Synthesis and iron sequestration equilibria of novel exocyclic 3-hydroxy-2-pyridinone donor group siderophore mimics

Article abstract of DOI:10.1021/ic902595c

The synthesis of a novel class of exocyclic bis- and tris-3,2- hydroxypyridinone (HOPO) chelators built on N² and N³ aza-macrocyclic scaffolds and the thermodynamic solution characterization of their complexes with Fe(III) are described. The chelators for this study were prepared by reaction of either piperazine or N,N′,N′′-1,4,7- triazacyclononane with a novel electrophilic HOPO iminium salt in good yields. Subsequent removal of the benzyl protecting groups using HBr/acetic acid gave bis-HOPO chelators N²(etLH)₂ and N²(prLH) ₂, and tris-HOPO chelator N³(etLH)₃ in excellent yields. Solution thermodynamic characterization of their complexes with Fe(III) was accomplished using spectrophotometric, potentiometric, and electrospray ionization-mass spectrometry (ESI-MS) methods. The pK_a's of N²(etLH)₂, N²(prLH)₂, and N ³(etLH)₃, were determined spectrophotometrically and potentiometrically. The Fe(III) complex stability constants for the tetradentate N²(etLH)₂ and N²(prLH)₂, and hexadentate N³(etLH)₃, were measured by spectrophotometric and potentiometric titration, and by competition with ethylenediaminetetraacetic acid (EDTA). N³(etLH)₃ forms a 1:1 complex with Fe(III) with log β₁₁₀ = 27.34 ± 0.04. N²(prLH)₂ forms a 3:2 L:Fe complex with Fe(III) where log β₂₃₀ = 60.46 ± 0.04 and log β₁₁₀ = 20.39 ± 0.02. While N²(etLH)₂ also forms a 3:2 L:Fe complex with Fe(III), solubility problems precluded determining log β₂₃₀; log β₁₁₀ was found to be 20.45 ± 0.04. The pFe values of 26.5 for N³(etLH)₃ and 24.78 for N²(prLH)₂ are comparable to other siderophore molecules used in the treatment of iron overload, suggesting that these hydroxypyridinone ligands may be useful in the development of new chelation therapy agents.

Full text of DOI:10.1021/ic902595c

8208 Inorg. Chem. 2010, 49, 8208–8221
DOI: 10.1021/ic902595c
Synthesis and Iron Sequestration Equilibria of Novel Exocyclic
3-Hydroxy-2-pyridinone Donor Group Siderophore Mimics
James M. Harrington,† Sumathi Chittamuru,‡ Suraj Dhungana,† Hollie K. Jacobs,‡ Aravamudan S. Gopalan,*,‡
and Alvin L. Crumbliss*,†
†Department of Chemistry, Duke University, Durham, North Carolina 27708-0346, and ‡Department of
Chemistry and Biochemistry, New Mexico State University, Las Cruces, New Mexico 88003-8001
Received December 31, 2009
The synthesis of a novel class of exocyclic bis- and tris-3,2-hydroxypyridinone (HOPO) chelators built on N2 and
N3 aza-macrocyclic scaffolds and the thermodynamic solution characterization of their complexes with Fe(III) are
described. The chelators for this study were prepared by reaction of either piperazine orN,N0,N00-1,4,7-triazacyclononane
with a novel electrophilic HOPO iminium salt in good yields. Subsequent removal of the benzyl protecting groups using
HBr/acetic acid gave bis-HOPO chelators N2(etLH)2 and N2(prLH)2, and tris-HOPO chelator N3(etLH)3 in excellent
yields. Solution thermodynamic characterization of their complexes with Fe(III) was accomplished using spectro-
photometric, potentiometric, and electrospray ionization-mass spectrometry (ESI-MS) methods. The pKa’s of N2(etLH)2,
N2(prLH)2, and N3(etLH)3, were determined spectrophotometrically and potentiometrically. The Fe(III) complex
stability constants for the tetradentate N2(etLH)2 and N2(prLH)2, and hexadentate N3(etLH)3, were measured by
spectrophotometric and potentiometric titration, and by competition with ethylenediaminetetraacetic acid (EDTA).
N3(etLH)3 forms a 1:1 complex with Fe(III) with log β110 = 27.34 ( 0.04. N2(prLH)2 forms a 3:2 L:Fe complex with
Fe(III) where log β230 = 60.46 ( 0.04 and log β110 = 20.39 ( 0.02. While N2(etLH)2 also forms a 3:2 L:Fe complex
with Fe(III), solubility problems precluded determining log β230; log β110 was found to be 20.45 ( 0.04. The pFe
values of 26.5 for N3(etLH)3 and 24.78 for N2(prLH)2 are comparable to other siderophore molecules used in
the treatment of iron overload, suggesting that these hydroxypyridinone ligands may be useful in the development of
new chelation therapy agents.
Introduction
Excess iron is toxic because it can potentially react with the
various oxidation states of oxygen in a biological system to
form reactive oxygen species through redox cycling, dama-
ging the organism and even causing death.8 An effective
method of treating iron overload disease and hemochroma-
tosis involves the use of iron chelation therapy. There are
currently two U.S. FDA-approved therapeutic agents for
iron-overload, Desferal (desferrioxamine B mesylate, DFB),
and Exjade (deferasirox), (see below).9-11 However, these
agents are both problematic. Desferal suffers from low
Virtually all forms of life depend on iron to carry out pro-
cesses necessary for their survival. However, in a pH 7 environ-
ment, iron is extremely insoluble ([Fe3þ
]
aq = ∼10-10 mol dm-3
)
and is usually found in the form of iron hydroxides or oxides
because of hydrolysis.1,2 To overcome problems of limited
availability, micro-organisms have developed methods of
safely and effectively extracting iron from the environment
through manipulating the inner coordination sphere of the
metal ion to form soluble complexes.3 Bacteria and fungi
produce small-molecule iron-specific chelators, siderophores,
to aid in the uptake of iron by sequestering the metal ion with
high stability and specificity (log β>25).4-7
(5) Crumbliss, A. L.; Harrington, J. M. In Advances in Inorganic Chemi-
stry; van Eldik, R., Hubbard, C. D., Eds.; Academic Press: Orlando, FL, 2009;
Vol. 61, pp 179-250.
(6) Raymond, K. N.; Dertz, E. A. In Iron Transport in Bacteria; Crosa,
J. H., Rey, A. R., Payne, S. M., Eds.; ASM Press: Washington, D.C., 2004;
pp 3-17.
*To whom correspondence should be addressed. E-mail: alc@chem.duke.
edu (A.L.C.), agopalan@nmsu.edu (A.S.G.). Fax: (919) 660-1605 (A.L.C.);
(575) 646-2649 (A.S.G.).
(1) Boukhalfa, H.; Crumbliss, A. L. Biometals 2002, 15, 325–339.
(2) Kragter, J. Atlas of Metal-Ligand Equilibria in Aqueous Solution;
Halstead Press: New York, 1978.
(3) Crichton, R. R. Inorganic Biochemistry of Iron Metabolism, 2nd ed.;
Ellis Horwood, Ltd.: Chichester, U.K., 2001.
(4) Albrecht-Gary, A. M.; Crumbliss, A. L. In Metal Ions in Biological
Systems; Sigel, A., Sigel, H., Ed.; Marcel Dekker: New York, 1998; Vol. 35,
pp 239-327.
(7) Harrington, J. M.; Crumbliss, A. L. Biometals 2009, 22, 679–689.
(8) Haber, F.; Weiss, J. Proc. Royal Soc. London, Ser. A 1934, 147,
332–351.
(9) Jacobs, A. J. In Development of Iron Chelators for Clinical Use; Martell,
A. E., Anderson, W. F., Badman, D. B., Eds.; Elsevier/North Holland: New York,
1981; pp 29-46.
(10) Lattmann, R.; Acklin, P.; U.S. Patent No. US 6,465,504 B1 (October 15,
2002).
(11) Nick, H. Curr. Opin. Chem. Biol. 2007, 11, 419–423.
r
pubs.acs.org/IC
Published on Web 08/17/2010
2010 American Chemical Society

