
u









-


U,

S-T









i-


t






i







67

















12 ? -- --- -- --- _^ -- ------ -- ___
,_ j. -









_- __ _0_ -
















_ 0
j0
^^ -- ^=== _.-._... _.--.- r. .















- *-
1 o











1- ----- ^ -- ---- U
1 ^-------^ -- ^---- I r













I

2
_. : ----- ----- -- I o
^_____~~ ~~ ^: __ ^._







- _______ __^ ** s-
_ --5 --- c, e


- sU



i (/"

I-4





a)

C,
- LL-


rL r
=--~ -




:,~~~ :-_--__ ___ ---____________







68












0

4
E

0



-o


I
o3












ro
0
ca




4r-
E

0
o




sr-
Co


0
.r-

cLJ










.0
i







4-.--

OE
E )



r-u














L.0
C :





















;.~c---c--~ -5 -- --------c
.._.____. .o.---- -


___------ -







t-- -
IC-- ---------- -~------------------2=
,__________________________ =.__|
-_ _'T




















-._ -_____ _.-, -0 7"o
r





-L -











"ti; '' -- I---- -z-- : ,J .






^-~ -- I- -
.. .


o~




(A
V;U







70










i



o a
N 0
C r-
(V E

>is-
x
0-

.0

EO
*r- --
W- *-
EO
1 x
C 0

S- -o
NO







I L




I r--
i-



s- ox

4-I

0
O
N e-

0 0


io


0 r-









rE
0)I
-E













o
L0










CI--





S-5
10
5-
I













0")
Ur-







71













o -o
N -
WE
-0 S-
x >

oD
0 3
Sc -0



SE



-
LM -Q


?1 0

to
I N
>-





1o 0
CII N




-oE
0 QO








X *.--

Q- I


.- C


0
I o
n-0








i-a
OO



















e. I
-) I
J,,
.o

1-o


I S


























ur-
L I-













s-


iZ







Anal. calcd. for C35H46N608: C, 61.93; H, 6.83; N, 12.38. Found:
C, 61.48; H, 6.79; N, 12.26.

N-[3-(2,3-dihdroxybenzamido)propyl]-N-[4-(2,3-dihydroxybenzamido)butylJ-
N-(N-2,3-dihydro-lH-imidazo]1,2-b]pyrazolo)carboxamido]butyramide

A solution of N-[3-(2,3-dimethoxybenzamido)propyl]-N-[4-(2,3-di-

methyoxybenzamido)butyl]-N-[4-(N-2,3-dihydro-lH-imidazo[1,2-b]pyrozolo)-

carboxamido]butyramide (0.19 g, 0.28 mmol) in 10 mL CH2CI2 was added

dropwise to BBr3 (1.05 g, 4.2 mmol) in 10 mL dry CH2C12 at 00C. The

resulting solution was allowed to warm to room temperature with continued

stirring under nitrogen.

After twenty-four hours total reaction time, the suspension was cooled
to 00C and reaction stopped by dropwise addition of 25 mL ice cold doubly

distilled, deionized water. If tap water is used, trace metals present

will cause contamination problems due to chelate formation.

The crude product was filtered off and washed with cold water and

CH2C12, then dried to yield 170 mg (86%) of a tan solid. Chromatography

on Sephadex LH-20 eluting with 20% methanol in benzene afforded clean,

white product: nmr: (CF3CO2H chemical shifts relative to internal methy-

lene chloride at 5.25 ppm) 2.17 (broad m, 8H), 3.03 (m, 4H), 3.90 (m, 8H),

4.90 (m, 4H) 6.5-7.67 (broad m, 7H), 8.1 (broad s, 2H); ir (KBr): 3320 (s),

1630 (s), 1545 (s), 1280 (s).

Anal. calcd. for C31H39N60gBr: C, 52.92; H, 5.55; N, 11.95. Found:

C, 52.56; H, 5.52; N, 11.73.

