Allosteric effects in polynuclear triple-stranded ferric complexes

Article abstract of DOI:10.1021/ja962472c

The synthesis and iron(III) coordination properties of three tripodal ligands (L¹, L², and L³) possessing hydroxamate coordination cavities are examined by various methods (ESMS, UV- vis, CD). The ligands rely on a trisamine as anchor, which is extended by an alternating sequence of variable spacers and hydroxamates as ion binding groups. This modular strategy of design is adopted for the compounds' preparation and enables modifications of each structural element independently. The coordination properties of these iron binding molecules and particularly the presence of allosteric effects are examined by classical spectrophotometric titrations in combination with electrospray mass spectrometric measurements (ESMS). A good match between these two methods is observed, as both indicate the formation of three species in thermodynamic equilibrium: mononuclear, binuclear, and trinuclear ferric complexes. The respective stability constants are determined at p[H] = 6.5 ± 0.1 in methanol, and the corresponding distribution curves clearly illustrate the variations from ligand to ligand. These findings demonstrate that subtle structural changes have a pronounced effect on these compounds' coordination properties. Moreover, among the binders studied representatives of opposite cooperative behavior is identified. The observed dependence of the ligands' coordination properties on their structural features are rationalized.

Full text of DOI:10.1021/ja962472c

4934
J. Am. Chem. Soc. 1997, 119, 4934-4944
Allosteric Effects in Polynuclear Triple-Stranded Ferric
Complexes
Sylvie Blanc,† Pnina Yakirevitch,‡ Emmanuelle Leize,§ Michel Meyer,
Jacqueline Libman,‡,| Alain Van Dorsselaer,§ Anne-Marie Albrecht-Gary,† and
Abraham Shanzer*,‡
Contribution from the Laboratoire de Physico-Chimie Bioinorganique, CNRS URA 405, ECPM,
1 rue Blaise Pascal, 67000 Strasbourg, France, Department of Organic Chemistry,
Weizmann Institute of Science, RehoVot 76100, Israe¨l, and Laboratoire de Spectrome´trie de
Masse Bio-Organique, CNRS URA 31, Faculte´ de Chimie, 1 rue Blaise Pascal,
67000 Strasbourg, France
ReceiVed July 17, 1996. ReVised Manuscript ReceiVed December 5, 1996X
Abstract: The synthesis and iron(III) coordination properties of three tripodal ligands (L1, L2, and L3) possessing
hydroxamate coordination cavities are examined by various methods (ESMS, UV-vis, CD). The ligands rely on a
trisamine as anchor, which is extended by an alternating sequence of variable spacers and hydroxamates as ion
binding groups. This modular strategy of design is adopted for the compounds’ preparation and enables modifications
of each structural element independently. The coordination properties of these iron binding molecules and particularly
the presence of allosteric effects are examined by classical spectrophotometric titrations in combination with
electrospray mass spectrometric measurements (ESMS). A good match between these two methods is observed, as
both indicate the formation of three species in thermodynamic equilibrium: mononuclear, binuclear, and trinuclear
ferric complexes. The respective stability constants are determined at p[H] ) 6.5 ( 0.1 in methanol, and the
corresponding distribution curves clearly illustrate the variations from ligand to ligand. These findings demonstrate
that subtle structural changes have a pronounced effect on these compounds’ coordination properties. Moreover,
among the binders studied representatives of opposite cooperative behavior is identified. The observed dependence
of the ligands’ coordination properties on their structural features are rationalized.
Introduction
helicates are much more abundant. Ru¨ttimann et al.8 reported
a dinuclear helicoidal complex with copper(I) and benzimida-
zole-pyridine. Lehn and co-workers synthesized double-
stranded (oligobipyridine)silver(I)9 and -copper(I)10 helicates
including up to five copper(I) ions.11-13 Dietrich-Buchecker
and Sauvage14 used a double helicoidal copper(I) complex with
two oligophenanthroline ligands as a precursor for the synthesis
of the first molecular knots. Potts et al.15 studied the occurrence
of interactions in double-stranded dicuprous helicates derived
from terpyridines. A stable trinuclear double helicoidal copper-
(II) complex of quinquepyridine was structurally characterized.16
Constable et al.17-20 used the oligobipyridine motif to form
homobinuclear helicoidal complexes with second- and third-
Allostery and cooperativity are interrelated phenomena that
play a major role in biological systems.1-3 They regulate the
activity of enzymes and the recognition properties of receptors,4
and occur when occupation of a given binding site changes the
complexation characteristics of the other binding site either by
improving or by reducing its binding efficiency.5 It occurred
to us that triple-stranded ligands could serve as ideal models
for the study of allosteric properties in helicoidal complexes
since they may be designed to possess distinct coordination sites
at predetermined positions.
A large variety of helicates have been described that differ
in their topology, the number of their ligands and of their guest
ions, and the nature of the metals and their properties. When
homopolynuclear helicates are classified according to the
number of ligands that are wrapped around the metal centers,
the first example consists of a single-stranded helicoidal complex
that possesses two ruthenium(II) ions.6,7 Double-stranded
(6) Thummel, R. P.; Herry, C.; Williamson, D.; Lefoulon, F. J. Am.
Chem. Soc. 1988, 110, 7894-7896.
(7) Cathey, L. J.; Constable, E. C.; Hannon, M. J.; Tocher, D. A.; Ward,
M. D. J. Chem. Soc., Chem. Commun. 1990, 621-622.
(8) Ru¨ttiman, S.; Piguet, C.; Bernardinelli, G.; Bocquet, B.; Williams,
A. F. J. Am. Chem. Soc. 1992, 114, 4230-4237.
(9) Garrett, T. M.; Koert, U.; Lehn, J. M. J. Phys. Org. Chem. 1992, 5,
529-532.
* To whom correspondence should be addressed.
(10) Pfeil, A.; Lehn, J. M. J. Chem. Soc., Chem. Commun. 1992, 838-
839.
† CNRS URA 405.
‡ Weizmann Institute of Science.
(11) Lehn, J. M.; Rigault, A.; Siegel, J.; Harrowfield, J.; Chevrier, B.;
Moras, D. Proc. Natl. Acad. Sci. U.S.A. 1987, 84, 2565-2569.
(12) Zarges, W.; Hall, J.; Lehn, J. M.; Bolm, C. HelV. Chim. Acta 1991,
74, 1843-1851.
§ CNRS URA 31.
Present address: Department of Chemistry, University of California,
Berkeley.
| Deceased March 1997.
(13) Schoentjes, B.; Lehn, J. M. HelV. Chim. Acta 1995, 78, 1-12.
(14) Dietrich-Buchecker, C. O.; Sauvage, J. P. Angew. Chem., Int. Ed.
Engl. 1989, 28, 189-192.
X Abstract published in AdVance ACS Abstracts, April 15, 1997.
(1) Lehn, J. M. Supramolecular Chemistry: Concepts and PerspectiVes;
VCH: Weinheim, 1995.
(15) Potts, K. T.; Keshavarz-K, M.; Tham, F. S.; Abrun˜a, H. D.; Arana,
C. R. Inorg. Chem. 1993, 32, 4450-4456.
(2) Koshland, D. E.; Neet, K. E. Annu. ReV. Biochem. 1968, 37, 359-
410.
(16) Baker, A. T.; Graig, D. C.; Dong, G. Inorg. Chem. 1996, 35, 1091-
1092.
(3) Hammes, G. G.; Wu, C. W. Annu. ReV. Biophys. Bioeng. 1974, 3,
1-33.
(17) Constable, E. C.; Ward, M. D. J. Am. Chem. Soc. 1990, 112, 1256-
1258.
(18) Constable, E. C.; Ward, M. D.; Tocher, D. A. J. Chem. Soc., Dalton
Trans. 1991, 1675-1683.
(4) Monod, J.; Changeux, J. P.; Jacob, F. J. Mol. Biol. 1963, 6, 306-
329.
(5) Perlmutter-Hayman, B. Acc. Chem. Res. 1986, 19, 90-96.
S0002-7863(96)02472-9 CCC: $14.00 © 1997 American Chemical Society

Allosteric Effects in Triple-Stranded Ferric Complexes
J. Am. Chem. Soc., Vol. 119, No. 21, 1997 4935
row transition metal elements (Cu(II), Fe(II), Pd(II), Ru(II),
Cd(II), Ni(II), Mn(II)). Recently, a synthetic analog of a
biscatecholate siderophore produced by the nitrogen-fixing
bacterium Azotobacter Vinelandii was shown to form a double
helical dioxomolybdenum(VI) complex.21 The diferric complex
formed at neutral pH with rhodotorulic acid, a ketopiperazine
bridged dihydroxamic siderophore, was the first characterized
triple helicate.22,23 The propeller-like arrangement of bidentate
chelating subunits in an octahedral environment has inspired
the design of various self-assembling systems forming triple
helicates. Several rhodotorulic acid mimics have been synthe-
sized by using 1,2-hydroxypyridinone,24 catechol,25,26 or tere-
phthalamide27 binding groups which favor hard metal cations
like iron(III) or gallium(III). Examples of multifunctional
strands incorporating soft donor atoms have also been
reported.28-33 This approach has been extended to ions of higher
coordination numbers (i.e., 9) and tridentate chelating subunits
by Piguet et al.,34-37 who synthesized a rich variety of
trihelicates with the aim to develop stable luminescent lanthanide
complexes that could serve as light converters.