Article
Inorganic Chemistry, Vol. 49, No. 18, 2010 8209
bioavailability, a short half-life in serum, and a treatment regime
that includes daily intravenous transfusions, which is a long,
painful, and expensive procedure leading to low patient com-
pliance. Exjade is perhaps a better choice, because of its increased
bioavailability and the ability to administer the drug orally.
However, the molecule lacks specificity for iron, which may
result in undesirable side effects because of sequestration of other
and more effective Fe(III) chelation by HOPO donor
groups, as reflected in the pFe values for catechol (15.1)
and 3,2-HOPO (16.26).17-20 The charge difference in the
donor moieties of catechol and HOPO also affect the
equilibrium constants for complexation. This is demon-
strated by comparing the stepwise formation constants for
the iron-catechol complex with those of the 3,2-HOPO
nþ
complex.18,19 The K1 for FeL(OH2)4 formation (where
K represents the proton independent stepwise formation
constant) is larger for L = catechol than for 3,2-HOPO, at
least partially because of the greater electrostatic force of
attraction between the þ3 iron and the catechol dianion.
However, apparently the higher negative charge for catechol
also has a destabilizing effect because of charge repulsion in
the higher order complexes. Normally, the stepwise equilib-
rium constants decrease in the order K1 > K2 > K3 for
statistical reasons. However, the observed decrease is
greater than expected on statistical grounds for catechol
(K1= 20.4, K2= 15.5, Δ1-2 = 4.9, K3 = 9.4, Δ2-3 = 6.1)
than for 3,2-HOPO (K1=11.48, K2=9.77, Δ1-2=1.71,
K3 = 8.01, Δ2-3 = 1.76) because of increased charge
repulsion.19,21
There are ostensibly two types of linkages that have been
used for the connection of multiple 3,2-HOPO donor groups
in the design of polydentate hydroxypyridinone siderophore
mimics. A key distinction is that in the HOPO system shown
below, the 3,2-HOPO ligand is attached to the scaffold
through the pyridinone nitrogen, while for the retroHOPO
system, an amide linkage on the 3,2-HOPO ring is used for
connection with the backbone. The orientation of the donor
metal ions. Thus, it is desirable to develop new synthetic
siderophores to treat iron overload diseases to find a
chelator with high bioavailability and iron specificity.
One high affinity iron(III) binding moiety that is rarely
observed in nature is the hydroxypyridinone (HOPO) donor
group. The only observed instance of a natural HOPO side-
rophore is cepabactin, a 1-hydroxy-2-pyridinone bidentate
siderophore produced by Burkholderia cepacia (above).12 Cepa-
bactin is one of a number of siderophores produced by the
bacteria, and often forms a mixed complex with pyochelin
(above) that is less stable than the tris-cepabactin-Fe(III) com-
plex, but more stable than the bis-pyochelin-Fe(III) complex.13
Hydroxypyridinones are similar to catechol donor groups
in their affinity for Fe(III) because of the similar electronic con-
figurations of the two moieties (eqs 1 and 2).14-16 However, their
groups in the siderophore mimics differ in these two
systems, and this difference plays a role in metal com-
plexation equilibria. The presence of an amide functio-
nal group conjugated to the pyridyl ring in the retro-
HOPO system has an effect on the acidity of the
hydroxypyridinone moiety. Comparison of the hydroxy-
pyridinone protonation constants for TRENHOPO and
TRISPYR (Figure 1) provides a demonstration of this
effect.15,22 The average pKa of the hydroxypyridinone
donor groups of TRISPYR, which features the HOPO
type donor group is 9.50, while the average pKa of the
retroHOPO donor groups of TRENHOPO is 6.99. This
greater acidity in the retroHOPO donor group is possibly
different number of donor moiety ionizable protons (one for
HOPO and two for catechol) results in less Hþ competition
(17) L’Heureux, G.; Martell, A. E. J. Inorg. Nucl. Chem. 1966, 28, 481–
491.
(12) Thomas, M. S. Biometals 2007, 20, 431–452.
(13) Klumpp, C.; Burger, A.; Mislin, G. L.; Abdallah, M. A. Bioorg. Med.
Chem. Lett. 2005, 15, 1721–1724.
(14) Scarrow, R. C.; Riley, P. E.; Abu-Dari, K.; White, D. L.; Raymond,
K. N. Inorg. Chem. 1985, 24, 954–967.
(15) Xu, J.; O’Sullivan, B.; Raymond, K. N. Inorg. Chem. 2002, 41, 6731–
6742.
(16) Abergel, R. J.; Raymond, K. N. Inorg. Chem. 2006, 45, 3622–3631.
(18) Martell, A. E.; Smith, R. M. Critical Stability Constant Database;
National Institute of Science and Technology (NIST): Gaithersburg, MD, 2003.
(19) Clarke, E. T.; Martell, A. E. Inorg. Chim. Acta 1992, 196, 185–194.
(20) Motekaitis, R. J.; Martell, A. E. Inorg. Chim. Acta 1991, 183, 71.
(21) Avdeef, A.; Sofen, S. R.; Bregante, T. L.; Raymond, K. N. J. Am.
Chem. Soc. 1978, 100, 5362–5370.
(22) Sun, Y.; Motekaitis, R. J.; Martell, A. E. Inorg. Chim. Acta 1998, 281,
60–63.

8210 Inorganic Chemistry, Vol. 49, No. 18, 2010
Harrington et al.
Figure 1. Representative synthetic 3,2-hydroxypyridinone siderophore mimics. Stability constants taken from refs 16, 22, 40, 60, and 61.
due to inductive effects arising from the proximity of the
amide connector to the pyridyl ring.
on the use of HOPO siderophores as a treatment for iron
overload disease have shown promising results with regard to
iron binding ability and lack of bioavailability of the iron(III)
complexes in mammalian systems.28,29 In addition to iron
overload chelation therapy agents, HOPO chelators and
other iron chelators have other potential medical applica-
tions. These include the development of novel antibacterial
agents, and cancer prevention and treatment agents.30
Hydroxypyridinone chelators also form strong complexes
with other hard metal ions leading to additional potential
applications. They have been examined for their ability to
achieve in vivo clearance of Pu(IV) ions, as potential extrac-
tion agents for Pu(IV), and have been attached to calix[4]-
arenes to yield useful Th(IV) extractants.31,32 Gadolinium
HOPO chelate complexes have been prepared and shown
to have potential as contrast agents in magnetic resonance
imaging (MRI).33-36 As chelators of Ga-67 and In-111,
The drug deferiprone, or 1,2-dimethyl-3-hydroxy-4-pyri-
dinone (above), is a bidentate 3,4-HOPO chelator that has
been approved for use in chelation therapy treatments in
Europe.23 However, questions have been raised with respect
to the toxicity of this drug, related to binding efficiency and
its possible role in the development of diseases, including
hepatic fibrosis and a variety of other disorders.24-26 There
are also questions related to the redox chemistry of its iron
complexes.27 The development of chelation therapy agents
requires that the metal free chelator have high bioavailabi-
lity to allow ease of uptake, but that the metal complex
formed have low bioavailability, to aid in removal of the
metal from the system.
The low bioavailability and high stability of previously
characterized Fe(III)-HOPO complexes suggests that the
development of synthetic siderophores with hydroxypyridi-
none donors may be a viable route for the development of
new treatments for iron overload disease. Preliminary studies
(28) Hider, R. C.; Kontoghiorghes, G. J.; Silver, J.; Brit. U.K. Patent
Appl.: Great Britain, 1984; p 17.
(29) Hider, R. C.; Kontoghiorghes, G. J.; Silver, J.; Stockton, M. A.; Brit.
U.K. Patent Appl.: Great Britain, 1984; p 19.
(30) Weinberg, E. D. Adv. Appl. Microbiol. 2003, 52, 187–208.
(31) Veeck, A. C.; White, D. J.; Whisenhunt, D. W., Jr.; Xu, J.; Gorden,
A. E. V.; Romanovski, V.; Hoffman, D. C.; Raymond, K. N. Solvent Extr.
Ion Exch. 2004, 22, 1037–1068.
(32) Lambert, T. N.; Dasaradhi, L.; Huber, V. J.; Gopalan, A. S. J. Org.
Chem. 1999, 64, 6097–6101.
(23) Capellini, M. D.; Pattoneri, P. Annu. Rev. Med. 2009, 60, 25–38.
(24) Kontoghiorghes, G. J. Expert Opin. Drug Saf. 2007, 6, 235–239.
(25) Olivieri, N. F.; Brittenham, G. M.; McLaren, C. E.; Templeton,
D. M.; Cameron, R. G.; McClelland, R. A.; Burt, A. D.; Fleming, K. A.
N. Eng. J. Med. 1998, 339, 417–423.
(26) Wu, S. F.; Peng, C. T.; Wu, K. H.; Tsai, C. H. Hemoglobin 2006, 30,
215–218.
(27) El-Jammal, A.; Templeton, D. M. Inorg. Chim. Acta 1996, 245,
199–207.
(33) Werner, E. J.; Datta, A.; Jocher, C. J.; Raymond, K. N. Angew.
Chem. 2008, 47, 85688580.
(34) Pierre, V. C.; Botta, M.; Aime, S.; Raymond, K. N. J. Am. Chem.
Soc. 2006, 128, 5344–5345.