Results
Potentiometric Determination of N1,Ns-bis(2,3-dihydroxybenzoyl)spermidine
pKa's

Preliminary evaluation of the data resulting from titration of the
ligand alone indicated that of the ligand's five potentially dissociable







protons, only two were being titrated within the pH range of the experi-

ments. The experimental curve was best fit by considering the follow-

ing step-wise equilibria, holding the equilibrium constants of reactions

(1), (2) and (3) at arbitrary, high values of 14.0 each:

L + H HL (1)

HL + H t H2L (2)

H2L + H 2 H3L (3)

H3L + H t H4L (4)

H4L + H Z HsL (5)

Setting the value of pKa at -13.787 as calculated by Sweeton, Mesmer and

Baes for an aqueous solution 0.1 M in KC1 at 25C,64 the values of equi-

librium constants (4) and (5) were refined by the program PHFIT as pre-

viously described. The average values thus obtained over seven titrations

were 6.60 0.08 (pKal) and 7.80 = 0.15 (pKa2) corresponding to logarith-

mic cumulative protonation constants of 56.40 and 49.80, respectively.

For each titration, the calculated curves based on these constants

correlated well with those observed. In virtually every case, the great-

est deviations between the two occurred in unbuffered regions of the curve

as might be expected, and in the highest pH region where electrode sensi-

tivity is reduced. In the latter case, deviations could not be attributed

to the titration of a third proton as such deviations tended to be posi-

tive i.e., the observed pH tended to be higher than that calculated.

Potentiometric Measurement of the Binding of NL,N8-bis(2,3-dihydroxyben-
zoyl)spermidine with Cu(II)

Titration of N1,N8-bis(2,3-dihydroxybenzoyl)spermidine trifluoroace-

tate in the presence of Cu(N03)2 introduced a number of species in addi-

tion to protonated ligand. The following metal hydrolysis equilibria

were included:







Hz0 + M M(OH) + H (6)
2H20 + M t M(OH)2 + 2H (7)
2H20 + 2M : M (OH)2 + 4H (8)

H0O + M M(OH), + 3H (9)
H20 + M M(OH)4 + 4H (10)
with literature constants of -8.2, -17.5, -10.6, -27.8 and -39.1 for

equilibria (6) through (10) respectively.65 Analyses of the titration
data based on these equilibria alone produced no reasonable fit. There-

fore, ligand metal binding equilibria were included as follows:

L + M LM (11)
L + H + M HLM (12)
L + 2H + M H H2LM (13)

L + 3H + M H3LM (14)
The best theoretical fit to the observed data was found with cumu-
lative formation constants (in logarithmic form) of 30.19 0.66, 39.85

0.51, 45.54 0.36 and 49.1 5.7 for equilibria (11) through (14) re-

spectively (Figure 40). The relatively high error associated with the

last of these values is attributed to the fact that at the initial pH
this species is already of little importance. In fact, when this species

was omitted completely from the analyses, the correlation between the
calculated and the observed titration curves was only mildly affected.

Deviations in the initial portions of the curves increased slightly while
the values of formation constants (11) through (13) were essentially
unchanged.

200 HMz 1H nmr assignments of N1,N8-bis(2,3-dihydroxybenzoyl)spermidine

In order to simplify the 'H nmr spectra of compound (II) we decided
to exchange the phenolic, amido and amino protons with deuterium by taking

















REPRESENTATIVE RESULTS:
COP3OU;D0 II COrPER BINDING TITRA1IO'l


observed"
calculated: ---

Figure 40. Potentiometric binding curve for N',N8-bis(2,3-
dihydroxbenzoyl)spermidine with Cu(II).


titrF 0.1977 1 NaOH/:,L







all spectra in CD30D. While the 100 MHz spectrum of the compound has

been reported in DMSO, the broad lines and poor resolution resulting

rendered detailed interpretation difficult.66 Methanol does not present

these problems and interpretation was straightforward.

The aromatic protons from 6.65-7.25 6 consist of three separate sets

of signals demonstrating nonequivalence in the aromatic rings (Figure 41).