The helicoidal systems described above illustrate the rich
variety of these structures and their possible applications for
the generation of functional, supramolecular assemblies, mostly
generated from single-stranded, symmetric ligands. However,
attachment of ligands to a tripodal anchor opens the possibility
to predetermine the coordinating binding cavities and to
drastically reduce the number of possible complexed species,
particularly when using nonsymmetric ligands. This approach
is synthetically more demanding, because it necessitates covalent
linkage of the ligating chains to a common anchor. Yet, the
use of tripodal, asymmetric ligands enables full physicochemical
characterization of these ligands’ coordination properties,
particularly the unambiguous characterization of each of the
ligands’ ion binding cavity, the preparation of site-directed,
heteronuclear complexes, and the generation of assemblies with
anisotropic properties. Motivated by these considerations, we
initiated a program aimed at the synthesis of chiral, triple-
Figure 1. Chemical formulas of the tripodal hydroxamate ligands: (a)
L1, m ) 0, R2 ) H; (b) L2, m ) 1, R1 ) i-Bu, R2 ) H; L3, m ) 1, R1
) i-Bu, R2 ) CONEt2.
stranded metal complexes and prepared the first homonuclear
representative back in 1987,38 and the related heteronuclear
derivatives thereafter.39
We describe here the synthesis, structural characteristics, and
coordination properties of a family of three triple-stranded,
polytopic ligands (Figure 1). These binders all possess hy-
droxamate binding cavities and bridges of the same length
between the coordination sites (Figure 1), yet they differ in the
spacers between the anchor and the closer hydroxamate group
(m ) 0 in L1, m ) 1 in L2 and L3) and in the substituents
attached to the spacers between the two hydroxamate functions
(R2 ) H in L1 and L2, R2 ) CONEt2 in L3). The consequences
of these subtle structural changes on the binders’ coordination
properties and particularly the presence of allosteric effects are
examined, by applying classical spectrophotometric titrations
in combination with electrospray mass spectrometric measure-
ments (ESMS).40-55 The combination of these two comple-
mentary methods enabled us to resolve the iron binding
(19) Constable, E. C.; Chotalia, R. J. Chem. Soc., Chem. Commun. 1992,
64-66.
(20) Constable, E. C.; Hannon, M. J.; Tocher, D. A. Angew. Chem., Int.
Ed. Engl. 1992, 31, 230-232.
(21) Duhme, A.-K.; Dauter, Z.; Hider, R. C.; Pohl, S. Inorg. Chem. 1996,
35, 3059-3061.
(22) Carrano, C. J.; Raymond, K. N. J. Am. Chem. Soc. 1978, 100, 5371-
5374.
(23) Carrano, C. J.; Cooper, S. R.; Raymond, K. N. J. Am. Chem. Soc.
1979, 101, 599-604.
(24) Scarrow, R. C.; White, D. L.; Raymond, K. N. J. Am. Chem. Soc.
1985, 107, 6540-6546.
(25) Enemark, E. J.; Stack, T. D. P. Angew. Chem., Int. Ed. Engl. 1995,
34, 996-998.
(26) Enemark, E. J.; Stack, T. D. P Inorg. Chem. 1996, 35, 2719-2720.
(27) Kersting, B.; Meyer, M.; Powers, R. E.; Raymond, K. N. J. Am.
Chem. Soc. 1996, 30, 7221-7222.
(38) Libman, J.; Tor, Y.; Shanzer, A. J. Am. Chem. Soc. 1987, 109,
5880-5881.
(39) Zelikovich, L.; Libman, J.; Shanzer, A. Nature 1995, 374, 790-
(28) Kra¨mer, R.; Lehn, J. M.; De Cian, A.; Fischer, J. Angew. Chem.,
Int. Ed. Engl. 1993, 32, 703-706.
792.
(40) Bruins, A. P.; Covey, T. R.; Henion, J. D. Anal. Chem. 1987, 59,
2642-2646.
(41) Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C.
M. Mass Spectrom. ReV. 1990, 9, 37-70.
(42) Bruins, A. P. Mass Spectrom. ReV. 1991, 10, 53-77.
(43) Bitsch, F.; Dietrich-Buchecker, C.; Khemiss, A. K.; Sauvage, J. P.;
Van Dorsselaer, A. J. Am. Chem. Soc. 1991, 113, 4023-4025.
(44) Katta, V.; Chowdhury, S. K.; Chait, B. T. J. Am. Chem. Soc. 1990,
112, 5348-5349.
(45) Ashton, P. R.; Brown, C. L.; Chapman, J. R.; Gallagher, R. T.;
Stoddardt, J. F. Tetrahedron Lett. 1992, 33, 7771-7774.
(46) Przybylski, M.; Glocker, M. O. Angew. Chem., Int. Ed. Engl. 1996,
35, 806-826.
(47) Jaquinod, M.; Leize, E.; Potier, N.; Albrecht, A. M.; Shanzer, A.;
Van Dorsselaer, A. Tetrahedron Lett. 1993, 34, 2771-2774.
(48) Leize, E.; Van Dorsselaer, A.; Kra¨mer, R.; Lehn, J. M. J. Chem.
Soc., Chem. Commun. 1993, 990-993.
(49) Russell, K. C.; Leize, E.; Van Dorsselaer, A.; Lehn, J. M. Angew.
Chem., Int. Ed. Engl. 1995, 34, 209-213.
(29) Kra¨mer, R.; Lehn, J. M.; Marquis-Rigault, A. Proc. Natl. Acad.
Sci. U.S.A. 1993, 90, 5394-5398.
(30) Williams, A. F.; Piguet, C.; Bernardinelli, G. Angew. Chem., Int.
Ed. Engl. 1991, 30, 1490-1492.
(31) Charbonniere, L.; Bernardinelli, G.; Piguet, C.; Sargeson, A. M.;
Williams, A. F. J. Chem. Soc., Chem. Commun. 1994, 1419-1420.
(32) Piguet, C.; Bernardinelli, G.; Bocquet, B.; Schaad, O.; Williams,
A. F. Inorg. Chem. 1994, 33, 4112-4121.
(33) Zurita, D.; Baret, P.; Pierre, J. L. New J. Chem. 1994, 18, 1143-
1146.
(34) Bernardinelli, G.; Piguet, C.; Williams, A. F. Angew. Chem., Int.
Ed. Engl. 1992, 31, 1622-1624.
(35) Piguet, C.; Williams, A. F.; Bernardinelli, G.; Bu¨nzli, J. C. Inorg.
Chem. 1993, 32, 4139-4149.
(36) Piguet, C.; Bu¨nzli, J. C.; Bernardinelli, G.; Hopfgartner, G.;
Williams, A. F. J. Am. Chem. Soc. 1993, 115, 8197-8206.
(37) Piguet, C.; Bu¨nzli, J. C.; Bernardinelli, G.; Bochet, C. F.; Froidevaux,
P. J. Chem. Soc., Dalton Trans. 1995, 83-97.

4936 J. Am. Chem. Soc., Vol. 119, No. 21, 1997
Blanc et al.
Scheme 1
Scheme 2
properties of the ligands considered in this work, to determine
the nature of their ferric species, to measure their respective
stability constants and absorption spectra, to ascertain the
distribution of the different species at any given ligand/iron
ratios, and to quantify the interactions between the two
trihydroxamate ion binding cavities.
intermittent 2 and terminal 3 monomers were linked together
to the dihydroxamate 6, and subsequently coupled to the anchor
8 and its extended congener 9. Ligands L1 and L2 were
prepared according to this sequence. In the second sequence
(Scheme 2) the ligating chains were assembled by the Merrifield
method of synthesis before condensation to the trisamine 8.
According to this sequence, leucine, intermittent monomer 10,
and terminal monomer 3 were successively linked to the
Merrifield resin, applying coupling methods commonly used
in peptide synthesis. The immobilized chain was then removed
from the solid support, purified, and condensed to the anchor
to provide ligand L3 after hydrogenation.
All intermediates and final products were fully characterized
by their IR and 1H NMR spectra, also partly by COESY spectra,
as described in the Experimental Section, and by their MS
spectra which are summarized in Table 1.
Electrospray Mass Spectra. As an example, the electro-
spray mass spectra recorded for ligand L1 at four different [L1]tot/
[Fe(III)]tot ratios are presented in Figure 2. The mass spectro-
metric study clearly showed the formation of three ferric
complexes in thermodynamic equilibrium. The pseudomolecu-
lar ions of the different species observed with L (L ) L1, L2,
and L3) in the ES mass spectra are given in Table 1.
Results
Synthesis. A modular strategy was adopted for the ligands’
synthesis, as outlined in Schemes 1 and 2. According to this
strategy the triple-stranded binders were synthesized from three
elements: an anchor, a hydroxamate-containing, variable in-
termittent monomer, and a hydroxamate group containing a
terminal monomer. As anchor we used the achiral tris(2-
aminoethyl)amine (8), which was optionally extended by a
leucyl residue as spacer to 9. As ion binding monomer we
employed either the achiral hydroxamate 2 assembled from
3-(N-hydroxyamino)propionic acid and â-alanine or its chiral
congener 10 derived from 3-(N-hydroxyamino)propionic acid
and aspartic acid. As terminal group we selected the benzoyl-
ated 3-(N-hydroxyamino)propionic acid 3.
The triple-stranded binders were accordingly synthesized in
either of two sequences. In the first sequence (Scheme 1) the
(50) Marquis-Rigault, A.; Dupont-Gervais, A.; Baxter, P. N. W.; Van
Dorsselaer, A.; Lehn, J. M. Inorg. Chem. 1996, 35, 2307-2310.
(51) Hopfgartner, G.; Piguet, C.; Henion, J. D.; Williams, A. F. HelV.
Chim. Acta 1993, 76, 1759-1766.
For L, LFe, and LFe2, only doubly-charged ions were
obtained. The major ionization was obtained by the diproto-
nation of the ligand. The minor peaks at -16 Da in the free
ligands and in the three corresponding ferric complexes could
be related to a minor reduction process of a single hydroxamate
function which has been widely observed by FAB mass
spectrometry in siderophores, but not fully elucidated.56-58 One
(52) Hopfgartner, G.; Piguet, C.; Henion, J. D. J. Am. Soc. Mass
Spectrom. 1994, 5, 748-756.
(53) Caudle, M. T.; Stevens, R. D.; Crumbliss, A. L. Inorg. Chem. 1994,
33, 843-844.
(54) Caudle, M. T.; Stevens, R. D.; Crumbliss, A. L. Inorg. Chem. 1994,
33, 6111-6115.
(55) Cheng, X.; Gao, Q.; Smith, R. D.; Simanek, E. E.; Mammen, M.;
Whitesides, G. M. J. Org. Chem. 1996, 61, 2204-2206.
