Fluorescently-labeled ferrichrome analogs as probes for receptor-mediated, microbial iron uptake

Article abstract of DOI:10.1021/ja9610646

Biomimetic analogs 1 of the microbial siderophore (iron carrier) ferrichrome were labeled with a fluorescent marker at a site which does not interfere with iron binding or receptor recognition to provide iron(III) carriers 5 and 10. These carriers were built from a tetrahedral carbon as an anchor which was symmetrically extended by three converging iron-binding chains and a single, exogenous anthracenyl residue. Carriers 5 varied in the nature of the amino acids (G = glycyl, A = alanyl and L = leucyl) linking the anchor with the iron-binding hydroxamate groups, while the alanyl derivative 10A differed from 5A in the spacer between the anthracenyl label and the anchor. Examination of these binders by ¹H NMR confirmed that their conformations were analogous to those of the nonlabeled parent compounds. Titration experiments using UV/vis and fluorescence spectroscopy demonstrated the quenching of these compounds' fluorescence upon iron(III) loading and its recovery upon iron(III)'s release to a competing chelator. The quenching process fits the Perrin model for static quenching and was more efficient in derivatives 5, where the label could approach the iron-binding domain, than in derivative 10A, where the label's approach was prohibited. These data are in compliance with an intramolecular quenching process. In vivo examination of the labeled derivatives 5 with Pseudomonas putida as the indicator organism established that their behavior parallels that of the nonlabeled analogs 1, with the added benefit of signaling microbial activity by fluorescence emission. Thus incubation of P. putida with iron(III)-loaded (and therefore nonfluorescent) 5G caused buildup of the label's fluorescence in the culture medium. These observations provide direct evidence for a shuttle-mechanism of iron delivery where the fluorescent, iron-unloaded carrier is released to the medium. Inhibition of both phenomena by natural ferrichrome or NaN₃ demonstrates the involvement of the microbial ferrichrome receptor and transport systems, respectively. On the other hand, 5A induced only modest iron(III) uptake by P. putida and failed to generate fluorescence in the culture medium, concurring with its action as an inhibitor. The fact that two strains of different Pseudomonas species did not respond to the ferrichrome analog 5G illustrates the specificity of these compounds. The performance of these carriers as structural and functional probes, paired with their high species specificity, encourage their consideration as diagnostic tools for the detection and identification of pathogenic bacteria and fungi.

Full text of DOI:10.1021/ja9610646

12368
J. Am. Chem. Soc. 1996, 118, 12368-12375
Fluorescently-Labeled Ferrichrome Analogs as Probes for
Receptor-Mediated, Microbial Iron Uptake
Haim Weizman,† Orly Ardon,‡ Brenda Mester,† Jacqueline Libman,† Oren Dwir,†
Yitzhak Hadar,*,‡ Yona Chen,‡ and Abraham Shanzer*,†
Contribution from the Department of Organic Chemistry, The Weizmann Institute of Science, and
Department of Plant Pathology and Microbiology, The Hebrew UniVersity of Jerusalem, Faculty
of Agriculture, RehoVot, 76100, Israel
ReceiVed April 1, 1996X
Abstract: Biomimetic analogs 1 of the microbial siderophore (iron carrier) ferrichrome were labeled with a fluorescent
marker at a site which does not interfere with iron binding or receptor recognition to provide iron(III) carriers 5 and
10 (Figure 1). These carriers were built from a tetrahedral carbon as an anchor which was symmetrically extended
by three converging iron-binding chains and a single, exogenous anthracenyl residue. Carriers 5 varied in the nature
of the amino acids (G ) glycyl, A ) alanyl and L ) leucyl) linking the anchor with the iron-binding hydroxamate
groups, while the alanyl derivative 10A differed from 5A in the spacer between the anthracenyl label and the anchor.
Examination of these binders by 1H NMR confirmed that their conformations were analogous to those of the nonlabeled
parent compounds. Titration experiments using UV/vis and fluorescence spectroscopy demonstrated the quenching
of these compounds’ fluorescence upon iron(III) loading and its recovery upon iron(III)’s release to a competing
chelator. The quenching process fits the Perrin model for static quenching and was more efficient in derivatives 5,
where the label could approach the iron-binding domain, than in derivative 10A, where the label’s approach was
prohibited. These data are in compliance with an intramolecular quenching process. In ViVo examination of the
labeled derivatives 5 with Pseudomonas putida as the indicator organism established that their behavior parallels
that of the nonlabeled analogs 1, with the added benefit of signaling microbial activity by fluorescence emission.
Thus incubation of P. putida with iron(III)-loaded (and therefore nonfluorescent) 5G caused buildup of the label’s
fluorescence in the culture medium. These observations provide direct evidence for a shuttle-mechanism of iron
delivery where the fluorescent, iron-unloaded carrier is released to the medium. Inhibition of both phenomena by
natural ferrichrome or NaN3 demonstrates the involvement of the microbial ferrichrome receptor and transport systems,
respectively. On the other hand, 5A induced only modest iron(III) uptake by P. putida and failed to generate
fluorescence in the culture medium, concurring with its action as an inhibitor. The fact that two strains of different
Pseudomonas species did not respond to the ferrichrome analog 5G illustrates the specificity of these compounds.
The performance of these carriers as structural and functional probes, paired with their high species specificity,
encourage their consideration as diagnostic tools for the detection and identification of pathogenic bacteria and fungi.
Introduction
deliver it to the cell as siderophore-iron complexes via binding
to specific membrane receptors and transport proteins.6-13 This
receptor-regulated process, in contrast to diffusion-controlled
processes, enables the accumulation of iron against unfavorable
concentration gradients. It should be noted that most bacteria
and fungi make use of siderophores, and that most possess
several siderophore receptors and transport systems, even though
they produce only one siderophore. This strategy enables the
organism to compete effectively for the scarcely available iron
by making use of any available siderophore. To date, more
than 200 siderophores have been isolated and characterized, and
these have been found to vary greatly in chemical composition
and topology. As a whole, they can be divided into two classes
that possess either catecholates or hydroxamates as binding sites
Iron plays a crucial role in the growth and development of
living systems: it functions as a catalyst in some of the most
fundamental enzymatic processes, such as oxygen metabolism,
electron transfer, and DNA and RNA synthesis. However,
adequate iron acquisition is not a trivial task due to two major
difficulties. Firstly, most of the iron in the earth’s crust exists
as ferric oxide, which is poorly accessible because of its low
water solubility. Secondly, intracellular iron levels need to be
meticulously controlled: a lack of iron causes severe deficiency
symptoms, whereas excess iron may be equally harmful or even
fatal. To overcome these problems, microorganisms have
developed ingenious iron acquisition strategies based on the use
of molecular vehicles, namely iron(III) carriers or siderophores.1-5
These iron carriers are excreted into the environment, where
they solubilize and scavenge the scarcely available iron and
(6) Bagg, A.; Neilands, J. B. Microbiology ReV. 1987, 51, 509-518.
(7) Emery, T.; Emery, L.; Olsen, R. K. Biochem. Biophys. Res. Commun.
1984, 119, 1191-1197.
(8) Winkelmann, G. FEBS Lett. 1979, 97, 43-46.
