Significance of asymmetric sites in choosing siderophores as deferration agents

Article abstract of DOI:10.1021/jm010019s

The syntheses of the microbial iron chelators L-fluviabactin, its unnatural enantiomer, D-fluviabactin, L-homofluviabactin, and L-agrobactin, are described. The key steps involve the selective bis-acylation of the terminal nitrogens of norspermidine, sper

Full text of DOI:10.1021/jm010019s

J . Med. Chem. 2001, 44, 2469-2478
2469
Sign ifica n ce of Asym m etr ic Sites in Ch oosin g Sid er op h or es a s Defer r a tion
Agen ts
Raymond J . Bergeron,* Mei Guo Xin, William R. Weimar, Richard E. Smith, and J an Wiegand
Department of Medicinal Chemistry, University of Florida, J . Hillis Miller Health Science Center, Gainesville, Florida 32610
Received J anuary 31, 2001
The syntheses of the microbial iron chelators L-fluviabactin, its unnatural enantiomer,
D-fluviabactin, L-homofluviabactin, and L-agrobactin, are described. The key steps involve the
selective bis-acylation of the terminal nitrogens of norspermidine, spermidine, or homosper-
midine with 2,3-bis(benzyloxy)benzoic acid in the presence of 1,1-carbonyldiimidazole, followed
by coupling of the N-hydroxysuccinimide ester of CBZ-protected L- or D-threonine with the
central nitrogen. The effectiveness of each of these ligands in supporting the growth of
Paracoccus denitrificans in a low-iron environment and the ability of these compounds to
promote iron uptake are evaluated. The stereochemical configuration of the oxazoline ring is
shown to be the major structural factor controlling both microbial growth stimulation and iron
uptake. L-Fluviabactin, L-homofluviabactin, and L-agrobactin all promoted growth and iron
uptake; D-fluviabactin was only marginally active. As with the microorganism’s native
siderophore, L-parabactin, all three ligands in the L-configuration investigated exhibited
biphasic, i.e., both high-affinity and low-affinity, kinetics. The high-affinity system (iron
concentration < 1 µM) yielded Km values between 0.11 and 0.23 µM and Vmax values from 157
to 129 pg-atoms Fe min-1 (mg of protein)-1, whereas the low-affinity scheme (iron concentration
> 1 µM) gave Km values from 0.53 to 3.5 µM and Vmax values between 96 and 413 pg-atoms Fe
min-1 (mg of protein)-1. Both L- and D-fluviabactin are very effective at clearing iron from the
bile duct-cannulated rodent; when given subcutaneously at a dose of 150 µmol/kg, both ligands
had iron clearing efficiencies of >13%, which is much greater than that of desferrioxamine in
this model. Thus, by altering the stereochemistry of certain microbial siderophores, it is possible
to generate deferration agents that are still effective at clearing iron from animals, yet do not
promote microbial growth.
In tr od u ction
to 0.25 to 0.40 mg of Fe kg-1 day-1 12
. Phlebotomy is not
an option, as the origin of the excess iron is the
transfused red blood cells. The only currently available
alternative is chelation therapy, in which the chelator
must be able to remove at least 0.25 to 0.40 mg of Fe/
kg daily. Left untreated, these patients generally die
in their second decade.
Many organisms are auxotrophic for Fe(III), and
nature’s response to the insolubility of the hydroxide
(Ksp ) 1 × 10 -38)1 formed under physiological conditions
has been the development of rather sophisticated iron
storage and transport systems. Whereas eukaryotes
tend to utilize proteins to transport iron (e.g., transfer-
rin) and store iron (e.g., ferritin), microorganisms utilize
low molecular weight ligands, siderophores.2
Iron metabolism in primates is characterized by a
highly efficient recycling process;3-6 no specific mech-
anism exists for eliminating this transition metal. As
“excess iron”7-9 cannot be cleared effectively, its intro-
duction into this closed metabolic loop leads to chronic
overload and, ultimately, to peroxidative tissue damage.
A number of situations, including high iron diet or
abnormally increased absorption of the metal, can result
in iron overload. In each of these scenarios, the patient
can be treated by phlebotomy.10 There are, however, iron
overload syndromes secondary to chronic transfusion
therapy (e.g., aplastic anemia and thalassemia).11 In
patients undergoing this regimen, erythropoiesis is
suppressed and iron absorption may be near normal,
but each unit of transfused erythrocytes contains 200
to 250 mg of iron. Most thalassemia major patients
require 200 to 300 mL kg-1 year-1 of blood, equivalent
The commercial introduction of the hydroxamate
siderophore desferrioxamine B (DFO)13 over 30 years
ago provided a life-saving treatment for iron overload
in chronically transfused thalassemia patients. The
subcutaneous (sc) infusion of DFO is regarded as the
method of choice for handling transfusional iron over-
load.14-16 The drug’s efficacy and long-term tolerability
are well documented, yet it suffers from a number of
shortcomings associated with its marginal oral activity,
its high cost of production, and its poor to moderate
efficiency. Efficiency is defined as the amount of iron
excreted by an animal or human upon administration
of a given chelator after subtracting the baseline iron
excretion, divided by the theoretical amount of iron
excretion that should be stimulated on the basis of the
dose and the compound’s coordination chemistry; this
quotient is expressed as a percentage. DFO can be
immunogenic.17-19 The ligand also has a very short half-
life in the body and must, therefore, be administered
by continuous sc infusion over long periods of time. At
the clinical level, the result is poor patient compliance.20
* To whom correspondence should be addressed. Tel. (352) 846-1956.
Fax: (352) 392-8406. E-mail: bergeron@mc.cop.ufl.edu.
10.1021/jm010019s CCC: $20.00 © 2001 American Chemical Society
Published on Web 06/21/2001

2470 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 15
Bergeron et al.
Sch em e 1. Synthesis of L- and D-Fluviabactin (15 and
16), L-Agrobactin (17), and L-Homofluviabactin (18)a
To circumvent these difficulties, the development of an
orally effective iron chelator has been a therapeutic
strategy for many years.
Although many synthetic iron chelators have been
studied in recent years as potential orally active thera-
peutics [e.g., hydroxypyridones21,22 and bis(o-hydroxy-
benzyl) ethylenediaminediacetic acid analogues23], none
has yet proven to be completely satisfactory. The sid-
erophores, microbial iron chelators, have remained
relatively unexplored in this search. The rate of their
isolation and structural elucidation has far surpassed
the progress in their evaluation as iron clearing agents.
Siderophores fall primarily into two structural classes,
hydroxamates or catecholamides.24-26 There are several
compounds that do not belong to either family, such as
pyochelin,27,28 rhizobactin,29 and 2-(3′-hydroxypyrid-2′-
yl)-4-methylthiazoline-4(S)-carboxylic acid (desferri-
thiocin).30-35 Of concern regarding the utilization of
siderophores as therapeutics for the treatment of iron
overload disease is that these ligands could serve to
stimulate the growth of pathogenic microorganisms in
humans. When the hexacoordinate catecholamide che-
lator enterobactin was tested as a possible deferrating
agent in rodents, a fulminant Escherichia coli infection
developed in these animals which ultimately resulted
in death due to septicemia.36 Although not all microor-
ganisms cross-utilize siderophores, the phenomenon can
occur.20,37-40 We have demonstrated that L-parabactin
both stimulated the growth of Paracoccus denitrificans
and promoted iron uptake, whereas D-parabactin did
neither, as the latter enantiomer was not recognized by
the bacterial iron transport apparatus.41 We also showed
that L-parabactin was very efficient at removing iron
from rodents and primates.42 A structurally similar
catecholamide, L-fluviabactin, was recently isolated from
Vibrio fluvialis.43 In the current study, we assembled
L- and D-fluviabactin, L-agrobactin, and L-homofluvia-
bactin, evaluated the cross-utilization of these ligands
by P. denitrificans, and compared the iron clearing
efficacy of D- and L-fluviabactin in a bile duct-cannulated
rodent model. The results further underscore the idea
that siderophores can serve as a platform in the design
of therapeutic deferration agents. Stereochemical modi-
fication of the natural product iron chelators can
profoundly diminish their capacity to stimulate micro-
bial growth without reducing their iron clearing proper-
ties in vivo.
a
Reagents: (a) 1,1-carbonyldiimidazole, CH2Cl2; (b) N-CBZ-(D-
or L-Thr)-OH, N-hydroxysuccinimide, 1,3-dicyclohexyldiimide (for
7-9) or N-CBZ-L-Thr-OSu (for 10); NEt3, CH2Cl2; (c) H2, 10% Pd-
C, CH3OH, HCl; (d) ethyl 2,3-dihydroxybenzimidate,44 CH3OH.