(56) Dell, A.; Hider, R. C.; Barber, M.; Bordoli, R. S.; Sedgwick, Tyler,
A. N.; Neilands, J. B. Biomed. Mass Spectrom. 1982, 9, 158-161

Allosteric Effects in Triple-Stranded Ferric Complexes
J. Am. Chem. Soc., Vol. 119, No. 21, 1997 4937
Table 1. ESMS Data Ferric L1, L2, and L3 Complex Pseudomolecular Ions in Methanol
could also notice the cationization with Na+ and K+ of the
different species, L, LFe, and LFe2. The mass difference
measured between these three species was about 52.7 Da,
demonstrating that the complexation by each of the ferric ions
was accompanied by loss of three protons from the hydroxamate
groups. On the other hand, the complexation of the third ferric
ion was not accompanied by loss of three protons, but by the
appearance of three positive charges, in agreement with the
LFe3, 3Cl- species. In this case, the major ionization did not
occur by protonation of the ligand L, but by loss of two or
three Cl- counterions. The complex LFe3 was thus character-
ized by distinct pseudomolecular ions of different charge states
such as [LFe3]3+, [LFe3 - H]2+, and [LFe3 + Cl]2+. In the
case of LFe3, cationization with K+ and an addition of methanol
were also observed. The peaks indicated by an asterisk in the
mass spectra were shown to derive from the FeCl3 solution.
Stability Constants. The global stability constants of the
ferric complexes formed with the polytopic binders L1, L2, and
L3 were determined by spectrophotometric titrations of the free
ligand with iron(III). As an example, a batch titration of
receptor L1 is shown in Figure 3 up to a 3.5-fold excess of
iron(III).
of the trinuclear complex could not be determined from these
experiments, since the ferric L1 solution used for titration with
CDTA contained 2 equiv of iron(III)/equiv of ligand. The
preliminary independent determination of the apparent stability
constant of CDTAFe(III) at p[H] ) 6.5 ( 0.1 in methanol gave
a log KCDTAFe value of 6.8 ( 0.3 at 25.0 ( 0.1 °C and I ) 0.05
M. Two different statistical methods (relying on the Letagrop-
Spefo62,63 and Specfit59-61 programs) were used to determine
the stability constants presented in Table 2. The most precise
data were obtained with the Letagrop program.62,63 The lack
of a pit-mapping adjustment in the recent Specfit software59-61
led to larger uncertainties. Yet, the key values obtained by either
method were in very good agreement (Table 2). For further
calculations, only the data fitted with the Letagrop program62,63
were used. In the case of L1, for which binding constants were
determined by both methods, a direct one and a competitive
one, only the average value will be used in the following
discussion.
Absorption UV-Vis Spectra. The electronic spectra cor-
responding to the L1, L2, and L3 mono-, di-, and triferric
complexes were calculated by both the Specfit59-61 and Leta-
grop-Spefo62,63 programs. The respective spectra are presented
in Figure 5. A typical charge-transfer band for trihydroxamate
ferric species is observed at 435 ( 10 nm for all the mono-
and diferric complexes. The formation of triferric species
induces a bathochromic shift of 5-25 nm, as expected for
dihydroxamate binding.
For each ligand, the best fit of the spectrophotometric data
was obtained with the model including three ferric species, a
mono-, a di-, and a trinuclear complex, in agreement with the
mass spectrometric observations. The values of the global
stability constants obtained for the three LFe, LFe2, and LFe3
complexes with the three ligands L1, L2, and L3 respectively,
are summarized in Table 2.
(59) Gampp, H.; Maeder, M; Meyer, C. J.; Zuberbu¨hler, A. D. Talanta
1985, 32, 95-101.
For ligand L1 the stability constants of the mono- and diferric
complexes were also determined by competition experiments
using CDTA as scavenger (Figure 4). The stability constant
(60) Gampp, H.; Maeder, M; Meyer, C. J.; Zuberbu¨hler, A. D. Talanta
1986, 33, 943-951.
(61) Gampp, H.; Maeder, M; Meyer, C. J.; Zuberbu¨hler, A. D. Comments
Inorg. Chem. 1987, 6, 41-60.
(57) De Hoffmann E.; Stroobant, V. Biol. Mass Spectrom. 1991, 20, 142-
152.
(58) Mohn, G.; Taraz, K.; Budzikiewicz, H. Z. Naturforsch. 1990, 45b,
1437-1450.
(62) Sillen, L. G.; Warnqvist, B. Ark. Kemi 1968, 31, 377-390.
(63) Arnek, R.; Sillen, L. G.; Wahlberg, O. ArkiV fo¨r Kemi 1968, 31,
353-363.

4938 J. Am. Chem. Soc., Vol. 119, No. 21, 1997
Blanc et al.
Figure 3. Spectrophotometric titration of the ligand L1 by iron(III):
solvent methanol; p[H] ) 6.5 ( 0.1; I ) 0.05; T ) 25.0 ( 0.1 °C; l
) 1 cm; [L1]tot ) 4.5 × 10-5 M; spectra 1-16, [Fe(III)]tot ) 0, 1.0 ×
10-5, 1.8 × 10-5, 2.7 × 10-5, 3.6 × 10-5, 4.5 × 10-5, 5.3 × 10-5, 6.3
× 10-5, 7.3 × 10-5, 8.0 × 10-5, 9.1 × 10-5, 1.08 × 10-4, 1.17 ×
10-4, 1.26 × 10-4, 1.44 × 10-4, 1.58 × 10-4 M, respectively.
methods, a batch method (Scheme 1) or a solid-state method
(Scheme 2). For the ligands prepared the two methods proved
comparable. Yet, when extension of these binders to polytopic
ion binders is considered, the solid-state, Merrifield method is
by far superior.64
FT-IR and 1H NMR spectroscopy of the free ligands and their
benzylated, protected precursors indicated the presence of
extensive H-bond networks. Thus, the stretching frequencies
of the amide NH groups were consistently close to 3300 cm-1
when measured in diluted chloroform solutions (<5 mM). The
1H NMR spectra of the chiral ligands L2 and L3 revealed
pronounced nonequivalence of their diastereotopic protons
-CH2NOCO- and -NCH2CH2- even in protic solvents. The
combination of these phenomena, short-frequency NH absorp-
tions, and pronounced nonequivalence of diastereotopic protons
has earlier been demonstrated to result from interstrand H-bonds
which restrict the compounds’ conformational freedom.65-67 The
IR spectra of the ferric complexes of all ligands also showed
low-frequency NH absorptions (see the Experimental Section),
indicating preservation (or even strengthening) of the ligands’
H-bond network.
Using both ESMS and UV-vis absorption spectrophotometry,
three distinct ferric complexes were identified upon titration of
the tripodal ligands L1, L2, and L3 in [Fe(III)]tot/[ligand]tot ratios
between 0.2 and 4.0, in methanol at p[H] ) 6.5. In order to
quantify the interactions between the two trihydroxamate metal
centers and to evaluate the competition between the dinuclear
and trinuclear complexes, we present in Table 3 the respective
stability constants. The latter were calculated by fitting
experimental data with the Letagrop program62,63 (Table 2), as
already discussed.
Figure 2. ES mass spectra of L1 ferric complexes in methanol (0,
L1; 9, L1Fe; 4, L1Fe2; b, L1Fe3; *, impurities coming from the FeCl3
solution): [L1]tot ) 2.0 × 10-5 M (a); [L1]tot ) 2.0 × 10-5 M, [Fe(III)]tot
) 1.4 × 10-5 M (b); [L1]tot ) 2.0 × 10-5 M, [Fe(III)]tot ) 3.4 × 10-5
M (c); [L1]tot ) 2.0 × 10-5 M, [Fe(III)]tot ) 5.0 × 10-5 M (d).
Discussion
In order to obtain triple-stranded, helicoidal ion binding
molecules that could accommodate two metal ions in their inner
space, a C3-symmetric trisamine as anchor has been selected
and extended by a sequence of two alternating, bidentate ion
binding sites and amide-containing spacers. The three strands,
possessing each of the two hydroxamate sites, create two binding
cavities for metals of octahedral coordination geometries. The
amide-containing spacers induce interstrand H-bonds such as
to minimize random coiling and to stabilize specific conforma-
tions. This type of assembly enables systematic modifications
by varying the nature of the ion binding sites, the length of the
spacers and their projecting substituents, and chiralities. The
latter possibilities were realized in this work with the synthesis
of ligands L1, L2, and L3.
Ferrioxamine B was used as a reference for the trihydrox-
amate monoferric complexes. Its stability constant in methanol
at p[H] ) 6.3 ( 0.1, I ) 0.1 M, and T ) 25.0 ( 0.1 °C was
determined to be log Kferrioxamine B ) 7.17 ( 0.03. Comparison
between the Kferrioxamine B value of ferrioxamine B and the K1
values for the monoferric complexes of the ligands L1, L2, and
(64) Stewart, J. M.; Young, J. D. Solid Phase Peptide Synthesis, 2nd
ed.; Pierce Chemical Co.: Rockford, IL, 1984; p 176.
(65) Tor, Y.; Libman, J.; Shanzer, A.; Felder, C. E.; Lifson, S. J. Am.
Chem. Soc. 1992, 114, 6653-6661.
(66) Tor, Y.; Libman, J.; Shanzer, A.; Felder, C. E.; Lifson, S. J. Am.
Chem. Soc. 1992, 114, 6661-6671.
(67) Dayan, I.; Libman, J.; Agi, Y.; Shanzer, A. Inorg. Chem. 1993, 32,
1467-1475.
The three polytopic ligands were prepared in a modular
fashion from several common elements by either of two

Allosteric Effects in Triple-Stranded Ferric Complexes
J. Am. Chem. Soc., Vol. 119, No. 21, 1997 4939
log â ( 3σ
Table 2. L1, L2, and L3 Ferric Complexes: Global Stability Constantsa
equilibrium
L1
L2
L3
L + Fe(III) {â
\
1} LFe(III)
5.95 ( 0.04b
(6.1 ( 0.9)c
6.60 ( 0.03b
(6.6 ( 0.8)c
6.75 ( 0.04b
(6.6 ( 0.5)c
5.70 ( 0.07b,d
(5.2 ( 1.2)c,d
L + 2Fe(III) {â
\
2} LFe(III)2
12.21 ( 0.02b
(12.2 ( 0.5)c
12.08 ( 0.04b,d
(12.1 ( 0.6)c,d
12.10 ( 0.05b
(12.1 ( 0.6)c
12.80 ( 0.05b
(12.6 ( 0.4)c
L + 3Fe(III) {â
\
3} LFe(III)3
16.8 ( 0.1b
17.20 ( 0.08b
(17.4 ( 0.8)c
17.6 ( 0.1b
(16.8 ( 0.5)c
(17.6 ( 0.4)c
a Solvent methanol; p[H] ) 6.5 ( 0.1; I ) 0.05 M; T ) 25.0 ( 0.1 °C. b Values fitted with the Letagrop program.62,63 c Values fitted with the
Specfit program59-61 and not used for further calculations. d Calculated from the competition titration with CDTA.