(9) Winkelmann, G.; Braun, V. FEBS Microbiol. Lett. 1981, 11, 237-
241.
† The Weizmann Institute of Science.
‡ The Hebrew University of Jerusalem.
X Abstract published in AdVance ACS Abstracts, November 15, 1996.
(1) Winkelmann, G.; van der Helm, D.; Neilands, J. B. Iron Transport
in Microbes, Plants and Animals; VCH Verlagsgesellschaft mbH, D-6940:
Weinheim, Germany, 1987.
(10) Neilands, J. B.; Erickson, T. J.; Rastetter, W. H. J. Biol. Chem.
1981, 256, 3831-3832.
(11) Braun, V. In The Enzymes of Biological Membranes; Martonosi,
A. N., Ed.; Plenum Press: NY, 1985; Vol. 3, pp 617-652.
(12) Rutz, J. M.; Liu, J.; Lyons, J. A.; Goranson, J.; Armstrong, S. K.;
McIntosh, M. A.; Feiz, J. B.; Klebba, P. E. Science 1992, 258, 471-475.
(13) Killmann, H.; Benz, R.; Braun, V. EMBO J. 1993, 12, 30007-
30016.
(2) Neilands, J. B. Struct. Bonding 1984, 58, 1-24.
(3) Matzanke, B. F.; Mu¨ller-Matzanke, G.; Raymond, K. N. Iron Carriers
and Iron Proteins; VCH Publishers: NY, 1989; Vol. 5, pp 1-121.
(4) Hider, R. C. Struct. Bonding 1984, 58, 25-87.
(5) Guerinot, M. L. Annu. ReV. Microbiol. 1994, 48, 743-772.
S0002-7863(96)01064-5 CCC: $12.00 © 1996 American Chemical Society

Fluorescent-Labeled Ferrichrome Analogs
J. Am. Chem. Soc., Vol. 118, No. 49, 1996 12369
and into three topological families that assume tripodal, mac-
rocyclic, or linear shapes.14,15 Regardless of their composition
or topology, they all are hydrophilic and function by the same
principles: they bind iron highly effectively by embedding it
into an octahedral ion-binding cavity and generate a molecular
envelope which fits the respective membrane receptor.
detection of iron27-29 and of other metal ions, including copper
and nickel.30-33 Moreover, studies from our laboratories have
demonstrated the potential of fluorescently labeled natural
siderophores for the investigation of iron exchange processes.34-37
Here we introduce bioactive ferrichrome analogs that are
labeled with anthracene as the fluorescent marker at a site which
interferes neither with iron binding nor receptor recognition.
We describe the design, synthesis, structures, and iron(III)-
binding properties of fluorescently labeled analogs 5 and 10.
We demonstrate that these compounds adopt the same confor-
mations as their parent congeners, both in the free state and as
their iron(III) complexes. We confirm that they convert from
a fluorescent to a nonfluorescent state when binding iron(III)
and that they regain their fluorescence upon iron(III)’s release
to a competing chelator. We further report on the in ViVo
behavior of these analogs 5 (Figure 1). We demonstrate that
the activity of these compounds as iron(III) carriers to P. putida
indeed parallels that of the nonlabeled parent compounds, with
the added benefit of signaling microbial iron(III) uptake via the
appearance of the compound’s fluorescence in the culture
medium. We also illustrate the strong specificity of these
compounds, which failed to recognize two strains of a different
Pseudomonas species.
The importance and ingenuity of these iron-acquisition
processes prompted us to consider all artificial siderophore
analogs as structural probes of microbial iron uptake pro-
cesses.16,17 To provide prototypical structures, we selected as
targets siderophore analogs that span both chemical classes,
catecholates and hydroxamates, and that include two of the three
topologies, tripodal and linear. Applying the principles of
biomimetic chemistry,18 we thus synthesized analogs of entero-
bactin (tripodal catecholate),19 ferrichrome (tripodal hydrox-
amate),20 and ferrioxamine and coprogen (linear hydrox-
amates),21 and examined their behavior both in Vitro and in ViVo.
For the in ViVo studies, Pseudomonas putida was used as the
indicator organism. P. putida possesses the desired range of
siderophore receptors, including those for ferrichrome, ferriox-
amine, and coprogen, whose responses to each of the sidero-
phore analogs can be readily tested and compared with that of
other organisms. These studies led us to identify ferrichrome22,23
(1 in Figure 1) as well as ferrioxamine analogs24,25 that
distinguish between analogous receptors of different organ-
isms,17,22 specifically analogous receptors of P. putida, Arthro-
bacter flaVescens, and Erwinia herbicola. The strikingly high
species specificity of some of the synthetic analogs is in contrast
to the behavior of the natural carriers, which do not distinguish
between related receptors of different organisms. This outstand-
ing feature of synthetic carriers encouraged us to examine their
potential as diagnostic tools for the detection and identification
of specific microorganisms.
Experimental Procedures, Materials and Methods38
Synthesis of Fluorescent Ferrichrome Derivatives 5G, 5A, and
5L. Synthesis of Tetraphenolate C(CH2OCH2CH2COOC6Cl5)4 (2).
Pentaerythritol (9.53 g, 0.07 mol) was treated dropwise with NaOH
(0.7 mL of 30% w/w) and then with acrylonitrile (20.3 mL, 0.44 mol)
such that the temperature did not exceed 30 °C. The mixture was stirred
overnight at room temperature, then neutralized with 1 N HCl, extracted
into EtOAc (200 mL), washed twice with water, dried over Na2SO4,
and concentrated to yield 22.9 g (0.0657 mol, 94% yield) of tetranitrile.
Tetranitrile (7.22 g, 0.021 mol) was treated with concentrated HCl (10
mL), stirred for 4 h at 95 °C, extracted into cold EtOAc (300 mL),
washed twice with water, dried over Na2SO4, and concentrated to yield
6.67 g (15.7 mmol, 75% yield) tetra acid. Tetra acid (11.5 g, 0.027
mol) was treated with SOCl2 (11.9 mL, 0.163 mol) overnight at 40 °C.
Excess thionyl chloride was distilled to provide a crude tetraacyl halide.
Crude tetraacyl chloride was dissolved in dry CHCl3, and pentachlo-
rophenol (28.76 g, 0.108 mol) was added. The mixture was cooled to
0 °C, treated with Et3N (15 mL, 0.108 mol), and then stirred at room
temperature. After the reaction was completed according to IR, the
mixture was concentrated and purified by flash chromatography (silica
gel, CHCl3). Residual pentachlorophenol was removed by filtration
over deactivated neutral alumina to yield 11.94 g (8.42 mmol, 31%
yield) of tetrapentachlorophenolate. IR (CDCl3) ν 1783 cm-1 (COOC6-
Cl5). 1H NMR 80 MHz, CDCl3 δ 3.81 (t, J ) 6 Hz, 8H, CCH2OCH2),
3.46 (s, 8H, CCH2O), 2.91 (t, J ) 6 Hz, 8H, CH2CN).
