Synthesis and iron(III) complexing ability of CacCAM, a new analog of enterobactin possessing a free carboxylic anchor arm. Comparative studies with TRENCAM

Article abstract of DOI:10.1039/b000229l

A new tricatecholate, CacCAM, has been synthesized by attachment of three catecholamide subunits to a COzH-functionalized triamine backbone. The solution coordination chemistry of the ligand and its iron(III) complex have been investigated by potentiometric, spectroscopic and kinetic methods. The results are compared to those obtained with TRENCAM, which possesses the same catecholamide subunits, but a TREN [tris(2-aminoethyl)amine] backbone. The stability constant (log β₁₁₀ = 42.9, pFe = 27.5) is close to that of TRENCAM (log β ₁₁₀ = 43.6, pFe = 27.8), showing that a change of backbone does not significantly alter the iron chelating ability of the ligand. Kinetics of iron exchange between Fe(III)-EDTA and CacCAM or TRENCAM involves similar rate constants (16 ± 1 and 14 ± 2 M^-1 s^-1, respectively). EPR data show that the iron(III) does not have the same distorted rhombic geometry in the two complexes. Fe(III)-CacCAM has been tested as a source of iron in nutritional experiments with Arabidopsis thaliana plant cells: its efficiency is good, comparable to that of Fe(III)-EDTA, the most widely used iron complex for plant nutrition.

Full text of DOI:10.1039/b000229l

View Article Online / Journal Homepage / Table of Contents for this issue
Synthesis and iron(III) complexing ability of CacCAM, a new analog
of enterobactin possessing a free carboxylic anchor arm. Comparative
studies with TRENCAM
Daniel Imbert,a Fabrice Thomas,a Paul Baret,a Guy Serratrice,a Didier Gaude,a
Jean-Louis Pierre*a and Jean-Pierre Laulhereb
a L aboratoire de Chimie Biomimetique (L EDSS, CNRS UMR 5616) Universite J. Fourier, BP 53,
38041 Grenoble cedex 9, France. E-mail: Jean-L ouis.Pierre=ujf-grenoble.fr
b CBCRB, DBMS/CB, CEA Grenoble, 17 av. des Martyrs, 38054 Grenoble cedex 9, France
Received (in Strasbourg, France) 29th December 1999, Accepted 18th February 2000
Published on the Web 13th April 2000
A new tricatecholate, CacCAM, has been synthesized by attachment of three catecholamide subunits to a CO H-
2
functionalized triamine backbone. The solution coordination chemistry of the ligand and its iron(III) complex have
been investigated by potentiometric, spectroscopic and kinetic methods. The results are compared to those
obtained with TRENCAM, which possesses the same catecholamide subunits, but a TREN [tris(2-aminoethyl)-
amine] backbone. The stability constant (log b \ 42.9, pFe \ 27.5) is close to that of TRENCAM (log b
\
110
110
43.6, pFe \ 27.8), showing that a change of backbone does not signiÐcantly alter the iron chelating ability of the
ligand. Kinetics of iron exchange between Fe(III)ÈEDTA and CacCAM or TRENCAM involves similar rate
constants (16 ^ 1 and 14 ^ 2 M~1 s~1, respectively). EPR data show that the iron(III) does not have the same
distorted rhombic geometry in the two complexes. Fe(III)ÈCacCAM has been tested as a source of iron in
nutritional experiments with Arabidopsis thaliana plant cells: its efficiency is good, comparable to that of
Fe(III)ÈEDTA, the most widely used iron complex for plant nutrition.
organizing unit (spacer), via amide linkers (Scheme 1). The
Introduction
substitution of the TREN unit by a C-functionalized tripodal
spacer necessarily induces geometrical changes and thus may
involve a di†erent (and undesirable) iron complexing ability.
In this paper we present the synthesis, the iron complexing
and iron nutritional properties towards plant cells of a new
iron chelator, CacCAM (Scheme 1). CacCAM bears a free car-
boxylic group that is located out of the coordination sphere of
the three catechol subunits. It uses the new tripodal spacer
depicted in Scheme 1 (insert). CacCAM may be considered as
the starting species for a series of chelators possessing exactly
the same coordination sphere; thus it is important to charac-
terize the iron complexing abilities of CacCAM.