Figure 4. Spectrophotometric competition titration of L1 ferric
complexes by CDTA: solvent methanol; p[H] ) 6.5 ( 0.1; I ) 0.05;
T ) 25.0 ( 0.1 °C; l ) 1 cm; [L1]tot ) 4.5 × 10-5 M; [Fe(III)]tot ) 9.0
× 10-5; spectra 1-18; [CDTA]tot ) 0, 0.9 × 10-5, 1.8 × 10-5, 2.7 ×
10-5, 3.7 × 10-5, 4.5 × 10-5, 5.4 × 10-5, 6.4 × 10-5, 7.3 × 10-5, 8.2
× 10-5, 9.1 × 10-5, 1.00 × 10-4, 1.09 × 10-4, 1.18 × 10-4, 1.27 ×
10-4, 1.36 × 10-4, 1.60 × 10-4, 1.82 × 10-4 M, respectively.
L3 under analogous conditions reveals destabilization of the L1,
L2, and L3 monoferric complexes of about 1 order of magnitude
(Table 3). This result could be explained by a more strained
tripodal structure compared with the flexible linear structure of
desferriferrioxamine B. The strongest destabilizing effect is due
to the shrinkage of the peptidic chains of ligand L1 compared
with ligands L2 and L3 (Table 3). The K1 values related to the
monoferric L2 and L3 complexes are not very sensitive to the
nature of the R2 substituents. This result suggests that the ferric
ion is coordinated by the hydroxamate cavity located in the
lower part of the ligands, close to the amine anchor.
Figure 5. Calculated electronic spectra of mono- (9), di- (4), and
triferric (b) complexes formed with L1 (a), L2 (b), and L3 (c): solvent
methanol; p[H] ) 6.5 ( 0.1; I ) 0.05; T ) 25.0 ( 0.1 °C.
Clearly, bulkier R2 substituents increase the stability of the
diferric complex, as revealed by comparison of the K2 values
for L3 and L2, which have identical spacers between the anchor
and the first hydroxamate binding site. The shortening of the
latter spacers in L1 relative to L2 and L3 also induces an increase
in the K2 values of the dinuclear L1 complex.
The K2/K1 ratio enables us to quantify the interactions between
the metal centers in the trihydroxamate helicoidal structures
(Table 3). These interactions are minimal for ligand L3, which
possesses long spacers between the anchor and the first
hydroxamate groups, and bulky R2 substituents: the two
trihydroxamate iron(III) binding cavities behave statistically.
Negative cooperativity is observed between the two centers of
ligand L2, which lacks R2 substituents. The transition from
negative cooperativity in L2 to statistical behavior in L3 mostly
derives from a higher K2 value in L3 than in L2 (Table 3). This
is ascribed to the presence of the aspartic linker in L3 instead
of the â-alanyl linker in L2. The aspartyl linker presumably
enhances K2 by forming a belt of H-bonds with the ligand’s
backbone amides. This explanation is supported by the
enhanced chiral preference of this ligand’s ferric complexes,
as reported in the Experimental Section. Positive cooperativity

4940 J. Am. Chem. Soc., Vol. 119, No. 21, 1997
Blanc et al.
Table 3. Interactions in the Dinuclear Complexesa,b
ligand log K1 ( 3σ log K2 ( 3σ log K3 ( 3σ
K2/K1 ( 3σ
L1
L2
L3
5.82 ( 0.05 6.32 ( 0.08
6.60 ( 0.03 5.50 ( 0.08
6.75 ( 0.04 6.05 ( 0.09
4.7 ( 0.1
5.1 ( 0.1
4.8 ( 0.1
3.2 ( 0.4
0.08 ( 0.01
0.20 ( 0.02
a Solvent methanol; p[H] ) 6.5 ( 0.1; I ) 0,05 M; T ) 25.0 ( 0.1
°C;
K2
L + Fe(III) {K
\ \} LFe(III)2;
1} LFe(III); LFe(III) + Fe(III) {
LFe(III)2 + Fe(III) {K
\
3} LFe(III)3
K2, K1, and K3 correspond to average values of the data fitted with the
Letagrop software.62,63 With an experimental error of (3σ and a
number of experimental points larger than 300, the confidence interval
is equal to 99.7%. b Cooperative effects5 are defined positive when K2/
K1 > 0.25, negative when K2/K1 < 0.25, and statistical when K2/K1 )
0.25.
in favor of the dinuclear complex is achieved with ligand L1,
which possesses short spacers between the anchor and the first
binding site, and shows a smaller stability constant for its
monoferric complex. Hence, the cooperative behavior of the
three diferric helicoidal complexes can be drastically tuned by
minor changes in the chemical structure of tripodal ligands.
The stability of ferric complexes with unsubstituted (RN
)
H) dihydroxamate ligands having from four to eight methylene
groups in the alkane bridge was previously studied.68 More
recently substituted dihydroxamic acids (RN ) CH3) with
spacers possessing from two to eight methylene groups were
considered.54 In the series of the unsubstituted compounds, a
shortening of the spacers results in a decrease of the overall
complex stability, and elongation of the spacers to seven or more
methylene groups results in a significant increase of complex
stability. In the substituted compounds, which are better models
for our ligands, the formation of monomeric dihydroxamate
ferric complexes occurs with spacers of at least six methylene
groups. The triple-stranded ligands L1, L2, and L3 possessing
spacers of six atoms between the two hydroxamate moieties do
also form triferric dihydroxamate complexes, as expected from
the latter studies. The close K3 values determined for the
triferric complexes of L1, L2, and L3 (Table 3) are thus in
agreement with the trends reported in the previous studies,68,54
which demonstrated the predominant influence of the length of
the spacers on the stability of the complexes. The small
variations in the K3 values are attributed to changes in the R1
and R2 substituents.
The distribution curves (Figure 6) of the three ferric com-
plexes formed with ligands L1, L2, and L3 at a given concentra-
tion of free ligand (10-4 M) and increasing concentrations of
iron(III) have been calculated using the Haltafall program,69 and
the respective stability constants are given in Table 3. The
formation of the triferric complex starts for the three ligands
considered at a [Fe(III)]tot/[ligand]tot ratio of about 1.2. When
iron(III) is in large excess (i.e., [Fe(III)]tot/[ligand]tot > 3 ), this
species is predominant. Figure 6 clearly shows that the
formation of the monoferric complex is strongly favored by the
chemical structure of ligand L2 and the formation of the
dinuclear helicate by that of ligand L1, while competition with
an open triferric species occurs in all three compounds.
It is important to emphasize that the assignment of the species
deduced from ESMS yielded the best fit with the spectropho-
tometric data.46-55 Moreover, the major species detected by
ESMS matched the major species observed by absorption
Figure 6. Distribution curves (0, L; 9, LFe; 4, LFe2; b, LFe3) of
the (a) L1, (b) L2, and (c) L3 ferric complexes: solvent methanol; p[H]
) 6.5 ( 0.1; I ) 0.05; T ) 25.0 ( 0.1 °C; [L]tot ) 10-4 M. The
stability constants are given in Table 3.
Table 4. ESMS Data: Distribution of the Ferric L1 Complexes in
Methanola
[Fe(III)]tot/[L1]tot
% L1
% L1Fe
% L1Fe2
% L1Fe3
0.5
0.7
1.2
1.5
1.7
2.0
2.5
5.0
17
11
12
0
0
0
56
58
24
0
5
0
27
31
63
100
95
88
57
0
0
0
0
0
0
12
43
100
0
0
0
0
a The proportion of the nth species was directly deduced from the
relative intensity In (n ) 1, 2, 3, 4) of its ESMS peak by the formula
In/∑n)1-4In.
spectrophotometry at given ligand:iron stoichiometries, such that
all quantitative trends were found to be identical (Table 4).
Accurate quantitative data were obtained from the spectropho-
tometric titrations.
The maximal absorbtivities of the various L1, L2, and L3 ferric
complexes are presented in Table 5. The broad charge-transfer
band is centered at 435 nm for all the mono- and diferric
complexes. The extinction coefficients of the monoferric
complexes (4300 ( 500 M-1 cm-1) show classical values for
ferric trihydroxamate complexes.70-72 The extinction coef-
ficients determined at the absorbance maxima are higher than
that of ferrioxamine B (Table 5) or those of various trihydrox-
(70) Konetschny-Rapp, S; Jung, G.; Raymond, K. N.; Meiwes, J.; Za¨hner,
H. J. Am. Chem. Soc. 1992, 114, 2224-2230.
(68) Evers, A.; Hancock, R. D.; Martell, A. E.; Motekaitis, R. J. Inorg.
Chem. 1989, 28, 2189-2195.
(69) Ingri, N.; Kakolowicz, W.; Sillen, L. G.; Warnqvist, B. Talanta 1967,
14, 1261-1286.
(71) Ng, C. Y.; Rodgers, S. J.; Raymond, K. N. Inorg. Chem. 1989, 28,
2062-2066.
(72) Wong, G. B.; Kappel, M. J.; Raymond, K. N.; Matzanke, B.;
Winkelmann, G. J. Am. Chem. Soc. 1983, 105, 810-815.

Allosteric Effects in Triple-Stranded Ferric Complexes
J. Am. Chem. Soc., Vol. 119, No. 21, 1997 4941
Table 5. L1, L2 and L3 Ferric Complexes: Maximal
These data are in compliance with the predominant formation
of a mononuclear complex at low iron(III) concentrations which
lacks chiral preference. At higher concentrations the diferric
complex is formed with some degree of helicity, and at still
higher concentrations the triferric complex which exhibits low
chiral preference becomes predominant. For L3, the molar
ellipticity increases almost linearly up to the addition of 1.5
equiv of iron(III), and falls off thereafter. These observations
are compatible with the formation of both mononuclear and
dinuclear complexes of helical nature74 up to 1.2 equiv of
iron(III). At higher concentrations, triferric complexes of little
chiral preference are formed.