To this aim we labeled the bioactive carriers with fluorescent
markers, to monitor their microbial activity. The use of
fluorescent markers appeared promising because of the estab-
lished versatility of fluorescently labeled alkali- and alkaline-
earth metal-ion binders for in ViVo studies of ion fluxes,26 and
the recent reports on fluorescently labeled binders for the
(14) van der Helm, D.; Jalal, M. A. F.; Hossain, M. B. In Iron Transport
in Microbes, Plants and Animals; Winkelmann, G., van der Helm, D.,
Neilands, J. B., Eds.; VCH Verlagsgesellschaft mbH, D-69451: Weinheim,
Germany, 1987; pp 135-166.
(15) Jalal, M. A. F.; van der Helm, D. In Handbook of Microbial Iron
Chelates; Winkelmann, G., Ed.; CRC Press: Boca Raton, Ann Arbor,
Boston, London, 1991; pp 235-269.
(16) Shanzer, A.; Libman, J. In Handbook of Microbial Iron Chelates;
Winkelmann, G., Ed.; CRC: Boca Raton, Ann Arbor, Boston, London,
1991; pp 309-338.
Synthesis of 9(Aminomethyl)anthracene (4). A solution of
9-anthracenecarbonitrile (8.13 g, 0.04 mol) and NaBH4 (1.02 g, 0.027
mol) in dry THF (100 mL, distilled over Na) was treated dropwise for
over 1 h with BF3‚Et2O (4.43 mL, 0.036 mol in 10 mL of THF). After
(17) Shanzer, A.; Libman, J; Yakirevitch, P.; Hadar, Y.; Chen, Y.;
Jurkevitch, E. Chirality 1993, 5, 359-365.
(18) Breslow, R. Chem. Soc. ReV. 1972, 1, 553-580.
(19) Tor, Y.; Libman, J.; Shanzer, A.; Felder, C. E.; Lifson, S. J. Am.
Chem. Soc. 1992, 114, 6661-6671.
(20) Dayan, I.; Libman, J.; Agi, Y.; Shanzer, A. Inorg. Chem. 1993, 32,
1467-1475.
(30) Czarnik, A. W. Acc. Chem. Res. 1994, 27, 302-308.
(31) Bissell, R. A.; Silva, A. P. d.; Gunaratne, H. Q. N.; Lynch, P. L.
M.; Maguire, G. E. M.; Sandanayake, K. R. A. S. Chem. Soc. ReV. 1992,
21, 187-195.
(32) Fabbrizzi, L.; Licchelli, M.; Pallavicini, P.; Perotti, A.; Sacchi, D.
Angew. Chem., Int. Ed. Engl. 1994, 33, 1975-1977.
(33) Fabbrizzi, L.; Poggi, A. Chem. Soc. ReV. 1995, 24, 197-202.
(34) Lytton, S. D.; Cabantchik, Z. I.; Libman, J.; Shanzer, A. Mol.
Pharmacol. 1991, 40, 584-590.
(35) Lytton, S. D.; Mester, B.; Libman, J.; Shanzer, A.; Cabantchik, Z.
I. Anal. Biochem. 1992, 205, 326-333.
(36) Bar-ness, E.; Hadar, Y.; Chen, Y.; Shanzer, A.; Libman, J. Plant
Physiol. 1992, 99, 1329-1335.
(37) Loyevsky, M.; Lytton, S. D.; Mester, B.; Libman, J.; Shanzer, A.;
Cabantchik, Z. I. J. Clin. InVestigations 1993, 91, 218-224.
(38) The molar extinction coefficients ꢀ are given in units of 103 cm2
M-1, and the circular dichroism values ∆ꢀ, defined as ꢀL - ꢀR, are given
in units of 103 cm2.
(21) Yakirevitch, P.; Rochel, N.; Albrecht-Gary, A.-M.; Libman, J.;
Shanzer, A. Inorg. Chem. 1993, 32, 1779-1787.
(22) Shanzer, A.; Libman, J; Lazar, R.; Tor, Y.; Emergy, T. Biochem.
Biophys. Res. Commun. 1988, 157, 389-394.
(23) Jurkevitch, E.; Hadar, Y.; Chen, Y.; Libman, J.; Shanzer, A. J.
Bacteriol. 1992, 174, 78-83.
(24) Berner, I.; Yakirevitch, P.; Libman, J.; Shanzer, A.; Winkelmann,
G. Biol. Metals 1991, 4, 186-191.
(25) Jurkevitch, E.; Hadar, Y.; Chen, Y.; Yakirevitch, P.; Libman, J.;
Shanzer, A. Microbiology-UK 1994, 140, 1697-1703.
(26) Tsien, R. Y. Ann. ReV. Neurosci. 1989, 12, 227-253.
(27) Bodenant, B.; Fages, F. Tetrahedron Lett. 1995, 36, 1451-1454.
(28) Barrero, M.; Ca´mara, C.; Pe´rez-Conde, M. C.; Jose´, C. S.; Ferna´ndez,
L. Analyst 1995, 120, 431-435.
(29) Barrero, J. M.; Morenobondi, M. C.; Pe´rez-Conde, M. C.; Ca´mara,
C. Talanta 1993, 40, 1619-1623.

12370 J. Am. Chem. Soc., Vol. 118, No. 49, 1996
Weizman et al.
stirring for an additional hour at room temperature, excess hydride was
destroyed by adding concentrated HCl. The reaction mixture was
concentrated to dryness, and 1 N NaOH (40 mL) was added. The
compound was extracted into CHCl3 (400 mL), concentrated, and
purified by flash chromatography (silica gel, 0-4% MeOH/CHCl3) to
provide 536 mg (2.59 mmol, 7% yield) of product with a mp of 98-
100 °C. 1H NMR 400 MHz, CD3OD δ 8.45 (s, 1H, ArH), 8.38 (d, J
) 8.88 Hz, 2H, ArH), 8.04 (d, J ) 8.4 Hz, 2H, ArH), 7.56 (t, J ) 7.2
Hz, 2H, ArH), 7.48 (t, J ) 7.4 Hz, 2H, ArH), 4.79 (m, 2H, AntCH2).
UV/vis (CHCl3)-λmax ) 334, 350, 368, and 389 nm, ꢀ ) 2906, 5744,
8719, and 783.
9-(chloromethyl)anthracene (5.1 g, 22.5 mmol), and NaHCO3 (3.78 g,
45 mmol) in DMF (20 mL) was stirred overnight at room temperature
under argon. The DMF was removed under high vacuum, and the
residue was dissolved in CHCl3 (200 mL). The solution was washed
with water, dried over Na2SO4, and concentrated. Chromatography
over silica gel (50-80% EtOAc-hexane) yielded 1.69 g of ant-CH2-
SCH2(CH2OH)3 (4.95 mmol, 22% yield). 1H NMR 270 MHz CD3OD
δ 8.43 (s, 1H, ArH), 8.41 (d, J ) 1.3 Hz, 2H, ArH), 8.02 (d, J ) 9.5
Hz, 2H, ArH), 7.54-7.45 (m, 4H, ArH), 4.83 (s, 2H, Ant-CH2S), 3.64
(s, 6H, CH2OH), 2.93 (s, 2H, SCH2C).