The two apparent downfield triplets centered at 6.70 and 6.72 respectively

integrate to two protons, and can be assigned as two nonequivalent meta

protons. The meta protons are split by the para protons and again by

the ortho protons. The doublet of doublets, when superimposed, appear

as a triplet. Moving further downfield, there are two sets of four lines

which correspond to the two protons assignable as the para protons. These

protons are split by the meta and again by the ortho protons resulting

in two nonequivalent doublets of doublets. Finally, what appears to be

two triplets integrating to two protons can be assigned to the ortho pro-

tons. The ortho protons are split by the meta and by the para protons

as well. This splitting would be expected to result in two sets of four

lines as with the para protons. However, the chemical shift nonequiva-

lence differences between these two sets of protons results in an overlap

producing two apparent triplets.

The remainder of the spectra can be divided into four sets of sig-

nals (Figure 41). The high field set centered at 1.75 6, four protons,

corresponds to the two internal methylenes in the four-methylene bridge

of spermidine. What appears to be a downfield quintet, two protons, cen-

tered at 1.98 6 corresponds to the internal methylene of the short three-

methylene bridge. Irradiation of the high field signal, 1.75 6 (Figure

42a), has no effect on the 1.98 6 quintet but causes a rather substantial


















N

0 I




-)
ro





u
U
a1

GI N
r-
0 S-

,-
z E


co V







S E
cD a)
C-



0





.-

00




C)
-a








-- I-.


04-







'aE
Ew
0.






-.-






N
* S -.
0)



U-





















C
o
0






FN



.c |
r)


L-
0
..

a)a
o
f- 0




LM -0


Sr--


0 -













LO
r .r-






L cm
L ^













S.-




am m
r o



r_-

I-

--
I 1.c_


-i








change in the downfield line shapes. The furthest downfield multiple,

3.4-3.55 6, four protons corresponding to the amide methylenes, reduces

from a five to a four line signal. This decoupling reduces an apparent

triplet on the long methylene bridge side to a singlet leaving the short

methylene bridge amino methylene as a triplet centered at 3.5 6. Irra-

diating the central methylene of the short bridge (Figure 43b) leaves

the long bridge amido methylene as a triplet and reduces the short bridge

amido methylene triplet to a singlet as expected.

Although both irradiations change the line shape of the amino methy-

lenes, analysis of these changes is made difficult by poor resolution of

this signal. Irradiation of the amino methylene (Figure 43c) reduces

the short bridge internal methylene quintet to a triplet. However, the

long bridge internal methylenes' signal analysis is also made difficult

by lack of resolution. Irradiation of the amido protons (Figure 43d)

produced identical changes in the line shapes of the short bridge inter-

nal methylenes but slightly different changes in the long bridge internal

methylenes.

Perhaps the most interesting feature of this spectrum is the nonequi-

valence of the aromatic protons. This nonequivalence is unlikely to

have come from the simple asymmetry of the methylene bridges but rather

from some preferred conformations.

Copper Induced 1H Line Broadening of NI,N8-bis(2,3-dihydroxybenzoyl)-
spermidine

Metal ion probes have been used extensively in the study of coordi-

nation complexation.67-69 Since paramagnetic relaxation is generally

quite efficient, it will be the predominant relaxation mechanism for

those nuclei within the metal's sphere of influence. Metal probes, such







80






m 0 ,

x
3 O
S*





4.J












cr
L
*r-









o






E
(\ .



















*r--
u N





r-
rc





03
1 C
L= c t
















-I


L )
Sn












UN
(t -




















o
LQ.
^ -lr i
*: ^














^-L ^ y
.J y 3 =s <
j ^ s w .

^ ^',


D.-







as copper, having long electron relaxation times, cause substantial

broadening for those nuclei close to the metal.

Preliminary studies of the NI,N8-bis(2,3-dihydroxybenzoyl)spermidine/

Cu(II) complex in aqueous solution revealed tight binding. In order to

verify that the site of chelation involved the catechol hydroxyls, a se-

ries of 1H nmr spectra of the ligand were acquired in CD30D with varying

concentrations of Cu(N03) Increased broadening of the N',N8-bis(2,3-

dihydroxybenzoyl)spermidine resonances was observed as the copper con-

centration was increased.

Interpretation of the broadening data in a quantitative way was not

attempted. As a qualitative indicator of the site of coordination the

observed line widths were adequate. Since paramagnetic effects diminish

as r-6, the metal's sphere of influence, is small and therefore only

protons very close to the site of association will be affected.