Absorbtivitiesa
λmax
(nm)
ꢀmax (M-1
cm-1
λmax
ꢀmax (M-1
cm-1
complexes
)
complexes (nm)
)
L1Fe
435 ( 2 4300 ( 500
430 ( 2 4100 ( 200
425 ( 2 4300 ( 300
L1Fe2
L2Fe2
L3Fe2
L1Fe3
L2Fe3
L3Fe3
435 ( 2 6500 ( 200
445 ( 2 7300 ( 300
435 ( 2 7200 ( 200
460 ( 2 6000 ( 500
450 ( 2 3700 ( 400
445 ( 2 5200 ( 200
L2Fe
L3Fe
ferrioxamine B 429 ( 2 2800 ( 200
a Solvent methanol; p[H] ) 6.5 ( 0.1; I ) 0.05 M; T ) 25.0 ( 0.1
°C.
Conclusion
The synthesis of three tripodal, polytopic binders is described.
Their iron(III) coordination properties have been investigated
by applying a combination of electron spray mass spectrometry
and UV-vis and CD absorption spectrophotometry. Quite
remarkably, subtle structural variations proved to tune these
compounds’ coordination properties and to provide representa-
tives of opposite cooperative behavior. The observation of
positive cooperativity for the achiral ligand L1 is of relevance
for the synthesis of extended helicates incorporating more than
two metal ions. Identification of the necessary structural
elements for achieving positive cooperativity has direct bearings
for the synthesis of heteronuclear metal complexes and structur-
ally related molecular redox switches.39 These possibilities are
under current consideration, as are further structural variations
to amplify positive cooperative behavior.
Figure 7. CD titration of ligands L2 and L3 with Fe(III): solvent
methanol; p[H] ) 6.3 ( 0.1 with dichloroacetic buffer, 5 × 10-2 M;
[L] ) 10-4 M; T ) 25.0 ( 0.1 °C; spectrum 1, L2; spectrum 2, L3.
The data given are absolute dichroic values measured at 375 nm for
L2 and at 460 nm for L3. The longer wavelength values were chosen
for L3, because they show minimal overlap with other chromophores.
The shorter wavelength values were chosen for L2, since its complexes’
small ellipticities at longer wavelength were conceived less reliable.
∆ꢀ ) ꢀ(L) - ꢀ(R), the difference of the absorption coefficient between
left and right circularly polarized light (M-1 cm-1).
Experimental Section
Synthesis. 1H NMR spectra were measured on a Bruker WH 270
or on a Bruker AMX 400 spectrometer, at concentrations of (1-2) ×
10-2 M. Chemical shifts are reported in parts per million on the δ
scale relative to tetramethylsilane (TMS) as internal standard. Infrared
spectra were recorded on a Nicolet 510 FT-IR spectrometer at
concentrations of (1-2) × 10-2 M. Absorption frequencies are given
in inverse centimeters. Circular dichroism (CD) spectra were measured
on a Jasco J-500C spectropolarimeter, and the dichroic values ∆ꢀ are
given in M-1 cm-1 units. Purifications were performed by column
chromatography, using silica gel 60 (70-230 mesh ASTM) or flash
chromatography using silica gel 60 (230-400 mesh ASTM). TLC
examinations were performed using Merck TLC precoated aluminum
sheets. Solvents and commercially available reagents were of analytical
grade. Protected amino acids were purchased from Sigma. The
following abbreviations have been used: Ar, aryl; Bn, benzyl; Boc,
tert-butyloxycarbonyl; DCC, N,N′-dicyclohexylcarbodiimide; DICD,
N,N′-diisopropylcarbodiimide; DMAP, 4-(dimethylamino)pyridine; Et,
ethyl; Ph, phenyl; TFA, trifluoroacetic acid.
General Coupling Procedure. To a cold solution of protected
amino acid (1.0 equiv) and 1-hydroxybenzotriazole (0.1 equiv) (Fluka)
in acetonitrile dried over basic alumina were added amine (1.1 equiv)
and N,N′-diisopropylcarbodiimide (1.1 equiv) (Fluka). The reaction
mixture was kept at 0-4 °C for 1-3 h and then at room temperature
overnight. The solution was concentrated and the product purified using
column chromatography.
General Procedure for Hydrolysis of Ethyl or Methyl Esters.
Ethyl or methyl ester (1 mmol) was dissolved in methanol (10 mL)
and treated with 1 M aqueous sodium hydroxide solution (1.25 mL) at
room temperature. The reaction mixture was monitored by TLC every
hour, and additional 1.25 mL portions of aqueous sodium hydroxide
solution were added until all starting material was consumed. The
mixture was cooled in an ice bath and acidified with KHSO4 to pH 2.
Methanol was evaporated, and the residue was extracted with ethyl
acetate. The organic fraction was washed with water, dried over
MgSO4, and concentrated to afford the acid.
amate species (425 nm < λmax < 450 nm; 2400 M-1 cm-1
<
ꢀmax < 3800 M-1 cm-1).70-72 This observation probably derives
from increased strain in the tripodal cavity close to the anchor,
when compared with the strain-free cavity formed by ferriox-
amine B or by bidentate hydroxamic acids. The extinction
coefficients of the diferric complexes are slightly smaller than
twice the value calculated for the monoferric species. This is
probably due to a slightly different coordination geometry of
both cations in the dinuclear helicates and to different electronic
properties of the hydroxamates’ substituents. A shift toward
higher wavelengths (460 nm) is observed for the triferric
complexes. This behavior is characteristic for dihydroxamate
ferric complexes (470 nm < λmax < 480 nm; 1700 M-1 cm-1
< ꢀmax < 2100 M-1 cm-1).23,73 The highest extinction coef-
ficient corresponding to the 3-fold classical dihydroxamate value
is observed for the triferric complex formed with the shortest,
nonsubstituted ligand L1 (Table 5). The extinction coefficients
of L2Fe3 and L3Fe3 are very sensitive to the R2 substituent
(Table 5): absence of the substituent drastically decreases the
ꢀmax value. One can well rationalize that folding of the arms
around three iron(III) cations is very sensitive to the chain length
and to the steric hindrance of lateral substituents.
CD titrations of the chiral ligands L2 and L3 with iron(III)
(Figure 7) qualitatively support the conclusions drawn from the
analysis of the spectrophotometric titration experiments. The
molar ellipticity is small for L2 up to 0.75 equiv of iron(III),
reaches a maximum around 1.75 equiv, and falls off thereafter.
(73) Harris, W. R.; Carrano, C. J.; Raymond, K. N. J. Am. Chem. Soc.
1979, 101, 2722-2727.
(74) van der Helm, D.; Baker, J. R.; Eng-Wilmot, D. L.; Hossain, M.
B.; Loghry, R. A. J. Am. Chem. Soc. 1980, 102, 4224-4231.

4942 J. Am. Chem. Soc., Vol. 119, No. 21, 1997
Blanc et al.
Procedure for Solid State Oligomerization. The synthesis was
performed using chloromethylated polystyrene-2% divinylbenzene
resin (Merck). The first unit was attached to the resin by mixing the
acid (3.4 mmol) in absolute ethanol (14 mL) with triethylamine (0.43
mL, 3 mmol) and polymeric resin (2.3 g) for 72 h at 80 °C. Subsequent
coupling steps on the polymeric support were carried out with freshly
prepared 1-hydroxybenzotriazole active esters of the monomer acids
in acetonitrile/methylene chloride (85/15 v/v) at room temperature
overnight. The active esters were prepared 2-3 h prior to each coupling
step by mixing monomer acid (2.0 mmol), 1-hydroxybenzotriazole (2.0
mmol), and N,N′-diisopropylcarbodiimide (2.2 mmol) in acetonitrile
at 0 °C.
CO), 1.19 (t, J ) 4.8 Hz, 3H, CH3). FT-IR (CDCl3, cm-1): ν 1728
(CdO ester), 1670 (CdO amide and hydroxamate).
Protected Ion Binding Chain 7. Hydrolysis of ester 6 according
to the general hydrolysis procedure provided the free acid. 1H NMR
(CDCl3): δ 7.60, 6.80 (ABq, J ) 8.7 Hz, 4H, anisoyl), 7.30-7.12,
(m, 10H, Ph), 6.65 (br t, 1H, NH), 4.76 (s, 2H, PhCH2O), 4.63 (s, 2H,
PhCH2O), 3.93 (m, 4H, CH2NO), 3.84 (s, 3H, OCH3), 3.49 (m, 2H,
CH2NH), 2.58-2.52 (m, 6H, CH2CO). FT-IR (CDCl3, cm-1): ν 1728
(CdO acid), 1670 (CdO amide and hydroxamate). Then, 2.0 g (3.6
mmol) of acid dissolved in 20 mL of acetonitrile was treated with 1.25
g (4.7 mmol) of pentachlorophenol, 450 mg of DMAP, and 0.8 mL
(5.1 mmol) of DICD overnight. The resulting mixture was concentrated
and chromatographed on silica gel (elution with hexane/chloroform,
chloroform, and chloroform/ethyl acetate mixtures) to provide 1.69 g
(1.95 mmol, 54% yield) of the active ester 7. 1H NMR (CDCl3): δ
7.68, 6.87 (ABq, J ) 8.7 Hz, 4H, anisoyl), 7.34, 7.28, 7.15 (m, 10H,
Ph), 6.50 (br t, 1H, NH), 4.82 (s, 2H, PhCH2O), 4.65 (s, 2H, PhCH2O),
4.01 (m, 4H, CH2NO), 3.84 (s, 3H, OCH3), 3.51 (m, 2H, CH2NH),
2.96 (t, J ) 6.9 Hz, 2H, CH2COOC6Cl5), 2.57 (m, 4H, CH2CO). FT-
IR (CDCl3, cm-1): ν 1780 (CdO active ester), 1663 (CdO amide and
hydroxamate).