Ant-CH2SCH2(CH2OH)3 (1.69 g, 4.95 mmol) was treated with NaOH
(50 mL of 30% w/w) and then added dropwise with acrylonitrile (1.3
mL, 19.8 mmol). The reaction mixture was stirred overnight, dissolved
in CHCl3, and washed with 1 N HCl and water. Concentration and
purification over silica gel (CH2Cl2) yielded 1.6 g of 7 (3.2 mmol, 64.7%
yield). IR (CDCl3) ν 2254 cm-1 (CN), 1118 cm-1 (CH2OCH2). 1H
NMR 270 MHz CDCl3 δ 8.41 (s, 1H, ArH), 8.34 (d, J ) 9.6 Hz, 2H,
ArH), 7.99 (d, J ) 7.9 Hz, 2H, ArH), 7.58 (m, 2H, ArH), 7.47 (m, 2H,
ArH), 4.77 (s, 2H, Ant-CH2S), 3.56 (t, J ) 6 Hz, 6H, CH2OCH2), 3.39
(s, 6H, CH2OCH2), 2.85 (s, 2H, SCH2C), 2.50 (t, J ) 6 Hz, CH2CN).
Ant-CH2SO2CH2(CH2OCH2CH2COOCH3)3 (8). A cold solution
(0 °C) of 7 (721 mg, 1.44 mmol) in dry CH2Cl2 (50 mL) was treated
with 3-chloroperbenzoic acid (680 mg of 80% pure, 3.95 mmol) for
10 min. The reaction mixture was diluted with CH2Cl2, washed with
water and 0.1 N NaOH, dried over Na2SO4, and concentrated.
Purification over silica gel (0-1% MeOH-CH2Cl2) yielded 523 mg
of Ant-CH2SO2CH2(CH2OCH2CH2CN)3 (0.98 mmol, 68% yield). IR
(CDCl3) ν 2254 cm-1 (CN), 1313 cm-1 (SO2), 1114 cm-1 (CH2OCH2),
1018 cm-1 (SO2). 1H NMR 400 MHz CDCl3 δ 8.55 (s, 1H, ArH),
8.35 (d, J ) 8.9 Hz, 2H, ArH), 8.05 (m, 2H, ArH), 7.65 (m, 2H, ArH),
7.52 (m, 2H, ArH), 5.43 (s, 2H, Ant-CH2SO2), 3.60 (s, 6H, CH2OCH2),
3.57 (t, J ) 6 Hz, 6H, CH2OCH2), 3.35 (s, 2H, SO2CH2C), 2.45 (t, J
) 6 Hz, CH2CN).
The trinitrile (608 mg, 1.41 mmol) was dissolved in a saturated
solution of HCl in dry MeOH (3 mL). The solution was heated under
reflux for 2 h and then stirred overnight at room temperature. The
solution was then concentrated, dissolved in CH2Cl2, washed with
saturated NaHCO3 and water, and dried over Na2SO4. Purification over
silica gel (0-1% MeOH-CH2Cl2) yielded 455 mg of 8 (0.72 mmol,
51% yield). IR (CDCl3) ν 1736 cm-1 (COOCH3). 1H NMR 400 MHz
CDCl3 δ 8.54 (s, 1H, ArH), 8.36 (d, J ) 8.2 Hz, 2H, ArH), 8.04 (d, J
) 8.4 Hz, 2H, ArH), 7.62 (m, 2H, ArH), 7.50 (m, 2H, ArH), 5.42 (s,
2H, Ant-CH2SO2), 3.653 (s, 9H, COOCH3), 3.647 (t, J ) 6.3 Hz, 6H,
CH2OCH2), 3.56 (s, 6H, CH2OCH2), 3.39 (s, 2H, SO2CH2C), 2.49 (t,
J ) 6.3 Hz, CH2COO).
Ant-CH2SO2CH2(CH2OCH2CH2COOC6Cl5)3 (9). The triester 8
(455 mg, 0.72 mmol) was dissolved in EtOH (3 mL), and NaOH (3
mL of 12% w/w) was added. The reaction mixture was stirred at room
temperature for 36 h and then EtOH was removed. The solution was
acidified to pH 2 with 10% HCl. A brown precipitate was formed
which was dissolved in CHCl3. Concentration of the solution yielded
347 mg (0.59 mmol, 82% yield) of the triacid. IR (CDCl3) ν 1715
cm-1 (COOH). 1H NMR 400 MHz (CDCl3-5% CD3OD) δ 8.38 (s,
1H, ArH), 8.22 (d, J ) 8.8 Hz, 2H, ArH), 7.90 (d, J ) 8.4 Hz, 2H,
ArH), 7.53 (m, 2H, ArH), 7.42 (m, 2H, ArH), 5.29 (s, 2H, Ant-CH2SO2),
3.56 (t, J ) 5.8 Hz, 6H, CH2OCH2), 3.48 (s, 6H, CH2OCH2), 3.33 (s,
2H, SO2CH2C), 2.43 (t, J ) 5.8 Hz, CH2COOH).
Synthesis of Ferrichrome Analog (5G). C14H9-CH2NHCOCH2-
CH2OCH2C(CH2OCH2CH2CONHCH2CON(OH)CH3)3. Tetra active
ester (2) (1.29 g, 0.91 mmol) and H2NCH2CON(OH)CH3 (3) (592 mg,
2.73 mmol) were dissolved in 100 mL of dry chloroform, and then
9-(aminomethyl)anthracene (4) (186 mg, 0.91 mmol) was added. The
solution was stirred overnight while maintaining pH ∼ 8 with Et3N.
The solution was concentrated and flash-chromatographed over silica
gel (0-15% MeOH-CHCl3). The compound was further purified on
PLC (Preparative Layer Chromatography) (9% MeOH-CHCl3) to give
110 mg (14% yield) of product. IR (CDCl3) ν 1634 cm-1 (CONH,
CONOH) 1H NMR 400 MHz, CD3OD δ 8.49 (s, 1H, ArH), 8.35 (d, J
) 8.6 Hz, 2H, ArH), 8.03 (d, J ) 8.2 Hz, 2H, ArH), 7.55 (m, 2H,
ArH), 7.48 (m, 2H, ArH), 5.38 (s, 2H, AntCH2), 4.10 (s, 2H, NHCH2),
3.59 (m, 2H, CH2O), 3.30 (m, 8H, OCH2C + OCH2CH2), 3.17 (s, 9H,
NCH3), 2.94 (s, 6H, CCH2O), 2.40 (t, J ) 4.9 Hz, 2H, COCH2), 2.32
(t, J ) 5.6 Hz, 6H, CH2CO). FAB MS 925.54 (M + Fe - 2H)+,
894.54 (M + Na)+, 872.53 (M + H)+. Fe(III) complex: UV/vis λmax
426 nm, ꢀ ) 2760.