Most living organisms require iron for their growth. Side-
rophores are low molecular weight iron(III) chelators excreted
by microorganisms for the uptake of iron.1 Interest in syn-
thetic biomimetic siderophores includes: (i) their potential
applications as clinical iron removal agents (iron overload is
one of the most common types of poisoning),2 (ii) the use of
their iron complexes in agriculture for prevention or treatment
of chlorosis3 and (iii) their design as structural probes and
diagnostic tools.4 Siderophores typically contain three cate-
cholate or three hydroxamate groups to bind2 a ferric ion,
giving octahedral complexes. For most applications the syn-
thetic siderophore (usully a tripodal derivative bearing one
bidentate ligand on each arm) and its iron complex have to be
water-soluble and, to achieve this, numerous abiotic iron chel-
ators are derivatized with three sulfonate groups.5 The sul-
fonation strategy leads to highly hydrophilic chelators and
complexes, unable to cross over lipophilic membranes.6
Results and discussion
Synthesis
On the other hand, the interest in grafted iron chelators
(analytical applications, biomedical applications with control
of the hydrophilic/lipophilic balance) or in Ñuorescent labeled
siderophore analogs involves the design of new ligands. Syn-
The synthetic pathway is depicted on Schemes 2 and 3. The
protected 2,3-hydroxybenzoic acid is activated with thionyl
chloride and coupled with 5 in order to obtain 6. Then, three
steps are needed to obtain protected CacCAM, 9. Deprotec-
tion of 9 gives CacCAM, 10. The key synthon 5 has been
obtained in Ðve steps with a global yield of 30% from the
starting commercial reagents.
theses of tripodal precursors bearing on the C axis of the
3
ligand a functionalized anchor group, thus allowing a connec-
tion with any desired substituent, seems to be a promising
strategy previously used by Shanzer and Libman.4
TRENCAM, previously described by Raymond et al.,7 and its
sulfonated derivative TRENCAMS,8 which are analogs of the
natural triscatechol derivative enterobactin, are among the
most powerful synthetic iron chelators. TRENCAM(S) are
built by connecting three catechol subunits to a TREN central
Ligand deprotonation constants
CacCAM possesses seven deprotonation sites (six phenolic
and one carboxylic groups) and is denoted LH . The poten-
7
tiometric titration of the fully protonated ligand with NaOH
DOI: 10.1039/b000229l
New J. Chem., 2000, 24, 281È288
281
This journal is ( The Royal Society of Chemistry and the Centre National de la Recherche ScientiÐque 2000

View Article Online
Scheme 1
in 0.1 M NaClO at 25 ¡C (Fig. 1, curve a) allowed the deter-
mination of only three deprotonation constants of the
catechol groups. The precipitation of the ligand below pH 5
(neutral LH species) precludes the use of these data in the
photometric titration was carried out over the pH range
10È12. The corresponding UV spectra exhibited isosbestic
points at j \ 209, 317 and 347 nm, indicating the presence of
only two absorbing species (Fig. 2). The data were processed
by the LETAGROP SPEFO13,14 program (absorbance values
at 10 wavelengths). The best Ðt [&(A [ A )2 \ 2 ] 10~2]
4
7
reÐnement program. Hence, the lowest pK corresponding to
a
the deprotonation of the carboxyl (pK ) cannot be deter-
a7
mined (pK \ 5). Analysis of the potentiometric titration
a
exp
calc
was obtained by considering a single proton step for the dep-
curve over the pH range 6È10.6 by the SUPERQUAD9
rotonation equilibrium, yielding a pK value of 11.16 ^ 0.02.
a
program yielded the pK values (106 points, p \ 2.2,
This value is close to that obtained for TRENCAM7
an
fit
pKa6 \ 7.19 ^ 0.01, pKa5 \ 7.98 ^ 0.01, pKa4 \ 8.64 ^ 0.01)
deÐned by eqns. (1) and (2). Charges of complexes are omitted
for sake of clarity all along the text [see also eqns. (3), (4) and
(9)].
(11.26 ^ 0.02). The two higher pK values cannot be obtained
a
since oxidation of the catechol is observed above pH 12.