(I) Ligand Synthesis. Intermittent Monomer 2. Condensation of
6.4 g (33.8 mmol) of N-Boc-â-alanine with 3-[N-(benzyloxy)amino]-
propionic acid ethyl ester75,76 (1) (8.29 g, 37.2 mmol) according to the
general coupling procedure provided 8.6 g (21.8 mmol, 64% yield) of
product 2. 1H NMR (CDCl3): δ 7.35 (m, 5H, Ph), 5.19 (br, 1H, NH),
4.78 (s, 2H, PhCH2O), 4.06 (q, J ) 7.1Hz, 2H, OCH2CH3), 3.95 (m,
2H, CH2NO), 3.36 (m, 2H, NCH2), 2.57 (m, 4H, CH2CO), 1.19 (t, J )
7.1 Hz, 3H, CH3). FT-IR (CDCl3, cm-1): ν 1715 (CdO ester), 1672
(CdO Boc and hydroxamate).
Ligand L1. A 225 mg (0.27 mmol) sample of active ester 7 was
dissolved in 10 mL of methylene chloride and stirred under argon with
0.010 mL (0.07 mmol) of tris(2-aminoethyl)amine (TREN, 8) and 20
mg of N-hydroxysuccinimide overnight. Then the mixture was
concentrated and chromatographed on silica gel with chloroform/
methanol as eluent to provide 84 mg (0.045 mmol, 64.2% yield) of
benzyl-protected ligand L1 (R ) Bn). 1H NMR (CDCl3): δ 7.7, 6.90
(ABq, J ) 8.7 Hz, 12H, anisoyl), 7.40-7.15 (m, 30H, Ph), 4.80 (m,
6H, PhCH2O), 4.75 (s, 6H, PhCH2O), 4.10 (m, 6H, CH2NO), 3.97 (m,
6H, CH2NO), 3.80 (s, 9H, OCH3), 3.50 (m, 6H, CH2CH2NH), 3.10
(m, 6H, CH2CH2NH), 2.6 (m, 6H, NCH2CH2NH), 2.50-2.45 (m, 12H,
CH2CO), 2.45 (m, 6H, CH2CO). FT-IR (CDCl3, cm-1): ν 1653 (CdO
amide and hydroxamate). A 200 mg (0.01 mmol) sample of protected
ligand L1 (R ) Bn) was dissolved in 50 mL of ethanol and
hydrogenated under atmospheric pressure in the presence of 75 mg of
Pd/C (10%) for 6 h. Filtration and concentration of the filtrate provided
104 mg of L1 (0.008 mmol, 80% yield). Mp: 140-145 °C. 1H NMR
(CDCl3/CD3OD): δ 7.7, 6.90 (ABq, J ) 8.7 Hz, 12H, anisoyl), 3.96
(m, 6H, CH2NO), 3.86 (m, 6H, CH2NO), 3.80 (s, 9H, OCH3), 3.43 (m,
6H, CH2CH2NH), 3.29 (m, 6H, CH2CH2NH), 2.6 (m, 6H, NCH2CH2-
NH), 2.57-2.51 (m, 18H, CH2CO). The assignment of the ligand’s
1H-NMR signals was confirmed by correlation spectroscopy (COSY),
which showed cross-peaks for each of the compound’s ethylene bridges
and for the two aromatic signals. FT-IR (KBr, cm-1): ν 1644 (CdO
amide and hydroxamate).
Ligand L2. Protected ligand L2 (R ) Bn) was prepared by
condensation of active ester 7 with trisamine 965 as described for the
preparation of protected ligand L1. 1H NMR (CD3OD, 10 mM): δ
7.56, 6.88 (ABq, J ) 8.7 Hz, 12H, anisoyl), 7.35-7.13 (m, 30H, Ph),
4.76 (m, 6H, PhCH2O), 4.66 (m, 6H, PhCH2O), 4.39 (m, 3H, CH-i-
Bu), 4.03 (m, 6H, CH2NO), 3.91 (m, 3H, CH2NO), 3.79 (s + m, 12H,
OCH3 + CH2NO), 3.37 (m, 12H, CH2CH2NH), 2.55 (m, 24H, CH2CO
and NCH2CH2NH), 1.60 (br, 9H, CH2{i-Bu} + CH{i-Bu}), 0.87-0.82
(m, 18H, CH3{i-Bu}). FT-IR (CDCl3 1.6 mM, cm-1): ν 3300 (NH),
1656, 1650, 1644 (CdO amide and hydroxamate). Hydrogenation of
protected ligand L2 (R ) Bn) as described for protected ligand L1
provided the free ligand L2. Mp: 105-110 °C. 1H NMR (CD3OD, 8
mM): δ 7.66, 6.93 (ABq, J ) 8.7 Hz, 12H, anisoyl), 4.36 (m, 3H,
CH-i-Bu), 3.97 (m, 6H, CH2NO), 3.86 (m, 6H, CH2NO), 3.82 (s, 9H,
OCH3), 3.65 (m, 3H, CH2CH2NH), 3.45 (m, 9H, CH2CH2NH), 2.67
(m, 6H, NCH2CH2NH), 2.60 (m, 18H, CH2CO), 1.65 (m, 9H, CH2{i-
Bu} + CH{i-Bu}), 0.93-0.88 (m, 18H, CH3{i-Bu}). FT-IR (KBr,
cm-1): ν 3284 (NH), 1657, 1650, 1643 (CdO amide and hydroxamate).
N-Boc-L-Aspartic Acid r-Diethyl Amide â-Benzyl Ester. Con-
densation of N-Boc-L-aspartic acid â-benzyl ester and diethylamine
according to the general coupling procedure, followed by column
chromatography using hexane/ethylacetate (1/1) as eluent, provided the
product in 41% yield. 1H NMR (CDCl3): δ 7.34 (s, 5H, Ph), 5.33 (d,
J ) 9.4 Hz, 1H, NH), 5.11 (m, 2H, CH2Ph), 4.95 (m, 1H, CH), 3.44
(m, 2H, NCH2CH3(trans)), 3.29 (m, 2H, NCH2CH3(cis)), 2.85 (dd, Jgem
) 15.8 Hz, Jvic ) 7.0 Hz, 1H, CHCH2), 2.65 (dd, Jgem ) 15.7 Hz, Jvic
Terminal Monomer 3. 3-[N-(Benzyloxy)amino]propionic acid ethyl
ester75,76 (1) (6.84 g, 30.6 mmol) was dissolved in 300 mL of dry toluene
and treated under ice, cooling first with 3.8 mL (30.4 mmol) of
triethylamine and then dropwise with 2.8 mL (26 mmol) of anisoyl
chloride dissolved in 50 mL of toluene. Then, the mixture was stirred
under cooling for 1 h and at room temperature for 1 h. The reaction
mixture was then washed with 1 M aqueous HCl, water, 1 M NaHCO3,
and water and dried over MgSO4. Concentration of the organic phase
and chromatography using hexane/ethyl acetate (7/3 v/v) as eluent
provided 7.45 g (22 mmol, 74.4% yield) of the ethyl ester of monomer
3. 1H NMR (CDCl3): δ 7.68, 6.85 (ABq, J ) 8.7 Hz, 4H, anisoyl),
7.30-7.26 (m, 5H, Ph), 4.66 (s, 2H, PhCH2O), 4.06 (q + t, 4H, OCH2-
CH3 and CH2NO), 3.84 (s, 3H, OCH3), 2.69 (t, J ) 6.6 Hz, 2H, CH2-
CO), 1.20 (t, J ) 6.6 Hz, 3H, CH3). FT-IR (CDCl3, cm-1): ν 1729
(CdO ester), 1631 (CdO hydroxamate). Hydrolysis of the ester
according to the general procedure afforded the monomer acid 3 in
49% yield. 1H NMR (CDCl3): δ 7.7, 6.88 (ABq, J ) 8.7 Hz, 4H,
anisoyl), 7.29-7.15 (m, 5H, Ph), 4.67 (s, 2H, PhCH2O), 4.06 (t, J )
6.8 Hz, 2H, CH2NO), 3.85 (s, 3H, OCH3), 2.74 (t, J ) 6.8 Hz, 2H,
CH2CO). FT-IR (CDCl3, cm-1): ν 1714 (CdO acid), 1629 (CdO
hydroxamate).
Protected Ion Binding Chain 6. Terminal monomer acid 3 was
activated by treating 2.2 g (7.35 mmol) of acid in 50 mL of acetonitrile
with 1.95 g (7.3 mmol) of pentachlorophenol, 150 mg of DMAP, and
1.2 mL (7.6 mmol) of DICD under cooling. After being stirred for
two days, the crude mixture was concentrated and chromatographed
first on neutral alumina (activity V) using chloroform as eluent and
then on silica gel (chloroform/hexane, 9/1 v/v) to provide 2.68 g (4.9
mmol, 64.9% yield) of activated ester 5. Mp: 103-106 °C. 1H NMR
(CDCl3): δ 7.75, 6.92 (ABq, J ) 8.7 Hz, 4H, anisoyl), 7.30-7.20 (m,
5H, Ph), 4.70 (s, 2H, PhCH2O), 4.19 (t, J ) 6.9 Hz, 2H, CH2NO),
3.87 (s, 3H, OCH3), 3.12 (t, J ) 6.9 Hz, 2H, CH2CO). FT-IR (CDCl3,
cm-1): ν 1780 (CdO active ester), 1633 (CdO hydroxamate).
Intermittent monomer 2 was deprotected by treating 2.5 g (6.34 mmol)
of compound with 4 mL of TFA in 8 mL of methylene chloride for 30
min at room temperature to provide 4. A 2.0 g sample of the
corresponding ammonium salt was dissolved in 1 mL of DMF,
neutralized with triethylamine, and then treated with 3.2 g (5.8 mmol)
of activated monomer 5 dissolved in 20 mL of methylene chloride in
the presence of 100 mg of imidazole. The reaction mixture was stirred
overnight at room temperature, concentrated, and chromatographed on
silica gel using chloroform as eluent to provide the ethyl ester 6. 1H
NMR (CDCl3): δ 7.75, 6.86 (ABq, J ) 8.7 Hz, 4H, anisoyl), 7.36-
7.27 (m, 10H, Ph), 6.5 (br t, 1H, NH), 4.74 (s, 2H, PhCH2O), 4.67 (s,
2H, PhCH2O), 4.07 (m, 4H, CH2NO and OCH2), 3.95 (m, 2H, CH2-
NO), 3.84 (s, 3H, OCH3), 3.45 (m, 2H, CH2NH), 2.55 (m, 6H, CH2-
(75) Yakirevitch, P.; Rochel, N.; Albrecht-Gary, A. M.; Libman, J.;
Shanzer, A. Inorg. Chem. 1993, 32, 1779-1787.