Resu lts
Syn th esis. The siderophores agrobactin (17)44 and
L-parabactin45 were first synthesized in this laboratory
by cyclocondensation of the appropriate imidate ester
with the hydrobromide salt of gem-amino alcohol 13
(Scheme 1). N4-Benzylspermidine was acylated with 2,3-
dimethoxybenzoyl chloride (2 equivalents); the central
nitrogen was deprotected46 and coupled with N-car-
bobenzoxy-L-threonine. Sequential removal of the block-
ing groups afforded synthon 13. In a subsequent route
to L-parabactin,47 N1,N8-bis[2,3-bis(benzyloxy)benzoyl]-
spermidine (5) was obtained directly from spermidine,
and the secondary nitrogen was acylated with N-tert-
butoxycarbonyl-L-threonine. Unmasking of the chiral
amino group, cyclization with ethyl 2-hydroxybenzimi-
date, and deprotection of the catecholamides completed
the natural product synthesis. The present route to
L-(15) and D-fluviabactin (16), L-agrobactin (17), and
L-homofluviabactin (18) exploits the preferential acyla-
tion of linear triamines, the simultaneous removal of
amine and phenol protecting groups, and cycloconden-
sation with ethyl 2,3-dihydroxybenzimidate.44
Accordingly, the syntheses of the hexacoordinate
chelators 15-18 began with the selective bis-acylation
of the primary nitrogens of norspermidine (1), spermi-
dine (2), or homospermidine (3)48 with 2,3-bis(benzy-
loxy)benzoic acid49 in the presence of 1,1-carbonyldiim-
idazole50 (Scheme 1), resulting in the corresponding
bis-amides, N1,N7-bis[2,3-bis(benzyloxy)benzoyl]-nor-
spermidine (4, 70%),51 N1,N8-bis[2,3-bis(benzyloxy)ben-
zoyl]spermidine (5, 73%),50-52 and N1,N9-bis[2,3-bis-

Asymmetric Sites in Choosing Siderophores
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 15 2471
F igu r e 2. Growth rate of P. denitrificans in the presence of
L-parabactin (“L-PB”) and ligands 15-18 (2.0 µM) in low-Fe
medium as measured by turbidity of the culture (OD660, y-axis)
vs time (x-axis).
F igu r e 1. Circular dichroism spectra of ferric L- (15, solid
line) and D-fluviabactin (16, dashed line). The chelates were
dissolved to a concentration of 400 µM in Tris-HCl buffer (100
mM Tris, pH 7.4); the cell path length was 1.00 cm.
(benzyloxy)benzoyl]homospermidine (6, 64%). Each of
these bis-amides (4-6) was reacted with an activated
ester of N-carbobenzyloxythreonine. The norspermidine
bis-amide (4) was coupled with N-carbobenzyloxy-L-
threonine or with N-carbobenzyloxy-D-threonine in the
presence of 1,3-dicyclohexyldiimide and N-hydroxysuc-
cinimide in triethylamine45 to generate N4-(N-carboben-
zyloxy-L-threonyl)-N1,N7-bis[2,3-bis(benzyloxy)benzoyl-
norspermidine (7, 52% yield) or its D-threonyl enantiomer
(8, 32%), respectively. The spermidine and homosper-
midine bis-amides (5 and 6) were also coupled with the
activated succinimidyl ester of N-carbobenzyloxy-L-
threonine in triethylamine to produce the corresponding
homologues (9 and 10) in 55 and 77% yield, respectively.
The carbobenzyloxy (CBZ) and benzyl (Bn) protecting
groups of 7-10 were simultaneously removed by hy-
drogenolysis (H2/Pd-C in methanolic HCl) to afford N4-
L-threonyl-N1,N7-bis(2,3-dihydroxybenzoyl)norspermi-
dine (11, 99% yield), N4-D-threonyl-N1,N7-bis(2,3-
dihydroxybenzoyl)norspermidine (12, 99%), and the
corresponding L-spermidine and L-homospermidine ho-
mologues (13 and 14) in 93 and 90% yield, respectively,
as their hydrochloride salts. Each of these compounds
was then condensed with ethyl 2,3-dihydroxybenzimi-
date44 to generate the final products L- and D-fluvia-
bactin (15 and 16, 65 and 63% yields, respectively),
L-agrobactin (17, 63%), and L-homofluviabactin (18,
64%). Finally, as expected from the enantiomeric rela-
tionship of the fluviabactins 15 and 16, the circular
dichroism spectra of their ferric complexes were anti-
symmetric (Figure 1).
Biologica l Eva lu a tion s. (a ) Micr obiology. The
impact of the asymmetric siderophores L- and D-fluvia-
bactin, L-parabactin, L-agrobactin, and L-homofluvia-
bactin on the growth of P. denitrificans was evaluated
at a ligand concentration of 2 µM in minimal salts
medium containing the nonutilizable iron chelator,
EDDA (Figure 2). Bacterial growth was followed over
48 h by measuring changes in the optical density of the
culture at 660 nm (OD660). All of the ligands in the
L-configuration (15, 17, and 18) stimulated the growth
of P. denitrificans, whereas D-fluviabactin did not. We
observed the same phenomenon with the enantiomers
of parabactin,41 i.e., the D-parabactin was inactive.
F igu r e 3. Uptake of [55Fe]ferric ligands by P. denitrificans.
Stock solutions of [55Fe]ferric chelates of L-parabactin (“L-PB”)
and compounds 15-18 (1.0 µM) were added to late log-phase
cultures of P. denitrificans. Samples were withdrawn from the
cell suspension at the times indicated, chilled, filtered, rinsed,
and subjected to scintillation counting.
The observed growth stimulation paralleled the abil-
ity to accumulate radiolabeled iron in the presence of
the corresponding [55Fe]ferric catecholamide complexes
(Figure 3). P. denitrificans is able to accumulate iron
from the ferric complexes of L-parabactin, L-fluviabactin
(15), L-agrobactin (17), and L-homofluviabactin (18), but
not from ferric D-fluviabactin (the complex of 16).
The dramatic differences in the transport kinetics of
ferric D-fluviabactin as compared to ferric L-fluviabactin
are illustrated in the Lineweaver-Burk plot (Figure 4).
These differences in velocity of transport are especially
pronounced at lower chelate concentrations (e.g., < 1
µM). In addition to illustrating marked differences in
the velocity of transport, Figure 4 suggests other
features of mechanistic interest. The kinetic data for
ferric D-fluviabactin fit a straight line, and thus obey a
simple Michaelis-Menten model (Km ) 3.5 ( 0.8 µM,
Vmax) 96 ( 63 pg-atoms of Fe/min-mg of protein). For
ferric L-fluviabactin, it is apparent that the data do not
show a linear relationship. The expanded view in Figure
5 emphasizes this nonlinearity. When analyzed sepa-
rately, the data from 1.0 to 5 µM ferric L-fluviabactin
fit one line (Km ) 2.2 ( 1.6 µM) implying a low-affinity

2472 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 15
Bergeron et al.
Ta ble 1. Kinetics of Accumulation of [55Fe]-Labeled Ferric
Complexes of Asymmetric Siderophores by P. denitrificansa
high-affinity
component
low-affinity
component
c c,d
c
c,d
[
55Fe]-chelateb
Km (µM)
Vmax
Km (µM)
Vmax
L-fluviabactin (3)e 0.23 ( 0.03 129 ( 2 2.2 ( 1.6 413 ( 149
D-fluviabactin (3)e
L-agrobactin (2)f
L-homo-
3.5 ( 0.8 96 ( 63
0.11, 0.11
157, 153 0.55, 0.53 245, 247
0.17 ( 0.04 138 ( 19 1.0 ( 0.1 229 ( 29
fluviabactin (3)e
a
Stock solutions of [55Fe]ferric chelates of the siderophores
listed were added to late log-phase cultures of P. denitrificans
grown in low-iron liquid medium. Samples were withdrawn from
the cell suspension every 20 s for 2 min, quickly chilled, filtered,
b
rinsed, and subjected to scintillation counting. The number in
parentheses is the number of assays at each chelate concentration
(0.1, 0.2, 0.5, 1.0, 2.0, and 5.0 µM). c The kinetic data were utilized
in the computer generation of plots, which were then used to
d
estimate Km and Vmax
.
Expressed as pg-atoms Fe min-1 (mg of
protein-1). e Mean ( standard error. f The data from the two
individual experiments are reported.
F igu r e 4. Lineweaver-Burk plot showing the kinetics of iron
transport of [55Fe]ferric L-(15, x-x) and D-fluviabactin (16,
squares) at concentrations ranging from 0.1 to 5.0 µM by P.
denitrificans. The data represent the mean of three experi-
ments at each concentration.
Ta ble 2. Iron Clearance Stimulated by L-Fluviabactin,
D-Fluviabactin, and L-Parabactina
efficiency
(%)b
distribution of
excreted Fe (%)
compound
L-fluviabactin (3)
D-fluviabactin (3)
L-parabactinc (8)
13.8 ( 0.6
16.0 ( 5.3
14.2 ( 2.0
84.6 bile, 15.4 urine
90.8 bile, 9.2 urine
90.3 bile, 9.7 urine
a
The ligands were administered to non-iron-overloaded, bile
duct-cannulated rats (n - value in parentheses) as a sc injection
at a dose of 150 µmol/kg. Bile samples were collected at 3-h
intervals for 48 h; urine samples were taken every 24 h over the
b
experimental period. The efficiency of each chelator was calcu-
lated by subtracting the iron excretion of control animals from the
iron excretion of treated animals. This number was then divided
by the theoretical output on the basis of a 1:1 ligand-iron complex;
the result is expressed as a percentage. c These results are from
ref 42.