Synthesis of Ferrichrome Analog (5A). C14H9-CH2NHCOCH2-
CH2OCH2C(CH2OCH2CH2CONHCH(CH3)CON(OH)CH3)3. Tet-
raphenolate C(CH2OCH2CH2COOC6Cl5)4 (2) (1.418 g, 1 mmol),
9-(aminomethyl)anthracene (4) (207 mg, 1 mmol), H2NCH(CH3)CON-
(OH)CH3 (3) (354 mg, 3 mmol), and imidazole (30 mg) in dry CHCl3
(20 mL) were stirred for 2 days at room temperature (maintaining pH
∼ 8 by addition of Et3N). The reaction mixture was concentrated and
purified by flash chromatography (silica gel) (0-6% MeOH-CHCl3)
to yield 183 mg (0.2 mmol, 20% yield) of the desired product. IR
(CDCl3) ν 1632 cm-1 (CONH, CONOH). 1H NMR 400 MHz, CDCl3
δ 10.15 (br, 3H, OH), 8.46 (s, 1H, ArH), 8.31 (d, J ) 8.3 Hz, 2H,
ArH), 8.03 (d, J ) 8.3 Hz, 2H, ArH), 7.55 (m, 2H, ArH), 7.49 (m, 2H,
ArH), 7.25 (NHCRH), 6.93 (m, 1H, CH2NH), 5.43 (d of ABq, J1 ) 28
Hz, J2 ) 14.5 Hz, J3 ) 4.8 Hz, 2H, AntCH2), 5.04 (m, 3H, CRH),
3.60 (m, 2H, CH2O), 3.25 (br, 8H, CH2C + OCH2) 3.19 (s, 9H, NCH3)
2.85 (ABq, J1 ) 29 Hz, J2 ) 8.9 Hz, 6H, CCH2O), 2.49 (m, 2H,
COCH2), 2.29 (m, 3H, CH2CO), 2.17 (m, 3H, CH2CO), 1.32 (d, J )
6.8 Hz, 9H, CRHCH3). FAB MS 914.22 (M + H)+. Fe(III) complex:
UV/vis λmax 425 nm, ꢀ ) 2610. CD λext 368, 422, 453 nm, ∆ꢀ )
-4.5, 0.0, +1.8.
Synthesis of Ferrichrome Analog (5L). C14H9-CH2NHCOCH2-
CH2OCH2C(CH2OCH2CH2CONHCH(C4H9)CON(OH)CH3)3. Tet-
raphenolate C(CH2OCH2CH2COOC6Cl5)4 (709 mg, 0.5 mmol), 9-(ami-
nomethyl)anthracene (103 mg, 0.5 mmol), H2NCH(iBu)CON(OH)CH3
(240 mg, 1.5 mmol), and imidazole (15 mg) in dry CHCl3 (15 mL)
were stirred for 2 days at room temperature (maintaining pH ∼ 8 by
addition of Et3N). The reaction mixture was concentrated and purified
by flash chromatography (silica gel, 0-4% MeOH/CHCl3) to yield 68
mg (0.065 mmol, 13 % yield) of product. IR (CDCl3) ν 1632 cm-1
(CONH,CONOH). 1H NMR 400 MHz, CD3OD δ 8.52 (s, 1H, ArH),
8.37 (d, J ) 8.3 Hz, 2H, ArH), 8.05 (d, J ) 8.4 Hz, 2H, ArH), 7.56
(m, 2H, ArH), 7.49 (m, 2H, ArH), 6.93 (m, 1H, CH2NH), 5.42 (s, 2H,
AntCH2), 5.11 (m, 3H, CRH) 3.60 (t, J ) 5.8 Hz, 2H, CH2CH2O), 3.32
(m, 6H, OCH2CH2), 3.18 (s, 2H, OCH2C), 3.16 (s, 9H, NCH3), 2.98
(s, 6H, CCH2O), 2.46 (m, 2H, COCH2), 2.29 (m, 6H, CH2CO), 1.65
(m, 3H, CH{iBu}), 1.53 (m, 6H, CH2{iBu}) 0.92 (m, 18H, CH3{iBu}).
Fe(III) complex: UV/vis λmax 425 nm, ꢀ ) 2800. CD λext 365, 412,
443 nm, ∆ꢀ ) -5.2, 0.0, +2.5.
A cooled (0 °C) solution of triacid (308 mg, 0.52 mmol), DMAP (5
mg), and C6Cl5OH (416 mg, 1.56 mmol) in dry THF (20 mL) was
treated with dicyclohexyl carbodiimide (321 mg, 1.56 mmol). The
solution was stirred for 1 day at room temperature, and then the
precipitated urea was filtered. The filtrate was concentrated and purified
by chromatography over silica gel (CHCl3) to yield 197 mg of 9 (0.147
mmol, 29% yield). IR (CDCl3) ν 1784 cm-1 (COOC6Cl5). 1H NMR
400 MHz CDCl3 δ 8.50 (s, 1H, ArH), 8.29 (d, J ) 8.3 Hz, 2H, ArH),
8.00 (d, J ) 8.4 Hz, 2H, ArH), 7.57 (m, 2H, ArH), 7.48 (m, 2H, ArH)
5.37 (s, 2H, Ant-CH2SO2), 3.76 (t, J ) 6.0 Hz, 6H, CH2OCH2), 3.62
(s, 6H, CH2OCH2), 3.38 (s, 2H, SO2CH2C), 2.80 (t, J ) 6.0 Hz, CH2-
COOH).
Synthesis of Ferrichrome Analog 10A. Ant-CH2SCH2(CH2-
OCH2CH2CN)3 (7). A solution of thiol (3.43 g, 22.5 mmol),39
Ant-CH2SO2CH2(CH2OCH2CH2CONHCH(CH3)CON(OH)-
CH3)3 (10). Triphenolate 9 (180 mg, 0.135 mmol) and H2NCH(CH3)-
CON(OH)CH3, 3A (64 mg, 0.54 mmol), were dissolved in dry CHCl3
(39) Tunaboylu, K.; Schwarzenbach, G. HelV. Chim. Acta 1971, 54,
2166-2184.

Fluorescent-Labeled Ferrichrome Analogs
J. Am. Chem. Soc., Vol. 118, No. 49, 1996 12371
Scheme 1
(10 mL) and stirred overnight at room temperature (maintaining pH ∼
8 by addition of Et3N). The reaction mixture was concentrated, washed
with 1 N HCl and water, and dried over Na2SO4. Purification by flash
chromatography (0-4% MeOH/CHCl3) yielded 40 mg (0.045 mmol,
33% yield) of product 10. IR (CDCl3) ν 1637 cm-1 (CONH, CONOH).
1H NMR 400 MHz CDCl3 δ 8.53 (s, 1H, ArH), 8.42 (d, J ) 8.9 Hz,
2H, ArH), 8.03 (d, J ) 8.3 Hz, 2H, ArH), 7.61 (m, 2H, ArH), 7.56 (m,
3H, NHCRH), 7.49 (m, 2H, ArH), 5.49 (ABq, J1 ) 48 Hz, J2 ) 14.6
Hz, 2H, AntCH2), 5.14 (m, 3H, CRH), 3.54 (m, 6H, OCH2), 3.43 (ABq,
J1 ) 32 Hz, J2 ) 13.8 Hz, 2H, CH2O), 3.41 (ABq, J1 ) 20 Hz, J2 )
11.0 Hz, 6H, CCH2O), 3.19 (s, 9H, NCH3), 2.49 (m, 3H, CH2CO),
2.25 (m, 3H, CH2CO), 1.38 (d, J ) 6.8 Hz, 9H, CRHCH3). FAB MS
974.41 (M + Fe - H)+, 891.45 (M + H)+. Fe(III) complex: UV/vis
55Fe-siderophore complex was added to a final concentration of 1 µM.