(76) Yakirevitch, P. Ph.D. Thesis, The Feinberg Graduate School of the
Weizmann Institute of Science, Rehovot, Israe¨l, 1992.

Allosteric Effects in Triple-Stranded Ferric Complexes
J. Am. Chem. Soc., Vol. 119, No. 21, 1997 4943
) 5.8 Hz, 1H, CHCH2), 1.42 (s, 9H, t-Bu), 1.23 (t, J ) 7.1 Hz, 3H,
NCH2CH3(cis)), 1.09 (t, J ) 7.1 Hz, 3H, NCH2CH3(trans)). FT-IR
(neat, cm-1): ν 3297 (NH), 1738 (CdO ester), 1710 (CdO Boc), 1641
(CdO amide).
was treated with triamine 8 (0.063 mL, 0.42 mmol), and stirred under
nitrogen atmosphere at room temperature overnight. N,N′-Dicyclo-
hexylurea (DCU) was filtered off and washed with cold tetrahydrofuran,
and the reaction mixture was evaporated to dryness under Vacuo.
Purification of the products by column chromatography using mixtures
of 3-8% methanol/chloroform as eluent afforded the protected ligand
L3 (R ) Bn) as a glassy solid in 48% yield. Mp: 46-48 °C. 1H
NMR (CDCl3, 20 mM): δ 7.66, 6.94 (ABq, J ) 8.7 Hz, 12H, anisoyl),
7.33-7.11 (m, 30H, Ph), 5.2 (m, 3H, CHCON), 4.76 (m, 6H, PhCH2O),
4.66 (br s, 6H, PhCH2O near anisoyl), 4.4 (m, 3H, CH-i-Bu), 4.05-
3.9 (m, 12H, CH2NO), 3.85 (s, 9H, OCH3), 3.24 (m, 18H, NCH2CH3
and CH2CH2NH), 2.9, 2.64, 2.53 (m, 24H, CHCH2, CH2CO and NCH2-
CH2NH), 1.60 (br, 6H, CH2{i-Bu}), 1.45 (m, 3H, CH{i-Bu}), 1.02 (m,
18H, NCH2CH3), 0.84 (br s, 18H, CH3{i-Bu}). FT-IR (CDCl3, cm-1):
ν 3306 (NH, OH), 1639 (CdO). A 0.27 g (1.1 mmol) sample of
protected ligand L3 was dissolved in methanol (40 mL) and chilled in
an ice-water bath. A 100 mg sample of Pd/C (10%) was added, and
the suspension was hydrogenated at atmospheric pressure for 2 h. The
catalyst was filtered off and washed with methanol. The solvent was
evaporated, affording the pure product L3 (0.18 g, 88% yield) as a
glassy powder. Mp: 133-134 °C. 1H NMR (CD3OD, 20 mM): δ
7.68, 6.94 (ABq, J ) 8.8 Hz, 12H, anisoyl), 5.23 (br t, 3H, CHCON),
4.31 (br t, 3H, CH-i-Bu), 3.99 (m, 6H, CH2NO), 3.82 (s + m, 15H,
OCH3 (9H) and CH2NO (6H)), 3.44 (m, 12H, NCH2CH3), 3.33 (m,
6H, CH2CH2NH), 3.05 (m, 3H, CHCH2), 2.73, 2.64, 2.56 (m, 21H,
CHCH2 (3H), NCH2CH2NH (6H) and CH2CH2CO (12H)), 1.68-1.59
(m, 9H, CH{i-Bu} and CH2{i-Bu}), 1.12-1.08 (br t, 18H, NCH2CH3),
0.94 (d, J ) 6.2 Hz, 9H, CH3{i-Bu}), 0.90 (d, J ) 6.1 Hz, 9H, CH3-
{i-Bu}). FT-IR (CDCl3 5 mM, cm-1): ν 3275 (NH, OH), 1630 (CdO).
(II) Preparation of Ferric Complexes of L1, L2, and L3. The
monoferric complexes were prepared by treating methanolic solutions
of ligand (3.9 × 10-6 mol in 2 mL of methanol) with 1 equiv of FeCl3
(0.65 mL of 6.0 mM FeCl3 in methanol) and addition of anhydrous
sodium acetate until the color changed to orange, pH > 8.0. The
reaction mixture was then stirred for 30 min, evaporated, and extracted
with methylene chloride (150 mL). The organic solution was washed
with water, dried over Na2SO4, and evaporated, affording the ferric
complexes as red powders. Yield: 90%. The spectral characteristics
of the ferric species using 1 equiv of FeCl3 are as follows. (With L1)
FT-IR (CDCl3, cm-1): ν 3318 (NH, OH). (With L2) FT-IR (CDCl3,
cm-1): ν 3285 (NH, OH). CD (MeOH): λext ) 455, 374 nm (∆ꢀ )
0.0, +0.6). CD (CDCl3): λext ) 455, 380 nm (∆ꢀ ) -0.76, +1.25).
(With L3) FT-IR (CDCl3, cm-1): ν 3295 (NH, OH). CD (MeOH):
λext ) 460, 380 nm (∆ꢀ ) -2.0, +3.7). CD (CDCl3): λext ) 460, 373
nm (∆ꢀ ) -2.8, +4.7). Higher loaded ferric complexes were obtained
according to the same procedure using 2 equiv of FeCl3 (1.33 mL of
6.0 mM FeCl3/methanol). The spectral characteristics of the corre-
sponding ferric species are as follows. (With L1) FT-IR (CDCl3, cm-1):
ν 3305 (NH). (With L2) FT-IR (CDCl3, cm-1): ν 3310 (NH). CD
(MeOH)SPCLN λext ) 470, 390 nm (∆ꢀ ) -0.6, +1.9). CD
(CDCl3): λext ) 450, 375 nm (∆ꢀ ) -2.6, +5.12). (With L3) FT-IR
(CDCl3, cm-1): ν 3295 (NH). CD (MeOH): λext ) 461, 370 nm (∆ꢀ
) -4.0, +6.0). CD (CDCl3): λext ) 460, 373 nm (∆ꢀ ) -4.7, +9.1).
N-Boc-L-aspartic Acid r-Diethyl Amide. To a solution of N-Boc-
L-aspartic acid R-diethyl amide â-benzyl ester (1.5 g, 4.017 mmol) in
absolute ethanol (100 mL) was added 300 mg Pd/C (10%, and the
reaction mixture was hydrogenated for 2 h at room temperature and
atmospheric pressure. The catalyst was filtered off, and the solvent
was evaporated, affording the free acid in 92% yield. 1H NMR
(CDCl3): δ 5.6 (br d, 1H, NH), 5.0 (m, 1H, CH), 3.71 (q, J ) 7.0 Hz,
2H, NCH2CH3(trans)), 3.42 (m, 2H, NCH2CH3(cis)), 2.71 (m, 2H,
CHCH2), 1.43 (s, 9H, t-Bu), 1.31-1.01 (m, 6H, NCH2CH3). FT-IR
(CDCl3, cm-1): ν 3299 (br, NH, OH), 1713 (CdO acid, Boc), 1634
(CdO amide).
Intermittent Monomer Acid 10. Boc-L-aspartic acid R-diethyl
amide and 3-[N-(benzyloxy)amino]propionic acid ethyl ester (1)75,76
were reacted according to the general coupling procedure, followed
by column chromatography using hexane/ethyl acetate (1/1) as eluent,
to afford the ethyl ester of 10 in 85% yield. 1H NMR (CDCl3): δ
7.34 (s, 5H, Ph), 5.24 (m, 1H, CH), 4.86 (s, 2H, PhCH2O), 4.04 (m,
4H, CH2NO and OCH2CH3), 3.42 (m, 4H, NCH2CH3), 2.78 (m, 2H,
CHCH2), 2.59 (br t, 2H, CH2COO), 1.42 (s, 9H, t-Bu), 1.20 (m, 9H,
NCH2CH3 and OCH2CH3). Hydrolysis according to the general
procedure afforded the acid 10 in 97% yield. 1H NMR (CDCl3): δ
7.37 (s, 5H, Ph), 5.60 (br d, 1H, NH), 4.88, 4.79 (ABq, J ) 10.2 Hz,
3H, CH2Ph (2H) and CH (1H)), 3.95 (br, 2H, CH2NO), 3.34 (m, 2H,
NCH2CH3(trans)), 3.21 (m, 2H, NCH2CH3(cis)), 2.78 (dd, Jgem ) 15.2
Hz, Jvic ) 6.6 Hz, 1H, CHCH2), 2.62 (br t + dd, 3H, CHCH2 (1H) and
CH2COO (2H)), 1.43 (s, 9H, t-Bu), 1.16-1.04 (m, 6H, NCH2CH3).
FT-IR (CDCl3, cm-1): ν 3433 (NH), 1710 (CdO acid), 1640 (CdO
amide).
Solid-Phase Synthesis of Protected Ion Binding Chain 11. The
solid synthesis was carried out according to the general oligomerization
procedure, starting with N-Boc-L-leucine monohydrate (1.87 g, 7.5
mmol) and 5.0 g of polymeric resin. The aspartyl-derived monomeric
unit 10 (2.3 g, 4.9 mmol) was added in the presence of DICD (0.8
mL, 5.2 mmol) in acetonitrile. The synthesis proceeded with the
addition of the terminating fragment 3 (1.6 g, 4.9 mmol) in the presence
of DICD (0.8 mL, 5.2 mmol) in CH3CN/CH2Cl2 (85/15 v/v) and
cleavage of the chain from the resin. The cleavage was performed by
suspension of the resin in anhydrous methanol (40 mL/1 g of resin),
addition of triethylamine (2.6 mL, 18.7 mmol/1 g of resin), and stirring
for 24 h at room temperature. Then the polymer was filtered off and
washed with methanol and the solution concentrated to dryness. Flash
chromatography (2-5% CH3OH/CH2Cl2) afforded 1.72 g of pure ester
11. 1H NMR (CDCl3, 25 mM): δ 7.68, 6.87 (ABq, J ) 8.8 Hz, 4H,
anisoyl), 7.35-7.14 (m, 10H, Ph), 7.06, 6.73 (two br d, 2H, NH), 5.20
(m, 1H, CaHCON), 4.81, 4.71 (ABq, J ) 10.1 Hz, 2H, PhCH2O), 4.66
(s, 2H, PhCH2O near anisoyl), 4.55 (m, 1H, CaH-i-Bu), 4.04 (m, 4H,
CH2NO), 3.84 (s, 3H, PhOCH3), 3.65 (s, 3H, COOCH3), 3.4-3.21 (m,
4H, NCH2CH3), 2.59 (m, 4H, CH2CO), 2.48 (t, J ) 6.3 Hz, 2H, CH2-
CO), 1.55 (br m, 3H, CH and CH2{i-Bu}), 1.06 (t, J ) 7.0 Hz, 6H,
NCH2CH3), 0.87 (d, J ) 6.1, 3H, CH3{i-Bu}), 0.85 (d, J ) 6.1 Hz,
3H, CH3{i-Bu}). FT-IR (CDCl3 10 mM, cm-1): ν 3432, 3334 (NH),
1740 (CdO ester), 1640 (CdO amide, CONOBn).
Hydrolysis of Ion Binding Chain 11. Methyl ester 11 (1.2 g, 1.5
mmol) was dissolved in methanol (7 mL) and treated with 2 M aqueous
NaOH (3.0 mL) for 2 h. Usual workup according to the general
hydrolysis procedure afforded the acid in 98% yield. 1H NMR
(CDCl3): δ 7.66, 6.85 (ABq, J ) 8.8 Hz, 4H, anisoyl), 7.33-7.12 (m,
10H, Ph), 5.14 (m, 1H, CHCON), 4.79, 4.70 (ABq, J ) 10 Hz, 2H,
PhCH2O), 4.66 (s, 2H, PhCH2O near anisoyl), 4.55 (m, 1H, CH-i-Bu),
4.1 (m, 4H, CH2NO), 3.83 (s, 3H, PhOCH3), 3.34-3.23 (m, 4H, NCH2-
CH3), 2.61-2.50 (m, 6H, CH2CO), 1.6 (m, 3H, CH and CH2{i-Bu}),
1.06 (m, 6H, NCH2CH3), 0.87 (d, J ) 6.1 Hz, 3H, CH3{i-Bu}), 0.85
(d, J ) 5.7 Hz, 3H, CH3{i-Bu}). FT-IR (CDCl3, cm-1): ν 3348 (NH),
1725 (w), 1637 (CdO).
Electrospray Mass Spectrometric Measurements. ES mass
spectra were obtained on a VG BioQ triple quadrupole with a mass to
charge (m/z) range of 4000 (VGBioTech Ltd., Altrincham, U.K.). The
ES interface was heated to 70 °C. The sampling cone voltage (Vc)
was 60 V. No fragmentation process was observed when the voltage
Vc was increased to 150 V. The calibration was performed using
multiprotonated ions from horse myoglobin. The resolution was about
600 at m/z ) 1000 (with a valley of 10%), and then average masses
were measured. Scanning was performed from m/z ) 400 to m/z )
1400 in 10 s. The data system was operated as a multichannel analyzer,
and several scans were summed to obtain the final spectrum. Non-
buffered solutions containing the L1 ligand (2 × 10-5 M) and 0.0, 0.5,
0.7, 1.2, 1.5, 1.7, 2.0, 2.5, and 5.0 equiv of FeCl3 in methanol,
respectively, were injected into the mass spectrometer source with a
syringe pump (Harvard type 55 1111, Harvard Apparatus Inc., South
Natick, MA) at a flow rate of 4 mL/min. The concentration of the
ligand L2 and the ligand L3 was 5 × 10-5 M, and the ligands were
tested only with 0.0, 0.5, 1.0, and 2.0 equiv of FeCl3.
Ligand L3. The above acid (1.22 g, 1.54 mmol) was activated with
1-hydroxybenzotriazole (0.21 g, 1.54 mmol) and DCC (0.35 g, 1.7
mmol) in cold THF (5 mL). The reaction mixture was kept at 0-4 °C
for 1 h and then at room temperature for 2 h. The active ester solution
Potentiometric and Spectrophotometric Measurements. Solid

4944 J. Am. Chem. Soc., Vol. 119, No. 21, 1997
Blanc et al.
samples were quantitatively dissolved in pure methanol (Merck, water-
free). Iron(III) solutions were prepared from FeCl3‚6H2O (Merck, p.a.)
and back-titrated with Th(NO3)4‚5H2O (Merck) in the presence of excess
EDTA (Merck, Titriplex III, 0.1 M) and xylenol orange (Merck) as
indicator. Dichloroacetic buffer77 (0.05 M) was used to maintain the
p[H] value at 6.5 ( 0.1. Tetrabutylammonium hydroxide (Fluka, 25%
in methanol) was used for neutralization of a titrated (ca. 0.6 M)
dichloroacetic acid (BDH, 99%) solution in methanol. Hydrogen ion
concentrations were measured with a combined glass electrode (Ingold,
high alkalinity). The Ag/AgCl reference electrode was filled with 0.05
M tetramethylammonium chloride (Fluka, p.a.) in pure methanol.
Potential differences were given by a Tacussel Isis 20,000 millivolt-
meter. Solutions with known hydrogen ion concentrations in
methanol78-81 were used to verify the linearity of the glass electrode.
The solutions were continuously swept with methanol-saturated argon.
Two methods were used to obtain spectrophotometric data: titration
of the free ligands (10-4 M) in 2 cm optical quartz cells (Hellma) by
concentrated iron(III) solutions (10-3 M) and a batch titration method
in 5 mL flasks. For the last method and in order to avoid precipitation
of iron(III), the reagents were added in the following order: ligand,
iron, and buffer. Although the solutions immediately turned red after
addition of iron, the buffer was introduced only 2 h later, which led to
an instantaneous color change from cherry-red to orange, indicating
the formation of the trihydroxamate complexes. At least 10 solutions
were prepared for each titration method, and the [Fe(III)]tot/[ligand]tot
ratios ranged between 0.2 and 4.0. The temperature was fixed at 25.0
( 0.1 °C with a Lauda thermostat. The stability of the various solutions
was carefully checked for hours after each addition of iron(III), and
UV-vis (300-650 nm) absorption spectra were recorded with a
Kontron Uvikon 860 spectrophotometer. In order to check the results
obtained from the direct spectrophotometric titration of the free ligand
L1 by iron(III), a competition experiment between the ferric L1 complex
and a tetracarboxylic acid, CDTA (N,N′-1,2-cyclohexanediylbis[N-
(carboxymethyl)glycine], Merck, Titriplex IV), was also undertaken
in buffered methanol (p[H] ) 6.5 ( 0.1). The ferric L1 complex
solution was prepared by mixing 1 equiv of ligand ([L1]tot ) 4.5 ×
10-5 M) with 2 equiv of iron(III) ([FeCl3]tot ) 9.0 × 10-5 M). The
dichloroacetic buffer77 concentration was 0.05 M. Methanolic solutions
with [L1]tot/[CDTA]tot ratios ranging between 0.2 and 4.0 were
spectrophotometrically (300-650 nm) examined using a Kontron
Uvikon 941 spectrophotometer and 1 cm Hellma quartz cells.
A
previous determination of the stability constant of the CDTA ferric
complex under identical experimental conditions was necessary to
calculate the stability of the L1 ferric species. A batch titration in 10
mL flasks of CDTA (8.2 × 10-4 M) by increasing amounts of FeCl3
(8 × 10-5 M e [Fe(III)]tot e 1.40 × 10-3 M) was carried out. When
equilibrium conditions were reached in all the solutions, UV-vis
absorption spectra were recorded using a Kontron Uvikon 860
spectrophotometer in the 250-450 nm range. As a reference for the
iron(III) binding properties by a trihydroxamate ligand, the stability
constant of the methanesulfonate salt of ferrioxamine B (Desferal, Ciba-
Geigy) was determined in methanol at p[H] ) 6.3 ( 0.1 fixed with
0.1 M dichloroacetic buffer by a direct spectrophotometric titration (0
e [Fe(III)]tot/[desferriferrioxamine B]tot e 1.5). The spectrophotometric
data were processed with both the Letagrop-Spefo62,63 and Specfit59-61
programs, which adjust the absorbtivities and the stability constants of
the species formed at equilibrium. Letagrop-Spefo uses the Newton-
Raphson algorithm to solve mass balance equations and a pit-mapping
method to minimize the errors and determine the best values of the
parameters. Specfit uses factor analysis to reduce the absorbance matrix
and to extract the eigenvalues prior to the multiwavelength fit of the
reduced data set according to the Marquardt algorithm. The model
including all ferric species identified by ESMS led to the best fit of
the spectrophotometric data. Apparent stability constants at p[H] )
6.5 ( 0.1 are given in this work. For the sake of simplicity charges
and protons are omitted in all the chemical equilibria given here.
Acknowledgment. This work has been supported by the
Centre National de la Recherche Scientifique (CNRS). E.L and
M.M. thank the CNRS and the Conseil Regional d’Alsace for
granting them a Ph.D. fellowship. The collaborative research
has been carried out with the help of the French-Israeli Keshet
Program. Financial support from the U.S.-Israel Binational
Foundation and the Israel Academy of Sciences and Humanities
is gratefully acknowledged. A.S. is holder of the Siegfried and
Irma Ullman Professorial Chair. This paper is dedicated to the
memory of Dr. Jacqueline Libman.
(77) Cox, B. G.; Truong, N. V. J. Chem. Soc., Perkin Trans. 2 1983,
515-521.
(78) De Ligny, C. L.; Luykx, P. F. M.; Rehbach, M.; Wieneke, A. A.
Recl. TraV. Chim. Pays-Bas 1960, 79, 699-726.
(79) Gelsema, W. J.; De Ligny, C. L.; Remijnse, A. G.; Blijleven, H. A.
Recl. TraV. Chim. Pays-Bas 1966, 85, 647-660.
(80) Alfenaar, M.; De Ligny, C. L. Recl. TraV. Chim. Pays-Bas 1967,
86, 1185-1190.
Supporting Information Available: Figures showing the
spectrophotometric titration data of CDTA and desferriferriox-
amine B by iron(III) (2 pages). See any current masthead page
for ordering and Internet access instructions.
(81) Rondinini, S.; Mussini, P. R.; Mussini, T. Pure Appl. Chem. 1987,
59, 1549-1560.
JA962472C