In the case of NI,N8-bis(2,3-dihydroxybenzoyl)spermidine, the aro-

matic protons rapidly broadened with increasing copper to chelate concen-

trations. At a ligand to metal ratio of 150, the lines collapsed into

the baseline, rendering line width measurements difficult. As expected,

the internal methylene resonances were observed to broaden at high copper

concentrations. Furthermore, when the solution at high metal to ligand

ratios was acidified, the original spectrum of N1,N8-bis(2,3-dihydroxy-

benzoyl)spermidine emerged, indicating the complex had dissociated.

In order to further substantiate the catechols were involved in

copper chelation, we examined a model system. N1,NB-bis(benzoyl-

spermidine is a structural analogue of NI,N8-bis(2,3-dihydroxybenzoyl)-

spermidine, lacking the phenolic hydroxyls. When a solution of this

model compound in CD30D was titrated with Cu(NO.)2 from a ligand to metal







ratio of 200 to 8, no aromatic line broadening was observed. However,

the internal methylenes exhibited approximately 1.5 Hz broadening at

high metal concentrations. This may be rationalized in terms of some

weak interaction between the secondary nitrogen of the spermidine back-

bone and Cu(II) in solution. Even when the metal concentration exceeded

that of substrate, no observable aromatic line broadening occurred. It

seemed clear, then, that chelation of Cu(II) in solution by Nl,N8-bis-

(2,3-dihydroxybenzoyl)spermidine was occurring at the phenolic oxygens.

On the basis of these results, it was concluded that in the presence

of Cu(II) and base, the phenolic oxygens of N1,NB-bis(2,3-dihydroxyben-

zoyl)spermidine would be protected, but the N4-nitrogen would be free to

react. These line broadening experiments further indicated that the com-

plex could be broken up simply by protonatingthe phenolic oxanions.

Spectrophotometric Determination of the Stoichiometry of Binding N1,N8-
bis(2,3-dihydroxy6benzoyl)spermidine with Copper at pH 7.6

In order to verify the validity of the complex species assumed in

the analysis of the potentiometric titration data, the stoichiometry of

the copper/ligand complex was experimentally established by means of

Job's plots.70

The ultraviolet spectrum of N',N8-bis(2,3-dihydroxybenzoyl)spermi-

dine trifluoroacetate salt at pH 7.6, I = 0.1, exhibits a maximum at

314 nm. Upon complexation with copper, this maximum shifts to 339 nm.

A plot of the difference in absorbance between the ligand alone and in

the presence of copper against mole fraction ligand revealed that the

complex was, in fact, 1:1 as assumed.

Potentiometric Measurement of the Binding of N1,N8-bis(2,3-dihydroxyben-
zoyl)spermidine with Fe(III)
When a solution of NI,N8-bis(2,3-dihydroxybenzoyl)spermidine tri-
fluoroacetate (8 x 10-4 M) and Fe(C104), (1 x 10-4 M) was rapidlytitrated







with NaOH, precipitation of the solution was observed at about pH 4.2.

Rapid titration is necessary in order to prevent reduction of Fe3 to

Fe2+ by prolonged exposure to acidic media. Assuming that the precipi-

tate was Fe(OH)3 formed resulting from incomplete ligand binding, the

Fe3+ concentration was decreased to 5 x 10-5 M and the titration repeated.

Again, precipitation of the solution was observed.

In order to circumvent the precipitation of Fe(OH)3 resulting from

ineffective ligand binding in acid, the titration was attempted by pre-

forming the complex at high pH and titrating with acid. However, when a

solution of 8 x 10-4 M NaOH was titrated with 0.2 N HC104, precipitation

was again observed at pH 6.8.

Finally, a solution 2 x 10-4 M NI,N8-bis(2,3-dihydroxybenzoyl)sper-
midine trifluoroacetate and 2 x 10-5 M Fe(C104)3 was titrated over the

pH range 3.5 to 10.1 with no obvious precipitation of any solute. The

solution underwent a color change from light purple (pH ~ 4.0) to bright

red (pH ~ 6.0). Also, substantial .drift in the pH meter was observed in

the high pH region. Furthermore, when the basic solution was acidified

to pH = 3.5 by addition of 20% (w/v) HC1, the original purple solution

was not regenerated, thus suggesting irreversible changes in the system

during the course of titration.