Article
Inorganic Chemistry, Vol. 49, No. 18, 2010 8211
HOPO ligands are potentially useful as radiopharmaceutical
imaging agents.37,38 The numerous potential applications of
HOPO chelators highlight the need to develop methods for
their synthesis that are flexible and enable their attachment to
a variety of molecular platforms.
The most common method for the preparation of 3,2-
HOPO chelators involves the coupling of an amine with an
activated carboxylic acid linker attached to the pyridinone
ring system.39 In the case of polyHOPO derivatives, this
method results in multiple amide bonds which can limit both
organic and aqueous solubility. Several examples of amide
linked tris-HOPO siderophores are shown in Figure 1.
TRENHOPO, 3,2-HOPOHL, and CP130, all have a similar
tripodal design with the three chelating arms attached to a
central atom using an amide-linked spacer group. The iron-
(III) affinity for these siderophores range from pFe = 26.8 to
32.23.22,40 Compared to the amide linked tripodal chelators,
the linear tris-HOPO chelator, 3,4-LI-(Me-3,2-HOPO), ex-
hibits a slightly lower complex stability, pFe = 25.5.16 The
tripodal chelator, TRISPYR, which does not have amide
linkages, was prepared by direct alkylation of the back-
bone amine resulting in low yields of the desired product.22
The very high iron(III) affinity of TRISPYR, pFe =32.23,
was attributed to the flexibility in the ether linkages to the
three hydroxypyridinones.22 While the presence of amide
linking groups may contribute to decreased ligand solubility
and increased susceptibility to enzymatic hydrolysis, their
presence can also contribute to complex stability through the
formation of hydrogen bonding networks, which may act to
preorganize the ligand binding site. Additionally, in natural
siderophores and synthetic mimics, the amide moieties such
asin catecholamide donor groups have been proposed to play
a role because of binding mode shifts over a range of pH
values.41-43
Very few poly-HOPO chelators built on a cyclic backbone
have been synthesized. HOPObactin, (Figure 1, pFe = 26.8)
is a 3,2-hydroxypyridinone analogue of the catechol side-
rophore enterobactin and is built on a trilactone core.40 This
ligand exhibits low solubility in water, and the trilactone core
is susceptible to hydrolysis above pH 8, as well as below pH 6.
By comparing the binding constants of HOPObactin with the
tripodal tris HOPO ligands, it is clear that the trilactone core
does not impart additional stabilization for the anchored
HOPO binding moieties. The trisHOPO chelator, TACN-1-
Me-3,2-HOPO (Figure 1), is built on an azamacrocyclic
platform with the three 3,2-HOPO chelating groups attached
by amide bonds.36 This chelator exhibits better solubility
Figure 2. Structures and measured pKa’s of the synthetic exocyclic
3-hydroxy-2-pyridinone donor group siderophores in this study: N2-
(etLH)2, N2(prLH)2, and N3(etLH)3. pKa values listed are from Table
1. Distinction between assignments for pKa1 and pKa4, and pKa2 and pKa3
for N2(etLH)2 and N2(prLH)2 are arbitrary. Distinction between assign-
ments for pKa2, pKa3, and pKa4 for N3(etLH)3 are arbitrary.
than HOPObactin, and its Gd(III) complex has been exam-
ined as a magnetic resonance contrast agent; however, its
iron(III) binding properties have not been evaluated. The
diHOPO cyclen chelator, I (Figure 1), was prepared, and
its Zn(II) and Cu(II) binding properties were evaluated.44
Unfortunately, the iron(III) binding properties were not
reported.
Previously, the synthesis of N2(prLH)2 (Figure 2) was
reported using methodology developed in the Gopalan
laboratory for the convenient attachment of 3,2-HOPO
groups without concomitant formation of an amide bond.45
It was thought that the azamacrocyclic platform might
provide an ideal backbone on which to anchor 3,2-HOPO
chelating units. Unlike the trilactone core of enterobactin, the
azamacrocyclic backbone is hydrolytically stable over a
range of solution pH values, including the physiologically
relevant 5-7.4 pH range. Further, the amine groups in the
backbone enhance water solubility.
Here we report the synthesis of a new class of bis- and tris-
exocyclic synthetic 3,2-HOPO siderophores built on an N2
and N3 azamacrocycle platform (Figure 2). The protona-
tion equilibria of these three siderophores were elucidated
through potentiometric and spectrophotometric titrations,
and the stability of their complexes with Fe(III) was deter-
mined through spectrophotometric titration and competition
reactionswith ethylenediaminetetraaceticacid(EDTA). Eva-
luation of the iron binding properties of such compounds
provides valuable information in the design of iron chelators
with desirable therapeutic properties.
(35) Thompson, M. K.; Misselwitz, B.; Tso, L. S.; Doble, D. M. J.;
Schmitt-Willich, H.; Raymond, K. N. J. Med. Chem. 2005, 48, 3874–
3877.
(36) Werner, E. J.; Avedano, S.; Botta, M.; Hay, B. P.; Moore, E. G.;
Aime, S.; Raymond, K. N. J. Am. Chem. Soc. 2007, 129, 1870–1871.
(37) Reichert, D. E.; Lewis, J. S.; Anderson, C. J. Coord. Chem. Rev. 1999,
184, 3–66.
(38) Miao, Y.; Hylarides, M.; Fisher, D. R.; Shelton, T.; Moore, H.;
Wester, D. W.; Fritzberg, A. R.; Winkelmann, C. T.; Hoffman, T.; Quinn,
T. P. Clin. Cancer Res. 2005, 11, 5616–5621.
(39) Rai, B. L.; Khodr, H.; Hider, R. C. Tetrahedron 1999, 55, 1129–1142.
(40) Meyer, M.; Telford, J. R.; Cohen, S. M.; White, D. J.; Xu, J.;
Raymond, K. N. J. Am. Chem. Soc. 1997, 119, 10093–10103.
(41) Abergel, R. J.; Warner, J. A.; Shuh, D. K.; Raymond, K. N. J. Am.
Chem. Soc. 2006, 128, 8920–8931.
Experimental Section
Materials. All solutions were prepared in deionized water.
Solid NaCl (>99%, Fisher Chemicals) was used to prepare the
background electrolyte solution. Standardized 1 N NaOH
solution (Fisher Chemicals) was used to prepare a 0.10 mol
(44) Ambrosi, G.; Dapporto, P.; Formica, M.; Fusi, V.; Giorgi, L.;
Guerri, A.; Lucarini, S.; Micheloni, M.; Paoli, P.; Pontellini, R.; Rossi, P.;
Zappia, G. New J. Chem. 2004, 28, 1359–1367.
(42) Stack, T. D. P.; Karpishin, T. B.; Raymond, K. N. J. Am. Chem. Soc.
1992, 114, 1512–1514.
(43) Raymond, K. N.; Dertz, E. A.; Kim, S. S. Proc. Natl. Acad. Sci. U.S.A.
2003, 100, 3584–3588.
(45) Lambert, T. N.; Chittamuru, S.; Jacobs, H. K.; Gopalan, A. S.
Tetrahedron Lett. 2002, 43, 7379–7383.

8212 Inorganic Chemistry, Vol. 49, No. 18, 2010
Harrington et al.
dm-3 solution. The base solution was standardized against
KHP (Aldrich, 99.95%) to the phenolphthalein end point.
Concentrated HCl (Mallinckrodt AR, 37%) was used to pre-
pare a 0.10 mol dm-3 solution, which was standardized against a
standard NaOH solution to the phenolphthalein end point.
Solid anhydrous FeCl3 was obtained from Sigma and used to
prepare 0.10 mol dm-3 stock solution that was standardized
titrimetrically by reduction with SnCl2, followed by titration
with K2Cr2O7.46 Competition reactions were performed with a
stock solution prepared from solid EDTA (Acros Organics,
99þ%). pKa titrations of each of the HOPO siderophores alone
were performed in NaClO4 background electrolyte, while titra-
tions in the presence of iron(III) were performed in NaCl back-
ground electrolyte because of low solubility in perchlorate above
pH 4. Hydroxypyridinone alcohol 1 and N2(prLH)2 were pre-
pared by published procedures.32,45
complex from pH 3.5 to pH 10.9. UV-vis spectra from the
titration were analyzed with the programs SPECFIT and
HYPERQUAD to determine the protonation constants of the
complexes.50,51 All determined complex stability constants are
shown in Table 1.
An acid titration, using 1.0 mol dm-3 HCl was conducted for
each Fe(III) siderophore system from pH 3.0 to ∼pH 0.3. For
these titrations, the Microsoft Excel program was used to
analyze the spectral shifts because of difficulty of solving for
some concentrations in both SPECFIT and HYPERQUAD.49
Electrode calibration in high-acid conditions was performed
assuming Nernstian behavior with a junction potential using the
program VLpH.52
EDTA Competition Titrations. UV-visible spectra were re-
corded using a Varian Cary 100 UV-visible spectrophotometer
measuring from 350-750 nm. For the Fe2(N2(prL)2)3 system,
stock solutions of the Fe(III)-siderophore complex were pre-
pared in 2.0 mL aliquots at [Fe3þ] = 2.47 ꢀ 10-4 mol dm-3 and
[N2(prLH)2] = 3.70 ꢀ 10-3 mol dm-3, and a range of concen-
trations of EDTA, from 0 to 25 equiv of EDTA relative to
Fe(III), were added while monitoring the pH (between 5.5 and
7.1). The competition reaction performed for the N2(prLH)2
system is shown in eqs 3-4.
Methods. Potentiometric Measurements. Ligand pKa’s were
determined by potentiometric and spectrophotometric titra-
tions. All pH measurements were made with an Orion 230 Aþ
pH/ion meter equipped with an Orion Ross pH electrode filled
with 3 mol dm-3 NaCl solution. The electrode was calibrated by
titration of standardized 0.10 mol dm-3 HCl with standardized
0.10 mol dm-3 NaOH, as in the “classical method,” and
calibration data were analyzed using the computer program,
GLEE.47,48 Ligand solutions were prepared at 6.7ꢀ10-4 mol
dm-3 for the N2(prLH)2 titration and at 4.0 ꢀ 10-4 mol dm-3 for
the N3(etLH)3 titration in acid and titrated with 0.010 mol dm-3
NaOH. Potentiometric data were analyzed with the Micro-
soft Excel program.49 Ferric complex solutions were pre-
pared at concentrations of [Fe3þ] = 4.00 ꢀ 10-4 mol dm-3
and [N2(prLH)2]=6.0 ꢀ 10-4 mol dm-3 for the Fe-N2(prLH)2
system and [Fe3þ]=3.98 ꢀ 10-4 mol dm-3 and [N3(etLH)3] =
4.0 ꢀ 10-4 mol dm-3 for the Fe-N3(etLH)3 system. These
solutions were also titrated with 0.010 mol dm-3 NaOH,
and potentiometric titration data were analyzed with
Excel.
3þ
-
Fe2ðHN2ðprLÞ2Þ3 þ 2HEDTA3 - þ Hþ / 2FeðEDTAÞ
þ 3HN2ðprLÞðprLHÞ
ð3Þ
ð4Þ
-
2
3
½FeðEDTAÞ ꢁ ½HN2ðprLÞðprLHÞꢁ
Kobs
¼
3þ
2
½Fe2ðHN2ðprLÞ2Þ3 ꢁ½HEDTA3 - ꢁ ½Hþꢁ
Solutions of the complexes were allowed to react for 24 h at
25 °C, until the spectra of the solutions were constant, and the
pH and UV-visible spectra were measured. Data from the
competition titration (Figure 3) were used in conjunction with
the published stability constant of the Fe(EDTA) complex to
determine the stability constant (Table 1) of the Fe2(N2(prL)2)3
complex using eqs 5-7.
Spectrophotometric Titrations. UV-visible spectra were re-
corded using a Cary-50 spectrophotometer equipped with an
external dip probe (Hellma, U.S.A.). For spectrophotometric
titrations of the siderophores alone, solutions of 1.0 ꢀ 10-5 mol
dm-3 siderophore were prepared in 0.10 mol dm-3 NaClO4
solutions. These solutions were titrated with standardized
0.010 mol dm-3 NaOH/0.090 mol dm-3 NaClO4 solutions.
The pH was measured after each addition, and the UV-
visible spectrum was measured in the wavelength region of
200-400 nm. Spectrophotometric data were analyzed by the
program SPECFIT to determine the protonation constants of
the ligands.50
Solutions for use in spectrophotometric titrations were all
prepared as described in the Supporting Information. Concen-
trations of metal and ligand are noted in the captions of the
figures showing the titrations. For all of the Fe(III)-siderophore
complex titrations, 0.10 mol dm-3 NaCl was used as a back-
ground electrolyte. The presence of chloride was taken into
account in the determination of the stability constants by
including the stability constants of the mono-, di-, and trichloro-
iron(III) complexes in the determinations of the stability con-
stants for the hydroxypyridinone complexes.18
2 -
2FeðH2OÞ63þ þ 3N2ðprLÞ2
/ Fe2ðN2ðprLÞ2Þ3 þ 6H2O
ð5Þ
½Fe2ðN2ðprLÞ2Þ3ꢁ
β230
¼
ð6Þ
ð7Þ
3þ 2
2 - 3
½FeðH2OÞ6 ꢁ ½N2ðprLÞ2
ꢁ
0
1
- 1
Kobs
@
A
β230
¼
2
ðβ110,EDTA
Þ
The computer program, HYPERQUAD was used to analyze
the spectral shifts in the titration of the Fe(III)-N2(prLH)2
system.51 The stability constants of the Fe(EDTA) complex
and the protonation constants of EDTA were taken from the
Critical Stability Constant Database and held constant in the
analysis, as were the siderophore protonation constants.18
A similar experiment was performed for the N3(etLH)3
system with the concentrations [Fe3þ] = [N3(etLH)3] = 4.0 ꢀ
10-4 mol dm-3 and the EDTA concentration ranging from
0 to 0.01 mol dm-3, and the solution pH ranging from 5.5 to
Two pH-dependent titrations (acid and base) were performed
for each Fe(III)-siderophore system. In the base titration, 0.10
mol dm-3 NaOH was used to titrate the Fe(III)-siderophore
(46) Vogel, A. I. A Text-Book of Quantitative Inorganic Analysis, 3rd ed.;
Longmans, Green, and Co Ltd: London, 1961.
(47) Martell, A. E.; Motekaitis, R. J. Determination and Use of Stability
Constants, 2nd ed.; VCH Publishers: New York, 1992.
(48) Gans, P.; O’Sullivan, B. Talanta 2000, 51, 33–37.
(49) Billo, E. J. EXCEL for Chemists; Wiley-VCH: New York, 2001.
(50) Gampp, H.; Maeder, M.; Meyer, C. J.; Zuberbuehler, A. D. Talanta
1985, 32, 257–264.
(51) Gans, P.; Sabatini, A.; Vacca, A. Talanta 1996, 43, 1739–1753.
(52) Gans, P., O’Sullivan, B. VLpH (see: http://www.hyperquad.co.uk/vlph.
htm), unpublished.