F igu r e 5. Expanded Lineweaver-Burk plot showing the
kinetics of iron transport of [55Fe]ferric L-fluviabactin (15) by
P. denitrificans. The triangles show the high-affinity compo-
nent of the transport (0.1 µM < [S] < 1.0 µM); the low-affinity
system (1.0 µM < [S] < 5.0 µM) is depicted by the squares.
system, whereas the data from 0.1 to 1.0 µM fit another
line (Km ) 0.23 ( 0.03 µM), suggesting a high-affinity
system. The ferric complexes of other catecholamides,
L-agrobactin (17) and L-homofluviabactin (18), which
also contain a chiral oxazoline derived from L-threonine,
exhibited biphasic kinetics with a high affinity compo-
nent (Km ) 0.1-0.2 µM; Table 1) as well. P. denitrificans
thus appears to have a stereospecific high-affinity
system to obtain iron from ferric catecholamides under
low-iron conditions (<1 µM) in which the stereochem-
istry of the chiral oxazoline ring plays a critical role.
(b) Clea r a n ce of Ir on fr om Bile Du ct-Ca n n u la ted
Rod en ts. When L- or D-fluviabactin was administered
sc to bile duct-cannulated rodents at a dose of 150 µmol/
kg, the iron clearing efficacy was very similar to that of
L-parabactin given at the same dose. The efficiency of
L-parabactin was 14.2 ( 2.0%;42 the efficiencies of 15
and 16 were 13.8 ( 0.6 and 16.0 ( 5.3%, respectively
(Table 2). The proportion of the iron that was excreted
in the bile was comparable in all three cases, between
F igu r e 6. Time course of mean biliary iron excretion in
normal rats after administration of L-parabactin (L-PB, n )
8, open circles),42 L-fluviabactin (15, n ) 3, filled circles), or
D-fluviabactin (16, n ) 3, filled squares). All chelators were
given sc at a dose of 150 µmol/kg.
84 and 91% (Table 2). Finally, the deferration kinetics
were also similar for the three ligands (Figure 6), i.e., a
protracted biliary excretion phase which returned to
baseline 48 h after dosing.
Discu ssion
The currently accepted drug for the treatment of
transfusional iron overload is DFO, a natural product

Asymmetric Sites in Choosing Siderophores
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 15 2473
iron chelator (siderophore) isolated from Streptomyces
pilosus.13 As a consequence of poor patient compliance,
investigators have continued to search for alternative
ligands, focusing on synthetic chelators. It is interesting
that, although numerous siderophores from various
microorganisms have been isolated and characterized,
very few have been pursued as therapeutics.
mechanism similar to those proposed for mycobactin-
mediated iron transport in Mycobacterium species65 and
more recently for amonabactin transport in Aeromonas
hydrophila 495A2.66 We have noted previously that an
attractive model involved a stereospecific high-affinity
binding step requiring the L-oxazoline ring, followed by
a nonstereospecific postreceptor processing involving
hydrolysis of the oxazoline ring of ferric L-parabactin
(E0′ ) -0.673 mV67) to the open-chain threonyl struc-
There is one potential problem implicit in utilizing a
siderophore to treat patients with transfusional iron
overload: the ligand can promote the growth of indig-
enous microorganisms. This phenomenon has been
observed in rodents which were given enterobactin and
subsequently died from septicemia36 as well as in
patients using DFO.20,37 Cross-utilization of other sid-
erophores by heterologous bacterial species is well
established.38,39 This is certainly the case in the present
work; the ferric complexes of L-agrobactin, isolated from
Agrobacterium tumefaciens,53 and of L-fluviabactin,
produced by V. fluvialis,43 were both utilized by P.
denitrificans.
ture of ferric parabactin A (E′ ) -0.400 mV67) from
0
which iron might be more readily removed, perhaps by
a nonstereospecific reductase. In this earlier work, we
assessed the contribution of parabactin A to iron
transport via a ligand exchange mechanism. Under the
conditions of the study, such exchange seemed insuf-
ficient to account for the Vmax observed for iron ac-
cumulation from ferric parabactin A.54 It is certainly
true that iron exchange can take place between heter-
ologous and native siderophores, promoting microbial
growth. Nevertheless, on the basis of the work of
Neilands et al. with enantiomeric enterobactin in E.
coli,56 our earlier studies with L- and D-parabactin in
P. denitrificans,41,54 and the results presented here, the
proof of concept is provided that changing the stereo-
chemistry of a siderophore compromises that sidero-
phore’s ability to transport iron and thereby promote
the growth of the organism which synthesized the
siderophore.
The iron clearing efficiencies of L- and D-fluviabactin
are quite high and similar to each other, 13.8 ( 0.6 and
16.0 ( 5.3%, respectively, and are also within error of
the efficiency of L-parabactin.42 These efficiencies are
also between five and six times that of sc-administered
DFO in the rodent model.42 Furthermore, a comparably
protracted biliary iron excretion was observed with all
three of the catecholamides. The current observations
suggest that, given the many siderophores that could
serve as platforms from which to contruct therapeutic
iron chelators, those with asymmetric centers are the
best choice. In these systems, changing the asymmetric
center(s) is likely to result in ligands that will not
support microbial growth, yet could still function as
deferration agents.
In this study, modification of the polyamine backbone
of the siderophore does not substantially affect either
the ability of the ligands to support bacterial growth or
the kinetics of iron uptake by P. denitrificans. Both the
growth curves and the Km values were similar for
L-fluviabactin (15), which possesses a norspermidine
backbone, L-agrobactin (17), which contains a spermi-
dine moiety, and L-homofluviabactin (18), which has a
homospermidine backbone. However, the relatively
small structural alteration of inverting the configuration
of two carbon atoms (4 and 5) within the oxazoline ring
is sufficient to render the siderophores unable to
promote microbial growth. This inversion changes the
stereochemistry of the ferric complex from Λ to ∆, as
inferred from the CD spectra, a configuration that
cannot be utilized by the microorganism. D-Fluviabactin
(16) does not promote the growth of P. denitrificans.
This further supports our previous findings with L- and
D-parabactin.41
Likewise, the kinetics of ferric L- and D-fluviabactin
transport mirrored the results from our earlier publica-
tion.54 The Λ complex of fluviabactin apparently is
recognized by a stereospecific, high-affinity mechanism
which is regulated by iron concentration. We have
isolated such a receptor from iron-starved P. denitrifi-
cans outer membranes.55 Such stereospecificity of iron
transport was reported earlier by Neilands et al. for
enterobactin; the Fe(III) complex of synthetic enan-
tioenterobactin (derived from D-serine) failed to promote
the growth of enterobactin-deficient E. coli mutants.56
High-affinity iron uptake systems are widespread among
prokaryotic and eukaryotic microorganisms.57-60
Exp er im en ta l Section
Gen er a l. N-Carbobenzyloxy-L-threonine and N-carboben-
zyloxy-D-threonine were purchased from Sigma Chemical Co.
(St. Louis, MO), N-carbobenzyloxy-L-threonine-1-(N-succinim-
idyl) ester was purchased from Bachem Bioscience Inc. (King
of Prussia, PA), and all other reagents were purchased from
Aldrich Chemical Co. (Milwaukee, WI). Fisher Optima grade
solvents (Fisher Scientific, Pittsburgh, PA) were used. DMF
was distilled under N2 and stored over molecular sieves.
Distilled solvents were employed for reactions involving che-
lators. Organic extracts were dried with sodium sulfate unless
otherwise indicated. Glassware for chelator reactions and
purification steps were soaked in 3 N HCl for 15 min, rinsed
with distilled water then distilled ethanol, and dried prior to
use. Silica gel 60 (70-230 mesh) obtained from EM Science
(Darmstadt, Germany) or Lipophilic Sephadex LH-20 from
Sigma was used as indicated for column chromatography.
Melting points were determined with a Fisher-J ohns melting
point apparatus and are uncorrected. Proton NMR spectra
were obtained on a Varian Unity 300 at 300 MHz in CD3OD
at ambient temperature unless otherwise indicated; chemical
shifts are given in parts per million downfield from an internal
tetramethylsilane standard with coupling constants (J ) in Hz.
On the other hand, utilization of iron from the ∆
complex by P. denitrificans likely is via a nonstereospe-
cific, low-affinity system, as postulated by Wee et al.61
These uptake systems, which do not appear to be subject
to regulation by iron concentration, have been charac-
terized in many microorganisms, both prokaryotic and
eukaryotic, in the past decade.58-60,62 Many of these
systems operate via a reductase; systems that utilize
ferrisiderophore reductases have been characterized in
microbial species as diverse as Mycobacterium smeg-
matis63 and Saccharomyces cerevisiae.64 A nonstereospe-
cific alternative to a reductase is a ligand exchange

2474 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 15
Bergeron et al.
High-resolution mass spectra were obtained utilizing FAB
ionization from a glycerol matrix on a Fennigan 4516. Optical
rotations were determined at 589 nm (sodium D line) with a
Perkin-Elmer 341 polarimeter using a 10-cm cell path length
in the indicated solvent; c is expressed as g of compound/100
mL. Elemental analyses were performed by Atlantic Microlabs
(Norcross, GA). Circular dichroism (CD) spectra were obtained
utilizing a J asco J 500C spectropolarimeter operating in analog
mode with a cell path length of 1.00 cm.