Spectrophotometric Determination of the Stoichiometry of N',N8-bis(2,3-
dihydroxybenzoyl)spermidine with Fe(III) at pH 7.4

The absorbance measured at 490 nm when plotted against mole fraction
of NI,N8-bis(2,3-dihydroxybenzoyl)spermidine trifluoroacetate exhibited

a maximum at about Xligand = 0.6 suggesting a stoichiometry of three

ligands per two iron atoms. At this wavelength, neither free ligand nor

Fe(III) has any absorption, only the colored complex absorbs. The solu-

tions corresponding to mole fractions of ligand 0.9 to 0.7 were bright







red in color. Increasing Fe(III) concentration (Xligand = 0.6) was

accompanied by a change in color to purple.

Spectrophotometric Measurement of Fe(III) Induced Changes inthe Spectrum
of NI,N Bbis(2,3-dihydroxybenzoyl)spermidine

Since the method of continuous variations for determining stoichio-

metry is only valid if one complex is being formed under the experimental

conditions imposed,70 the effect of added iron on the spectrum of the

ligand was examined. The free ligand has no absorption in the spectral

region but in the presence of Fe(III) at concentrations where the ligand

to metal ratio was 3 and 2, the complex showed a maximum at 496 nm. As

the Fe(III) was increased, corresponding to ligand to metal ratios of 1.5,

1 and 0.5, the maximum shifted to 510, 530 and 544 nm respectively. As

the maximum was shifted to longer wavelength, no isosbestic point was

observed. These results strongly suggested that in the presence of ex-

cess iron, Nl,Nd-bis(2,3-dihydroxybenzoyl)spermidine is capable of form-

ing a series of association species. Therefore, the actual stoichiometry

is not simple but is dependent upon the relative concentrations of ligand

and iron present.

Potentiometric Determination of N4-[N-(2-hydroxybenzoyl)qlycyl]-NI,N8-
bis(2,3-dihydroxybenzoyl)spermidine pKa's

Evaluation of the titration data indicated that,of the ligand's five

potentially titratable protons, only three were dissociating within the

pH range of the experiment. Thus, the following equilibria were consid-

ered, holding (15) and (16) at values of 14.0 each:
L + H H HL (15)
HL + H H2L (16)
H,L + H H L (17)
H L + H H L (18)
S+ 4
H4L + H H H5L (19)







Using the value of pKw of -13.78764 and using the program PHFIT as

previously described, values of the equilibrium constants (17), (18) and

(19) were refined. The average values thus obtained were 6.52 + 0.22

(pKaj), 7.24 0.12 (pKa2) and 8.07 0.05 (pKa3).

The calculated curves based on the constants correlated well with

those observed (Figure 44) with, again, the largest deviations in the

high pH region. The titrations were complicated by precipitation of the

ligand in the lower pH regions necessitating the use of reasonably dilute

solutions.

Spectrophotometric Measurement of the pH Induced Changes in the Spectrum
of N-[LN-(2-hydroxybenzoyl)qlycyi]-N ,N -bis(2,3-dihydroxybenzoyl)sper-
midine

The initial spectrum of N4-[N-(2-hydroxybenzoyl)glycyl]-N1,N8-bis-
(2,3-dihydroxybenzoyl)spermidine in the presence of equimolar Fe(III) at

pH 11 showed an absorption maximum at 505 nm. As the pH was lowered to

8.2 by addition of HN03, no change was observed in the spectrum. However,

as the pH was decreased below 8.2, the absorbance maximum shifted to 575

nm at pH 4.2. The complex in this pH region also exhibited an isosbestic

point at 554 nm. Furthermore, the color of the solutions changed from red

(basic pH) to purple as the pH was lowered to 6.5. This behavior is con-

sistent with some type of association change in the lower pH region.