Article
Inorganic Chemistry, Vol. 49, No. 18, 2010 8213
Table 1. Complete Proton Dissociation Equilibria for HOPO Ligands and Their Iron(III) Complexes
equilibriuma
K or β
log(K or β)b
eq.
N2(etLH)2
2 þ
þ
H2N2ðetLHÞ2
/ HN2ðetLHÞ2 þ Hþ
Ka1
Ka2
Ka3
Ka4
-3.7 ( 0.1c
-6.1 ( 0.1c
-7.50 ( 0.08c
-8.95 ( 0.02c
1-1
1-2
1-3
1-4
þ
HN2ðetLHÞ2 / HN2ðetLÞðetLHÞ þ Hþ
-
HN2ðetLÞðetLHÞ / HN2ðetLÞ2 þ Hþ
-
2 -
HN2ðetLÞ2 / N2ðetLÞ2 þ Hþ
3 þ
3 þ
þ
2FeðH2OÞ6 þ 3HN2ðetLHÞ2 / Fe2ðHN2ðetLÞ2Þ3 þ 6Hþ þ 12H2O
K230
β230
K2
d
1-5
3 þ
2 -
3
2FeðH2OÞ6 þ 3N2ðetLÞ2
/ Fe2ðN2ðetLÞ2Þ þ 6H2O
d
1-6
3 þ
2 þ
þ
2FeðHN2ðetLÞ2ÞðH2OÞ2 þ HN2ðetLHÞ2 / Fe2ðHN2ðetLÞ2Þ3 þ 2Hþ þ 4H2O
d
1-7
3 þ
2 -
þ
FeðH2OÞ6 þ N2ðetLÞ2
/ FeðN2ðetLÞ2ÞðH2OÞ2 þ 4H2O
β110
K110
KaFe
β111
K
21.08 ( 0.01e
7.5f
1-8
3 þ
þ
2 þ
FeðH2OÞ6 þ HN2ðetLHÞ2 / FeðHN2ðetLÞ2ÞðH2OÞ2 þ 2Hþ þ 4H2O
1-9
3 þ
2 þ
FeðHN2ðetLÞðetLHÞÞðH2OÞ4
/ FeðHN2ðetLÞ2ÞðH2OÞ2 þ Hþ þ 2H2O
-0.96 ( 0.09c
22.1 ( 0.1f
8.5f
1-10
1-11
1-12
1-13
3 þ
2 -
2 þ
FeðH2OÞ6 þ N2ðetLÞ2 þ Hþ / FeðN2ðetLÞðetLHÞÞðH2OÞ4 þ 2H2O
3 þ
þ
3 þ
FeðH2OÞ6 þ HN2ðetLHÞ2 / FeðHN2ðetLÞðetLHÞÞðH2OÞ4 þ Hþ þ 2H2O
3 þ
-
2 þ
FeðH2OÞ6 þ N2ðetLÞðetLHÞ / FeðN2ðetLÞðetLHÞÞðH2OÞ4 þ 2H2O
β
14.6f
N2(prLH)2
2 þ
þ
H2N2ðprLHÞ2
/ HN2ðprLHÞ2 þ Hþ
Ka1
Ka2
Ka3
Ka4
-3.8 ( 0.1c
1-14
1-15
1-16
1-17
þ
HN2ðprLHÞ2 / HN2ðprLÞðprLHÞ þ Hþ
-5.91 ( 0.09c
-7.94 ( 0.05c
-9.21 ( 0.02c
-
HN2ðprLÞðprLHÞ / HN2ðprLÞ2 þ Hþ
-
2 -
HN2ðprLÞ2 / N2ðprLÞ2 þ Hþ
3 þ
3 þ
þ
2FeðH2OÞ6 þ 3HN2ðprLHÞ2 / Fe2ðHN2ðprLÞ2Þ3 þ 6Hþ
K230
β230
K2
18.91f
1-18
1-19
1-20
1-21
1-22
1-23
1-24
1-25
1-26
FeðHN3ðetLÞ3Þþ / FeðN3ðetLÞ3Þ þ Hþ
60.46 ( 0.02e
5.71f
3 þ
2 þ
þ
2FeðHN2ðprLÞ2ÞðH2OÞ2 þ HN2ðprLHÞ2 / Fe2ðHN2ðprLÞ2Þ3 þ 2Hþ þ 4H2O
3 þ
þ
2 þ
FeðH2OÞ6 þ HN2ðprLHÞ2 / FeðHN2ðprLÞ2ÞðH2OÞ2 þ 2Hþ þ 4H2O
K110
β110
KaFe
β111
K
6.60f
3 þ
2 -
þ
FeðH2OÞ6 þ N2ðprLÞ2
/ FeðN2ðprLÞ2ÞðH2OÞ2 þ 4H2O
20.45 ( 0.04g
-1.1 ( 0.1c
21.5 ( 0.1f
7.65f
3 þ
2 þ
FeðHN2ðprLÞðprLHÞÞðH2OÞ4
/ FeðHN2ðprLÞ2ÞðH2OÞ2 þ Hþ þ 2H2O
3 þ
2 -
2 þ
FeðH2OÞ6 þ N2ðprLÞ2 þ Hþ / FeðN2ðprLÞðprLHÞÞðH2OÞ4 þ 2H2O
3 þ
þ
3 þ
FeðH2OÞ6 þ HN2ðprLHÞ2 / FeðHN2ðprLÞðprLHÞÞðH2OÞ4 þ Hþ þ 2H2O
3 þ
-
2 þ
FeðH2OÞ6 þ N2ðprLÞðprLHÞ / FeðN2ðprLÞðprLHÞÞðH2OÞ4 þ 2H2O
β
13.56f
N3(etLH)3
2 þ
þ
H2N3ðetLHÞ3
/ HN3ðetLHÞ3 þ Hþ
Ka1
Ka2
Ka3
Ka4
Ka5
-3.79 ( 0.07c
-5.7 ( 0.1c
-7.50 ( 0.02c
-8.84 ( 0.03c
-10.40 ( 0.04c
1-27
1-28
1-29
1-30
1-31
þ
HN3ðetLHÞ3 / HN3ðetLÞðetLHÞ2 þ Hþ
-
HN3ðetLÞðetLHÞ2 / HN3ðetLÞ2ðetLHÞ þ Hþ
-
2 -
HN3ðetLÞ2ðetLHÞ / HN3ðetLÞ3 þ Hþ
2 -
3 -
HN3ðetLÞ3
/ N3ðetLÞ3 þ Hþ
3 þ
3 -
FeðH2OÞ6 þ N3ðetLÞ3
/ FeðN3ðetLÞ3ÞH1 - þ Hþ þ 6H2O
β11-1
KaFe2
β110
17.66 ( 0.09f
-9.68 ( 0.08c
27.34 ( 0.04e
5.3g
1-32
1-33
1-34
1-35
1-36
1-37
-
FeðHN3ðetLÞ3Þþ / FeðN3ðetLÞ3Þ þ Hþ
3 þ
3 -
FeðH2OÞ6 þ N3ðetLÞ3
/ FeðN3ðetLÞ3Þ þ 6H2O
FeðH2OÞ6 þ HN3ðetLHÞ3 / FeðHN3ðetLÞ3Þþ þ 3Hþ þ 6H2O
K1
3 þ
þ
3 þ
3 -
þ
FeðH2OÞ6 þ N3ðetLÞ3 þ Hþ / FeðN3ðetLÞ2ðetLHÞÞðH2OÞ2
β111
30.44 ( 0.08g
-3.10 ( 0.07f
FeðHN3ðetLÞ2ðetLHÞÞðH2OÞ2
/ FeðHN3ðetLÞ Þþ þ Hþ þ 2H2O
KaFe1
2 þ
3