Aliquots (0.5 mL) were taken in duplicate, layered onto a mixture of
dibutyl/octyl phthalate (1/1, v/v Sigma), and centrifuged. The super-
natant was removed, and the Eppendorf tube tips were cut. Radioactiv-
ity in the tips of the tubes containing the bacterial cells was counted
on a Beckman LS1801 counter.
Fluorescence Studies. The Fe-siderophore complex was added to
the bacterial suspension to a final concentration of 5 µM. Aliquots of
1 mL were centrifuged, and the supernatant was collected. Fluorescence
was measured in an SLM Instruments fluorometer (Model 4800).
Results and Discussion
λmax 432 nm, ꢀ ) 2850. CD λ ext 358, 373, 387, 431, 456 nm, ∆ꢀ )
Design and Synthesis. To obtain ferrichrome analogs that
could serve as diagnostic tools, we attached a chemically inert
fluorescent marker to the bioactive glycyl 1G (R ) H) and
alanyl 1A (R ) Me) derivatives (Figure 1), at a site that does
not interfere with receptor binding.22,23 These two congeners
were selected because they had previously been proven to act
as agonist42 and antagonist,42 respectively, towards P. putida.23
The glycyl derivative 1G fully mimicks the natural ferrichrome
as an iron(III) carrier and growth promoter, whereas the alanyl
derivative 1A inhibits the action of the natural iron(III) carrier
by competing for a specific site in the uptake system. We
selected a fluorescent marker as a label because fluorescently
labeled ferrioxamines have been shown to provide powerful
probes for iron(III)-exchange processes.35,43 We specifically
chose anthracene as the label because it combines chemical
inertness with highly intense absorption, high fluorescence yield,
and short singlet lifetime.44 Moreover, the absorption wave-
length of anthracene does not overlap with those of ferric-
hydroxamate45 or the bacterial culture. These spectral charac-
teristics enabled both in Vitro monitoring of iron(III)-binding
and fluorescence quenching and in ViVo follow-up of the
microbial iron(III) uptake process. The fulfillment of these
-4.5, -5.3, -5.1, 0.0, +1.1.
Ligand Titration with Ferric Ions. Aliquots of stock solutions of
the ligands 5L, 5A, 5G, or 10A and increasing amounts of FeCl3 (both
in spectroscopic grade methanol) were mixed, and after 15 min diluted
with methanol/0.1 N aqueous NaOAc40 (4/1) to a constant ligand
concentration of 0.3 mM for UV/vis titrations, and of 6 µM for
fluorescence titration. Measurements were performed after equilibration
for 24 h at room temperature. UV/vis absorption spectra were recorded
on a Hewlett-Packard diode array 8450A spectrophotometer, and
fluorescence spectra were recorded on a Shimadzu RF-540 spectro-
fluorometer.
Bacterial Strains and Growth Conditions. The sid- mutants
Pseudomonas putida JM218, Pseudomonas fluorescens S680, and
Pseudomonas fluorescens WCS 3742 were provided by L. C. van Loon,
Utrecht, The Netherlands. Such mutants are ideal for the study of Fe-
siderophore utilization, since the absence of native siderophores
eliminates possible interference through ligand exchange.
Bacteria were grown in LMKB medium41 overnight at 28 °C on a
rotary shaker at 180 RPM. The bacterial cultures were then centrifuged
for 15 min at 2500 RPM, resuspended in fresh half-strength standard
succinate medium (SSM) to a final o.d.620 of 0.6, and incubated for 60
min in a water bath at 28 °C. When appropriate, NaN3 was added to
a final concentration of 5 mM 30 min prior to the addition of the
siderophores.
Microbial 55Fe Uptake Studies. 55Fe uptake studies were conducted
according to Jurkevitch et al.25 with minor modifications. The labeled
(42) Here we use the term “agonist” for a siderophore analog that fully
mimics the natural compound by acting as a microbial iron(III) carrier,
and the term “antagonist” for a siderophore analog that inhibits the action
of the natural compound by competitively binding to the receptor.
(43) Weizman, H. M.Sc. Thesis, Feinberg Graduate School, The Weiz-
mann Institute of Science, 1993.
(40) We used NaOAc buffer to maintain the solutions’ pH values at 7.4.
The use of other buffers that would have been superior for this pH range
was prohibited either because they quench the probes’ fluorescence (amino
groups containing buffers) or because they compete with the ligands for
iron(III) (phosphate buffers).
(44) Birks, J. B. Photophysics of Aromatic Molecules; Wiley-Inter-
science: London, New York, Sydney, Toronto, 1970.
(41) King, E. O.; Ward, M. K.; Raney, D. E. J. Lab. Clin. Med. 1954,
44, 301-307.
(45) Raymond, K. N.; Mueller, G.; Matzanke, B. F. Top. Cur. Chem.
1984, 123, 49-102.

12372 J. Am. Chem. Soc., Vol. 118, No. 49, 1996
Weizman et al.
Table 1. 1H NMR Shift Differences [∆(δ5,10-δ1)] between the
Scheme 2
Fluorescently Labeled Ligands 5 and 10 and Their Parent Ligands 1
ligand
5G
CD3OD
5A
CDCl3
5A
CD3OD
5L
CDCl3
5L
CD3OD
10A
CDCl3
10A
CD3OD
C-CH2O
-0.26
OCH2
-0.34
CH2CO
-0.17
NCH3
0.00
-0.32
-0.33
-0.32
-0.21
+0.24
+0.38
-0.33
-0.23
+0.02
-0.27
-0.04
+0.05
-0.23, -0.16
-0.17
-0.07
+0.10
-0.12
-0.08
-0.07
-0.03
-0.21, -0.17
-0.14
-0.03, -0.08
0.00
anthracene 4 (Scheme 1). The sulfon derivative 10A was
prepared by coupling the labeled tris-phenolate 9 with the
hydroxamate-bearing alanyl residue 3A (Scheme 2).
Structures and Iron(III)-Binding Properties.38 The struc-
tures of all the ferrichrome analogs were confirmed by their
spectroscopic properties, such as IR, NMR, UV, and FAB MS.
The 1H NMR data were particularly useful, providing informa-
tion on the compounds conformation and the average location
of the fluorescent label relative to the iron(III)-binding domain.
Three parameters were compared: (i) the number of sets of
signals, to establish binders symmetry; (ii) the extent of
nonequivalence of the diastereotopic protons, to estimate the
ligands’ conformational constraints; and (iii) the chemical shifts
of the binders’ backbone protons in comparison with those of
the parent ligands, to assess the extent to which the anthracene
label could approach the ion-binding cavity.
requirements was crucial since other labels, such as the weakly
fluorescent quinoline, failed to provide reliable results. As an
attachment point for the label we selected the carbon anchor of
the ferrichrome analogs 1. Attachment to this site was expected
to not interfere with iron(III) binding or with recognition by
ferrichrome-specific membrane receptors (such as the FhuA
receptor in E. coli), as receptor recognition has been shown to
occur around the iron(III)-binding domain.22,23 For in Vitro
comparison, we also prepared the leucyl derivative 5L and the
alanyl derivative 10A. Derivative 5L, with its bulky side chain,
was chosen to establish whether steric screening of the iron-
binding domain would influence its quenching efficiency.