Spectrophotometric Determination of the Stoichiometry of Binding N4-[N-
(2-hydroxybenzoyl )glycy1-N ,N8-bis(2,3-dihydroxybenzoyl)spermidine with
Fe(III)

The absorbance measured at 500 nm for each solution measured at pH=
9.0 was plotted against the mole fraction of N4-[N-(2-hydroxybenzoyl)gly-

cyl]-NI,N8-bis(2,3-dihydroxybenzoyl)spermidine and exhibited a maximum at

Xligand = 0.5. This corresponds to a 1:1 stoichiometry for metal-ligand
binding as would be expected at high pH.


















































titre 0.0995 N HNO3(microliters)


Figure 44.


Potentiometric titration for N' [N-(2-hydroxybenzoyl)-
glycyl]-N1,Nf-bis(2,3-dihydroxybenzoyl)spermidine.







However, when the same analysis was employed at pH 7.4, the maximum

appeared at Xligand = 0.66.

Discussion
The isolation by Tait of the spermidine siderophores and the subse-

quent preliminary analysis of their iron clearing properties were impor-

tant contributions to the already extensive search for clinically viable

iron chelators. However, as was pointed out early on, the major obsta-

cle to the use of these compounds was that of accessibility. The biolo-

gical yields of both NI,N8-bis(2,3-dihydroxybenzoyl)spermidine and N"-[N-

(2-hydroxybenzoyl)threonyl]-NI,N8-bis(2,3-dihydroxybenzoyl)spermidine

were very low and their isolation and purification difficult at best.

Nevertheless, their promising potential for therapeutic use suggested

that they be evaluated synthetically.

The approach chosen for the synthesis of these compounds was defined

by certain boundary conditions. The synthetic sequence had to consist

of as few high yield steps as possible. This was necessary in order that

scale-up to the quantities required for clinical use be easily facilitated.

A corollary condition was that purification be limited to simple tech-

niques applicable to reasonably large quantities of material. Further-

more, it was desired that the synthesis be general enough such that struc-

tural or functional modifications could be introduced if desired.

The synthetic scheme decided upon was to attach the desired function-

ality to the NI and N8 primary nitrogens of spermidine, then to affix

the appropriate group to the NI-secondary nitrogen. This sequence is

outlined in Figure 3.

The first problem dealt with was that of selective acylation of

spermidine's primary and secondary nitrogens. This was accomplished by







the preparation of the key synthon N4-benzylspermidine as previously

described.26,27 With this reagent in hand, selective functionalization

of spermidine's primary nitrogens was possible. Subsequent reaction of

the secondary nitrogen with a different acylating agent was then possible

if so desired.

Once the target compounds were prepared, the feasibility of any one

for potential therapeutic use had to be determined. First, some sort of

preliminary toxicity study was necessary. No matter how spectacular a

particular compound might be as an iron chelator, if it exhibited toxic-

ity at levels equivalent to an effective dose, it would clearly be useless

for therapy.

Once the relative safety of the compound had been determined, some

estimation of its efficacy for removing iron from the organism was desired.

This information, while seemingly straightforward, in practice often is

not. For instance, measurement of the thermodynamic association constant

for a particular compound with iron in vitro frequently bears no correla-

tion to its ability to sequester iron in vivo.71 The problem of accessi-

bility to the bound iron in the organism appears to be of great importance.

However, certain lower limits to the ability to complex iron may be de-

fined. If a candidate is thermodynamically capable of removing iron from

transferring (the body's iron transport protein) then the possibility of

mobilizing normally inaccessible storage iron exists. Since transferring

has an unlimited access to the various iron pools in the body,11 it may

be employed as an indirect agent for the removal of cell bound iron. As

the plasma transferring is stripped of iron by an administered chelator,

it will regain the lost iron from some cellular storage pool which is

frequently inaccessible to the chelator directly. Therefore. an association







constant for iron of at least 1024 transferring'ss iron binding constant)72

is a prerequisite for screening.

Once the favorability of iron association has been established, the

actual ability to mobilize iron in vivo remains to be ascertained. There

are a variety of accepted experimental techniques for these determina-

tions.73-7I Again, it must be emphasized that many compounds which have

been screened very successfully have turned out to be less than satisfac-

tory when subjected to actual clinical trial. However, since there ap-

pears to be no absolutely reliable method for determining the final clin-

ical utility of any given compound, the successful performance through

these preliminary screens is accepted as demonstrating some promise for

therapeutic trial.