8214 Inorganic Chemistry, Vol. 49, No. 18, 2010
Table 1. Continued
Harrington et al.
equilibriuma
K or β
log(K or β)b
eq.
3 þ
2 -
þ
FeðH2OÞ6 þ N3ðetLÞ2ðetLHÞ
/ FeðN3ðetLÞ2ðetLHÞÞðH2OÞ2 þ 4H2O
βtet
21.58f
1-38
1-39
1-40
1-41
3 þ
þ
2 þ
FeðH2OÞ6 þ HN3ðetLHÞ3 / FeðHN3ðetLÞ2ðetLHÞÞðH2OÞ2 þ 2Hþ þ 4H2O
K111
K112
βbis
8.4f
h
3 þ
þ
3 þ
FeðH2OÞ6 þ HN3ðetLHÞ3 / FeðHN3ðetLÞðetLHÞ2ðH2OÞ4Þ þ Hþ þ 2H2O
3 þ
-
2 þ
FeðH2OÞ6 þ N3ðetLÞðetLHÞ2 / FeðN3ðetLÞðetLHÞ2ðH2OÞ4Þ þ 2H2O
h
a Equilibria are shown as proton dissociation reactions except in log β111 reactions. Conditional equilibrium constants are given the symbol Kand written to show
the protonation state of the ligand in solution, including the protonation of the central ring system. The protonation state of the central ring is unimportant to the
equilibrium constants determined, as protons will only be released from the HOPO binding moieties of the molecule upon chelation. The overall stability constants,
log β, are shown as reactions involving completely deprotonated ligands, regardless of the protonation state of the central ring system, to simplify the representation
and calculation of the stability constants. b All values were determined at μ=0.10M,T=25°C. Errors shown in Table 1 for equilibrium constant values are derived
from a number of sources. For equilibrium constants determined directly or indirectly by competition reaction or spectrophotometric titration, the error represents
the standard deviation of the experimentally determined value. For equilibrium constants determined indirectly by linear combination of equilibria and their
equilibrium constants, the error was determined by propagation of error of the standard deviations of the summed equilibrium constants. c Values and estimated
error limits are from direct determination by spectrophotometric titration. Potentiometric titration results are in agreement within the error limits of the titrations.
d Not determined because of solubility problems, see text. e Indirect determination by competition reaction . f Indirect determination by linear combination of
experimental results. g Indirect determination by spectrophotometric titration. h Not determined because of lack of protonation constant; see text.
7.1, following eqs 8-10. The spectra measured during this
experiment are shown in Supporting Information, Figure S1A.
FeðHN3ðetLÞ3Þþ þ HEDTA3 - / FeEDTA-
-
þ HN3ðetLÞ2ðetLHÞ
ð8Þ
ð9Þ
-
½FeEDTA- ꢁ½HN3ðetLÞ2ðetLHÞ
½FeðHN3ðetLÞ3Þþꢁ½HEDTA3 -
ꢁ
Kobs
¼
ꢁ
!
- 1
Figure 3. Spectrophotometric measurement of the competition reaction
betweentheFe2(N2(prL)2)3 complexandEDTA. [Fe3þ] =2.5ꢀ 10-4 mol
dm-3,[N2(prLH)2] =3.8ꢀ 10-4 mol dm-3,pH=5.5-7.1, μ=0.10(NaCl),
[EDTA] = 0-2.5 ꢀ 10-3 mol dm-3. Arrow indicates the direction in
which the spectrum changes with addition of EDTA.
Kobs
β110
¼
ð10Þ
β110,EDTA
The stability constant, β110 was determined from the measured
spectra using the program, SPECFIT.
For the Fe/(N2(etLH)2) system, a stock solution of Fe(EDTA)
was prepared at [Fe3þ] = [EDTA] = 1.04 ꢀ 10-4 mol dm-3 and
separated into 2.0 mL aliquots. The competition study was con-
ducted by treating the stock aliquots of Fe(EDTA) complex with
concentrations of N2(etLH)2 ranging from 0 mol equiv to 10 mol
equiv of N2(etLH)2:Fe, as described by eqs 11 and 12, and the
solutions were allowed to equilibrate for 24 h before measuring
the UV-visible spectra (Supporting Information, Figure S1B)
Scheme 1. Synthesis of Electrophilic HOPO Imidate Salt, 3
FeðEDTAÞ þ H2N2ðetLHÞ22þ þ Hþ þ 2H2O
-
/ FeðH2N2ðetLÞ2ÞðH2OÞ23þ þ H3EDTA-
ð11Þ
ð12Þ
3þ
½FeðH2N2ðetLÞ2ÞðH2OÞ2 ꢁ½H3EDTA-
ꢁ
Kobs
¼
-
2þ
Information. The synthesis of an electrophilic 3,2-HOPO
imidate salt, 4 (Scheme 1), which can be used for the
facile incorporation of the HOPO moiety, was recently
disclosed.45 This reagent reacted with a variety of nucleo-
philes, including amines, to provide siderophores with a
three carbon tether between the nucleophile and the
HOPO ring without formation of an amide linkage.53
Of particular interest is the facile reaction of 4 with
secondary amines, which makes it an ideal reagent for the
attachment of HOPO binding groups to cyclic polyamines.
½FeðEDTAÞ ꢁ½H2N2ðetLHÞ2 ꢁ½Hþꢁ
and the final pH (2.3 for all solutions). Using the condition-
dependent equilibrium constant (eq 12) and the stability constant
of the Fe(EDTA) complex, it was possible to determine the β110
3þ
(eq 13) for the formation of Fe(H2N2(etL)2)(H2O)2 using
the program HYPERQUAD.51
β110 ¼ ðKobsÞðβ110,EDTAÞ
Results and Discussion
ð13Þ
HOPO Ligand Synthesis. Details of the HOPO ligand
synthesis and characterization are given in Supporting
(53) Chittamuru, S.; Lambert, T. N.; Martinez, G.; Jacobs, H. K.;
Gopalan, A. S. Tetrahedron Lett. 2007, 48, 567–571.

Article
Inorganic Chemistry, Vol. 49, No. 18, 2010 8215
Figure 4. Potentiometric titration of two hydroxypyridinone-Fe3þ systems. (A) Fe-N2(prLH)2 system. [Fe3þ] = 4.00 ꢀ 10-4 mol dm-3, [N2(prLH)2] =
6.0 ꢀ 10-4 mol dm-3, μ = 0.10 (NaCl). The trace on the left represents the ligand only titration ([N2(prLH)2] = 6.7 ꢀ 10-4 mol dm-3), while the trace onthe
right represents the Fe-N2(prLH)2 complex titration. (B) Fe(N3(etL)3) system. [Fe3þ] = 3.98 ꢀ 10-4 mol dm-3, [N3(etLH)3] = 4.0 ꢀ 10-4 mol dm-3, μ =
0.10 (NaCl). The trace on the left represents the ligand only titration ([N3(etLH)3] = 4.0 ꢀ 10-4 mol dm-3), while the trace on the right represents the
potentiometric titration performed in the presence of 1 equiv of metal.
Scheme 2. Synthesis of Siderophore Mimics N2(etLH)2 and N3(etLH)3
N2(prLH)2 and other polyHOPO chelators were pre-
pared using 4.45
after chromatographic purification (Scheme 2). Debenzyla-
tion with HBr/AcOH (1:1) gave the dihydroxypyridinone
N2(etLH)2 in 93% yield as its dihydrobromide salt. Similarly,
treatment of N,N0,N00-1,4,7-triazacyclononane (1 equiv) with
the mesylate salt 3 (4.5 equiv) and triethylamine (6 equiv) in
acetonitrile at room temperature led to the formation of
6 which underwent debenzylation with HBr/AcOH to give
N3(etLH)3 as a trihydrobromide salt (Scheme 2). It is
noteworthy that all three siderophore mimics, N2(etLH)2,
N2(prLH)2, and N3(etLH)3 are soluble and stable in water
over a wide pH range.
HOPO Ligand Protonation Constants. Protonation
constants for N2(etLH)2, N2(prLH)2, and N3(etLH)3
(Figure 2) in aqueous solution were determined by po-
tentiometric and spectrophotometric titration. Potentio-
metric titration data for the latter two are shown in Figure 4.
Representative spectrophotometric titration data for
N2(prLH)2 are shown in Figure 5. The spectral shifts
observed in the spectrophotometric titrations correspond
to the deprotonation reactions of the hydroxypyridinone
moieties of the molecules. As the hydroxypyridinone
groups are separate but undergoing identical deprotona-
tion reactions, the spectral shift appeared as a single large
shift corresponding to all hydroxypyridinone proton-
ation constants instead of three spectral shifts with
separate isosbestic points. In the N2(prLH)2 titration,
To prepare HOPO analogues in which the tether
between the nucleophile and the HOPO ring is two
carbons, the synthesis of HOPO imidate salt 3 was
developed (Scheme 1). Preparation of the imidate salt
3 was performed in a manner similar to the method
developed for the synthesis of the homologous imidate
salt, 4. Treatment of the known alcohol 1 with methane-
sulfonic anhydride in dichloromethane in the presence of
triethylamine gave the desired bicyclic HOPO imidate salt 3
along with some of the intermediate mesylate 2 (<10%) as
determined by 1H NMR spectral analysis. Complete conver-
sion to the bicyclic salt 3 was achieved by stirring the crude
product mixture from the mesylation in chloroform at room
temperature. After removal of the solvent, the product was
triturated with ethyl acetate. The desired HOPO imidate salt
3was isolated as a pale white solid in 92% yield in high purity.
The HOPO imidate salt 3 was found to be more reactive
than its homologue 4 when reacted with nucleophiles such as
amines, alcohols, phenols, thiols, and thiophenols.54 When
piperazine (1 equiv) was treated with the mesylate salt 3
(3 equiv) and triethylamine (4 equiv) in acetonitrile at room
temperature the protected HOPO 5was obtained in 97% yield
(54) Chittamuru, S.; Martinez, G.; Jacobs, H. K.; Gopalan, A. S.,
unpublished results.