These pK values can be assumed to be close to those (12.9
a
and 12.1) used for TRENCAM7 and enterobactin.11
LH H LH
n~1
K \ [LH ][H`]/[LH ]
an n~1
] H`
(1)
(2)
n
Stability constants of the ferric complexes
n
The equilibria of the metal complexes were studied by poten-
tiometric and spectrophotometric titrations. The poten-
tiometric equilibrium curve of Fe(III)ÈCacCAM (Fig. 1 curve
b) shows a large pH jump at m \ 7, indicating that the
protons of the six catechol and of the carboxyl group are rel-
eased when the ligand binds to ferric ion at pH 7.5. The pre-
cipitation of the complex under experimental conditions (0.3
mM) below pH 6.5 precluded the use of the titration data in
the reÐnement software. A spectrophotometric titration of the
ferric complex (0.09 mM) of CacCAM (Fig. 3) has been carried
out from pH 8.44 to 4.65. At pH below 4.5 precipitation
occurred. Between pH 8.44 and 5.69 [Fig. 3(A)] the spectra
display an isosbestic point at 552 nm and a band in the visible
range corresponding to a LMCT; its wavelength evolved from
495 (pH 8.44) to 519 nm (pH 5.69). Decreasing the pH from
5.43 to 4.65 resulted in the appearance of a new isosbestic
These three pK values reÑect a roughly statistical separation
(log 3 or 0.48), which is consistent with non-interacting arms.
The average of these deprotonation constants is 7.94, which
a
is between the deprotonation constant, pK \ 8.42 of N,N-
a
dimethyl-2,3-dihydroxybenzamide10 and the average of the
deprotonation constants, pK B 7.4, of triscatechol ligands
a
(TRENCAM,7 MECAM and derivatives)11,12 and entero-
bactin.11 The lowest pK of the hydroxy group is roughly one
a
unit higher for CacCAM than that for triscatechol ligands (5.6
for TRENCAM,7 5.9 for MECAM,11 6.2 for MMECAM12).
This may be ascribed to an H-bond between the deprotonated
carboxylate oxygen and the hydroxy group of the catechol.
The three higher pK values of the catechol groups cannot
a
be accurately determined from potentiometric titrations, since
deprotonation of these groups occurs at pH [ 11. A spectro-
Scheme 2 (a) K CO , toluene, 0 ¡C; (b) acrylonitrile, 4 h, 170 ¡C; (c) HClÈH O, reÑux; (d) NaBH , methanol under N ; (e) TBDMSCl, DMAP,
2
3
2
4
2
Et N, CH Cl under N ; ( f ) Raney nickel, hydrazine, ethanol.
3
2
2
2
282
New J. Chem., 2000, 24, 281È288

View Article Online
Scheme 3 (a) Thionyl chloride; (b) Et N, CH Cl under N ; (c) TBAF, THF; (d) oxalyl chloride, DMSO, CH Cl , Et N; (e) sulfamic acid,
3
2
2
2
2
2
3
NaClO , THFÈH O; ( f ) BBr , CH Cl , 0 ¡C.
2
2
3
2 2
point at 620 nm, while j
evolved from 519 to 531 nm [Fig.
The stability constant b
of the ferric complex FeL was
max
110
3(B)]. These absorbance data were reÐned with the nonlinear
least squares program LETAGROP SPEFO. The best reÐne-
ment [&(A [ A )2 \ 10~3] was obtained with the model
determined by spectrophotometric competition experiments
with a 1 to 200-fold excess of EDTA (denoted as Y), over the
pH range 6.81 to 6.62. The competition equilibrium can be
expressed by the following equations:
exp
calc
including three species: [FeL]4~, [FeLH]3~, [FeLH ]2~.
2
The calculated spectra are depicted in Fig. 4. The deprotona-
FeL ] YH ] 5H` H FeY ] LH
(4)
tion constants pK
values obtained for the equilibria (3)
6
FeLHn
K \ [FeY][LH ]/[FeL][YH][H`]5
FeLH H FeLH ] H`
(3)
\ 6.59 ^ 0.08. The
6
n
n~1
\ b
K@ /b
K Ka2 Ka3 Ka4 Ka5 K
(5)
110(FeY) a1 110(FeL) a1
a6
are pKFeLH2 \ 5.15 ^ 0.16 and pK
FeLH
species distribution curves are shown in Fig. 5.
where LH represents the ligand CacCAM with a deprotonat-
6
ed carboxylic group, K ÈK the deprotonation constants of
a1 a6
Fig. 1 Potentiometric titration curves for (a) ligand CacCAM, (b)
1 : 1 CacCAM ] Fe(III); m \ mole of base added per mole of ligand.
Fig. 2 UVÈvisible absorption spectra of CacCAM as a function of
[CacCAM] \ 0.32 mM, T \ 25 ¡C, I \ 0.1 M (NaClO ), 2% MeOH
pH: [CacCAM] \ 0.02 mM, I \ 0.1 M (NaClO ), 2% MeOH by
4
4
by volume. The dashed lines indicate precipitation.
volume. Curve 1 (pH 8.44), curve 2 (pH 11.1).