To this end, two compounds were prepared and tested, the natural

product NI,N8-bis(2,3-dihydroxybenzoyl)spermidine and the model compound,,

N4 -N-(2-hydroxybenzoyl)glycyl]-N',N8-bis(2,3-dihydroxybenzoy)spermidine.

The results for these two compounds will be discussed in detail.

Preparation of Nl,N_-bis(2,3-dihydroxybenzoyl )spermidine

The synthesis of this natural product was accomplished in high yield

and purity in three steps from the key synthon N4-benzylspermidine. How-

ever, before attempting this synthesis, some relatively simpler compounds

were prepared in order to work out the necessary conditions for the gen-

eral scheme.

N4-benzylspermidine was smoothly acylated with two equivalents of

benzoyl chloride to give the N',N8-bisbenzoyl-N4-benzylspermidine adduct

in excellent yield. While this compound lacks any chelating functionality,

it successfully demonstrated that the protected spermidine derivative

could indeed be selectively acylated in high yield. This compound was







further useful in that the changes in the 1H nmr spectrum of the spermi-
dine backbone observed upon condensation served as a model for all the

acylating agents employed in this work.

In order that the secondary nitrogen of the N1,N8-bis-acylated sper-

midines be accessible for reaction, the benzyl protecting group had to

be removed. For the model bisbenzoyl compound, this was easily accom-

plished by means of catalytic hydrogenolysis. Thus, N1,N8-bisbenzoyl

spermidine acetate was smoothly generated at room temperature and pres-

sure in acetic acid over palladium.76

Having successfully devised this sequence, it was then applied to

the synthesis of N',N8-bis(2,3-dihydroxybenzoyl)spermidine. When N4-

benzylspermidine was reacted with excess 2,3-diacetoxybenzoyl chloride27

in the presence of 1,8-(dimethylamino)naphthalene as base, the resulting

N1,N8-bis(2,3-diacetoxybenzoyl)spermidine was produced as the hydrochlo-

ride salt in 98% yield. The masking of the phenolic hydroxyls of the acid

chloride as acetates77 was simply to prevent self condensation potentially

arising from the mixed functionality present. However, the resulting con-

densed adduct proved to be exceptionally difficult to purify. It would

not crystallize, and attempted chromatography on conventional supports

resulted in partial decomposition of the compound. However, eventual

purification on v Styrogel demonstrated that the crude product was typi-

cally about 95% pure and required no further clean-up. Furthermore, it

was later ascertained that the major impurity could be easily removed at

a later step in the synthesis.

There were, next, two conceivable routes to the desired NI,N8-bis-

(2,3-dihydroxybenzoyl)spermidine. Either the catechol acetoxy protecting

groups could be hydrolyzed off followed by hydrogenolysis of the benzyl
group or, alternatively, the benzyl group could be removed first and then







the acetates. In practice, the former sequence proved to be the more

desirable. Catechols are known to beextremely susceptible to oxidation

under basic conditions.78 Therefore, unless the most stringent precau-

tions are taken, the basic hydrolysis of the acetate protecting groups

may be expected to result in some oxidation byproducts. These oxidation

products would then have to be separated. However, if the hydrolysis is

performed before the catalytic debenzylation, any oxidation products

would be expected to revert back to their reduced form under the hydro-

genolysis conditions. Furthermore, as will be discussed in more detail

later, when the benzyl group was first removed and the resulting secon-

dary amine subjected to basic conditions in the presence of the acetoxy

moieties, an acyl transfer of the acetates from the hydroxyls to the cen-

tral nitrogen was observed. However, if the benzyl group was left intact

during the hydrolysis step, no such transacylation could occur.

Therefore, when N1,N8-bis(2,3-diacetoxybenzoyl)-N4-benzylspermidine

was first deacylated with sodium methoxide in methanol followed by hydro-

genolysis in trifluoroacetic acid over palladium, the resultant N1,N8-

bis(2,3-dihydroxybenzoyl)spermidine was isolated in 93% yield as the tri-

fluoroacetate salt. The use of trifluoroacetic acid was necessary as the

hydrogenolysis in acetic acid proceeded quite slowly. The final purifi-

cation was effected by chromatography of the product on Sephadex LH-20

eluting with a benzene-ethanol mixture.