8216 Inorganic Chemistry, Vol. 49, No. 18, 2010
Harrington et al.
observed for the 3-hydroxy-2-pyridinone donor group
alone, 8.66.14 It is interesting to note that the average pKa
values observed for the hydroxypyridinone donor groups
of N2(etLH)2 and N2(prLH)2, 6.80 and 6.92 respectively,
are lower than expected and in fact are much closer to the
pKas observed for retroHOPOs such as TRENHOPO.
For comparison, in the bisHOPO cyclen I the average
pKa of the HOPO donor groups is 8.57 This may be due to
steric or electrostatic considerations. The average pKa
separation for two identical acidic moieties is 0.60 in the
absence of any intramolecular interactions, less than the
separation observed here (1.4 for N2(etLH)2 and 2.0 for
N2(prLH)2). This suggests the influence of intramolecular
interactions on the pKa’s of the bishydroxypyridinone
molecules.58
Figure 5. Spectrophotometric titration of the synthetic exocyclic bishy-
droxypyridinone siderophore N2(prLH)2 with 0.010 mol dm-3 NaOH
over the pH range of 3.25 to 10.90. [N2(prLH)2] = 1.0 ꢀ 10-5 mol dm-3
,
The pKa values observed for N3(etLH)3 follow a similar
pattern as that observed for the bishydroxypyridinone
siderophores (Figure 2 and Table 1). pKa1 and pKa5
correspond to the central ring system and are significantly
separated because of electrostatic repulsion.59 The third
proton on the central ring has a pKa too low to be
observed, also because of steric interactions and electro-
static repulsion from the 2þ doubly protonated ring
system. As with the bishydroxypyridinone chelators, it
is not possible to make an unambiguous assignment of
protonation constants pKa2, pKa3, and pKa4 of the
N3(etLH)3 HOPO donor groups. However, it is possible
to assign pKa1 to the proton of the central ring that is
bound to a single amine group and pKa5 to the proton that
is initially accepting electron density from two amine
donors on the central ring. The hydroxypyridinone
pKa2-4 are 5.7, 7.5, and 8.84 (significant figures are
determined based on the standard deviation of the values,
as shown in Table 1). The observed pKa’s are separated by
more than the predicted separation for three identical
acidic moieties, 0.48 log units, implying a contribution
from intramolecular interactions.58 It is interesting that
the 3,2-HOPO protonation constants (average pKa =
7.3) are greater than 2 orders of magnitude more acidic
than the 3,2-HOPO protonation constants (average
pKa =9.5) of TRISPYR (Figure 1).22 This likely arises
because of electrostatic effects resulting from protonation
of the central ring system. The protonation constants are also
slightly more basic than those observed in TRENHOPO
(Figure 1), likely because of the lack of the amide oxygen
atoms in the structure of the HOPO donor group arms.15
Fe(III)-HOPO Complex Stability and Protonation Con-
stants. General Observations. Iron(III) may be bound by
the bidentate HOPO moieties in three modalities as mono,
bis, and tris (bidentate, tetradentate, and hexadentate) Fe-
(III)-HOPO complexes. These complexes exist in solution as
a dynamic equilibrium system where mono, bis, and tris
coordinated iron(III) complexes interconvert with changing
pH because of competition between Hþ and Fe3þ for the
HOPO oxygen sites (as shown in eqs 16-18, where HOPO-
μ = 0.10 (NaClO4), T = 25 °C. Arrows indicate the direction of spectral
changes upon addition of base.
isosbestic points were observed at 273 and 298 nm, indicat-
ing a transition between two light absorbing species in all
recorded spectra (Figure 5). A similar spectral shift was
observed in the N3(etLH)3 titration, with isosbestic points
observed at 276 and 300 nm (Supporting Information,
Figure S2A). In contrast, the N2(etLH)2 titration showed
similar spectral shifts as in the other two experiments, but
only a single isosbestic point was observed at 277 nm
(Supporting Information, Figure S2B).
The proton dissociation equilibria are described by
eqs 14 and 15 below. The measured deprotonation con-
stants for all three HOPO ligands are shown in Table 1
and Figure 2.
n - 3
n - 4
LHn
/ LHn - 1
þ Hþ
ð14Þ
n - 4
½LHn - 1
½LHn
ꢁ½Hþꢁ
Kan
¼
ð15Þ
n - 3
ꢁ
For the bishydroxypyridinones, N2(etLH)2 and N2(prLH)2,
four separate deprotonation constants were observed
(Figure 2 and Table 1), where pKa1 and pKa4 correspond
to the deprotonation of the ring N atoms and pKa2 and
pKa3 correspond to the hydroxypyridinone donor groups.
The deprotonation constants of the ring N atoms are
consistent with the observed literature values of [6]ane-N2
ring systems.55 The lower pKa (≈3.7) corresponds to loss
of a proton from the doubly protonated ring. This pKa is
greatly depressed because of electrostatic repulsion and
steric hindrance.56 The higher pKa (9-10) is typical for
the deprotonation of the singly protonated [6]ane-N2
ring. However, it is important to point out that pKa1
and pKa4 are structurally indistinguishable because of the
symmetry of the molecule. After deprotonation of the
first polyamine ring proton, the remaining proton will
shift to accept electron density from both amine groups.
The same is true for pKa2 and pKa3, as the deprotonation
of one HOPO group will be indistinguishable from the
other. The pKa values that have been assigned to the
hydroxypyridinone donor groups are lower than those
(57) Ambrosi, G.; Formica, M.; Fusi, V.; Giorgi, L.; Guerri, A.; Lucarini,
S.; Micheloni, M.; Paoli, P.; Rossi, P.; Zappia, G. Inorg. Chem. 2005, 44,
3249–3260.
(58) Noszal, B. In Biocoordination Chemistry: Coordination equilibria in
Biologically Active Systems; Burger, K., Ed.; Ellis Horwood: New York, 1990;
pp 18-55.
(59) Geraldes, C. F. G. C.; Alpoim, M. C.; Marques, M. P. M.; Sherry,
A. D.; Singh, M. Inorg. Chem. 1985, 24, 3876–3881.
(55) Khalili, F.; Henni, A.; East, A. L. L. J. Chem. Eng. Data 2009, 54,
2914–2917.
(56) Luckay, R.; Hancock, R. D.; Cukrowski, I.; Reibenspies, J. H. Inorg.
Chim. Acta 1996, 246, 159–169.

Article
Inorganic Chemistry, Vol. 49, No. 18, 2010 8217
Table 2. Wavelengths of Maximum Absorbance and Molar Absorptivities for Fe-HOPO Complexes in Aqueous Solutiona
complex
N3(etLH)3
N2(etLH)2
N2(prLH)2
3-hydroxy-2(1H)pyridinoneb
mono-coordinated
bis-coordinated
N/A
546
(2650 ( 20)
413
∼600c
∼600c
600
548
(3690)
415
547
(3580 ( 20)
419
545
(3650 ( 50)
417
(1760 ( 40)
507
(3820 ( 10)
417
(2400 ( 100)
(2300 ( 200)
503
(4910 ( 20)
418
(1880)
502
(5160)
417
(4080)
tris-coordinated
508
d
418
d
(3320 ( 10)
(4000 ( 20)
a Molar absorptivity values (in dm3 mol-1 cm-1) from an average of four measurements. b Reference 14. c Values were measured at high dilution and
may be inaccurate. d Values were not determined because of low solubility of complex at higher pH.
represents a generic anionic bidentate hydroxypyridinone
donor moiety).
2þ
FeðH2OÞ 3þ þ ðHOPOÞH / FeðHOPOÞðH2OÞ
6
4
þ Hþ þ 2H2O
ð16Þ
ð17Þ
ð18Þ
FeðHOPOÞðH2OÞ 2þ þ ðHOPOÞH
4
þ
/ FeðHOPOÞ2ðH2OÞ2 þ Hþ þ 2H2O
FeðHOPOÞ ðH2OÞ þ þ ðHOPOÞH
2
2
Figure 6. Spectrophotometric titration of the Fe-N2(prLH)2 system,
pH 3.5 to 7.5. [Fe3þ] = 1.99 ꢀ 10-4 mol dm-3, [N2(prLH)2] = 3.0 ꢀ 10-4
mol dm-3, μ = 0.10, T = 25 °C. Arrows indicate the direction of spectral
changes upon addition of base.
/ FeðHOPOÞ3 þ Hþ þ 2H2O
Changes in the Fe(III) inner coordination sphere with pH
produce observable changes in the UV-visible spectrum of
the solution. The spectral changes observed in this study are
in agreement with those previously reported for Fe(III)-3,2-
HOPO systems (Table 2).14
The determined equilibrium constants for all possible
complexation and deprotonation reactions for all three
hydroxypyridinone siderophore mimics are shown in
Table 1. The values presented were obtained from a
combination of direct determination, and indirect deter-
mination from competition experiments. The equilibria
determined through direct and indirect methods are
related through linear combinations of constants deter-
mined by other methods and can be used as an internal
check for consistency in our equilibrium model.
potentiometric and spectrophotmetric data (see Support-
ing Information, Figure S3B) were obtained for the Fe-
(III)-N2(etLH)2 system. The observed changes over the
pH range of ∼0.2 to 2.5 correspond to conversion of
the complex from the bis-coordinate to tetra-coordinate
complex, eqs 1-10 and 1-23. Using the spectra measured
2þ
at low pH, the stability constant of the FeXH(H2O)4
complex (where X = N2(etLH)2, shown in eq 1-11 or
where X = N2(prLH)2, shown in eq 1-24) was calcu-
lated. The value obtained indirectly for the pH indepen-
dent stability constant of the bishydroxypyridinone com-
plexes, log β110, calculated by spectrophotometric com-
petition titratþions (Figure 3) was 20.45 for Fe(HN2-
(prL)2)(H2O)2 (eq 1-22) and log β110 = 21.08 for Fe-
(HN2(etL)2)(H2O)2þ (eq 1-8).
Bishydroxypyridinone Ligands: N2(prLH)2 and N2-
(etLH)2. Initial complex formation occurs via deprotona-
tion of one hydroxypyridinone donor moiety, which
coordinates to the metal center via displacement of two
water ligands, as shown in eqs 1-12 and 1-25 in Table 1
(all eqs 1-X designations refer to equilibria listed in
Table 1). A gradual increase in pH leads to deprotona-
tion of the second hydroxypyridinone donorþgroup and forma-
tion of the tetracoordinate FeX(H2O)2 complex (X =
N2(prLH)2 or N2(etLH)2), shown in eqs 1-10 and 1-23.
The potentiometric titration of the Fe(III)-N2(prLH)2
system (Figure 4A) showed an initial three-proton buffer
region, one proton corresponding to the removal of a
single proton from the central ring system and the other
two corresponding to the deprotonation of the two
HOPO donor groups upon chelation. Figure S3A in the
Supporting Information shows the low pH spectrophoto-
metric titration of the Fe(III)-N2(prLH)2 system. Similar
At low pH, N2(prLH)2 and N2(etLH)2) exhibit similar
iron(III) complexation equilibria forming Fe(HN2-
(prL)2)(H2O)2 and Fe(HN2(etL)2)(H2O)2þ. However,
þ
at pH values above 3, their aqueous solution behavior
diverges, as discussed below. In the Fe(III)-N2(prLH)2
system, as the pH increased above where the tetradentate
species predominates, a further deprotonation of the free
ligand in solution occurred as evidenced by a second
buffer region observed in the potentiometric titration
(Figure 4A). This second buffer region, from 3 to 5 equiv
of base added, corresponds to the formation of the
hexacoordinate complex Fe2(HN2(prL)2)33þ, as shown
in eq 1-20. Spectrophotometric titration of the Fe(III)-
N2(prLH)2 system over the pH range of 3.5 to 7.5 exhibits
a spectral change with an isosbestic point at 557 nm
(Figure 6), indicating a single protonation equilibrium.