New J. Chem., 2000, 24, 281È288
283

View Article Online
Fig. 5 Species distribution curves for Fe(III)ÈCacCAM from pH 4.5
to 10: [CacCAM] \ [Fe3`] \ 1 mM.
lated from the mass balance equation and pH:
[Fe] \ b [FeL] ] a [FeY]
tot FeL FeY
(6)
(7)
(8)
[L] \ b [FeL] ] a [L]
tot FeL
L
[Y] \ b [FeY] ] a [Y]
tot FeY
Y
where the a are the usual Ringbom coefficients.16 Using the
known formation constant of FeY (log b \ 25.1),15 the
110
Fig. 3 UVÈvisible absorption spectra of Fe(III)ÈCacCAM as a func-
tion of pH: [CacCAM] \ 0.09 mM, [Fe] \ 0.09 mM, I \ 0.1 M
average formation constant of the ferric CacCAM was deter-
mined to be log b \ 42.5(3). To measure the efficiency of
110
(NaClO ). (A) Curve 1 (pH 8.44), curve 2 (pH 5.69); (B) curve 3 (pH
the ligand CacCAM for iron(III) at physiological pH we have
4
5.43), curve 4 (pH 4.65).
calculated the concentration of free Fe3` under standard con-
ditions (pH \ 7.4, [L] \ 10~5 M, [Fe] \ 10~6 M), which
tot
tot
LH and K@ the deprotonation constant of HY (10~10.2).15
is expressed as pFe (pFe \ [log [Fe3`]). The obtained value
for CacCAM is 27.5, which is very close to that of
TRENCAM (27.8).7 The substitution of the TREN
[(N(CH CH ) ] for the Cac [HO CC(CH CH CH ) ] back-
6
a1
The concentration of FeL was calculated from the absorbance
at 500 nm where FeL is the only absorbing species. The con-
centrations of the other species in eqns. (4) and (5) were calcu-
2
2 3
2
2
2
2 3
bone does not change signiÐcantly the complexing efficiency
of the ligand.
The visible spectrum of the FeL species (j \ 495 nm) is
max
similar to that of the iron complexes of triscatecholate ligands
or of enterobactin (Table 1). However, the extinction coeffi-
cient (5900 M~1 cm~1) is higher than that generally measured
for other triscatecholate complexes (4900 M~1 cm~1) but
close to that of the ferric enterobactin (5700 M~1 cm~1)11 and
the ferric MECAMS (5900 M~1 cm~1).17 The UVÈvisible
spectra of [FeLH]3~ and [FeLH ]2~ species exhibited shifts
2
in j
from 495 (pH 8.44) to 519 (pH 5.69) and 531 nm (pH
max
4.65). Similar shifts have been observed for catecholate com-
plexes and have been attributed to a change from a catechol-
ate coordination mode to a salicylate coordination mode
(coordination by one hydroxyl and O-amide for each arm of
the tripodal ligand) induced by stepwise protonations of the
m-hydroxy groups, that is a mono-salicylate, bis-catecholate
coordination in [FeLH]3~ and a mono-chatecholate, bis-
Fig. 4 Calculated extinction coefficient spectra of Fe(III)ÈCacCAM
complexes: [CacCAM] \ 0.09 mM, [Fe] \ 0.09 mM, I \ 0.1 M
(NaClO ).
4
Table 1 Equilibrium constants and UVÈvisible spectral characteristics of ferric complexes with triscatechol ligands
/nm (e/103 M~1 cm~1)
j
max
Ligand
pK
pK
log b
pFe
FeL
FeLH
FeLH
FeLH2
FeLH
110
2
CacCAM
5.15
È
4.2
6.03
4.10
3.52
6.59
5.59
5.5
7.2
5.74
4.95
42.5
43.2
41.3
43
41
49
27.5
27.8
29.6
29.4
29.4
35.5
495(5.9)
496(4.9)
488(5.3)
492(4.7)
485(5.9)
498(5.7)
519(5.0)
510(4.1)
500(4.6)
515(3.9)
490(5.1)
514(4.6)
531(4.0)
È
È
È
È
TRENCAMa
TRENCAMSb
MECAMc
MECAMSd
Enterobactinc
518(3.5)
a Ref. 7. b Ref. 8. c Ref. 11. d Ref. 17.
284
New J. Chem., 2000, 24, 281È288

View Article Online
a slightly rhombic distorted geometry around the iron center.
Spectra of Fe(III)ÈTRENCAM and Fe(III)ÈTRENCAMS are
similar, as shown in Fig. 6 (the di†erential spectrum is very
close to a baseline). They are both centered at g \ 4.28. The
spectrum of Fe(III)ÈCacCAM shows a signal centered at
g \ 4.25 with
a larger linewidth than that for Fe(III)È
TRENCAM. The larger linewidth may be interpreted as due
to a less tightly bonded Fe(III) with CacCAM than with
TRENCAM: in the former the carboxylic group is located
outside of the ““umbrellaÏÏ of the three arms enclosing the iron
atom, while in the latter the nitrogen electron pair is inside the
““umbrellaÏÏ of the TREN backbone and can establish hydro-
gen bonds with amide protons, leading to a tight cavity
around the ferric ion.
Kinetic studies
The iron exchange between the ferric EDTA complex and
CacCAM or TRENCAM has been measured at pH 7.0
(pseudo-Ðrst-order conditions with ligand concentration in
excess with respect to the complex) by monitoring the increase
in absorbance at 500 nm on the minute time scale. A mono-
exponentional signal was observed and could be Ðtted to
Fig. 6 ESR spectra of Fe(III)ÈTRENCAMS (spectrum 1), Fe(III)È
TRENCAM (spectrum 2) and Fe(III)ÈCacCAM (spectrum 3): 1 mM
concentration in 10% glycerol, 0.1 M Tris-Cl bu†er, pH 7.2, T \
100 K.
obtain the pseudo-Ðrst-order rate constants k . The plot of
obs
k
vs. the ligand concentration, presented in Fig. 7, shows a
obs
linear dependence. From the slope, the second-order rate con-
stants k can be calculated from:
e
salicylate coordination in [FeLH ]2~. Table 1 summarizes the
d[FeL]/dt \ k [FeEDTA]
2
obs
tot
spectral characteristics and the pK of several catecholate
a
\ k [LH ] [FeEDTA]
(9)
complexes and their protonated species. The Ðrst two proto-
nation constants of the Fe(III)ÈCacCAM complex (6.59 and
e
6 tot
tot
corresponding to the global reaction:
5.15 for FeLH and FeLH ) are lower than those for Fe(III)È
2
ke
MECAM11 (7.02 and 6.03) but signiÐcantly higher than those
FeEDTA ] LH ÈÈÈ FeL ] YH ] 5H`
(10)
6
for enterobactin11 (4.95 and 3.52). The relatively high pK
a
The values of k are 16 ^ 1 and 14 ^ 2 M~1 s~1 for CacCAM
and TRENCAM, respectively.
values for FeÈCacCAM reÑects the stability of the bis-
e
salicylate complex ([FeLH ]2~) at high pH. It should be
2
At pH 7, the reaction proceeds through four paths when
taking into account the hydrolysis reaction of the ferric EDTA
complex (K \ 10~7.37)15 and the LH ~/LH 2~ equi-
noted that the di†erence between the value of pK
and
FeLH2
pK
varies over the range 1.2È1.6 for the various complex-
es. This indicates that the equilibrium constant between the
FeLH
OH
6
5
librium (Ka6 \ 10~7.19). The determination of the rate con-
mono-[FeLH]3~ and the bis-salicylate [FeLH ]2~ complexes
2
stants for the di†erent paths requires the measurement of k
depends very little on the tripodal backbone of the ligand.
obs
at various pH, but this was not done in this study. A value of
3.3 M~1 s~1 was determined under the same conditions for
EPR spectroscopy
O-TRENSOX,18
a
tripodal
ligand
containing
8-
The X-band EPR spectra of frozen aqueous solutions (with
10% glycerol) of iron(III) complexes of TRENCAM, TREN-
CAMS and CacCAM were recorded at 100 K (pH 7.2). The
spectra exhibit one single dissymetric signal at a g value close
to 4.3, characteristic of a high-spin octahedral ligand Ðeld with
hydroxyquinoline chelating units. Our results show that the
exchange rate of Fe(III) between FeÈEDTA and tripodal
ligands are higher for ligands with catechol groups.
Plant nutrition studies
The growth and greening of Arabidopsis thaliana plant cells
fed with Fe(III)ÈCacCAM and Fe(III)ÈTRENCAMS as the
only source of iron has been studied. In one control experi-
ment, iron was provided with Fe(III)ÈEDTA and in another,
no iron source was added. Results obtained for CacCAM are
shown in Fig. 8 and 9. Growth and resistance to chlorosis are
equivalent when Fe(III)ÈCacCAM or Fe(III)ÈTRENCAMS are
used instead of Fe(III)ÈEDTA. The efficiency of Fe(III)È
CacCAM and Fe(III)ÈTRENCAMS as an iron source for A.
thaliana is good, comparable to that of Fe(III)ÈEDTA, the
most commonly used iron source for plant nutrition.