All chemical components of the liquid culture media were
purchased from Fisher Scientific (Pittsburgh, PA). Trypticase
soy agar and trypticase soy broth were procured from Becton-
Dickinson (Cockeysville, MD). The optical density of bacterial
cultures was measured at 660 nm (OD660) on a Perkin-Elmer
Lambda 3B spectrophotometer (Norwalk, CT). Ethylenedi-
amine-di(o-hydroxyphenyl acetic acid) (EDDA) and trisodium
nitriloacetate (NTA) were obtained from Sigma. [55Fe]Ferric
chloride (specific activity, 41.8 Ci g-1) in 0.5 M HCl and
Biofluor Scintillation Cocktail were purchased from DuPont/
New England Nuclear (Boston, MA). Chelex 100 resin was
purchased from Bio-Rad (Hercules, CA). The C18 silanol
reversed-phase chromatographic support was procured from
J . T. Baker (Phillipsburg, NJ ), and the polyethyleneimine
cellulose thin-layer ion exchange chromatographic support was
obtained from Merck (Cherry Hill, NJ ). The glass fiber filters
used in the accumulation and kinetic assays (Whatman GF/F
Cat. No. 1825-025) were obtained from Fisher. L-Parabactin,
the reference compound in the bacterial growth and iron
accumulation studies (Figures 2 and 3), was synthesized by a
known method.45
Cremophor RH-40 was obtained from BASF (Parsippany,
NJ ). Sprague-Dawley rats were purchased from Charles River
(Wilmington, MA). Nalgene metabolic cages, rat jackets, and
fluid swivels were obtained from Harvard Bioscience (South
Natick, MA). Intramedic polyethylene tubing PE-50 was
obtained from Fisher.
N1,N7-Bis[2,3-b is(b en zyloxy)b en zoyl]n or sp er m id in e
(4).51 The reagents 2,3-bis(benzyloxy)benzoic acid49 (3.34 g, 10
mmol) and 1,1-carbonyldiimidazole (1.62 g, 10 mmol) were
dissolved in CH2Cl2 (100 mL) and stirred for 1 h at room
temperature under a nitrogen atmosphere.50 A solution of 1
(0.67 g, 5.1 mmol) in CH2Cl2 (10 mL) was added, and the
mixture was stirred overnight. The resulting solution was
washed with 2% NaOH (100 mL), H2O (100 mL), and brine
(100 mL) and then dried over MgSO4 and filtered. Solvent
removal in vacuo followed by flash chromatography of the
residue on silica (5% EtOH in EtOAc) gave 4 (3.57 g, 70%) as
a colorless oil: 1H NMR (CDCl3) δ 1.52-1.63 (m, 4 H), 2.45-
2.50 (m, 4 H), 3.28-3.36 (m, 4 H), 5.07 (s, 4 H), 5.15 (s, 4 H),
7.10-7.49 (m, 26 H), 7.64-7.68 (m, 2 H), 8.04-8.08 (br, 1 H).
N4-(N-Ca r b ob en zyloxy-L-t h r eon yl)-N1,N7-b is[2,3-b is-
(ben zyloxy)ben zoyl]n or sp er m id in e (7). A solution of 1,3-
dicyclohexylcarbodiimide (0.34 g, 1.30 mmol) in CH2Cl2 (20
mL) was added to a solution of N-carbobenzyloxythreonine
(0.33 g, 1.50 mmol) and N-hydroxysuccinimide (0.19 g, 1.65
mmol) in CH2Cl2 (20 mL) and stirred at room temperature for
18 h. The mixture was filtered; 4 (1.0 g, 1.3 mmol) and
triethylamine (0.20 g, 2.0 mmol) were added. The resulting
mixture was stirred 24 h, concentrated in vacuo, then dissolved
in EtOAc (100 mL). The solution was washed with H2O (2 ×
50 mL), 10% citric acid (50 mL), and H2O (100 mL), dried, and
filtered. Solvent removal in vacuo followed by flash chroma-
tography on silica (1:1 EtOAc CHCl3) gave 7 (0.68 g, 52%)
as a viscous oil: 1H NMR δ 1.08-1.11 (m, 3 H), 1.51-1.72
(m, 4 H), 2.92-3.31 (m, 8 H), 3.83-3.92 (m, 1 H), 4.38-4.42
(m, 1 H), 5.02-5.21 (m, 10 H), 7.11-7.51 (m, 31 H). Anal.
(C60H62N4O10) C, H, N.
N4-L-Th r eon yl-N1,N7-bis(2,3-dih ydr oxyben zoyl)n or sper -
m id in e Hyd r och lor id e (11). A mixture of 7 (0.75 g, 0.75
mmol) and 10% Pd-C (0.10 g) in HCl-saturated CH3OH (50
mL) was stirred at room temperature for 1 h under a H2
atmosphere at ambient pressure. The mixture was filtered
through acid-washed Celite and concentrated in vacuo to give
0.40 g (99%) of 11 and used without further purification. An
analytical sample was obtained by column chromatography on
LH-20 (10-15% EtOH in toluene) and gave 11 as a white
solid: 1H NMR δ 1.25-1.28 (m, 3 H), 1.84-2.02 (m, 4 H), 3.38-
3.78 (m, 8 H), 4.02-4.10 (m, 1 H), 4.18-4.21 (m, 1 H), 6.68-
6.75 (m, 2 H), 6.91-6.95 (m, 2 H), 7.18-7.21 (m, 2 H). Anal.
(C24H33ClN4O8) C, H, N.
N4-[2-(2,3-Dih yd r oxyp h en yl)-(4S,5R)-tr a n s-5-m eth yl-2-
oxa zolin e-4-ca r b oxa m id o]-N1,N7-b is(2,3-d ih yd r oxyb en -
zoyl)n or sp er m id in e (L-F lu via ba ctin ; 15). Ethyl 2,3-dihy-
droxybenzimidate44 (0.22 g, 1.21 mmol) was added to a solution
of 11 (0.16 g, 0.30 mmol) in CH3OH (20 mL). The reaction
mixture was heated at reflux under a N2 atmosphere for 30 h
and then concentrated in vacuo. Column chromatography on
LH-20 (10% EtOH in toluene) gave 15 (0.12 g, 65%) as a white
glass: 1H NMR (CD3OD, 50 °C) δ 1.41 (d, 3 H, J ) 6.4), 1.85-
1.96 (m, 2 H), 2.03-2.14 (m, 2 H), 3.40-3.92 (m, 8 H), 4.80 (d,
1 H, J ) 6.5), 5.21-5.30 (m, 1 H), 6.61-6.76 (m, 3 H), 6.86-
6.97 (m, 3 H), 7.12-7.22 (m, 3 H); HRMS m/z calcd for
C
31H35N4O10 623.2353 (M + H), found 623.2339; [R]D24
+89.18 (c ) 1.00). Anal. (C31H34N4O10‚0.5 H2O) C, H, N.
)
N4-(N-Ca r b ob en zyloxy-D-t h r eon yl)-N1,N7-b is[2,3-b is-
(ben zyloxy)ben zoyl]n or sp er m id in e (8). A solution of N-
carbobenzyloxy-D-threonine (0.76 g, 3.00 mmol) and 1,3-
dicyclohexylcarbodiimide (0.63 mg, 3.04 mmol) in CH2Cl2 (40
mL) and a solution of 4 (1.50 g, 1.96 mmol) and N-hydroxy-
succinimide (0.35 g, 3.04 mmol) in CH2Cl2 (40 mL) were
combined; triethylamine (0.31 g, 3.06 mmol) was added.
Reaction conditions, workup, and purification were as de-
scribed for 7 and afforded 8 (0.62 g, 32%) as a viscous oil: 1H
NMR δ 1.08-1.11 (d, 3 H, J ) 7), 1.51-1.72 (m, 4 H), 2.92-
3.31 (m, 8 H), 3.83-3.92 (m, 1 H), 4.38-4.42 (m, 1 H), 5.02-
5.21 (m, 10 H), 7.11-7.51 (m, 31 H). Anal. (C60H62N4O10) C,
H, N.
N4-D-Th r eon yl-N1,N7-bis(2,3-dih ydr oxyben zoyl)n or sper -
m id in e Hyd r och lor id e (12). A mixture of 8 (0.30 g, 0.30
mmol) and 10% Pd-C (0.15 g) in HCl saturated CH3OH (50
mL) was reacted under a H2 atmosphere as described for 11.
The resulting white solid (12, 0.16 g, 99%) was used without
further purification. An analytical sample was obtained by
column chromatography on LH-20 (15% EtOH in toluene) and
gave 12 as a white solid: 1H NMR δ 1.25-1.28 (d, 3 H, J ) 7),
1.84-2.02 (m, 4 H), 3.38-3.78 (m, 8 H), 4.02-4.10 (m, 1 H),
4.18-4.21 (m, 1 H), 6.68-6.75 (m, 2 H), 6.91-6.95 (m, 2 H),
7.18-7.21 (m, 2 H). Anal. (C24H33ClN4O8) C, H, N.
N4-[2-(2,3-Dih yd r oxyp h en yl)-(4R,5S)-tr a n s-5-m eth yl-2-
oxa zolin e-4-ca r b oxa m id o]-N1,N7-b is(2,3-d ih yd r oxyb en -
zoyl)n or sp er m id in e (D-F lu via ba ctin ; 16). Utilizing the
conditions, procedures, and workup described for 15, ethyl 2,3-
dihydroxybenzimidate44 (0.12 g, 0.66 mmol) was reacted with
12 (0.15 g, 0.28 mmol). Column chromatography of the
resulting residue on LH-20 (15% EtOH in toluene) gave 16
(0.11 mg, 63%) as a white glass: 1H NMR (CD3OD, 50 °C) δ
1.41 (d, 3 H, J ) 6.4), 1.85-1.96 (m, 2 H), 2.03-2.14 (m, 2 H),
3.40-3.92 (m, 8 H), 4.80 (d, 1 H, J ) 6.5), 5.21-5.30 (m, 1 H),
6.61-6.76 (m, 3 H), 6.86-6.97 (m, 3 H), 7.12-7.22 (m, 3 H);
HRMS m/z calcd for C31H35N4O10 623.2353 (M + H), found
623.2334; [R]D24 ) -85.86 (c ) 1.00). Anal. (C31H34N4O10‚0.5
H2O) C, H, N.
N1,N8-Bis[2,3-bis(ben zyloxy)ben zoyl]sp er m id in e (5).50
A mixture of 2,3-bis(benzyloxy)benzoic acid (2.60 g, 7.78 mmol),
1,1-carbonyldiimidazole (1.26 g, 7.77 mmol), and 2 (0.56 g, 3.86
mmol) in CH2Cl2 (10 mL) was reacted and worked up as
described for 4. Flash chromatography of the resulting residue
on silica (10% CH3OH in CHCl3) gave 5 (2.2 g, 73%) as a pale
yellow oil: 1H NMR δ 1.37-1.44 (m, 4 H), 1.56-1.66 (m, 2 H),
2.38-2.44 (m, 2 H), 2.44-2.55 (m, 2 H), 3.20-3.26 (m, 4 H),
5.08 (s, 4 H), 5.16-5.17 (m, 4 H), 7.10-7.50 (m, 26 H).
N4-(N-Ca r b ob en zyloxy-L-t h r eon yl)-N1,N8-b is[2,3-b is-
(ben zyloxy)ben zoyl]sp er m id in e (9). Utilizing the proce-
dures described for 7, N-carbobenzyloxy-L-threonine (0.76 g,
3.00 mmol), 1,3-dicyclohexylcarbodiimide (0.63 g, 3.04 mmol)
in dry CH2Cl2 (40 mL), 5 (1.5 g, 1.93 mmol), and N-hydroxy-
succinimide (0.35 g, 3.04 mmol) in dry CH2Cl2 (40 mL) and
triethylamine (0.31 g, 3.06 mmol) were combined and reacted.

Asymmetric Sites in Choosing Siderophores
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 15 2475
After workup, flash chromatography of the resulting residue
on silica (1:1 EtOAc CHCl3) gave 9 (1.08 g, 55%) as a viscous
oil: 1H NMR δ 1.13-1.16 (m, 3 H), 1.56-1.76 (m, 6 H), 2.94-
3.35 (m, 8 H), 3.72-3.86 (m, 1 H), 4.32-4.41 (m, 1 H), 5.06-
5.25 (m, 10 H), 7.10-7.50 (m, 31 H). Anal. (C61H64N4O10) C,
H, N.
N4-L-Th r eon yl-N1,N8-bis(2,3-d ih yd r oxyben zoyl)sp er m i-
d in e Hyd r och lor id e (13). A mixture of 9 (0.31 g, 0.31 mmol)
and 10% Pd-C (0.15 g) in HCl-saturated CH3OH (30 mL) was
reacted under a H2 atmosphere as described for 11. The
resulting white solid (13, 0.16 g, 93%) was used without further
purification. An analytical sample was obtained by column
chromatography on LH-20 (15% EtOH in toluene) and afforded
13 as a white solid: 1H NMR δ 1.07-1.10 (m, 3 H), 1.52-1.73
(m, 6 H), 2.91-3.32 (m, 8 H), 3.82-3.91 (m, 1 H), 4.36-4.44
(m, 1 H), 6.67-6.73 (m, 2 H), 6.90-6.93 (m, 2 H), 7.19-7.23
(m, 2 H). Anal. (C25H35ClN4O8) C, H, N.
N4-[2-(2,3-Dih yd r oxyp h en yl)-(4S,5R)-tr a n s-5-m eth yl-2-
oxa zolin e-4-ca r b oxa m id o]-N1,N8-b is(2,3-d ih yd r oxyb en -
zoyl)sp er m id in e (L-Agr oba ctin ; 17). Ethyl 2,3-dihydroxy-
benzimidate44 (0.12 mg, 0.66 mmol) was added to a solution
of 13 (0.14 g, 0.25 mmol) in CH3OH (30 mL); reaction
conditions and workup were as described for 15. Column
chromatography of the resulting residue on LH-20 (15% EtOH
in toluene) gave 17 (0.10 g, 63%) as a white glass: 1H NMR
(CD3OD, 50 °C) δ 1.40 (d, 1.26 H, J ) 6.4) 1.45 (d, 1.74 H, J
) 6.4), 1.56-1.96 (m, 5 H), 2.00-2.13 (m, 1 H), 3.34-3.90 (m,
8 H), 4.78 (d, 0.46 H, J ) 6.2) 4.85 (d, 0.54 H, J ) 6.5), 5.20-
5.29 (m, 1 H), 6.61-6.76 (m, 3 H), 6.86-6.97 (m, 3 H), 7.11-
7.24 (m, 3 H); HRMS m/z calcd for C32H37N4O10 637.2510 (M
+ H), found 637.2510. Anal. (C32H36N4O10‚0.5 H2O) C, H, N.
H), 7.12-7.23 (m, 3 H); HRMS m/z calcd for C33H39N4O10
651.2666 (M + H), found 651.2664. Anal. (C33H38N4O10‚0.5
H2O) C, H, N.
Low -F e Defin ed Min im a l Sa lts Med iu m . Glassware was
presoaked in 3 N HCl for 15 min and rinsed well with water;
all water used was distilled in a Mega-Pure still (Corning
Costar, Corning, NY) and passed through a deionizing car-
tridge (Sybron-Barnstead, Milwaukee, WI) before use.
For a 12-L batch of culture medium, succinic acid (70.8 g),
KH2PO4 (48.0 g), Na2HPO4 (60.6 g), and NH4Cl (19.2 g) were
dissolved in H2O (10 L); the solution was autoclaved and was
then stored at 4 °C for 3-4 days to allow Fe salts to
coagulate.68 After filtration through a 0.2-µm membrane and
adjustment of the pH to 7.0, the solution was passed through
a column of Chelex 100 resin (1500 g). The pH was again
adjusted to 7.0, Chelex-treated Tween 80 (300 mL) was added,
and the final volume was adjusted to 12 L (final Tween
concentration, 0.5%). The following were then added to the
medium to provide divalent cations and trace metals which
were removed during Chelex treatment (final concentration
in parentheses): MgSO4 (1.7 mM), Ca2+ (182 µM), Mn2+ (10
µM), Zn2+ (1 µM), Cu2+ (0.1 µM), and Co2+ (0.01 µM).54 Atomic
absorption analysis revealed that the liquid culture medium
had an Fe concentration of < 0.1 µM; thus, the degree of low-
iron stress could then be controlled by addition of a known
quantity of ferric(NTA)2 to the medium. “Transport buffer” is
the culture medium without additional Fe.
P r ep a r a tion of La beled Ch ela tes. Stock solutions of
labeled chelates of compounds 15-18 were prepared according
to the ligand exchange method described previously for [55Fe]-
ferric parabactin with ferric (NTA)2.54 Briefly, [55Fe]ferric-
(NTA)2 was prepared from 55FeCl3 in 0.5 M HCl and trisodium
NTA in 10% NTA excess in 0.1 N HCl; the pH was adjusted
to 7.0 with a Tris buffer solution (500 mM). Ligand exchange
was accomplished employing a 10% molar excess of ligand in
EtOH and ferric (NTA)2; Tris hydrochloride (500 mM, pH 7.4)
was used to dilute the labeled chelate solution to 300 µM.
Purification on a C18 silanol reversed-phase column (2:5 CH3-
OH/H2O) yielded the labeled chelate solution, which was then
adjusted with Tris hydrochloride (100 mM, pH 7.4) to a
concentration of 100-300 µM as determined spectrophoto-
metrically at the visible λmax of the chelates: ferric fluviabactin
(ꢀ508 ) 3.8 × 103 M-1 cm-1); ferric homofluviabactin (ꢀ516 ) 3.1
× 103 M-1 cm-1); ferric agrobactin (ꢀ507 ) 3.6 × 103 M-1 cm-1).
The purified chelates were homogeneous by thin-layer ion
exchange chromatography on polyethyleneimine cellulose (2:
2:1 THF/CH3OH/10 mM tetrabutylammonium phosphate, pH
7.0); their respective Rf values were 0.59-0.63, 0.60-0.64, and
0.60-0.64. When the ferric fluviabactin complex was dissoci-
ated by addition of pH 3.0 phosphate buffer and extracted with
EtOAc, the organic layer contained only the intact oxazoline,
fluviabactin, as determined by thin-layer chromatography on
silica gel (94:6 CHCl3/CH3OH), Rf ) 0.40-0.44.
N 1,N 9-Bis[2,3-b is(b e n zyloxy)b e n zoyl]h om osp e r m i-
d in e (6). A mixture of 2,3-bis(benzyloxy)benzoic acid (2.10 g,
6.28 mmol), 1,1-carbonyldiimidazole (1.02 g, 6.29 mmol), and
348 (0.5 g, 3.14 mmol) in CH2Cl2 (10 mL) was reacted and
worked up as described for 4. Flash chromatography of the
resulting crude material on silica (10% CH3OH in CHCl3) gave
6 (1.6 g, 64%) as a pale yellow oil: 1H NMR δ 1.41-1.48 (m,
8 H), 2.51-2.58 (m, 4 H), 3.22-3.28 (m, 4 H), 4.85 (s, 4 H),
5.09 (s, 2 H), 5.18 (s, 2 H), 7.01-7.50 (m, 26 H).
N5-(N-Ca r b ob en zyloxy-L-t h r eon yl)-N1,N9-b is[2,3-b is-
(ben zyloxy)ben zoyl]h om osp er m id in e (10). A mixture of
6 (1.50 g, 1.89 mmol), N-carbobenzyloxy-L-threonine-1-(N-
succinimidyl) ester (1.00 g, 2.85 mmol), and triethylamine (0.20
g, 1.98 mmol) in CH2Cl2 (100 mL) was stirred overnight. The
resulting solution was washed with H2O (100 mL), 10% citric
acid (100 mL), and brine (100 mL) and then dried and filtered.
Solvent removal in vacuo followed by flash chromatography
of the residue on silica (1:1 EtOAc/CHCl3) gave 10 (1.49 g, 77%)
as a viscous oil: 1H NMR δ 1.09-1.11 (d, 3 H, J ) 7), 1.31-
1.54 (m, 8 H), 3.01-3.28 (m, 8 H), 3.84-3.93 (m, 1 H), 4.42-
4.45 (m, 1 H), 5.07-5.16 (m, 10 H), 7.09-7.48 (m, 31 H). Anal.
(C62H66N4O10) C, H, N.
N5-L-Th r eon yl-N1,N9-bis(2,3-d ih yd r oxyben zoyl)h om o-
sp er m id in e Hyd r och lor id e (14). A mixture of 10 (0.60 g,
0.59 mmol) and 10% Pd-C (0.20 g) in HCl-saturated CH3OH
(50 mL) was reacted under a H2 atmosphere as described for
11. The resulting white solid (14, 0.30 g, 90%) was used
without further purification. An analytical sample was ob-
tained by column chromatography on LH-20 (15% EtOH in
toluene) to give 14 as a white solid: 1H NMR δ 1.19-1.21 (d,
3 H, J ) 7), 1.41-1.54 (m, 8 H), 3.11-3.36 (m, 8 H), 4.44-
4.53 (m, 1 H), 4.92-5.12 (m, 1 H), 6.84-7.28 (m, 6 H). Anal.
(C26H37ClN4O8) C, H, N.
N5-[2-(2,3-Dih yd r oxyp h en yl)-(4S,5R)-tr a n s-5-m eth yl-2-
oxa zolin e-4-ca r b oxa m id o]-N1,N9-b is(2,3-d ih yd r oxyb en -
zoyl)h om osp er m id in e (L-Hom oflu via ba ctin ; 18) Ethyl 2,3-
dihydroxybenzimidate44 (0.20 g, 1.1 mmol) and 14 (0.30 g, 0.53
mmol) in CH3OH (50 mL) were reacted and worked up as
described for 15. Column chromatography of the resulting
residue on LH-20 (15% EtOH in toluene) gave 18 (0.22 g, 64%)
as a white glass: 1H NMR (CD3OD, 50 °C) δ 1.44 (d, 3 H, J )
6.4), 1.55-1.88 (m, 8 H), 3.36-3.79 (m, 8 H), 4.83 (d, 1 H, J )
6.5), 5.19-5.29 (m, 1 H), 6.63-6.76 (m, 3 H), 6.85-6.98 (m, 3
P r ep a r a tion of Un la beled Ch ela tes for CD Sp ectr a .
Unlabeled ferric L-fluviabactin and ferric D-fluviabactin solu-
tions utilized for the CD spectral studies employed unlabeled
Fe(NTA)2 and were prepared as described above for the labeled
chelates.
Ba cter ia l Str a in s. P. denitrificans (ATCC strain 17741)
was maintained on trypticase soy agar plates. Unless other-
wise indicated, all liquid cultures were incubated at 30 °C with
rotary shaking (120 rpm). Individual bacterial colonies were
inoculated into trypticase soy broth (20 mL) in 250-mL culture
flasks and incubated for 24 h. Inoculations were then made
into the defined medium containing 1 µM Fe(III) (50 mL) to
give a starting OD of 0.015 and incubated for 16 h. This
culture, in turn, was used to inoculate a 250-mL culture flask
containing defined medium with 0.5 µM Fe(III) (50 mL) to give
a starting OD of 0.040. After a 12-h incubation, the cells were
in late logarithmic growth (3.0 < OD660 < 3.5), and the medium
was positive for catechols.54 Cells were harvested by centrifu-
gation (15 min at 1100g) at 15-20 °C, washed twice with
transport buffer, and resuspended in fresh transport buffer
to an OD660 reading of 1.000.

2476 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 15
Bergeron et al.
P r otein Deter m in a tion . The protein concentration in the
cell suspensions was estimated by the method of Lowry et al.69
with the modifications for membrane protein samples as given
by Markwell et al.70 using bovine serum albumin as the
standard. The cell density of the suspensions was estimated
spectrophotometrically as described previously;54 1 OD660 unit
refers to the amount of cells yielding an OD660 of 1.000 when
suspended in transport buffer at room temperature. Under the
growth conditions described, 1 OD660 unit corresponded to
∼0.25 mg of protein.
from the duodenum and tied in place. After threading through
the shoulder, the cannula was passed from the rat to the swivel
inside a metal torque-transmitting tether, which was attached
to a rodent jacket around the animal’s chest. The cannula was
directed from the rat to a Gilson microfraction collector
(Middleton, WI) by a fluid swivel mounted above the metabolic
cage. Bile samples were collected at 3-h intervals for 48 h.
Urine samples were taken every 24 h. Sample collection and
handling were as previously described.31 The efficiency of each
chelator was calculated by subtracting the iron excretion of
control animals from the iron excretion of treated animals.
This number was then divided by the theoretical output on
the basis of a 1:1 ligand-iron complex; the result is expressed
as a percentage.
Gr ow th Stu d ies. The effects of D- and L-fluviabactin,
L-homofluviabactin, and L-agrobactin on the growth of P.
denitrificans in iron-deficient media were assessed. Iron-free
ligands (2 µM in CH3OH) were added to empty, sterile flasks,
evaporated to dryness under nitrogen [which was passed
through a 0.22-µm pore size filter (Millipore)], and redissolved
in sterile medium. For these studies, low-iron defined minimal
salts medium was formulated as described above except that
EDDA was added (1.1 mM final concentration) to the me-
dium.41 Bacterial suspension was added to the medium such
that the starting OD660 was 0.017. Growth rates were deter-
mined by monitoring OD660 readings vs time (t ) 5, 10, 20,
24, 36, and 48 h; Figure 1). The results of the growth studies
are reported as OD660 vs time.
Accu m u la tion of Ra d iola beled Liga n d s. A cell suspen-
sion with an initial OD660 ) 1.00 (50 mL) was added to each
250-mL flask and incubated for 15 min before addition of stock
solutions of labeled chelates (1 mL; final concentration 1 µM;
specific activity, 0.033 µCi mL-1). Samples of cell suspensions
(5 mL) were aseptically removed at t ) ca. 0, 5, 10 min, and
at 10-min intervals through 60 min and immediately diluted
with cold transport medium (5 mL) to stop the reaction.
Whatman GF/F fiber glass filters, which had been presoaked
for 24 h in 500 µM unlabeled chelate and rinsed with chilled
transport buffer (5 mL), were used to separate cells from
medium by rapid filtration. After additional rinsing with
transport buffer (2 × 5 mL), the filters were dried at 70 °C
and placed in a scintillation vial containing Biofluor (10 mL).
Liquid scintillation counting was performed in a Beckman
LS6000IC counter (Beckman, Palo Alto, CA). Control values
of labeled chelates adsorbed to the filters in the absence of
cells were measured and subtracted from the values obtained
in the presence of cells. Uptake of labeled chelates by P.
denitrificans is reported as the net amount of radiolabeled Fe
per mg of cell suspension.
Atom ic Absor p tion Ir on Deter m in a tion s. Bile and urine
samples were analyzed on a Perkin-Elmer 5100 PC atomic
absorption spectrophotometer fitted with a model AS-51 au-
tosampler using a quartz sampling probe as previously de-
scribed.30
Ack n ow led gm en t. We appreciate the constructive
critiques of Dr. J ames S. McManis, the technical as-
sistance of Elizabeth M. Nelson, and the aid of Dr.
Eileen Eiler-McManis in the organization and editing
of this manuscript.
Refer en ces
(1) Raymond, K. N.; Carrano, C. J . Coordination Chemistry and
Microbial Iron Transport. Acc. Chem. Res. 1979, 12, 183-190.
(2) Bergeron, R. J . Iron: A Controlling Nutrient in Proliferative
Processes. Trends Biochem. Sci. 1986, 11, 133-136.
(3) Finch, C. A.; Deubelbeiss, K.; Cook, J . D.; Eschbach, J . W.;
Harker, L. A.; Funk, D. D.; Marsaglia, G.; Hillman, R. S.;
Slichter, S.; Adamson, J . W.; Ganzoni, A.; Giblett, E. R. Ferro-
kinetics in Man. Medicine (Baltimore) 1970, 49, 17-53.
(4) Finch, C. A.; Huebers, H. A. Perspectives in Iron Metabolism.
N. Engl. J . Med. 1982, 306, 1520-1528.
(5) Finch, C. A.; Huebers, H. A. Iron Metabolism. Clin. Physiol.
Biochem. 1986, 4, 5-10.
(6) Hallberg, L. Bioavailability of Dietary Iron in Man. Annu. Rev.
Nutr. 1981, 1, 123-147.
(7) O’Connell, M. J .; Ward, R. J .; Baum, H.; Peters, T. J . The Role
of Iron in Ferritin- and Haemosiderin-Mediated Lipid Peroxi-
dation in Liposomes. Biochem. J . 1985, 229, 135-139.
(8) Seligman, P. A.; Klausner, R. D.; Huebers, H. A. Molecular
Mechanisms of Iron Metabolism. The Molecular Basis of Blood
Diseases; W. B. Saunders: Philadelphia, 1987; p 219.
(9) Thomas, C. E.; Morehouse, L. A.; Aust, S. D. Ferritin and
Superoxide-Dependent Lipid Peroxidation. J . Biol. Chem. 1985,
260, 3275-3280.
(10) Weintraub, L. R.; Conrad, M. E.; Crosby, W. H. The Treatment
of Hemochromatosis by Phlebotomy. Med. Clin. North Am. 1966,
50, 1579-1590.
(11) Hoffbrand, A. V. Transfusion Siderosis and Chelation Therapy.
Iron in Biochemistry and Medicine, II; Academic Press: London,
1980; p 499.
Kin etic Tr a n sp or t Assa y. Polystyrene conical-bottom
tubes (50-mL) were used as incubation vessels. The cell
suspensions (OD660 ) 1.00, 3.45 mL/tube) were incubated with
shaking in a water bath (30 °C) for 15 min. After the addition
of radiolabeled chelate (50 µL), samples (500 µL) were with-
drawn every 20 s for 2 min; the reaction was stopped by adding
the sample to 5 mL of chilled (0 °C) transport buffer. Triplicate
100-µL portions of each incubation mixture were counted to
verify the amount of label added. Preparation and counting
of the samples were as described for the accumulation assay
above.
(12) Brittenham, G. M. Pyridoxal Isonicotinoyl Hydrazone (PIH):
Effective Iron Chelation After Oral Administration. Ann. N. Y.
Acad. Sci. 1990, 612, 315-326.
The data were manipulated as described in an earlier
publication.54 Briefly, the six datum points taken during the
first 2 min for each assay tube were fitted to a regression line
with a slope equal to the initial reaction velocity (v0). These
(13) Bickel, H.; Hall, G. E.; Keller-Schierlein, W.; Prelog, V.; Vischer,
E.; Wettstein, A. Metabolic Products of Actinomycetes. XXVII.
Constitutional Formula of Ferrioxamine B. Helv. Chim. Acta
1960, 43, 2129-2138.
(14) Brittenham, G. M.; Griffith, P. M.; Nienhuis, A. W.; McLaren,
C. E.; Young, N. S.; Tucker, E. E.; Allen, C. J .; Farrell, D. E.;
Harris, J . W. Efficacy of Deferoxamine in Preventing Complica-
tions of Iron Overload in Patients with Thalassemia Major. N.
Engl. J . Med. 1994, 331, 567-573.
(15) Graziano, J . H.; Markenson, A.; Miller, D. R.; Chang, H.; Bestak,
M.; Meyers, P.; Pisciotto, P.; Rifkind, A. Chelation Therapy in
Beta-Thalassemia Major. I. Intravenous and Subcutaneous
Deferoxamine. J . Pediatr. 1978, 92, 648-652.
(16) Pippard, M. J .; Callender, S. T. The Management of Iron
Chelation Therapy. Br. J . Haematol. 1983, 54, 503-507.
(17) Bousquet, J .; Navarro, M.; Robert, G.; Aye, P.; Michel, F. B.
Rapid Desensitization for Desferrioxamine Anaphylactoid Reac-
tions. Lancet 1983, 2, 859-860.
initial rates, expressed as pg-atoms 55Fe min-1 (mg of protein)-1
,
were then used to generate Lineweaver-Burk plots by using
Cleland’s Fortran program71 to estimate Km, Vmax, and their
standard errors. Rates varied linearly with the number of cells
present (OD660 from 0.5 to 2.000). Data were corrected for cell
growth that had occurred during the experiment.
Ca n n u la tion of Bile Du ct in Ra ts. The cannulation has
been described in detail in earlier publications.31,72 Briefly,
male Sprague-Dawley rats averaging 400 g were housed in
Nalgene plastic metabolic cages during the experimental
period and were given free access to water. The animals were
anesthetized using sodium pentobarbital (55 mg/kg) given ip.
The bile duct was cannulated using 22-gauge polyethylene PE-
50 tubing. The cannula was inserted into the duct about 1 cm
(18) Miller, K. B.; Rosenwasser, L. J .; Bessette, J . A. M.; Beer, D. J .;
Rocklin, R. E. Rapid Desensitisation for Desferrioxamine Ana-
phylactic Reaction. Lancet 1981, 1, 1059.

Asymmetric Sites in Choosing Siderophores
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 15 2477
(19) Shalit, M.; Tedeschi, A.; Miadonna, A.; Levi-Schaffer, F. Desferal
(desferrioxamine)-A Novel Activator of Tissue-Type Mast Cells.
J . Allergy Clin. Immunol. 1991, 88, 854-860.
(20) Kirking, M. H. Treatment of Chronic Iron Overload. Clin. Pharm.
1991, 10, 775-783.
(21) Kontoghiorghes, G. J .; Aldouri, M. A.; Sheppard, L.; Hoffbrand,
A. V. 1,2-Dimethyl-3-hydroxypyrid-4-one, an Orally Active Ch-
elator for Treatment of Iron Overload. Lancet 1987, 1, 1294-
1295.
(22) Uhlir, L. C.; Durbin, P. W.; J eung, N.; Raymond, K. N. Specific
Sequestering Agents for the Actinides. 21. Synthesis and Initial
Biological Testing of Octadentate Mixed Catecholate-Hydroxy-
pyridinonate Ligands. J . Med. Chem. 1993, 36, 504-509.
(23) Grady, R. W.; Hershko, C. HBED: A Potential Oral Iron
Chelator. Ann. N. Y. Acad. Sci. 1990, 612, 361-368.
(24) Bergeron, R. J . Synthesis and Solution Structures of Microbial
Siderophores. Chem. Rev. 1984, 84, 587-602.
(25) Bergeron, R. J .; Wiegand, J .; McManis, J . S.; Perumal, P. T.
Synthesis and Biological Evaluation of Hydroxamate-Based Iron
Chelators. J . Med. Chem. 1991, 34, 3182-3187.
(26) Bergeron, R. J .; McManis, J . S. Synthesis and Biological Activity
of Hydroxamate-Based Iron Chelators. The Development of Iron
Chelators for Clinical Use; CRC: Boca Raton, 1994; pp 237-
273.
(42) Bergeron, R. J .; Streiff, R. R.; King, W.; Daniels, R. D., J r.;
Wiegand, J . A Comparison of the Iron-Clearing Properties of
Parabactin and Desferrioxamine. Blood 1993, 82, 2552-2557.
(43) Yamamoto, S.; Okujo, N.; Fujita, Y.; Saito, M.; Yoshida, T.;
Shinoda, S. Structures of Two Polyamine-Containing Catecho-
late Siderophores from Vibrio fluvialis. J . Biochem. (Tokyo) 1993,
113, 538-544.
(44) Bergeron, R. J .; McManis, J . S.; Dionis, J . B.; Garlich, J . R. An
Efficient Total Synthesis of Agrobactin and Its Gallium(III)
Chelate. J . Org. Chem. 1985, 50, 2780-2782.
(45) Bergeron, R. J .; Kline, S. J . Short Synthesis of Parabactin. J .
Am. Chem. Soc. 1982, 104, 4489-4492.
(46) Bergeron, R. J .; Kline, S. J .; Stolowich, N. J .; McGovern, K. A.;
Burton, P. S. Flexible Synthesis of Polyamine Catecholamides.
J . Org. Chem. 1981, 46, 4524-4529.
(47) Buckley, G. M.; Pattenden, G.; Whiting, D. A. New Synthetic
Probes of the Iron Transport System of Paracoccus denitrificans.
Tetrahedron 1994, 50, 11781-11792.
(48) Bergeron, R. J .; Feng, Y.; Weimar, W. R.; McManis, J . S.;
Dimova, H.; Porter, C.; Raisler, B.; Phanstiel, O. A Comparison
of Structure-Activity Relationships between Spermidine and
Spermine Analogue Antineoplastics. J . Med. Chem. 1997, 40,
1475-1494.
(49) Rastetter, W. H.; Erickson, T. J .; Venuti, M. C. Synthesis of Iron
(27) Cox, C. D.; Graham, R. Isolation of an Iron-Binding Compound
from Pseudomonas aeruginosa. J . Bacteriol. 1979, 137, 357-
364.
(28) Cox, C. D.; Rinehart, K. L., J r.; Moore, M. L.; Cook, J . C., J r.
Pyochelin: Novel Structure of an Iron-Chelating Growth Pro-
moter for Pseudomonas aeruginosa. Proc. Natl. Acad. Sci. U.S.A.
1981, 78, 4256-4260.
(29) Smith, M. J . Total Synthesis and Absolute Configuration of
Rhizobactin, a Structurally Novel Siderophore. Tetrahedron Lett.
1989, 30, 313-316.
(30) Bergeron, R. J .; Streiff, R. R.; Wiegand, J .; Vinson, J . R. T.;
Luchetta, G.; Evans, K. M.; Peter, H.; J enny, H.-B. A Compara-
tive Evaluation of Iron Clearance Models. Ann. N. Y. Acad. Sci.
1990, 612, 378-393.
Chelators. Enterobactin, Enantioenterobactin, and a Chiral
Analogue. J . Org. Chem. 1981, 46, 3579-3590.
(50) J oshua, A. V.; Scott, J . R. A Simple Method for the Direct Bis-
Acylation of the Primary Amino Groups in Spermidine and
Other Linear Triamines. Tetrahedron Lett. 1984, 25, 5725-5728.
(51) Peterson, T.; Falk, K.-E.; Leong, S. A.; Klein, M. P.; Neilands,
J . B. Structure and Behavior of Spermidine Siderophores. J . Am.
Chem. Soc. 1980, 102, 7715-7718.
(52) Miyasaka, T.; Nagao, Y.; Fujita, E.; Sakurai, H.; Ishizu, K.
Synthesis of Parabactin Analogues and Formation of Transition
Metal Complexes of Parabactin and Related Compounds. J .
Chem. Soc., Perkin Trans. 2 1987, 1543-1549.
(53) Ong, S. A.; Peterson, T.; Neilands, J . B. Agrobactin, a Sidero-
phore from Agrobacterium tumefaciens. J . Biol. Chem. 1979, 254,
1860-1865.
(54) Bergeron, R. J .; Weimar, W. R. Kinetics of Iron Acquisition from
Ferric Siderophores by Paracoccus denitrificans. J . Bacteriol.
1990, 172, 2650-2657.
(31) Bergeron, R. J .; Wiegand, J .; Dionis, J . B.; Egli-Karmakka, M.;
Frei, J .; Huxley-Tencer, A.; Peter, H. H. Evaluation of Desfer-
rithiocin and its Synthetic Analogues as Orally Effective Iron
Chelators. J . Med. Chem. 1991, 34, 2072-2078.
(32) Bergeron, R. J .; Streiff, R. R.; Creary, E. A.; Daniels, R. D., J r.;
King, W.; Luchetta, G.; Wiegand, J .; Moerker, T.; Peter, H. H.
A Comparative Study of the Iron-Clearing Properties of Desfer-
rithiocin Analogues with Desferrioxamine B in a Cebus Monkey
Model. Blood 1993, 81, 2166-2173.
(33) Bergeron, R. J .; Liu, C. Z.; McManis, J . S.; Xia, M. X. B.; Algee,
S. E.; Wiegand, J . The Desferrithiocin Pharmacophore. J . Med.
Chem. 1994, 37, 1411-1417.
(34) Bergeron, R. J .; Wiegand, J .; Weimar, W. R.; Vinson, J . R. T.;
Bussenius, J .; Yao, G. W.; McManis, J . S. Desazadesmethyldes-
ferrithiocin Analogues as Orally Effective Iron Chelators. J . Med.
Chem. 1999, 42, 95-108.
(35) Bergeron, R. J .; Wiegand, J .; McManis, J . S.; McCosar, B. H.;
Weimar, W. R.; Brittenham, G. M.; Smith, R. E. Effects of C-4
Stereochemistry and C-4′ Hydroxylation on the Iron Clearing
Efficiency and Toxicity of Desferrithiocin Analogues. J . Med.
Chem. 1999, 42, 2432-2440.
(36) Guterman, S. K.; Morris, P. M.; Tannenberg, W. J . K. Feasibility
of Enterochelin as an Iron-Chelating Drug: Studies with Human
Serum and a Mouse Model System. Gen. Pharmacol. 1978, 9,
123-127.
(37) Nouel, O.; Voisin, P. M.; Vaucel, J .; Dartois Hoguin, M.; Le Bris,
M. [Yersinia enterocolitica Septicemia Associated with Idiopathic
Hemochromatosis and Deferoxamine Therapy. A Case]. Presse
Med. 1991, 20, 1494-1496.
(38) Carson, S. D. B.; Klebba, P. E.; Newton, S. M. C.; Sparling, P.
F. Ferric Enterobactin Binding and Utilization by Neisseria
gonorrhoeae. J . Bacteriol. 1999, 181, 2895-2901.
(39) Coulanges, V.; Andre, P.; Vidon, D. J .-M. Effect of Siderophores,
Catecholamines, and Catechol Compounds on Listeria spp.
Growth in Iron-Complexed Medium. Biochem. Biophys. Res.
Commun. 1998, 249, 526-530.
(55) Bergeron, R. J .; Weimar, W. R.; Dionis, J . B. Demonstration of
Ferric L-Parabactin Binding Activity in the Outer Membrane
of Paracoccus denitrificans. J . Bacteriol. 1988, 170, 3711-3717.
(56) Neilands, J . B.; Erickson, T. J .; Rastetter, W. H. Stereospecificity
of the Ferric Enterobactin Receptor of Escherichia coli K-12. J .
Biol. Chem. 1981, 256, 3831-3832.
(57) Neilands, J . B. Siderophores: Structure and Function of Micro-
bial Iron Transport Compounds. J . Biol. Chem. 1995, 270,
26723-26726.
(58) Leong, S. A.; Winkelmann, G. Molecular Biology of Iron Trans-
port in Fungi. Met. Ions Biol. Syst. 1998, 35, 147-186.
(59) Braun, V.; Hantke, K.; Koster, W. Bacterial Iron Transport:
Mechanisms, Genetics, and Regulation. Met. Ions Biol. Syst.
1998, 35, 67-145.
(60) Ratledge, C.; Dover, L. G. Iron Metabolism in Pathogenic
Bacteria. Annu. Rev. Microbiol. 2000, 54, 881-941.
(61) Wee, S.; Hardesty, S.; Madiraju, M. V. V. S.; Wilkinson, B. J .
Iron-Regulated Outer Membrane Proteins and Non-Siderophore-
Mediated Iron Uptake by Paracoccus denitrificans. FEMS
Microbiol. Lett. 1988, 51, 33-36.
(62) Byers, B. R.; Arceneaux, J . E. Microbial Iron Transport: Iron
Acquisition by Pathogenic Microorganisms. Met. Ions Biol. Syst.
1998, 35, 37-66.
(63) McCready, K. A.; Ratledge, C. Ferrimycobactin Reductase Activ-
ity from Mycobacterium smegmatis. J . Gen. Microbiol. 1979, 113,
67-72.
(64) Lesuisse, E.; Labbe, P. Reductive and Non-Reductive Mecha-
nisms of Iron Assimilation by the Yeast Saccharomyces cerevi-
siae. J . Gen. Microbiol. 1989, 135, 257-263.
(65) Ratledge, C.; Patel, P. V.; Mundy, J . Iron Transport in Myco-
bacterium smegmatis: The Location of Mycobactin by Electron
Microscopy. J . Gen. Microbiol. 1982, 128, 1559-1565.
(66) Stintzi, A.; Barnes, C.; Xu, J .; Raymond, K. N. Microbial Iron
Transport via a Siderophore Shuttle: A Membrane Ion Trans-
port Paradigm. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 10691-
10696.
(40) Kingsley, R. A.; Reissbrodt, R.; Rabsch, W.; Ketley, J . M.; Tsolis,
R. E. M.; Everest, P.; Dougan, G.; Baumler, A. J .; Roberts, M.;
Williams, P. H. Ferrioxamine-Mediated Iron(III) Utilization by
Salmonella enterica. Appl. Environ. Microbiol. 1999, 65, 1610-
1618.
(41) Bergeron, R. J .; Dionis, J . B.; Elliott, G. T.; Kline, S. J .
Mechanism and Stereospecificity of the Parabactin-Mediated
Iron-Transport System in Paracoccus denitrificans. J . Biol.
Chem. 1985, 260, 7936-7944.
(67) Robinson, J . P.; McArdle, J . V. Electrochemistry of Ferric
Complexes of Parabactin and Parabactin A. J . Inorg. Nucl.
Chem. 1981, 43, 1951-1953.
(68) Rosenberg, H. Transport of Iron into Bacterial Cells. Methods
Enzymol. 1979, 56.

2478 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 15
Bergeron et al.
(69) Lowry, O. H.; Rosebrough, N. J .; Farr, A. L.; Randall, R. J .
Protein Measurement with the Folin Phenol Reagent. J . Biol.
Chem. 1951, 193, 265-275.
(70) Markwell, M. K.; Haas, S. M.; Bieber, L. L.; Tolbert, N. E. A
Modification of the Lowry Procedure to Simplify Protein Deter-
mination in Membrane and Lipoprotein Samples. Anal. Biochem.
1978, 87, 206-210.
(71) Cleland, W. W. The Statistical Analysis of Enzyme Kinetic Data.
Adv. Enzymol. 1967, 1-32.
(72) Bergeron, R. J .; Wiegand, J .; Brittenham, G. M. HBED:
A
Potential Alternative to Deferoxamine for Iron-Chelating Therapy.
Blood 1998, 91, 1446-1452.
J M010019S