8218 Inorganic Chemistry, Vol. 49, No. 18, 2010
Harrington et al.
experiments (Supporting Information, Figure S1B) and
low pH spectrophotometric titrations (Supporting Infor-
mation, Figure S3). The species distribution diagram for
the Fe(III)-N2(etLH)2 system was not calculated because
of the insolubility of the complex at pH>3.
The tetradentate chelators N2(etLH)2 and N2(prLH)2
can reasonably encapsulate a single Fe3þ ion or form a
bridge between two Fe3þ ions. This impacts the possible
structures for the 1:1 and 2:3 Fe:X (X = ligand) stoichi-
ometry complexes described here. Specifically, the 1:1
complex can exist in monomeric form (FeX(H2O)2þ) or
2þ
as a doubly bridged dimer (H2O)2Fe(μ-X)2Fe(H2O)2
(X = N2(etLH)2 or N2(prLH)2). The coordinatively
saturated 2:3 complex may exist as a singly ((X)Fe(μ-
X)Fe(X)) or triply bridged (Fe(μ-X)3Fe) species. To ad-
dress this question we used ESI-MS to characterize solu-
tions of complexes at the Fe:X ratio and pH conditions
where the only species present was FeX as indicated by
the speciation plot. The isotopic ratio of 56Fe (91.18%
natural abundance) to 57Fe (2.1% natural abundance) is a
useful tool to probe the nature of the iron complexes.
There is a peak observed in the Fe(III)-N2(etLH)2 ESI-
MS spectrum that corresponds to the 1:1 complex with-
out coordinated water molecules, Fe(N2(etL)2)þ (m/z =
414.1, Supporting Information, Figure S5). The isotopic
peaks of the 1:1 complex signal are separated by a whole
m/z unit, which implies that the 1:1 complex is present iþn
solution as the monomeric form Fe(N2(etL)2)(H2O)2
and not as the ligand-bridged dimer form (see below). A
similar isotopic separation is observed in the mass spec-
trum of the Fe(III)-N2(prLH)2 1:1 L:Fe complex system,
where the 442.1 m/z peak represents Fe(N2(prL)2)þ with-
out coordinated water molecules and exhibits an isotopic
Figure 7. Species distribution diagram for the completely characterized
systems of N2(prLH)2 and N3(etLH)3 in aqueous solution, showing
percent total Fe3þ vs pH of solution. Values were calculated from deter-
mined stability constants of complexes and hydrolysis constants for Fe3þ
.
T =25°C, μ =0.10. (A) Fe-N2(prLH)2 system, with [Fe3þ
]
total =2.00ꢀ
10-4 mol dm-3, and [N2(prLH)2]total = 3.00 ꢀ 10-4 mol dm-3. a = free
Fe3þ, b = Fe(HN2(prL)2)(H2O)2 complex, c = Fe2(HN2(prL)2)3
2þ
3þ
complex, d = Fe(OH)4-, e = Fe(HN2(prLH)(prL))3þ, and f = Fe-
(OH)2þ. (B) Fe-N3(etLH)3 system, [Fe3þ] = 1.00 ꢀ 10-4 mol dm-3
,
[N3(etLH)3] = 1.00 ꢀ 10-4 mol dm-3. a = free Fe3þ, b = Fe(HN3-
(etL)2(etLH))2þ, c = Fe(HN3(etL)3)þ, d = Fe(N3(etL)3, e = Fe(OH)4
,
-
f = Fe(OH)2þ, and g = Fe(OH)2
.
þ
The observed shift in λmax is consistent with conversion of
the tetracoordinate Fe(HN2(prL)2)(H2O)2 complex to the
hexacoordinate Fe2(HN2(prL)2)3 complex (Table 2 and
eq 1-20).14 The equilibrium constant for eq 1-18 is
sufficiently high that direct measurement is not possible.
Consequently, an indirect stability constant determina-
tion by competition with EDTA (see Experimental
Section) yielded a log β230 value of 60.46 for N2(prLH)2
(eq 1-19). The species distribution diagram for the
Fe(III)-N2(prLH)2 system (Figure 7A) shows that at the
chosen concentrations the Fe(III) is present as a hexa-
peak separation of 1 m/z unit, also suggesting that the
complex is present in aqueous solution as the monomeric
þ
FeX(H2O)2 form. If the complexes were found in the
ligand bridged dimeric form, (H2O)2Fe(μ-X)2Fe(H2O)2
2þ
,
the observed isotopic separation would be 0.5 m/z unit, as
the difference in mass-to-charge ratio of the complex
arising from the presence of a single 57Fe metal center
with the same overall charge of the complex would be 0.5
m/z unit. No peak that corresponds to a possible dimer is
observed in the mass spectrum. Peak intensities provide
additional evidence for the assignment of the peaks to the
FeXþ complex. Theoretical calculations of the relative
peak intensity of the FeXþ1þ peak for the Fe(N2(prL)2)þ
complex show that the corresponding isotopic peak
should represent 23.5% of the entire peak intensity for
all observed isotopes. The observed relative intensity of
the FeXþ1þ peak for the Fe(N2(prL)2)þ system was
24.6%, showing very good agreement with the predicted
isotopic splitting pattern. A similar calculation for the
Fe(N2(etL)2)þ system predicted 21.5% of the total
peak intensity, and the observed relative intensity of the
coordinate Fe2(HN2(prL)2)3
range of approximately 5 to 10.
complex over the pH
3þ
Similar behavior is observed in the Fe(III)-N2(etLH)2
system shown in eqs 1-5 through 1-7; however, the
Fe2(HN2(etL)2)3 complex precipitates above pH 3, mak-
ing it difficult to accurately determine log β 230. The
complex λmax can be estimated from the remaining ab-
sorbance, as shown in Table 2. Comparison to the deter-
mined λmax values of the Fe(III)-N2(prLH)2 complex
shows that the Fe(III)-N2(etLH)2 complex exhibits a
similar inner coordination sphere as the N2(prLH)2 com-
plex at pH 7. Equilibrium constants for eqs 1-10 through
1-13 involving chelation by a single hydroxypyridinone
donor group were calculated through EDTA competition

Article
Inorganic Chemistry, Vol. 49, No. 18, 2010 8219
FeXþ1þ peak was 20%, also showing very good agree-
ment with the predicted isotopic splitting pattern.
On the basis of these observations, we propose that the
iron(III) sequestration equilibria for the tetradentate HOPO
ligands N2(etLH)2 and N2(prLH)2 proceed through the
formation of the following structures (R = et or pr).
Table 3. Calculated pFe Values for a Number of Siderophores
pFea
ligand
1-methyl-3-hydroxypyridin-2-one
Deferiprone
Deferasirox
Transferrin
N2(prLH)2
3,4-LI-(Me-3,2-HOPO)
N3(etLH)3
Deferrioxamine B
TRENHOPO
HOPObactin
16.0b
19.4c
23.5d
23.6e
24.78f
25.5g
26.5f
26.6h
26.7i
27.4j
27.6k
32.23l
þ
FeðH2OÞ63þ þ HN2ðRLHÞ2
/ FeðHN2ðRLÞ2ÞðH2OÞ22þ þ 2Hþ þ 4H2O ð19Þ
CP130
TRISPYR
þ
2FeðHN2ðRLÞ2ÞðH2OÞ22þ þ HN2ðRLHÞ2
a pH = 7.4, [Fe3þ] = 1.0 ꢀ 10-6 mol dm-3, [L] = 1.0 ꢀ 10-5 mol
3þ
/ ðHN2ðRLÞ2ÞFeðμ-HN2ðRLÞ2ÞFeðHN2ðRLÞ2Þ þ 2Hþ
dm-3 b Ref 63. c Ref 64. d Ref 65. e Ref 66. f This work. g Ref 16. h Ref 67.
.
i Ref 40. j Ref 15. k Determined using equilibrium constants and proton-
ation constants reported in ref 61. l Determined using equilibrium
constants and protonation constants reported in ref 22.
ð20Þ
Trishydroxypyridinone Ligand: N3(etLH)3. Initial com-
plex formation of the trishydroxypyridinone (N3(etLH)3)
complex occurs in a similar manner as the bishydroxy-
pyridinone complexes, where addition of iron(III) to the
ligand results in displacement of two protons to form the
tetracoordinate complex (eq 1-39, Table 1). This com-
plex is formed at a low pH, and the two deprotonation
steps are not resolvable, as evidenced by the low pH
spectrophotometric and potentiometric titrations. Over
the pH range of ∼0.2 to 2.5, the spectra that are observed
correspond to the formation of the tetracoordinate com-
plex without observing λmax values that correspond to the
biscoordinate complex (Supporting Information, Figure
S3C). Also, the potentiometric titration (Figure 4B) ex-
hibits an initial 4-proton buffer region that corresponds
to the removal of a proton from the central ring, followed
by the removal of two protons from hydroxypyridinone
donor groups to form the tetracoordinate complex and
subsequent removal of a third hydroxypyridinone proton
to form the trishydroxypyridinone complex at higher
solution pH (eqs 1-39 and 1-37). Spectrophotometric
titration of the complex characterized the transition from
slowly dissociate, as evidenced by the gradual decrease of
the spectral intensity to the baseline.
The species distribution diagram for the Fe(III)-N3-
(etLH)3 system as a function of pH is shown in Figure 7B.
The plot shows that the Fe(III) is present exclusively as a
hexadentate Fe(HN3(etL)3)þ complex over the pH range
of approximately 5 to 8.
Comparison of Bis- and Tris-HOPO Ligands. A de-
monstration of the internal consistency of our solution
equilibria model shown in Table 1 can be made by com-
paring the stability constants of complexes of similar
denticity. Through linear combination of the protona-
tion constant corresponding to the first hydroxypyridi-
none moiety of N3(etLH)3 (eq 1-30) and the log β111
(eq 1-36), one can obtain the equilibrium and stability
constant shown in eq 1-38. The equilibrium constant for
reaction of the doubly deprotonated trishydroxypyridi-
none ligand N3(etLH)3 with iron(III) is log βtet = 21.58
(eq 1-38), which is very close to the log β110 of N2(etLH)2,
21.08 (eq 1-8). The small increase in the log βtet of
N3(etLH)3 compared to the log β110 of N2(etLH)2 may
be due to a chelate effect in N3(etLH)3 or more favorable
steric factors. The similarity of the two values demon-
strates that our equilibrium constant determinations are
consistent between ligand systems. A similar comparison
2þ
Fe(HN3(etL)2(etLH))(H2O)2 to Fe(HN3(etL)3)þ over
the pH range of ∼2.5 to 8, exhibiting a spectral change
consistent with the deprotonation of the third donor
group to form the hexacoordinate trishydroxypyridinone
complex (Supporting Information, Figure S4A). The
isosbestic point observed is indicative of a single proton-
ation equilibrium, suggesting deprotonation of a hydro-
xypyridinone donor group over that pH range (eq 1-37).
The measured stability constant for N3(etLH)3 was found
by indirect determination via EDTA competition to be
log β110 = 27.34, eq 1-34.
Between pH 8 and 10.4, another deprotonation event,
involving the dissociation of a fourth proton is ob-
served (Supporting Information, Figure S4B) and exhi-
bits an isosbestic point at λmax=391 nm. The most likely
assignment of this deprotonation event is deproto-
nation of the central polyamine ring (eq 1-33). This is
evidenced by the small changes in absorbance and
may be made for the N2(prLH)2 system, where log β110
20.45 (eq 1-22).
=
pFe values are a useful criterion to compare iron
chelators of different denticities and pKa’s. Table 3 con-
tains pFe values for N3(etLH)3 and N2(prLH)2, along
with a representative group of comparable chelators
listed in ascending order. The high Fe(III) affinity of
N3(etLH)3 and N2(prLH)2 is reflected by their large pFe
values, which are higher than 1-methyl-3-hydroxypyri-
din-2-one, consistent with their higher denticity and a
modest chelate effect. The pFe value observed for N3-
(etLH)3 is slightly less than the other tris-HOPO chelators
in Table 3, suggesting that the triazacyclononane back-
bone does not confer significant additional stabilization
in the arrangement of the HOPO chelating groups for
iron sequestration.
λ
max which suggest a minimal change in the inner coordi-
nation sphere of the iron(III). Further, this is consistent
with the pKa, for reaction 1-33, which is ∼0.7 log units
lower than the pKa assigned to the less acidic proton
dissociation (pKa = 10.4) of the central ring system of
H2N3(etLH)32þ. Above pH 10.4, the complex began to
Comparison of the stability of the hexacoordinate
complex of N3(etLH)3 with other HOPO chelators can
give an indication of the structural factors that are of