While the compound isolated in this manner had certain advantages

such as fairly appreciable water solubility, little is known of the toxic-

ity of trifluoroacetate salts. Therefore, the anion was exchanged for

chloride. This was first accomplished by dissolving the compound in 20%

(w/v) aqueous HC1 followed by evaporation of the solvent. This conversion







certainly underscores the acid stability of the compound, and is in marked

contrast to desferrioxamine which is rapidly decomposed in acid.79 There-

fore, NI,N8-bis(2,3-dihydroxybenzoyl)spermidine has the potential to be

orally effective since it would be stable in the stomach, whereas des-

ferrioxamine is not. Further support for this claim was provided in an

investigation of the absorption properties of the compound.28

However, the exchange still involved rather drastic conditions which,

preferably, could be avoided. In later work, it was found that simply

sonicating the trifluoroacetate salt in 3% (w/v) HC1 would suffice.

While the trifluoroacetate salt was appreciably soluble, the hydrochlo-

ride was not, and readily precipitated as a white powder. The product

could then be collected by filtration. This exchange process is simple

enough to accommodate large quantities of material. Also, since chroma-

tographic purification of the exchange product demonstrated it to be in

excess of 98% pure, it was frequently used in the subsequent investiga-

tions as generated.

Clearly, the sequence is general enough that a variety of acylating

agents may be affixed to spermidine's primary nitrogens. This was one of

the boundary conditions for the general synthesis at the outset. Further-

more, the product is arrived at in three steps from the key synthon, N4-

benzylspermidine (Figure 4). Finally, all are high yield steps and puri-

fication of the final product is straightforward.

Preparation of Nr-IN-(2-hydroxybenzyl)glycyl]-NI,N8-(2,3-dihydroxybenzoyl)-
spermidine

The problem of appending an acyl group to the secondary nitrogen of

the NI,N8-bis(acylated)spermidines was again approached through the use

of model compounds. The actual natural product contains a salicyloyl

threonine group at the central nitrogen, either in the open or closed








form (Figure 1). However, it has been suggested that the coordination

sites of the metal filled by this moiety in the open form are the phe-

nolic hydroxyl and amido carbonyl oxygen.22 In the closed form, the

coordinating functionality probably is that of the phenolic oxygen and

oxazole nitrogen.80 Thus, the threonine side chain functionality prob-

ably has little, if anything, to do with the siderophore's ability to

chelate iron. Therefore, it was decided that a somewhat structurally

simpler acylating agent be tried in the initial experiments. The gly-

cine analogue of salicyloyl threonine, N-(2-hydroxybenzoyl)glycine was

chosen as a representative model. Use of salicyloyl threonine as a synthon

in Scheme 1 is complicated by the mixed functionalities present. Coupling

of the acid group in the presence of the phenolic hydroxyl and threonine

side chain hydroxyl would require potentially complicated protection steps.

The glycine analogue lacks the aliphatic hydroxyl. Thus, the problem

was simplified to dealing with two functionalities instead of three.

Furthermore, the distance between the spermidine backbone and purported

coordination sites would remain the same (Figure 45). Thus, a comparison

of this derivative's iron sequestering properties with those of the natu-

ral product would help substantiate the claim for threonine's lack of

importance. Finally, if, indeed, this coupling product possessed the

ability to sequester iron effectively and mobilize it in an organism,

it would be a potential candidate for chelation therapy which was readily

available synthetically.

The mechanism of activation of the acylating agent was next approached.

While there are numerous methods in the literature for forming amidebonds,

the compounds we wished to activate were multifunctional. Clearly, if

the carboxyl group of N-(2-hydroxybenzoyl)glycine is activated in the















0
\_ II
o o







SI -0
0 0




I 9 ;







C I I) o

O p
'Ul



-E- -C








I(
.-,


I00 z


I z
-o

















I I
C J

r- -

LE
-r- I E



(/r,-




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 EUQ48LHGS_MUWIRG INGEST_TIME 2012-03-02T22:50:02Z PACKAGE AA00009117_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES