Fluorescent 3-hydroxy-4-pyridinone hexadentate iron chelators: Intracellular distribution and the relevance to antimycobacterial properties

Article abstract of DOI:10.1007/s00775-010-0650-1

We report the synthesis and characterization of a fluorescent iron chelator (4), shown to be effective in inhibiting the growth of Mycobacterium avium in macrophages, together with the synthesis and characterization of two unsuccessful analogues selected

Full text of DOI:10.1007/s00775-010-0650-1

J Biol Inorg Chem (2010) 15:861–877
DOI 10.1007/s00775-010-0650-1
ORIGINAL PAPER
Fluorescent 3-hydroxy-4-pyridinone hexadentate iron chelators:
intracellular distribution and the relevance to antimycobacterial
properties
•
Ana Nunes Maria Podinovskaia Andreia Leite
•
•
•
•
•
Paula Gameiro Tao Zhou Yongmin Ma Xiaole Kong
•
•
Ulrich E. Schaible Robert C. Hider Maria Rangel
•
Received: 19 January 2010 / Accepted: 19 March 2010 / Published online: 3 April 2010
Ó SBIC 2010
Abstract We report the synthesis and characterization of
a fluorescent iron chelator (4), shown to be effective in
inhibiting the growth of Mycobacterium avium in macro-
phages, together with the synthesis and characterization of
two unsuccessful analogues selected to facilitate identifi-
cation of the molecular properties responsible for the
antimicrobial activity. Partition of the chelators in lipo-
somes was investigated and the compounds were assessed
with respect to uptake by macrophages, responsiveness to
iron overload/iron deprivation and intracellular distribution
by flow cytometry and confocal microscopy. The synthesis
of the hexadentate chelators is based on a tetrahedral
structure to which three bidentate 3-hydroxy-4-pyridinone
chelating units are linked via amide bonds. The structure is
synthetically versatile, allowing further addition of func-
tional groups such as fluorophores. Here, we analyse the
non-functionalized hexadentate unit (3) and the corre-
sponding rhodamine B (4) and fluorescein (5) labelled
chelators. The iron(III) stability constant was determined
for 3 and the values log b = 34.4 and pFe3? = 29.8
indicate an affinity for iron of the same order of magnitude
as that of mycobacteria siderophores. Fluorescence prop-
erties in the presence of liposomes show that 4 strongly
interacts with the lipid phase, whereas 5 does not. Such
different behaviour may explain their distinct intracellular
localization as revealed by confocal microscopy. The flow
cytometry and confocal microscopy studies indicate that 4
is readily engulfed by macrophages and targeted to cytosol
and vesicles of the endolysosomal continuum, whereas 5 is
differentially distributed and only partially colocalizes with
4 after prolonged incubation. Differential distribution of
the compounds is likely to account for their different effi-
cacy against mycobacteria.
A. Nunes and M. Podinovskaia contributed equally to the manuscript.
Electronic supplementary material The online version of this
article (doi:10.1007/s00775-010-0650-1) contains supplementary
material, which is available to authorized users.
A. Nunes ꢀ A. Leite ꢀ P. Gameiro
´
REQUIMTE, Departamento de Quımica, Faculdade de Ciencias,
Universidade do Porto, 4069-007 Porto, Portugal
ˆ
M. Podinovskaia ꢀ U. E. Schaible
Department of Infectious and Tropical Diseases,
London School of Hygiene and Tropical Medicine,
Keppel Street, London WC1E 7HT, UK
T. Zhou ꢀ Y. Ma ꢀ X. Kong ꢀ R. C. Hider
Division of Pharmaceutical Sciences, King’s College London,
Franklin-Wilkins Building, London SE1 9NH, UK
T. Zhou
College of Food and Biotechnology,
Zhejiang Gongshang University,
Hangzhou 310035, People’s Republic of China
Keywords Fluorescent hexadentate iron(III) chelators ꢀ
Synthesis ꢀ Affinity constants ꢀ Partition in liposomes ꢀ
Confocal microscopy
U. E. Schaible
Department of Molecular Infection Biology,
Research Center Borstel, 23845 Borstel, Germany
Introduction
M. Rangel (&)
ˆ
´
REQUIMTE, Instituto de Ciencias Biomedicas de Abel Salazar,
Universidade do Porto, 4099-003 Porto, Portugal
e-mail: mcrangel@fc.up.pt
Iron is an essential micronutrient for pathogenic bacteria as
well as for the host organisms, thereby causing a competitive
123

862
J Biol Inorg Chem (2010) 15:861–877
relationship between host and pathogen [1–5]. To acquire
iron from their environment, microbes have developed
high-affinity iron-scavenging molecules, known as sidero-
phores, which are hexadentate chelators that strongly com-
plex iron and efficiently deliver the element inside the
pathogen. Host defence strategies depend on iron for the
generation of reactive oxygen and nitrogen intermediates.
Iron distribution in the host is tightly controlled and iron
sequestration by binding proteins such as lactoferrin and
inflammation-triggered iron redistribution, as controlled by
hepcidin, contribute to antimicrobial host defence. Conse-
quently, pharmacological intervention by iron chelators has
been considered as a therapy against bacterial infections
[6–10]. The design of iron chelators capable of competing
for iron with bacterial siderophores and which possess the
required features to target infection sites will undoubtedly
contribute to the implementation of novel therapeutic
strategies.
In the particular case of mycobacteria, it is well known
that these pathogens synthesize two types of siderophores:
(1) mycobactins, which remain cell-wall-associated, and
(2) carboxymycobactins and exochelins, which are released
to the extracellular medium [15, 16]. These molecules have
chelating units based on the phenyloxazolidine ring, orni-
thine-derived hydroxamates and salycilates. In the case of
Mycobacterium avium, mycobactins and carboxymyco-
bactins possess identical iron(III) binding sites and differ
only in side chain characteristics, such that a lipophilic
character is conferred on mycobactins and a hydrophilic
character on carboxymycobactins [2, 4]. With this knowl-
edge, we designed chelators based on 3-hydroxy-4-pyridi-
none ligands, which are known for their use in the
treatment of iron-overloaded patients and which exhibit
higher affinity for iron when compared with hydroxamates.
The results from a systematic investigation centred on 15
iron chelators indicate that the 3-hydroxy-4-pyridinone
hexadentate chelator (4), Fig. 1, is effective at inhibiting
the growth of M. avium, an opportunistic pathogen able to
survive in macrophages [11]. The data demonstrate that the
antimicrobial activity is associated with the chelating
properties of 4 and its ability to penetrate cell membranes
to access the bacteria. Closely related compounds, with an
identical affinity for iron, although exhibiting a positive
response on the inhibition of bacterial growth in axenic
media, were found not to be effective in the M. avium
model [11].
Fig. 1 Structure of hexadentate chelator 4 and its iron(III) complex
[12]. Thus, the complete understanding of the iron acqui-
sition pathways of bacteria will facilitate the identification
of the optimal physicochemical properties for effective
antimicrobial drugs. Several studies on structure and
transport properties of amphiphilic siderophores such as
acinetoferrin indicate that their hydrophilic/lipophilic
properties are critical for their membrane permeation. Also,
it has been proposed that the hydrophobic components of
amphiphilic siderophores facilitate flip–flop diffusion
through lipid membranes [13–16]. Mycobacteria scavenge
extracellular iron from host proteins using hydrophilic
siderophores such as carboxymycobactins [3, 5] and
extraction of intracellular iron from macrophages is
achieved by lipophilic mycobactins, the resulting myco-
bactin–iron complex accumulating in macrophage lipid
droplets. It is thus clear that iron acquisition represents a
realistic target for the control of mycobacterial infection
[17].
In the present work we report the synthesis and prop-
erties of the rhodamine B derived chelator (4) and its
fluorescein analogue (5). In an attempt to begin to under-
stand the mode of action of 4, we have investigated the
distribution of the fluorescent chelators in both liposomes
and macrophages.
Compound 4, which results from the coupling of a
hexadentate chelating unit and rhodamine B, appears to
possess key molecular features that allow penetration
through cell membranes. Rhodamine B conjugated pep-
tides have previously been shown to have antimicrobial
activity, the process being related to their surface activities
mainly associated with the lipophilic rhodamine B residue
Materials and methods
For all the reactions, chemicals were obtained from Sigma–
Aldrich (grade puriss, p.a.) and were used as received
unless otherwise specified.
123

J Biol Inorg Chem (2010) 15:861–877
863
tripod possesses a free nitro group which can be reduced
with freshly prepared Raney nickel to produce the corre-
sponding primary amine (2). The starting anchor was
acylated with acetyl chloride to produce di-tert-butyl 4-
acetamido-4-(3-tert-butoxy-3-oxopropyl)heptanedioate (6).
The three protecting tert-butyl groups of the resulting
compound were removed using formic acid, producing 4-
acetamido-4-(2-carboxyethyl)heptanedioic acid (7). Com-
pound 1 was linked to the anchor structure by coupling the
free amino group of the pyridinone to the acid groups of the
anchor to give 8, followed by deprotection with BCl3 to
yield the hexadentate chelator 3.
Di-tert-butyl 4-acetamido-4-(3-tert-butoxy-3-oxopropyl)
heptanedioate
To a stirred solution of amine 2 (12 mmol) and triethyl-
amine (12 mmol) in dichloromethane (40 mL), cooled with
an ice-bath, acryl chloride (10 mmol) was added. Stirring
was continued overnight. The reaction mixture was washed
successively with aqueous hydrochloric acid, aqueous
sodium hydrogen carbonate and water, dried over Na2SO4
and concentrated. The residue was subjected to column
chromatography on silica gel (EtOAc/hexane) to give 6
(71%) as a white solid. m.p. 118–120 °C, 1H NMR
(360 MHz, CDCl3) d: 1.44 (s, CH3, 27H), 1.92 (s, CH3,
3H), 1.97 (m, CH2, 6H), 2.23 (m, CH2, 6H), 5.92 (s, NH,
1H). 13C NMR (90 MHz, dimethyl-d6 sulfoxide, DMSO-
d6) d: 24.3 (CH3CO), 28.1 (CH3), 29.8 (CCH2CH2), 30.0
(CCH2CH2), 57.4 (NHC), 80.7 (OCMe3), 169.6 (CH3CO),
172.9 (COOBu-t). Electrospray ionization (ESI) mass
spectrometry (MS): m/z 458.17 ([M?H]?).
Scheme 1 Formulae and numbering of hexadentate chelators and
their precursors. Compounds 3, 4 and 5 were designated as CP256,
CP777 and CP852 in the study regarding the evaluation of their
antimicrobial properties [11]
Synthesis of hexadentate chelators
The synthesis of the 3-hydroxy-4-pyridinone hexadentate
chelators is based on the coupling of three bidentate 3-
hydroxy-4-pyridinone moieties (1) to a tripodal anchor (2)
as outlined in Scheme 1.
Tripodal hydroxypyridinone
4-Acetamido-4-(2-carboxyethyl)heptanedioic acid
The synthesis of tripodal hydroxypyridinone (3) is outlined
in Scheme 2. The 2-aminomethyl-3-benzyloxy-1,6-dime-
thyl-1H-pyridin-4-one (1) was synthesized as previously
described [18, 19]. The anchor molecule (2) was synthe-
sized by a method described by Newkome et al. [20]
starting from nitromethane and tert-butyl acrylate. The
A solution of tert-butyl ester 6 (10 mmol) in formic acid
(96%, 10 mL) was stirred at 25 °C overnight. After con-
centration, toluene (5 mL) was added and the solution was
again evaporated in a vacuum to remove azeotropically any
residual formic acid. The residue was then precipitated in
R'O
Scheme 2 Synthesis of 3.
HOBt N-hydroxybenzotriazole,
DCCI dicyclohexylcarbodiimide,
DMF dimethylformamide
O
(Z)
O
COOR
COOR
N
H
O
COOBu-t
COOBu-t
N
(Z)
HN
O
HN
O
c
a
H2N
OR'
(Z)
N
H
O
O
COOR
COOBu-t
N
HN
N
(Z)
(Z)
OR'
6 R = Bu-t
7 R = H
8 R' = Bn
3 R' = H
b
d
(Z)
O
(a) acetyl chloride, Et3N, CH2Cl2; (b) HCOOH; (c) HOBt, DCCI, 1,6-dimethyl-2-aminomethyl-3-
benzyloxy-pyridin-4(1H)-one, DMF; (d) BCl3, CH2Cl2.
123

864
J Biol Inorg Chem (2010) 15:861–877
Hexadentate chelator 3 hydrochloride
acetone (50 mL) to afford 4-acetamido-4-(2-carboxy-
ethyl)heptanedioic acid (7; 95%) as a yellow solid, which
could be used for the next step without further purification.
De-benzylation of 8 was achieved by hydrogenation at
30 psi H2 in the presence of 5% Pd/C (10% w/w of 8) in
ethanol for 6 h. The catalyst was removed by filtration and
the combined filtrates were acidified to pH 1 with con-
centrated hydrochloric acid. After the solvents had been
removed to dryness, the residues were purified by recrys-
tallization from methanol/acetone to give hexadentate
chelator 3 hydrochloride (92%) as a white power. 1H NMR
(360 MHz, DMSO-d6) d: 1.76 (s, CH3, 3H), 1.79 (m, CH2,
6H), 2.08 (m, CH2, 6H), 2.57 (s, CH3, 9H), 3.88 (s, CH3,
9H), 4.56 (d, J = 5.0 Hz, 6H), 7.28 (s, NH, 1H), 7.29 (s,
pyridinone C5-H, 3H), 8.93 (t, J = 5.0 Hz, NH, 3H). 13C
NMR (360 MHz, DMSO-d6) d: 21.0 (CH3), 23.8 (CH3CO),
29.6 (CCH2CH2), 30.3 (CCH2CH2), 35.1 (NHCH2), 39.5
(NCH3), 57.0 (NHC), 113.1 (C-5H in pyridinone), 140.3
(C-2 in pyridinone), 143.0 (C-3 in pyridinone), 148.9 (C-6
in pyridinone), 159.9 (C-4 in pyridinone), 169.3 (CH3CO),
173.8 (CONH). MS: calculated for C36H50N7O10: 740.3618
(monoisotopic molecular weight M?H), found: matrix-
assisted laser desorption/ionization time of flight MS:
740.4 [M?H]? (a-cyano-4-hydroxycinnamic acid as the
matrix); high-resolution MS–ESI: 740.3621[M?H]?. Ele-
mental analysis for (C36H48N7O14ꢀ3HClꢀ6H2O). Calculated
(found): C 47.55 (46.93); H 6.49 (6.56); N 10.49 (10.64).
1
m.p. 124–126 °C, H NMR (360 MHz, DMSO-d6) d: 1.78
(s, CH3, 3H), 1.81 (m, CH2, 6H), 2.10 (m, CH2, 6H), 7.19
(s, NH, 1H). 13C NMR (90 MHz, DMSO-d6) d: 23.8 (CH3),
28.4 (CCH2CH2), 29.3 (CCH2CH2), 56.7 (NHC), 169.5
(CH3CO), 174.8 (COOH). ESI–MS: m/z 290.12
([M?H]?).
Protected hexadentate ligand
A mixture of triacid 7 (0.867 g, 3 mmol), 1,6-dimethyl-2-
aminomethyl-3-benzyloxy-pyridin-4(1H)-one (2.79 g, 10.8
mmol),
dicyclohexylcarbodiimide
(DCCI;
2.22 g,
10.8 mmol), N-hydroxybenzotriazole (HOBt; 1.46 g,
10.8 mmol) and dimethylformamide (DMF; 30 mL) was
stirred for 2 days. After filtration and removal of solvent
under reduced pressure, the residue was chromatographed
in a silica gel column using methanol/chloroform (1:4) as
an eluent to afford protected hexadentate ligand 8 (2.23 g,
1
73% yield) as a white solid. H NMR (360 MHz, DMSO-
d6) d: 1.74 (s, CH3, 3H), 1.79 (m, CH2, 6H), 2.03 (m, CH2,
6H), 2.25 (s, CH3, 9H), 3.39 (s, CH3, 9H), 4.33 (d,
J = 4.8 Hz, 6H), 5.06 (s, CH2, 6H), 6.18 (s, pyridinone
C5-H, 3H), 7.19 (s, NH, 1H), 7.35 (m, ArH, 9H), 7.41 (m,
ArH, 6H), 8.09 (t, J = 4.8 Hz, NH, 3H). 13C NMR
(90 MHz, DMSO-d6) d: 20.4 (CH3), 23.9 (CH3CO), 29.7
(CCH2CH2), 30.3 (CCH2CH2), 34.8 (NHCH2), 36.0
(NCH3), 57.0 (NHC), 72.5 (PhCH2O), 117.8 (C-5H in
pyridinone), 128.2 (CH in Ar), 128.6 (CH in Ar), 128.8
(CH in Ar), 138.0 (C in Ar), 140.6 (C-2 in pyridinone),
146.0 (C-3 in pyridinone), 148.1 (C-6 in pyridinone), 169.2
(CH3CO), 172.4 (CO), 172.5 (C-4 in pyridinone). MS:
calculated for C57H68N7O10: 1,010.5027 (monoisotopic
molecular weight M?H), found: ESI–MS: 1,010.5
[M?H]?, 505.8 [M?2H]2?; high-resolution MS–ESI:
1,010.5020 [M?H]?.
Rhodamine-labelled tripodal hydroxypyridinone
The synthesis of the rhodamine-labelled hexadentate
compound is outlined in Scheme 3. The starting molecule 9
was obtained by hydrazinolysis of the amino group of the
corresponding phthalimido tripodal precursor [21]. Com-
pound 9 was coupled with rhodamine B isothiocyanate in
the presence of anhydrous DMF and triethylamine to pro-
duce 10, followed by deprotection with BCl3 to yield
rhodamine-labelled tripodal hydroxypyridinone (4).
Scheme 3 Synthesis of 4
123

J Biol Inorg Chem (2010) 15:861–877
865
COOBu-t
COOBu-t
Scheme 4 Synthesis of 5. NHS
N-hydroxysuccinimide
COOBu-t
COOBu-t
O
HN
O
a
HN
O
N
H2N
COOBu-t
COOBu-t
O
11
12
R'O
O
(Z)
O
COOR
COOR
N
H
O
HO
O(E)
O
N
(Z)
HN
HN
d
b
(Z)
(Z)
OR'
(Z)
HN
Fl
O
HN
Fl
O
N
H
COOH
O
COOR
O
Fl=
N
HN
N
(Z)
(Z)
O
13 R = Bu-t
14 R = H
OR'
c
(Z)
O
15 R = Bn
5 R = H
e
(a) NH2NH2, EtOH; (b) HOBt, DCCI, 5(6)-carboxy-fluorescein (c) HCOOH, (d) DCCI, NHS, 1,6-
dimethyl-2-aminomethyl-3-benzyloxy-pyridin-4(1H)-one, DMF; (e) BCl3, CH2Cl2.
Compound 10
2.00 (d, 6H, CH2); 2.27 (d, 6H, CH2); 2.64 (s, 9H, CH3);
3,32 (d, 2H, CH2); 3.68 (q, 8H, CH2); 3.99 (s, 9H, CH3);
To a solution of 100 mg of rhodamine B isothiocyanate
(1.87 mmol) in anhydrous DMF(20 mL), 26 lL
(1.87 mmol) of triethylamine and 290 mg (2.8 mmol) of 9
were added and the mixture was stirred at room tempera-
ture overnight. The product was purified by gradient col-
umn chromatography, eluting with methanol/chloroform
3.99 (t, 2H, CH2); 4.64 (t, 6H, CH2N); 6.86 (s, 3H, H5);
6.97 (m, 4H, Ar); 7.06 (d, 2H, Ar); 7.25 (d, 1H, Ar); 7.26
(d, 2H, Ar). 13C NMR (90 MHz, DMSO-d6) d: 21.0 (CH3),
23.8 CH3CO); 29.6 (CCH2CH2), 30.3 (CCH2CH2), 35.1
(NHCH2), 39.5 (NCH3), 57.0 (NHC), 113.1 (C-5 in py-
ridinone), 140.3 (C-2 in pyridinone), 143.0 (C-3 in pyrid-
inone), 148.9 (C-6 in pyridinone), 159.9 (C-4 in
pyridinone), 169.3 (CH3CO), 173.8 (CONH). MS: calcu-
lated for C66H83N11O13 MS: 1269.6: (monoisotopic
molecular weight M?H), found: matrix-assisted laser
desorption/ionization time of flight MS: 1269.6 [M?H]?.
Elemental analysis for (C66H83N11O13Sꢀ4HClꢀ6H2O). Cal-
culated (found): C 50.12 (50.46); H 8.56 (8.63); N 5.74
(5.82).
1
(1:1) and ammonia to afford 10 (4%) as a purple solid. H
NMR (360 MHz, CDCl3) d: 1.26 (t, 12H, CH3) 1.96 (d,
6H, CH2); 2.17 (d, 6H, CH2); 2.31 (s, 9H, CH3); 2.58 (s,
2H, CH2); 3.32 (q, 8H, CH2); (s, 9H, CH3); 3.67 (s, 9H,
CH3); 3.38 (t, 2H, CH2); 4.40 (t, 6H, CH2N); 5.12 (d, 6H,
CH2Ar); 6.47 (s, 3H, H5) 6.87–6.96 (m, 4H, Ar); 7.17 (d,
2H, Ar); 7.3–7.4 (m, 5H, Ar); 7.57 (d, 1H, Ar); 7.93 (d, 2H,
Ar).
Compound 4
Fluorescein-labelled tripodal hydroxypyridinone
Compound 10 (23 mg, 0.027 mmol) was dissolved in
anhydrous dichloromethane (50 mL), under argon and
cooled to 0 °C. BCl3 1 mL (0.0241 mmol) was added
dropwise and the reaction mixture was kept overnight with
stirring at room temperature. Methanol (50 mL) was added
and the mixture was stirred for 1 h. The solid product that
formed was removed by filtration and the solvent was
removed under reduced pressure to afford the crude prod-
uct. Recrystallization of the crude product from methanol/
acetone (4:6) afforded 4 (36%) as a crystalline purple
solid.1H NMR (360 MHz. DMSO-d6) d: 1.32 (t, 12H, CH3)
The synthesis of fluorescein-labelled tripodal hydroxypy-
ridinone (5) is outlined in Scheme 4. The phthalimido tri-
podal ester (11) was prepared as described by Zhou et al.
[21]. Deprotection of the amino group by hydrazinolysis
yielded the amine (12), which was coupled with 5(6)-car-
boxyfluorescein in the presence of DCCI/HOBt to produce
the triester (13). Compound 13 was hydrolysed by formic
acid to generate the triacid 14, which was coupled with 2-
aminomethyl-3-benzyloxy-1,6-dimethyl-1H-pyridin-4-one
to yield 15, followed by deprotection with BCl3 to yield the
hexadentate chelator 5.
123

866
J Biol Inorg Chem (2010) 15:861–877
Compound 12
concentrated under reduced pressure, The residue obtained
was purified by column chromatography (methanol/chloro-
form, 1:9), affording 15 (54%) as a yellow solid. m.p. 195–
216 °C. 1H NMR (DMSO-d6) d: 1.79 (m, 6H), 2.02 (m, 6H),
2.24 (s, 9H), 2.38 (m, 2H), 3.38 and 3.39 (2 s, 9H), 3.45 (m,
2H), 4.32 (s, 6H), 5.06 (s, 6H), 6.18 (s, 3H), 6.42–6.60 (m,
6H), 7.23–8.85 (m, 21H). m/z 1,397 [(M?1)]?.
To a solution of 11 (3 mmol) in ethanol (40 mL) was added
5.5% aqueous hydrazine (3 mL). After it had been refluxed
for 3 h, the reaction mixture was cooled to 0 °C, acidified to
pH 1 with concentrated hydrochloric acid and filtered. The
filtrate was concentrated in a vacuum, and the residue was
dissolved in distilled water (40 mL), adjusted to pH 12 with
10 mol NaOH and extracted with chloroform (59 100 mL).
The combined organic extracts were dried over anhydrous
Na2SO4 under reduced pressure to furnish 12 (85%) as a
pale-brown oil. 1H NMR (CDCl3) d: 1.43 (s, 27H), 1.94–1.99
(m, 6H), 2.19–2.24 (m, 6H), 2.29 (t, J = 5.9 Hz, 2H), 2.45
(br s, 2H), 3.00 (t, J = 5.9 Hz, 2H), 7.34 (br s, 1H).
Compound 5
A solution of 15 (3 mmol) in dichloromethane (50 mL) was
flushed with nitrogen. After the flask had been cooled to
0 °C, BCl3 (1 mol in dichloromethane, 20 mL) was added
dropwise and the reaction mixture was stirred at room tem-
perature for 1 day. The excess BCl3 was eliminated at the
end of the reaction by the addition of methanol (10 mL) and
the mixture was stirred for another 10 min. After removal of
the solvents under reduced pressure, the residue was purified
by recrystallization from methanol/acetone to afford 5 (40%)
Compound 13
A
mixture of 12 (5 mmol), 5(6)-carboxyfluorescein
(5 mmol), DCCI (6 mmol) and HOBt (5.5 mmol) in DMF
(50 mL) was stirred at room temperature for 1 day. After
filtration, the solvent was removed under reduced pressure,
and the residue was purified on a silica gel column using
methanol/chloroform (1:9) as an eluent to afford 13 (59%)
as a yellow foam.1H NMR (CDCl3) d: 1.39 and 1.40 (2 s,
27H), 1.88–1.99 (m, 6H), 2.14–2.25 (m, 6H), 2.39 and 2.51
(2 m, J = 7.5 Hz, 2H), 3.59 and 3.70 (2 m, J = 7.5 Hz,
2H), 6.51–8.43 (m, 9H).
1
as a yellow solid. m.p. 198–213 °C. H NMR (DMSO-d6)
1.81 (m, 6H), 2.09 (m, 6H), 2.40 (m, 2H), 2.56 (s, 9H), 3.34
(m, 2H), 3.87 and 3.88 (2 s, 9H), 4.55 (s, 6H), 6.56–8.93 (m,
12H), 10.25 (brs, 1H), 10.81 (brs, 1H). m/z 1,127 (M?1)?.
Elemental analysis for (C58H62N8O16ꢀ3HClꢀH2OꢀCHCl3).
Calculated (found): C 51.58 (51.59); H 4.99 (5.23); N 8.16
(8.31).
Determination of pKa and log b values
Compound 14
All solutions were prepared with double-deionized water
(conductivity less than 0.1 lS cm-1). Solutions of NaOH
were prepared in deionized water previously purged with
argon while boiling to reduce carbonate impurity by dilu-
tion of a saturated stock solution. Hydrochloric acid
(1 mM) solution was prepared by dilution of standard
solution of hydrochloric acid (Tritisol). Hydrochloric acid
(Titrisol) and all other chemicals were purchased from
Merck (pro analysis grade).
A solution of tert-butyl ester 13 (10 mmol) in formic acid
(96%, 10 mL) was stirred at 25 °C overnight. After con-
centration, toluene (5 mL) was added and the solution was
again evaporated in a vacuum to remove azeotropically any
residual formic acid. The residue was then precipitated in
acetone (50 mL) to afford 14 (95%) as a yellow solid,
which could be used for the next step without further
purification. m.p. 189–201°C. 1H NMR (DMSO-d6) d:
1.76–1.85 (m, 6H), 2.03–2.12 (m, 6H), 2.33 and 2.42 (2t,
J = 6.9 and 7.1 Hz, 2H), 3.30–3.50 (m, 2H), 6.53–6.69 (m,
6H), 7.23–8.88 (m, 6H), 10.20 (brs, 1H).
All potentiometric measurements were carried out with
a Crison 2002 pH meter and a Crison 2031 burette con-
trolled by a computer. The electrode assembly consisted of
an Orion 900029 double-junction AgCl/Ag reference
electrode and a Russell SWL07 glass electrode. System
calibration was performed by the Gran method [22] in
terms of hydrogen ion concentration, by titrating solutions
of strong acid with strong base. A calibration was per-
formed before each run. All titrations were carried out
under an argon atmosphere in a thermostat-controlled
double-walled glass cell; the temperature was maintained
at 25.0 0.1 °C, and the ionic strength was adjusted to
0.10 M with sodium chloride.
Compound 15
To a solution of 14 (10 mmol) in dry DMF (100 mL), DCCI
(12 mmol) and N-hydroxysuccinimide (12 mmol) were
added. The mixture was stirred for 2 h before 2-amino-
methyl-3-benzyloxy-1,6-dimethyl-1H-pyrindin-4-one
(10 mmol) was added, and the reaction mixture was stirred at
room temperature overnight. The precipitate was removed
by filtration, and the solvent was evaporated. The residue was
dissolved in dichloromethane and washed three times with
0.1 N NaOH and brine, dried over anhydrous Na2SO4 and
The dissociation constants of 3 were obtained by titra-
tion of 20 mL of an acidified aqueous solution of the
123

J Biol Inorg Chem (2010) 15:861–877
867
compound (1 mM; 1.5 mM HCl), with approximately
0.03 mol NaOH, under an argon atmosphere. The values of
the equilibrium constants defined by Eq. 1 were refined by
least-squares calculation using the computer program Hy-
perquad [23].
metal ions were acquired [Fe(NO3)3, Ca(NO3)2, Cu(NO3)2,
Mg(NO3)2, Zn(NO3)2, Mn(NO3)2 and Ni(NO3)2] from
Sigma–Aldrich. A saturated solution of 4 in DMSO was
prepared and used as a stock solution. To prepare the
solution for fluorescence measurements, a known volume
of the stock solution was diluted with MOPS buffer to
achieve a final concentration of 4 of 3 9 10-6 M. The
percentage of the DMSO stock solution was always less
than 1% in the final volume. The probe solution was then
mixed with increasing amounts of the metal stock solution,
in a range of molar ratios from 1:0.2 to 1:2 of 4 to metal
ion. All measurements were performed using a Varian Cary
Elipse spectrofluorimeter.
Spectrophotometric measurements
Absorption spectra were recorded with a UNICAM UV-
300 spectrophotometer equipped with a constant-tempera-
ture cell holder. Spectra were recorded at 25.0 °C in 1-cm
quartz cuvettes with a slit width of 2 nm in the range 230–
400 nm. Typical solution concentrations were 0.1 mol
NaCl, 1 mmol ligand, 1 mmol N,N0-bis(2-hydroxyben-
zyl)ethylenediamine-N,N0-diacetic acid) (HBED), and
1 mmol Fe3?, the pH being fixed at a given value for each
run. The pH values were set by addition of trace amounts
of strong acid or base to a volume of 15 mL of solution and
the solutions were allowed to equilibrate for 24 h. In each
experiment, the UV/vis spectra of the iron complexes as a
function of pH were obtained from a single solution. A
spectrum was recorded after each pH adjustment. The
values of the stability constants are the average values of
three experiments. The overall metal-complex stability
constant was determined by competition studies with the
hexadentate ligand HBED by spectrophotometric methods
and calculated using the program pH AB 2006 [23]. Dis-
tribution diagrams were plotted using the program Hyss
2006 [24].
The values of the fluorescence quantum yield of the
fluorescent chelators 4 and 5 and rhodamine B were
determined according to the method of Williams et al. [25]
using fluorescein as the standard. To minimize reabsorption
effects, the values of the absorbance in the fluorescence
cuvette were maintained below 0.1 at and above the exci-
tation wavelength.
Liposome preparation
Liposomes were prepared by evaporation to dryness with a
stream of argon of a lipid solution in chloroform (dimyr-
istoylphosphatidylcholine; DMPC) or in chloroform/meth-
anol (1:1) (dimyristoylphosphatidylglycerol; DMPG). The
lipids DMPC and DMPG were purchased from Avanti Polar
Lipids (Alabaster, AL, USA). The films were maintained
under a vacuum, for a minimum of 3 h to remove all traces
of the organic solvent. The resulting dried lipid films were
dispersed with 10 mM MOPS buffer (0.1 M NaCl, pH 7.4),
and the mixture was vortexed above the phase-transition
temperature (37 0.1°C) to produce multilamellar lipo-
somes. The multilamellar liposomes were then subjected to
the following cycle five times: freeze vesicles in liquid
nitrogen and thaw in a water bath at 37 0.1 °C. Lipid
suspensions were equilibrated at 37 0.1 °C for 30 min
and extruded ten times through polycarbonate filters
(100 nm) to produce large unilamellar vesicles. Extrusion
of liposomes was performed with a Lipex Biomembranes
(Vancouver, Canada) extruder attached to a circulating
water bath. The size distribution of extruded DMPC
liposomes, with and without added compound, was
determined by dynamic light scattering analysis using a
Malvern Instruments Zetasizer nano ZS. Lipid concen-
trations in vesicle suspensions were determined by
phosphate analysis, using a modified version of the Fiske
and Subbarow [26] method. Chelator solutions were
added to liposome suspensions and vortexed for 5 min
followed by incubation at 37 °C for 30 min. Two sets of
nine samples (1.5 mL) were typically used in each
experiment.
Fluorescence measurements
Fluorescence measurements were performed with a Varian
Cary Elipse spectrofluorometer equipped with a Varian
Peltier single cell holder, in 1-cm cuvettes. All the spectra
were recorded at 37 0.1 °C, with excitation and emis-
sion slit widths of 0.05 m and with an exposure time of
4,000 s, in the range from 500 to 700 nm for emission and
at 496 nm for excitation for 5 and from 565 to 750 nm for
emission and at 561 nm for excitation for 4.
Stock solutions of the different compounds were
obtained by preparing a saturated solution of the compound
in dimethyl sulfoxide (DMSO). Samples for fluorescence
measurements were prepared by dilution of a known vol-
ume of the DMSO stock solution in 10 mM 3-(N-mor-
pholino)propanesulfonic acid (MOPS) buffer solution. The
percentage of the DMSO stock solution was always less
than 1% in the final volume.
Fluorescence quenching
Fluorescence quenching measurements were performed in
MOPS buffer at 37 °C. Stock solutions of the different
123

868
J Biol Inorg Chem (2010) 15:861–877
Partition constants
resuspended in cold PBS. Measurements were carried out
using an FACSCalibur flow cytometer (BD Biosciences,
Oxford, UK). Analysis proceeded via Cell-Quest and
FlowJo software programs. Gates were based on dot plots
of untreated cell populations. The median fluorescence of
at least 10,000 events was recorded and corrected for cell
autofluorescence. Mean values were calculated for medians
of three independent experiments.
Kp values for the liposome distribution of fluorophores
were determined using the equation DI = (DImax[lipid]/
(1/Kpc) ? [lipid]), in which I is the fluorescence intensity
of the fluorophore in the presence of lipid vesicles and I0 is
the fluorescence intensity in its absence, DI = I - I0,
DImax = I? - I0 is the maximum difference in fluores-
cence intensity, I? represents the limiting value of I, and c
is the molar volume of the lipid [27].
Confocal microscopy
Values of clog P were calculated using the Molinspi-
ration software suite (version 2.2) (http://www.
molinspiration.com).
For live cell imaging of intracellular probe distribution,
BMMØ were plated onto sterile cover slips in six-well
plates at 105 cells per cover slip and left to settle overnight.
For lysosomal labelling, fluorescein-conjugated or Texas
Red conjugated dextran (Invitrogen) was added at 100 lg/
mL for 4 h, and subsequently chased overnight with cell
medium. For phagosomal labelling, cells were incubated
with fluorescein-conjugated zymosan (Invitrogen) or with
DynabeadsÒ M-280 Tosylactivated (Invitrogen) for
20 min. Cells were washed with PBS and incubated for
20 min with 10 lmol 4 or 100 lmol 5 in PBS, simulta-
neously with fluorescein-conjugated or Texas Red conju-
gated dextran for endosomal labelling, or with
MitoTrackerÒ Green FM or MitoTrackerÒ Deep Red 633
(Invitrogen) for mitochondria, or with Cy3-conjugated
transferrin (transferrin-Cy3) (Invitrogen) for early and
recycling endosomes. The cells were washed with cold
PBS. For surface labelling, cells were incubated in cold
block buffer for 20 min, followed by 20-min incubation
with Cy5-conjugated anti-mouse CD11b (Abcam, Cam-
bridge, UK, 1:200, in block buffer) with a subsequent wash
with PBS.
Cell culture and reagents
Bone-marrow-derived macrophages (BMMØ) were iso-
lated from the femurs and tibias of 6–8-week-old C57BL/6
mice (Harlan Bioproducts, Bicester, UK) using standard-
ized procedures. Cells were plated in Dulbecco’s modified
Eagle’s medium (Invitrogen, Paisley, UK) supplemented
with 10% heat-inactivated fetal calf serum (Sigma–
Aldrich, UK), 5% horse serum (Invitrogen), 1% L-gluta-
mine (Invitrogen) and 20% L929 cell culture supernatant as
a source of macrophage colony-stimulating factor. Cells
were cultivated at 37 °C and 7.5% CO2. Mature adherent
macrophages were harvested after 7 days of culture and
stored at -80 °C in 10% DMSO to be replated when
required. Thawed cells were cultured overnight and non-
adherent cells were washed off prior to manipulations. No
antibiotics were used in the cell culture medium. Phos-
phate-buffered saline (PBS) (pH 7.4, Invitrogen) and block
buffer (1% bovine serum albumin, 5% heat-inactivated
horse serum, 0.05% sodium azide in PBS, pH7.4) were
treated with Chelex-100 (sodium form, Sigma–Aldrich,
Gillingham, UK) to remove iron contamination.
Fluorescence was visualized using an inverted Zeiss
confocal fluorescence microscope and a 963 oil immersion
objective. Images were acquired by Zeiss LSM Image
Browser.
Flow cytometry
For flow cytometry measurements, non-adherent cells were
washed off from cultivated cell monolayers, which were
subsequently incubated for 20 min in cold PBS. The cells
were scraped off the plates, washed with PBS and split into
flow cytometry tubes at 105 cells per tube. For cellular iron
loading and deprivation, cells were incubated for 50 min at
37 °C and 7.5% CO2 in 500 lg/mL iron(III) ammonium
citrate (Sigma–Aldrich) and in 100 lmol deferoxamine
mesylate (DFO; Sigma–Aldrich), deferasirox (DFR; Exj-
adeÒ) or deferiprone (DFP; FerriproxÒ), respectively.
Following washing with cold PBS, the cells were incubated
with 4, provided as a 10 mM stock solution in DMSO and
diluted in PBS to a final concentration of 10 lM, for
20 min at 37 °C and 7.5% CO2. The cells were washed and
Results
Design of iron chelators
The hexadentate chelators reported in the present work
were designed to be evaluated as iron deprivation agents to
target infection processes induced by Mycobacteria. Also,
when designing the potential antimicrobial agents, we
considered the possibility of introducing into the chelator
framework fluorescent labels that could allow its pathways
in the cell to be followed by fluorescence microscopy
a priority. To synthesize the hexadentate chelators we
linked three bidentate 3-hydroxy-4-pyridinone moieties (1)
to a selected tripodal anchor based on a tetrahedral carbon
123

J Biol Inorg Chem (2010) 15:861–877
869
100
80
60
40
20
0
atom (2) (Scheme 1). The use of this particular tripodal
molecule permits the addition of other functional groups
such as fluorophores. Compound 2 had previously been
conjugated via amide bonds to produce iron-binding den-
drimers [21] and polymers [28], and in the present work an
analogous procedure was adopted to prepare iron chelators
3, 4 and 5 (Scheme 1). The hexadentate chelating unit,
common to 3, 4 and 5, is built from 3-hydroxy-4-pyridi-
none molecules which provide for iron(III) an octahedral
coordination sphere, as shown in Fig. 1 for 4. The choice of
ligand (1) that binds to the anchor (2) via the ortho position
was made to ensure that the resulting hexadentate molecule
unit possesses a high affinity for iron(III) [29, 30]. Com-
pound 3 is essentially the hexadentate chelating unit which
has been blocked from further substitution to facilitate the
physicochemical characterization of the hexadentate unit.
Compound 4 is a fluorescent hexadentate chelator prepared
by coupling the chelating unit with rhodamine B, a fluo-
rophore which has a lipophilic character and contrasts with
green-labelled bacteria used in the infection model studies.
Compound 5 is an analogous fluorescent hexadentate
chelator, which contains the fluorescein core.
H6L3+
-
H3L
L3
HL2
-
H5L2+
H4L+
-
H2L
2
6
10
pH
Fig. 2 Speciation plot for hexadentate chelator 3, as a function of pH
hexadentate chelators which have been prepared using
different molecular frameworks [32]. The neutral species
H3L dominates over the pH range 4–9 and its molar
fraction is 96% at pH 7.4 (Fig. 2) [24].
The high affinity of this ligand for iron(III) is confirmed
by the UV/vis spectra of solutions of the ligand in the
presence of iron(III) which show a characteristic ligand to
metal charge transfer band of the iron(III) complex even at
low pH values (Fig. S1). The overall affinity constant of 3
for iron was determined by competition studies against the
ligand HBED using UV/vis spectroscopy as previously
described [33–36]. The log b(Fe3?) and pFe3? values
(calculated as -log[Fe3?] at pH 7.4 with a total ligand
concentration of 10-5 M and a total Fe(III) concentration
of 10-6 M [37]) obtained for 3 were 34.4 0.1 and 29.8,
respectively. Different values for log b(Fe3?) and pFe3?
have been previously reported [21], and were based on
erroneously estimated pKa values. The values of log b and
pFe3? of 3 are compared with those of a range of sidero-
phores in Table 1. It is clear that chelator 3 and related
compounds are good candidates to deprive bacteria of iron.
Indeed the very strong capacity to chelate iron is a major
determinant for the antibacterial activity of 4 since it has
been demonstrated that the iron(III) complex is an inef-
fective antimycobacterial agent [11].
Speciation and affinity for iron
Information relating to the predominant ligand species at
physiological pH values and the affinity constant for iron
are essential for a clear comprehension of the distribution
of iron chelating molecules. Owing to interference arising
from the fluorophores in both potentiometric and spectro-
photometric measurements, we isolated the hexadentate
chelating unit 3, common to all the chelators, to determine
the dissociation constants of the ligand and its affinity
constant for iron(III). Compound 3 was obtained as a
hydrochloride which in strong acidic solution can be rep-
resented as H6L3? and which has six dissociable protons
that are displaced according to Eq. 1 for 6 B n B 1.
HnLꢁ3þn $ Hnꢁ1Lꢁ4þn þ Hþ
ÀÂ
Ã
Á Â
Ã
ð1Þ
Kn ¼ Hnꢁ1Lꢁ4þn ½Hþꢂ = HnLꢁ3þn
The acid–base properties of 3 were investigated by both
potentiometric and spectrophotometric methods. Fitting
analysis of the titration curves was achieved using the
program Hyperquad [25] and led to the determination of
six pKa values: pKa1 = 2.44 0.01; pKa2 = 3.08 0.04;
Fluorescence properties
The fluorescence properties of 4 and 5 were studied in
aqueous solution over the pH range 5–9 (Table 2). The
values of the fluorescence quantum yields (U) determined
for chelators 4 and 5 are lower than those of the corre-
spondent fluorophores. For 4 we investigated the influence
of Fe3?, Zn2?, Cu2?, Ni2?, Mn2?, Ca2? and Mg2? on its
fluorescence intensity. The results obtained in 3 lM
pKa4 = 3.78 0.07;
pKa4 = 8.57 0.09;
pKa5 =
9.21 0.11; pKa6 = 9.80 0.18. These were found to
be consistent for the two methods. The values obtained can
be grouped into two sets, centred at the values 3 and 9,
which are the typical pKa values found for bidentate
3-hydroxy-4-pyridinones [31]. The values are in close
agreement with those of other 3-hydroxy-4-pyridinone
123

870
J Biol Inorg Chem (2010) 15:861–877
Table 1 Affinity constants and pFe3? values for chelator 3, deferoxamine and siderophores produced by mycobacteria
Siderophore
Ligand architecture
log b pFe3? Reference
3
Tripodal, 3-hydroxy-4-pyridinone
34.40 29.84 This work
a
Mycobactin
Linear, bishydroxamate, N-(oxazoline) and OH(salycilic acid) 31
29.0
–
Cell-wall-associated mycobacteria siderophores
Carboxymycobactin
a
–
Linear, bishydroxamate, N-(oxazoline) and OH(salycilic acid) 31
29.0
Extracellular mycobacteria siderophores
Exochelin MN
Linear, bishydroxamate, threo-b-hydroxylhistidine
Linear, trishydroxamate
39.10 31.10 [44]
28.90 25.00 [35]
31.00 26.60 [45]
Mycobacterium neoaurum siderophore
Exochelin MS
Mycobacterium smegmatis siderophore
Deferoxamine
Linear, trishydroxamate
Actinobacter Streptomyces pilosus siderophore
a
Calculated with the software suite Marvin & Marvin version 5.1.4 [43]
solutions of chelator 4 together with a 1:1 molar ratio of
metal ion are shown in Fig. 3. The values are the mean values
obtained in three independent experiments and the associ-
ated standard deviation is 0.1%. From analysis of the figure it
is clear that Ca2? and Mg2? do not influence the fluorescence
intensity, whereas Fe3?, Zn2?, Cu2?, Ni2? and Mn2? induce
fluorescence quenching. Although fluorescence quenching
of 4 in the presence of Cu2? and Zn2? is appreciable at
equivalent concentrations, under physiological conditions
the cytoplasmic levels of low molecular weight complexes of
Cu2? and Zn2? are much lower than those of Fe3?, these
values typically being less than 10-15 M Cu2? [38], less
than 10-9 M Zn2? [39] and approximately 10-6 M Fe3?
[40]. Thus, in living cells, we believe that any observed
contribution of metal ion chelation to the fluorescence
quenching of 4 is likely to be associated with the presence of
chelatable iron. Evidence that formation of the correspond-
ing metal complexes is a significant contribution to fluores-
cence quenching is provided by experiments performed with
different concentrations of the metal ion. For iron(III), the
maximal fluorescence quenching is achieved at a 1:1 molar
ratio (Fig. S2). A profile of fluorescence quenching in the
presence of different concentrations of the biologically
relevant metal ions Fe3?, Zn2? and Cu2? is depicted in
Fig. S3. The quenching profiles evidence the different
behaviour of the paramagnetic metal ions Fe3? and Cu2? and
diamagnetic Zn2? concerning the contribution to fluores-
cence quenching.
Table 2 Spectroscopic data of fluorescent compounds
UV/vis
Fluorescence
kmax (nm)
e (M-1 cm-1
)
kmax (nm)
U
4
561
496
543
498
2.80 9 104
6.34 9 104
1.06 9 105
9.23 9 104
582
520
550
425
0.027
0.171
0.310
0.950
5
Rhodamine B
Fluorescein
Fig. 3 Influence of Fe3?, Zn2?, Cu2?, Ni2?, Mn2?, Ca2? and Mg2?
on the fluorescence intensity of chelator 4. The results were obtained
in 10 mM 3-(N-morpholino)propanesulfonic acid solution (pH 7.4) of
3 lM chelator 4 together with a 1:1 molar ratio of different metal ions
more hydrophobic than the fluorescein-labelled chelator
(5), the clog P values being 1.731 and 0.042, respectively.
In an attempt to investigate this difference in greater
detail, partition experiments were undertaken in liposome
suspensions, using DMPC and DMPG. For both the
rhodamine-labelled chelator (4) and rhodamine B, the
fluorescence intensity was observed to be significantly
Partition studies
Conventional partition coefficients were not directly
determined because of the difficulty of the measurement of
the pKa values of both 4 and 5, thus preventing reliable
conversion of log D7.4 values to log P values. The clog P
values revealed that the rhodamine-labelled chelator (4) is
123

J Biol Inorg Chem (2010) 15:861–877
871
250
100
80
60
40
20
0
200
150
100
DMPC
50
2.0x10-4 4.0x10-4 6.0x10-4 8.0x10-4 1.0x10-3 1.2x10-3
0
0.0
600
650
700
750
λ(nm)
450
400
350
300
250
200
150
100
50
300
250
200
150
100
50
0
DMPG
0.0
2.0x10-4 4.0x10-4 6.0x10-4 8.0x10-4 1.0x10-3 1.2x10-3
0
600
650
700
750
λ(nm)
Fig. 5 Fitting of fluorescence intensity parameters of chelator 4
according to the equation DI = (DImax[lipid]/(1/Kpc) ? [lipid]), in
which I is the fluorescence intensity of the fluorophore in the presence
of lipid vesicles and I0 is the fluorescence intensity in its absence,
Fig. 4 Fluorescence intensity observed for 2.85 lM solutions of
chelator 4 in the presence of suspensions of dimyristoylphosphati-
dylcholine (DMPC) and dimyristoylphosphatidylglycerol (DMPG)
large unilamellar vesicles prepared with lipid concentrations indicated
by the following colour code: black 0.0 lM, red 134 lM, green
267 lM, blue 401 lM, cyan 536 lM, magenta 670 lM, yellow
803 lM, dark yellow 933 lM, navy 1,080 lM
DI
= I - I0, DImax = I? - I0 is the maximum difference in
fluorescence intensity, I? represents the limiting value of I, and c is
the molar volume of the lipid [27]. Log KP values for 4: DMPC, 3.71;
DMPG, 3.79
preparations of 4 are presented in Fig. 5. The log KP values
in DMPC and DMPG were determined as 3.71 and 3.79.
These values are similar to the corresponding values
obtained for rhodamine B, namely 3.76 and 3.84, which
confirms that 4 and rhodamine B interact strongly with both
zwitterionic and negatively charged lipid membranes. Rho-
damine B has a marginally greater degree of partition in the
membrane phase when compared with the partition of 4,
indicating that it is this portion of the molecule which enjoys
the favourable interaction with the lipid surface.
enhanced in the presence of large unilamelar vesicles
(Fig. 4). The fluorescence intensity of rhodamine B is known
to be perturbed by the presence of other molecules able to
interact with the xanthene chromophore. In a study of the
fluorescence properties of rhodamine B in the presence of
cyclodextrins, it was shown that the fluorescence intensity
can be either quenched or enhanced depending on which of
two distinct binding sites is preferred by cyclodextrins [41].
The observation of a significant enhancement of the fluo-
rescence intensity of chelator 4 is thus assigned to its inter-
action with the lipid phase provided by liposomes. No such
enhancement was observed with either the fluorescein-
labelled chelator (5) or fluorescein itself, presumably
because they did not interact with the liposome preparations.
This finding is in agreement with the two estimated log P
values. The fluorescence data obtained for liposome
Macrophage studies
Intracellular distribution of fluorescent labelled chelators
The intracellular distribution of the fluorescent chelators
was analysed in BMMØ from C57BL/6 mice. These
123

872
J Biol Inorg Chem (2010) 15:861–877
Fig. 6 Representative confocal microscopy images of intracellular
distribution and colocalization studies of chelators 4 and 5 in bone-
marrow-derived macrophages (BMMØ). Cells were incubated with
individual solutions of 4 or 5 or with combined chelators in iron-free
phosphate-buffered saline (PBS). From titration experiments where
cells were incubated with increasing chelator concentrations, compa-
rable fluorescence intensity was observed for 4 (10 lM) and 5
(100 lM). Cells were incubated simultaneously with 4 (10 lM) and 5
(100 lM) for 20 min, and the intracellular chelator distribution was
monitored immediately (a) and upon a 20-min chase with PBS (b). To
characterize the access of chelators 4 and 5 to phagosomes,
colocalization studies were performed using Dynabeads as phago-
some markers. Cells were incubated with DynabeadsÒ M-280
Tosylactivated (Invitrogen) for 20 min, washed with PBS and
incubated for 20 min with 10 lM 4 or 100 lM 5 in PBS (c).
Fluorescence of stained cells was measured in the green as well as in
the red channel, and data were merged for evaluation of colocaliza-
tion in yellow. Fluorescence data are shown in comparison with the
corresponding bright-field images. Magnification of merged fluores-
cence data show detailed colocalization
adherent cells were incubated with individual solutions of 4
or 5 or with combined chelators in iron-free PBS. From
titration experiments where cells were incubated with
increasing chelator concentrations, comparable fluores-
cence intensity was observed for 4 (10 lM) and 5
(100 lM). Considering the quantum yields determined for
the two compounds in buffer solution at pH 7.4 (Table 2),
this result was not to be expected.
Cells were pulsed simultaneously with 4 (10 lM) and 5
(100 lM), and the intracellular chelator distribution was
monitored immediately (Fig. 6a) and upon a 20-min chase
with PBS (Fig. 6b). After 20 min of incubation, 4 was
123

J Biol Inorg Chem (2010) 15:861–877
873
Fig. 7 Representative confocal
microscopy images of
intracellular colocalization
studies of chelator 4 and
phagosomal (a), mitochondria
(b), endosomal (c) and
lysosomal markers (d, e) in
BMMØ. For phagosomal,
mitochondrial, and endosomal
labelling, cells were incubated
for 20 min with fluorescein-
labelled zymosan (zymosan-
FITC), MitoTracker Green FM
and fluorescein-labelled dextran
(dextran-FITC), respectively.
For lysosomal labelling,
dextran-FITC (Invitrogen) was
added at 100 lg/mL for 4 h, and
subsequently chased overnight
with cell medium. Cells were
washed with PBS and incubated
with chelator 4 (10 lM) for
20 min. Fluorescence of stained
cells was measured in the green
as well as in the red channel,
and data were merged for
evaluation of colocalization in
yellow. Fluorescence data are
shown in comparison with the
corresponding bright-field
images
readily visualized within the macrophages, whereas 5 was
found at the periphery of the cells, revealing the distinctly
different intracellular distribution of both chelators early
after uptake. Only after a chase of an additional 20 min did
both chelators partially colocalize within the cells, indi-
cating differential pathways and/or kinetics of internaliza-
tion for both chelators.
to bead-containing phagosomes, 4 also accessed fluores-
cein-conjugated zymosan phagosomes (Fig. 7a). Colocal-
ization of 4 with the zymosan surface suggested chelator
accumulation in the yeast cell wall fragments (zymosan),
which was confirmed by incubation of free zymosan with 4
strongly labelling the zymosan particles.
Colocalization patterns acquired after 20-min incubation
of macrophages with chelators and simultaneously with
MitoTracker Green FM or MitoTracker Deep Red 633
revealed that neither chelator accesses mitochondria
(Figs. 7b, 8a). Simultaneous incubation of macrophages
with 4 and the endosomal tracer fluorescein-labelled dex-
tran showed that 4 accessed endosomes and subsequently
lysosomes (Fig. 6c vs. d, e). In contrast, only a small
proportion of 5 colocalized with Texas Red conjugated
dextran labelled endosomes and lysosomes (Fig. 8b–d).
Compound 5 accumulated predominantly in vesicles at the
To characterize the intracellular compartments targeted
by the chelators, colocalization studies were performed
using (1) fluorescein-labelled zymosan and Dynabeads as
phagosome markers, (2) fluorescein-conjugated and Texas
Red conjugated dextran for endosomes and lysosomes and
(3) MitoTracker Green FM and MitoTracker Deep Red 633
as mitochondrial markers. Both 4 and 5 have access to
bead-containing phagosomes, following 20-min incubation
with the chelators (Fig. 6c), indicating at least partial
intersection with the endosome/phagosome system. Similar
123

874
J Biol Inorg Chem (2010) 15:861–877
Fig. 8 Representative confocal
microscopy images of
intracellular colocalization
studies of chelator 5 and
mitochondria (a), endosomes
(b) and lysosomes (c, d) in
BMMØ).The procedure is as
described in the legend to Fig. 6
but used a concentration of
chelator 5 of 100 lM and
MitoTracker Deep Red 633 and
Texas Red conjugated
dextran.For the imaging of early
and recycling endosomes, cells
were incubated with Cy3-
conjugated transferrin (e). For
surface labelling, cells were
incubated in cold block buffer
for 20 min, followed by 20-min
incubation with Cy5-conjugated
anti-mouse CD11b (f).
Fluorescence of stained cells
was measured in the green as
well as in the red channel, and
data were merged for evaluation
of colocalization in yellow.
Fluorescence data are shown in
comparison with the
corresponding bright-field
images
cell surface, whereas 4 spread throughout the endosomal/
lysosomal system of the cells.
probe is further chased into lysosomes, whereas the
majority of 5 remains in vesicles or membrane moieties at
the cell surface.
The cellular distribution, kinetics of uptake and colo-
calization patterns of the fluorescein-labelled hexadentate
chelator 5 are similar to those previously observed for a
fluorescein-labelled bidentate pyridinone probe [42]. Both
compounds appear to remain localized at the cell surface
after 20-min incubation; however, the bidentate pyridinone
The intracellular localization of 5 was further studied
using the early/recycling endosome tracer transferrin-Cy3
and the surface marker CD11b (Fig. 8e, f). After a 20-min
incubation, 5 did not colocalize with early endosomes
(Fig. 8e) but colocalize partially with the marker CD11b
123

J Biol Inorg Chem (2010) 15:861–877
875
Fig. 9 a Representative flow cytometry histograms of chelator 4 in
BMMØ. Cells were incubated with 4, at a final concentration of
10 lmol, for 20 min at 37 °C and 7.5% CO2. The cells were washed
and resuspended in cold PBS. For cellular iron loading, cells were
incubated for 50 min at 37 °C and 7.5% CO2 in 500 lg/mL iron(III)
ammonium citrate following washing with cold PBS. Cells were
incubated with 4, at a final concentration of 10 lmol, for 20 min at
37 °C and 7.5% CO2. The cells were washed and resuspended in cold
PBS. We used the designations ‘‘unstained cells’’ for macrophages,
‘‘no treatment’’ for macrophages incubated with 4 and ‘‘FeCit’’ for
macrophages treated with iron(III) citrate and incubated with 4. Data
were acquired in the red channel. Gates were based on dot plots of
untreated cell populations. b Responsiveness of intracellular chelator
4 in iron-overloaded and iron-deprived BMMØ. For cellular iron
loading and deprivation, cells were incubated for 50 min at 37 °C and
7.5% CO2 in 500 lg/mL iron(III) ammonium citrate and in 100 lmol
deferoxamine mesylate (DFO), deferasirox (DFR) and deferiprone
(DFP), respectively, following washing with cold PBS. Cells were
incubated with 4, at a final concentration of 10 lmol, for 20 min at
37 °C and 7.5% CO2. The cells were washed and resuspended in cold
PBS. Data were acquired in the red channel. Gates were based on dot
plots of untreated cell populations
(Fig. 8f), suggesting association with membrane moieties
distinct from endocytic entry sites.
This observation confirms that the chelator has access to
citrate-delivered iron and cellular compartments suscepti-
ble to fluctuations in iron levels.
Intracellular iron sensitivity of 4 under iron overload
and deprivation
Iron deprivation by the chelators DFR and DFP caused a
significant decrease in fluorescence intensity of 4, whereas
treatment with DFO resulted in a less significant fluores-
cence decrease. To explain the differential responsiveness
of 4, both thermodynamic and kinetic arguments have to be
considered. According to the pFe3? values (29.8 for 4, 26.6
for DFO, 22.5 for DFR, 19.5 for DFP), chelator 4 has the
highest affinity for iron, followed by DFO, DFR and DFP.
Thermodynamically, chelator 4 is expected to displace
iron(III) from the three complexes. However, from the
kinetic point of view it is known that complexes formed by
tridentate and bidentate ligands are much more labile than
those formed by hexadentate chelators. This tendency may
explain the fact that in the same period of time of
The intracellular fluorescence of 4 was analysed in vitro by
flow cytometry using BMMØ from C57BL/6 mice. All
flow cytometry analyses were carried out in Chelex-treated
PBS to avoid external iron contamination. As shown in
Fig. 9a, 4 is efficiently endocytosed by macrophages,
leading to an increase in intracellular fluorescence when
compared with unlabelled cells. Chelator sensitivity was
investigated under conditions of iron overload by adding
iron citrate or iron deprivation by DFO, DFR and DFP. Iron
overload induced a significant decrease of fluorescence
intensity when compared with untreated cells (Fig. 9a, b).
123

876
J Biol Inorg Chem (2010) 15:861–877
Acknowledgments Financial support from FCT-Fundac¸a˜o para a
ˆ
Ciencia e Tecnologia, Portugal, through projects POCI/QUI/56214/
2004 and PTDC/QUI/67915/2006 is gratefully acknowledged.
incubation with chelator 4 the decrease in fluorescence is
much more evident for DFR and DFP. The flow cytometry
results indicate that, within the intracellular environment,
chelator 4 is able to displace iron(III) from other complex
forms.
References
Discussion
1. Marx JJ (2002) Best Pract Res Clin Haematol 15:411–426
2. Ratledge C (2004) Tuberculosis 84:110–130
3. Ratledge C, Dover LG (2000) Annu Rev Microbiol 54:881–941
4. De Voss JJ, Rutter K, Schroeder BG, Barry CE (1999) J Bacteriol
181:4443–4451
5. Monfeli RR, Beeson C (2007) Infect Disord Drug Targ 7:213–
220
6. Vergne AF, Waolz AJ, Miller MJ (2000) Nat Prod Rep 17:99–116
7. Xu Y, Miller MJ (1998) J Org Chem 63:4314–4322
8. Roosenberg JM II, Lin YM, Lu Y, Miller MJ (2000) Curr Med
Chem 7:159–197
9. Cronje L, Bornman L (2005) Int J Tuberc Lung Dis 9:2–9
10. Cronje L, Edmondson N, Eisenach KD, Bornman L (2005) FEMS
Immunol Med Microbiol 45:103–112
11. Fernandes SS, Nunes A, Gomes AR, Castro B, Hider RC, Rangel
M, Appelberg R, Gomes MS (2010) Microbes Infect 12:287–294
12. Bucki R, Pastore JJ, Randhawa P, Vegners R, Weiner DJ, Janmey
PA (2004) Antimicrob Agents Chemother 48:1526–1533
13. Luo M, Fadeev EA, Groves JT (2005) J Am Chem Soc
127:1726–1736
14. Fadeev EA, Luo M, Groves JT (2004) J Am Chem Soc
126:12065–12075
15. Xu G, Martinez JS, Groves JT, Butler A (2002) J Am Chem Soc
124:13408–13415
The speciation plot of 3 demonstrates that the hexadentate
moiety is neutral over the pH range 4–9 (Fig. 2), thus
indicating that the net charge of chelator probes 4 and 5 is
dominated by the respective chromophores, chromophore 4
being zwitterionic and 5 being negatively charged. Probes
4 and 5 contain the hexadentate chelating unit of 3 and thus
have similar affinity for iron(III). The latter is of the same
magnitude as that exhibited by mycobacterium sidero-
phores (Table 1). Both 4 and 5 are selective for iron(III)
under biological conditions.
As demonstrated by the values of clog P and the lipo-
some partition studies, probe 4 is strongly lipophilic in
contrast to 5. Thus, from consideration of the physico-
chemical characterization of 4 and 5, neither of the com-
pounds would be predicted to readily cross membranes by
non-facilitated diffusion but 4 is predicted to be strongly
surface active.
In macrophages, the intracellular distribution of 4 and 5
was markedly different. Probe 4 became widely distributed
in the cell within 20 min, accumulating in phagosomes,
endosomes, and finally predominantly in lysosomes but not
in mitochondria. Probe 5 localized to vesicles and mem-
brane moieties proximal to the plasma membrane but not to
phagosomes, endosomes, lysosomes nor mitochondria. It is
likely that probe 4 remains largely membrane bound when
distributing throughout the cell endocytic pathway. How-
ever, the chelating portion of the molecule is predicted to
be exposed to the lumen of the organelles to allow iron
scavenging from the aqueous phase. This property was
confirmed by a cell sorting study, where it was demon-
strated that within macrophages probe 4 was able to
remove iron from both DFP and DFR. Competition with
DFO was much less effective. However, this was to be
expected as competition between hexadentate chelators is
remarkably slow even at pH 5.5.
16. Pahl PMB, Yan XD, Hodges YK, Rosenthal EA, Horwitz MA,
Horwitz LD (2000) J Biol Chem 275:17821–17826
17. Luo M, Fadeev EA, Groves JT (2005) Nat Chem Biol 1:149–153
18. Tilbrook GS, Hider RC (1998) In: Sigel A, Sigel H (eds) Metal
ions in biological systems, vol 35. Marcel Dekker, Inc., New
York, pp 691–730
19. Piyamongkol S, Liu ZD, Hider RC (2001) Tetrahedron 57:3479–
3486
20. Newkome GR, Behera RK, Moorefield CN, Baker GR (1991) J
Org Chem 56:7162–7167
21. Zhou T, Neubert H, Liu DY, Liu ZD, Ma Y, Kong X, Luo W,
Mark S, Hider RC (2006) J Med Chem 49:4171–4182
22. Gran G (1952) Analyst 77:661–671
23. Gans P, Sabatini A, Vacca A (1996) Talanta 43:1739–1753
24. Alderighi L, Gans P, Ienco A (1999) Coord Chem Rev 184:311–
318
25. Williams ATR, Winfield SA, Miller JN (1983) Analyst
108:1067–1071
26. Fiske CH, Subbarow Y (1925) J Biol Chem 66:375–400
27. Santos NC, Prieto M, Castanho MARB (2003) BBA Biomembr
1612:123–135
28. Zhou T, Kong X, Liu DY, Liu ZD, Hider RC (2008) Biomac-
romolecules 9:1372–1380
29. Liu ZD, Kayyali R, Hider RC, Porter JB, Theobald AE (2002) J
Med Chem 45:631–639
30. Liu ZD, Hider RC (2002) Coord Chem Rev 232:151–171
31. Burgess J, Rangel M (2008) Adv Inorg Chem 60:167–243
Thus, it is likely that probe 4 will be able to successfully
compete with carboxymycobactins and mycobactins for
iron, thereby offering a mechanism for the observed anti-
mycobacterial effect. Once iron is bound to probe 4, natural
siderophores would be predicted to experience consider-
able difficulty in scavenging this iron. Competitive inter-
ference with mycobacterial siderophores by exogenous iron
chelators offers an effective antimicrobial strategy with
potential to combat infections, including tuberculosis.
ˇ
´
32. Grazina R, Gano L, Sebestık J, Santos MA (2009) J Inorg Bio-
chem 103:262–273
33. Mies KA, Gebhardt P, Mollmann U, Crumbliss AL (2008) J Inorg
Biochem 102:850–861
34. Dhungana S, Harrington JM, Gebhardt P, Mllmann U, Crumbliss
AL (2007) Inorg Chem 46:8362–8371
123

J Biol Inorg Chem (2010) 15:861–877
877
35. Dhungana S, Ratledge C, Crumbliss AL (2004) Inorg Chem
43:6274–6283
41. Liu Y, Chen Y, Liu SX, Guan XD, Wada T, Inoue Y (2001) Org
Lett 3:1657–1660
36. Dhungana S, Heggemann S, Gebhardt P, Mllmann U, Crumbliss
AL (2003) Inorg Chem 42:42–50
42. Fakih S, Podinovskaia M, Kong X, Collins HL, Schaible UE,
Hider RC (2008) J Med Chem 51:4539–4552
37. Raymond KN, Muller G, Matzanke BF (1984) Top Curr Chem
123:49–102
38. Rae TD, Schmidt PJ, Pufahl RA, Culotta VC, O’Halloran TV
(1999) Science 284:805–808
43. ChemAxon (2007) Marvin & Marvin version 5.1.4. ChemAxon,
Budapest. http://www.chemaxon.com/products/marvin/
44. Dhungana S, Miller MJ, Dong L, Ratledge C, Crumbliss AL
(2003) J Am Chem Soc 125:7654–7663
39. Maret W (2009) Biometals 22:149–157
40. Glickstein H, Ben El R, Shvartsman M, Cabantchik ZI (2005)
Blood 106:3242–3250
45. Schwarzenbach G, Schwarzenbach K (1963) Helv Chim Acta
46:1390–1400
123