8220 Inorganic Chemistry, Vol. 49, No. 18, 2010
Harrington et al.
importance in complex formation. Interestingly, compar-
ison of the stability constants would suggest the lack of
chelate effect in these macrocyclic hydroxypyridinone
siderophore mimics. The measured log β110 for N3-
(etLH)3 (27.34) can be compared to the log units of
complex stability per iron center of the Fe2(N2(etL)2)3
complex(log β2301/2 = 30.23), which would suggest higher
stability in the bis-hydroxypyridinone complex. How-
ever, using the stability constant β values as the standard
of comparison ignores the influence that ligand proton-
ation constants play in aqueous solution complex forma-
tion. Therefore, in assessing the effective presence of the
chelate effect, it is more appropriate to consider the ligand
pFe values. Comparison of the pFe values determined for
the N3(etLH)3 system (26.5) to the N2(prLH)2 system
(24.78) seems to suggest a chelate effect in these exocyclic
polyhydroxypyridinone molecules. This may have to do
with the structure of the central ring systems as well as
different length spacer arms (N3(etLH)3 has a two carbon
chain spacer and N2(prLH)2 has a three carbon chain
spacer), which may contribute to the difference in stabi-
lity. The determined value of log β110 for N3(etLH)3 also
agrees very well with a previously reported value of log
HOPO system with amide linkages in the backbone.
Structurally, the closest siderophore mimic to N3(etLH)3
is the retroHOPO 3,4-LI-(Me-3,2-HOPO), pFe 25.5, in
which the 3,2-HOPO donor groups are linked through
amide bonds to an acyclic polyamine platform. Compari-
son of the pFe values for these chelators suggests that the
amide bonds are not critical for iron binding and the
overall structure motif/design is more important.
One can also compare N3(etLH)3 to CP130 to gain
insight into the effect of amide bonds on complex stabi-
lity. The calculated pFe value for CP130 using the re-
ported protonation constants and the equilibrium
constant for iron(III) chelation is 28.8, which is signifi-
cantly higher than that determined for N3(etLH)3. It is
possible that the longer arms of CP130 allows for more
stable chelation of iron(III) than does the exocyclic
architecture of N3(etLH)3 because of less steric strain
upon binding. The hydroxypyridinone protonation con-
stants of N3(etLH)3 are slightly more acidic than those of
CP130, suggesting that the acidity of the donor groups in
N3(etLH)3 do not provide enough of a thermodynamic
advantage to the exocyclic siderophore mimic to over-
come the advantage of the less sterically constrained
tripodal structure of CP130. A small amount of this
difference in stability may also be an artifact of the
methods used to calculate pFe values. The protonation
constants of the Fe-CP130 complex were not determined,
meaning that any protonated form of the complex that
would be present in the system are ignored when calculat-
ing the complex speciation, which suggest that the pFe
values as calculated feature a small error. However, as
pFe values are calculated at pH 7.4, it is likely that the
protonated forms of the complexes will constitute a
relatively minor fraction of total complexed iron at those
conditions.
Finally, the pFe values for the chelators prepared in this
study make them viable candidates for application as iron
overload drugs. Their pFe values are greater than trans-
ferrin and superior to the currently approved iron chela-
tion therapy agents Deferasirox and Deferiprone. The
pFe value for N3(etLH)3 is comparable to that of Desferal
(desferrioxamine B), the other FDA approved thera-
peutic for this application. However, there are multiple
factors involved in the effectiveness of a chelating agent in
iron chelation therapy.62 Less favorable thermodynamics
of iron chelation may be overcome by increasing chelator
plasma concentration, and the kinetics of removal of iron
from transferrin is also related to the efficiency of a
molecule as an iron chelation therapy agent. Previous
studies have shown that hydroxypyridinones can rapidly
remove iron from transferrin, suggesting that the syn-
thetic HOPO siderophores studied here exhibit iron
β
110 = 27.6 for a tripodal tris-hydroxypyridinone side-
rophore mimic HOPOHL with Fe(III) (Figure 2).60 It is
interesting to note that HOPOHL features amide oxygens
in the donor group connector arms. This suggests that the
absence of amide bonds in N3(etLH)3 does not adversely
affect the Fe(III) complex stability.
It is additionally important to acknowledge the poten-
tial contribution of amide donor groups to the complex
stability in some cases. As mentioned previously, amide
groups in the spacer arms of the ligand will decrease the
aqueous solubility of the molecule. However, at the same
time, amide oxygens can contribute to complex stability
through the formation of hydrogen bonding networks,
resulting in preorganization of the binding cavity. One
way of assessing the effect of the amide linkages on
complex stability in similar molecules is to compare
ligand structures where one features the amide linkages
and the other lacks them. Comparison of the pFe values
of TRISPYR (pFe = 32.23) and CP130 (pFe = 27.6)
(Figure 1), both hexadentate trishydroxypyridinone side-
rophore mimics, shows more stable chelation by TRIS-
PYR, which lacks the amide functional groups.22,61 It is
possible that in CP130 the amide moieties result in less
flexible donor group arms and more unfavorable steric
interactions than in TRISPYR. Conversely, the donor
arms of TRISPYR do not feature amide functional
groups, so the less rigid arms and decreased steric inter-
actions result in more stable complex formation with
iron(III). The thermodynamic characterizations of these
two chelators were conducted in different solvent sys-
tems; however, such a great difference in stability is not
likely to be entirely due to solvent effects.
(62) Turcot, I.; Stintzi, A.; Xu, J.; Raymond, K. N. J. Biol. Inorg. Chem.
2000, 5, 634–641.
(63) Liu, Z. D.; Hider, R. C. Coord. Chem. Rev. 2002, 232, 151–171.
(64) Liu, Z. D.; Khodr, H. H.; Liu, D. Y.; Lu, S. L.; Hider, R. C. J. Med.
Chem. 1999, 42, 4814–4823.
(65) Steinhauser, S.; Heinz, U.; Bartholoma, M.; Weyhermuller, T.; Nick,
H.; Hegetschweiler, K. Eur. J. Inorg. Chem. 2004, 2004, 4177–4192.
(66) Martin, R. B.; Savory, J.; Brown, S.; Bertholf, R. L.; Wills, M. R.
Clin. Chem. 1987, 33, 405–407.
(67) Harris, W. R.; Carrano, C. J.; Raymond, K. N. J. Am. Chem. Soc.
1979, 101, 2722–2727.
A similar comparison is not necessarily possible for
N3(etLH)3, as a suitable amide containing HOPO model
system has not been studied. No binding constants have
been reported for the TACN-1-Me-3,2-HOPO, a retro-
(60) Katoh, A.; Kudo, H.; Saito, R. Heterocycles 2005, 66, 285–297.
(61) Streater, M.; Taylor, P. D.; Hider, R. C.; Porter, J. B. J. Med. Chem.
1990, 33, 1749–1755.

Article
Inorganic Chemistry, Vol. 49, No. 18, 2010 8221
binding capabilities in the range that could be of interest
in the development of treatments for iron overload.62
use as iron chelation therapy agents. This could prove useful
in the further investigation of new designs for chelation
therapy agents.
Conclusions
Acknowledgment. A.L.C. thanks the NSF for financial
support (CHE 0809466), the Duke Center for Biomolecular
and Tissue Engineering for partial support of J.M.H., and
E. M. Tristani and R. D. Hancock for helpful discussions.
A.S.G. thanks the National Institutes of Health (PHS
Grants S06 GM08136 and 1SC3GM084809-01).
The investigation of synthetic siderophores can be useful in
determining factors of ligand design that lead to increased
stability and selectivity in siderophores for Fe(III). This
can, in turn, be used for developing new molecules for the
treatment of iron overload diseases. One type of binding
group that shows promise as a potential candidate for
chelation therapy agents is the HOPO donor group. Meth-
odology for the facile synthesis of three synthetic HOPO
donor group siderophores has been developed. Further, these
synthetic siderophores have shown the ability to form stable
complexes with iron at physiological pH, with relative
stability comparable to other molecules that are currently in
Supporting Information Available: General synthetic meth-
ods, procedures and characterization of all compounds; details
of spectrophotometric titration methods and selected spectro-
photometric titrations; ESI-MS spectra for bishydroxypyridi-
none-Fe(III) complexes. This material is available free of charge
via the Internet at http://pubs.acs.org.