Derivative 10A, with the anthracene attached via a shorter linker,
was chosen to trace the effect of the proximity between the label
and iron(III) on the quenching process.
Table 1 highlights the differences in chemical shifts between
the labeled derivatives 5 and 10 and the nonlabeled congeners
1. Similar to the parent ligands 1, the labeled derivatives 5
and 10 exhibited single sets of signals, confirming average C3-
symmetry on the NMR scale. Moreover, the diastereotopic
protons C-C1H2-O and -C1H2CO- of the chiral derivatives
5A, 5L, and 10A were magnetically nonequivalent in CDCl3
but equivalent (or nearly equivalent) in CD3OD, as in the parent
ligands 1. The magnetic nonequivalence of these protons was
Synthesis of the labeled compounds 5 involved coupling of
the C3-symmetric tetraphenolate 2 with 3 equiv of hydroxamate
bearing amino acid residues 320 and 1 equiv of aminomethyl-
Figure 1. Schematic representation of ferrichrome analogs 1 and of fluorescent ferrichrome analogs 5 and 10.

Fluorescent-Labeled Ferrichrome Analogs
J. Am. Chem. Soc., Vol. 118, No. 49, 1996 12373
Figure 2. UV/vis (left) and fluorescence (right) iron(III)-titration experiments with labeled ferrichrome 5 analogs. Aliquots of stock solutions of
ligand 5A in methanol were treated with aliquots of methanolic solutions of FeCl3 (0.0, s, 0.2 s‚s, 0.6 ‚‚, 0.8 - -, 1.0 equiv ) and diluted with
methanol (80%)-0.1 N aqueous NaOAc (20%) to a final ligand concentration of 0.3 mM for UV/vis measurements and 6 µM for fluorescence
measurements (final pH 7.4); UV/vis (left insert) and fluorescence (right insert) titration experiments of ligand 5A-Fe(III) complex with EDTA.
Solutions of ligand-5A-Fe(III), 1:1 complex (0.3 mM for UV/vis and 6 µM for fluorescence measurements) in methanol (80%)-0.1 N aqueous
NaOAc (20%), were treated with aliquots of aqueous EDTA (0.0, s, 1.0, - -, 3.0 eq. s‚s). All samples were equilibrated overnight before
measurements were performed.
Figure 3. Plots for dynamic quenching (left) and static quenching (right) of the fluorescence of labeled ferrichrome analogs 5 (5G, O; 5A, ]; 5L,
9) and 10 (2) by iron(III). The conditions were as described in legend to Figure 2.
attributed to the conformational constraints of the tripodal ion-
binding cavity brought about by a network of intra- and
interstrand H-bonds. The latter are only expressed in apolar
solvents and are broken in polar solvents, as in the parent ligands
1. Comparison between the chemical shifts of the backbone
protons in the parent ligands 1 and their labeled derivatives 5
revealed pronounced up-field shifts (Table 1). These up-field
shifts were particularly pronounced in the glycyl derivative 5G
and alanyl derivative 5A, less so in the leucyl derivative 5L.
They were attributed to the proximity of the anthracenyl ring
which bends toward the ligand backbone, as schematically
depicted in Figure 1. The nonequivalence of the anthracenyl
protons Ant-CH2- in 5A confirm this type of arrangement. In
derivative 5L, the isobutyl side chains screen the backbone,
thereby diminishing the effect of the aromatic ring current. In
the alanyl derivative 10A, where the label is close to the anchor,
such bending is prohibited, and the differences between the
chemical shifts of 10A and 1A were negligible, except for the
anchor’s -C-C1H2-O protons which resonated at significantly
lower field, probably because of the inductive effects exerted
by the sulfone group.
UV/vis titration of the labeled derivatives 5 and 10 with FeCl3
established the formation of Fe(III) complexes with a 1:1
stoichiometry, as reflected by their absorptions around 420 nm,
which is characteristic for iron(III) trishydroxamates.45 More-
over, the iron(III) complexes of the chiral derivatives 5A, 5L,
and 10A revealed two Cotton effects in the CD-spectra of
opposite absolute sign in the visible region, with the positive
Cotton effect appearing at a higher wavelength. These data
demonstrate preferential Λ-cis configuration for the fluorescent
derivatives, analogous to the chiral preference observed earlier
for the parent compounds.20

12374 J. Am. Chem. Soc., Vol. 118, No. 49, 1996
Weizman et al.
These data confirm the close structural analogy between the
iron-binding domain of the parent ligands 1 and their labeled
derivatives 5 and 10, both in the free state and complexed with
iron. The structural analogy around the binding cavity is a
prerequisite for recognition of the labeled derivatives by the
microbial receptor sites. Subtle differences were, however,
observed for the average location of the anthracenyl label in
derivatives 5 and 10. In the former, the label may bend toward
the ion-binding domain, whereas in the latter the label may not
reach the binding cavity. The latter structural differences were
reflected in the compounds’ photophysical properties.
All the labeled ligands, 5 and 10, showed the characteristic
absorption maxima of anthracene at 333, 348, 366, and 386
nm (ꢀ ) 2430, 4850, 7150, and 6610, respectively), and high
fluorescence intensity with maxima at 395, 412, 435 and 464
nm when excited at 348 nm. Upon titration with iron(III), the
molecules’ fluorescence intensity decreased with increasing iron-
hydroxamate absorption in the UV/vis spectrum (Figure 2). The
observed fluorescence quenching, φ/φo, was not linear with
quencher (iron(III)) concentration, and thus failed to obey the
Stern-Volmer model of dynamic quenching. However, the lnφ/
φo values were linear with quencher concentration, in agreement
with the Perrin model of static quenching, where immobilized
quenchers act over a significant distance (Figure 3).44 The
fitting of the quenching data to the Perrin model may be
rationalized by assuming that (i) all added quencher is bound
to the ligand, and (ii) the effect of the bound iron on the
chromophore’s fluorescence depends on the probability of the
chromophore lying within the quencher’s active sphere.
Figure 4. Appearance of anthracene fluorescence (λexc ) 375 nm, λem
) 415 nm) in the culture media of Pseudomonas putida JM218 upon
treatment with labeled ferrichrome analogs 5G and 5A in the absence
(5G, 0; 5A, ]) and presence (5G, O; 5A, 4) of NaN3 as inhibitor.
Insert: 55Fe uptake with the ferrichrome analogs 5G (0) and 5A (]).
with a value of 2.84 × 10-5 M-1. This illustrates the effect of
the proximity (or accessibility) of the label to the iron(III)
hydroxamate on the quenching process.
A priori, fluorescence-quenching of anthracene by the iron-
(III)-hydroxamate chromophore could occur by heavy-atom-
induced intersystem crossing, energy transfer, or electron
transfer. Indirect evidence from our laboratory43 as well as
others31,32,47 support the involvement of electron transfer in such
processes. Earlier studies also seem to suggest that close contact
between the components is advantageous for electron transfer.48
The higher quenching efficiency in 5, where the fluorophore
can reach the iron(III) center, than in 10, where the approach
of the fluorophore to the iron(III) center is sterically prohibited,
may reflect the contribution of electron transfer to the observed
anthracenyl quenching processes.
a. Dynamic quenching (Stern-Volmer Model)
1M* + 1Q a 1(M.Q)*
φFM
1
)
(φFM 0
)
1 + k[Q]
Microbial Activity. Incubation of P. putida JM218 with the
iron(III) complex of the labeled glycyl derivative 5G resulted
in the appearance of anthracene fluorescence in the cultures
growth media. (As emphasized in the Experimental Section,
the organisms were removed by centrifugation before measuring
fluorescence.) This fluorescence was drastically reduced by the
addition of NaN3, an inhibitor of energy-dependent processes
(Figure 4), or by the addition of natural ferrichrome as a
competitor (data not shown). Incubation of P. putida JM218
with the alanyl derivative 5A failed to induce any significant
fluorescence (Figure 4). These experiments were complemented
by 55Fe-uptake experiments (Figure 4, inserts). Again, deriva-
tive 5G proved superior to derivative 5A in facilitating 55Fe
uptake. NaN3 negated 55Fe(III) uptake by both compounds. In
contrast to the behavior of P. putida, incubation of P. fluorescens
S680, or P. fluorescens WCS3742 with 5G or 5A did not result
in the generation of fluorescence in the culture media (Figure
5).
b. Static quenching (Perrin Model)
1M + 1Q′ a 1(M.Q)′ f 1(M.Q)*
φFM
) e-k′[Q]
(φFM 0
)
The addition of EDTA to solutions of the ligand-5-Fe(III)
complex caused release of the hydroxamate-bound iron(III), as
evidenced by the absorption decrease at 430 nm, and resultant
gradual recovery of the fluorescence intensity (Figure 2,
inserts).46 These observations demonstrate that fluorescence
quenching occurs predominantly, if not exclusively, intramo-
lecularly.
Some differences in the fluorescence-quenching properties
of the two types of derivatives, 5 and 10, were observed. The
fluorescence-quenching efficiencies k′ in the Perrin equation for
the leucyl-, alanyl-, and glycyl derivatives 5L, 5A, and 5G,
These results indicate that the labeled glycyl derivative 5G
acts as a specific agonist,42 delivering iron(III) to the microor-
ganism in an energy-driven process via a ferrichrome-specific
iron(III) uptake system, then returning to the medium. The
intactness of the released 5G molecule is indicated by the
quenching of its fluorescence upon addition of iron(III).
respectively, were almost identical (4.75 × 10-5, 4.39 × 10-5
,
and 4.63 × 10-5 M-1, respectively). However, the quenching
efficiency of the alanyl derivative 10A was substantially lower,
(46) Although the binding constants of hydroxamate siderophores to iron-
(III) are higher than those of EDTA (log K ∼30 and 25, respectively), the
apparent binding constant of the synthetic ferrichrome analogs 5 is
comparable to that of EDTA under our experimental conditions, with pFe
values of 15. EDTA may thus effectively compete with the ferrichrome
analogs for iron(III).
(47) Nakamura, T.; Kira, A.; Imamura, M. J. Phys. Chem. 1983, 87,
3122-3125.
(48) Donckt, E. V.; vanVooren, C. J. Chem. Soc. Faraday Trans. I 1978,
74, 827-836.

Fluorescent-Labeled Ferrichrome Analogs
J. Am. Chem. Soc., Vol. 118, No. 49, 1996 12375
labels for signaling. In these assemblies, the labels did not
interfere with either the ligands’ iron-binding processes or with
the complexes’ microbial activity. Iron binding of these
molecules caused quenching of their labels’ fluorescence.
Quenching was more efficient when the label could approach
the iron-binding domain and is likely to involve electron transfer
from the chromophore to the metal ion. The bioactive deriva-
tives of these fluorescent compounds can therefore serve as
versatile biological probes. They enabled us to monitor optically
microbial iron-uptake processes and to trace the fate of the iron
carriers after iron delivery. Concentrating on P. putida as an
indicator organism, we demonstrated that the behavior of the
labeled derivatives parallel that of the nonlabeled analogs, with
the added benefit of signaling iron delivery via the appearance
of fluorescence in the culture media. Moreover, these com-
pounds were found to exhibit high specificity by distinguishing
between different Pseudomonas species. This high specificity,
and the possibility of incubating cultures with a series of
differently labeled siderophore analogs that fit a given species’
characteristic range of receptors and thereby provide a uniquely
defined microbial “fingerprint”, encourage the consideration of
fluorescent analogs as diagnostic tools for the detection and
identification of microorganisms. The energy dependence of
such siderophore-based, diagnostic methods guarantees their
uniqueness to live organisms, in contrast to DNA probes, which
fail to distinguish between live and dead cells. Current
investigations are aimed at preparing differently labeled sid-
erophore analogs and applying them to the study of iron
metabolism by fluorescence-imaging methods.49,50 In addition,
replacement of the fluorescent labels with toxic residues is
envisioned to provide antimicrobial agents.
Figure 5. Appearance of anthracene fluorescence (λexc ) 375 nm, λem
) 415 nm) in the culture media of Pseudomonas putida JM218 upon
treatment with labeled ferrichrome analogs 5G (0) and 5A (]).
Anthracene fluorescence (λexc ) 375 nm, λem ) 415 nm) in the culture
media of Pseudomonas fluorescens S680 and Pseudomonas fluorescens
WCS3742 upon treatment with labeled ferrichrome analogs 5G (9 and
*, respectively) and 5A (] and diamond with right half shaded,
respectively).
Derivative 5A failed to deliver iron(III), acting as an antagonist42
and remaining bound to the membrane as an intact complex.
These results are in full agreement with our earlier findings that
the glycyl-based ferrichrome analog 1G functions as an iron-
(III) carrier in P. putida JM218, whereas the corresponding
alanyl derivative 1A merely acts as an inhibitor of the natural
ferrichrome by binding to the corresponding membrane recep-
tor.23 They further suggest that the synthetic ferrichrome
analogs behave as shuttles and demonstrate that the presence
of the rather bulky anthracene residue does not compromise
the analogs’ microbial activity.
Acknowledgment. The authors thank Mrs. Rahel Lazar for
her skillful technical assistance and the Israel Science Founda-
tion and Edith Reich Foundation for financial support. A.S. is
the holder of the Siegfried and Irma Ullman Professorial Chair.
JA9610646
Conclusions
(49) Because of the small size of bacteria, fluorescence microscopy of
bacterial cultures (even at 1000× magnifications) does not enable distin-
guishing between fluorescence inside and outside the organisms and is thus
of little diagnostic value with regards to the fluorescent siderophore analogs
described here. In contrast to bacteria, fungi lend themselves to fluorescence
microscopy studies. The latter possibilities are under current investigation.
(50) Ardon, O.; Nudelman, R.; Weizman, H.; Hadar, Y.; Libman, J.;
Shanzer, A. Preliminary results, 1996.
Fluorescently labeled biomimetic analogs of the ferrichrome
siderophore were prepared by applying a modular strategy of
design and synthesis. These analogs possessed three crucial
elements: hydrophilic cavities for iron binding, molecular
envelopes for receptor recognition, and projecting fluorescent