Conclusions
The solution coordination chemistry studies described in this
paper have shown that CacCAM, an analog of enterobactin,
possesses properties similar to those of TRENCAM, a known
and efficient iron chelator. Moreover, preliminary biological
studies have shown that CacCAM possesses good efficiency
for iron plant nutrition. The fact that the new C-
functionalized spacer involved in the structure of CacCAM
Fig. 7 Rate constants k
(s~1) for iron-exchange reactions from
obs
Fe(III)ÈEDTA to CacCAM and TRENCAM (2% MeOH by volume):
[Fe(III)ÈEDTA] \ 0.02 mM, [CacCAM, TRENCAM] \ 0.08È0.8
mM; solvent: [HEPES] \ 0.05 M, I \ 0.1 M (NaClO ), pH 7.0,
4
T \ 25 ¡C.
New J. Chem., 2000, 24, 281È288
285

View Article Online
studies have been conducted using a ESP 300E Bruker appar-
atus with a variable temperature unit. Spectra were treated by
using the WinEPR software (Bruker). Each aqueous sample
(200 lL) contained 10% glycerol, 0.1 M Tris-Cl bu†er, pH 7.2,
10~3 M FeÈCacCAM and FeÈTRENCAMS, 0.5 ] 10~3 M
FeÈTRENCAM. The temperature was 100 K and the Ðeld
was scanned from 1000 to 5000 gauss.
Synthesis of CacCAM
4-(2-Cyanoethyl)-4-cyclohexyliminomethylheptane-1,7-
dinitrile (1). Acetaldehyde (16.9 mL) was slowly added to a
solution of cyclohexylamine (34.3 g, 0.3 mmol) in 15 ml of dry
toluene at 0 ¡C, over 20 min. K CO (2.5 g) was added and the
2
3
reaction mixture was stirred for 10 min and then allowed to
warm to room temperature. The organic layer was placed in
an autoclave and 68.5 mL of acrylonitrile was added. The
solution was stirred for 4 h at 170 ¡C. The black reaction
mixture was cooled to [20 ¡C for 6 h. The crystalline product
was washed with diethyl ether to give 41 g (white crystals,
48%) of 1, which was used without further puriÐcation. Mp
99È100 ¡C. IR (KBr) l(cm~1): 2248, 1663. 1H NMR (300
MHz, CDCl ): d 1È2 (m, 16H, CH CH CN ] CH cyc), 2.37
Fig. 8 Growth of A. thaliana cells using Fe(III)ÈEDTA or Fe(III)È
CacCAM as siderophores: 25 ¡C, 220 rpm, 18 h of light per day.
3
2
2
2
(t, 6H, J \ 7.9 Hz, CH CN), 3.01 (m, 1H, CH cyc), 7.45 (s, 1H,
2
CH2NÈ). 13C NMR (300 MHz, CDCl ): d 11.67, 29.34
3
(CH CH CN), 24.16, 25.26, 34.02 (CH cyc), 43.07 (Cq), 69.29
2
2
2
(CH cyc), 119.13 (CN), 161.23 (CH).
4-(2-Cyanoethyl)-4-formylheptane-1,7-dinitrile (2). Com-
pound 1 (10 g, 35.2 mmol) was dissolved in a solution of con-
centrated HCl (5 mL) and water (130 mL). The resulting
mixture was reÑuxed for 30 min, Ðltered hot and allowed to
cool to 0 ¡C. The yellow precipitate was collected, washed
with water, dried and puriÐed by recrystallisation from meth-
anol (white crystals, 6.6 g, 92%). Mp 108 ¡C. IR (KBr) l
Fig. 9 Growth of A. thaliana cells using Fe(III)ÈEDTA or Fe(III)È
CacCAM as siderophores. (Curve 1) 50 lM Fe(III)ÈCacCAM, (curve
2) 50 lM Fe(III)ÈEDTA, (curve 3) 10 lM Fe(III)ÈEDTA, (curve 4) 10
lM Fe(III)ÈCacCAM, (curve 5) without iron: 25 ¡C, 220 rpm, 18 h of
light per day.
(cm~1): 2250, 1716. 1H NMR (300 MHz, DMSO-d ): d 1.84
6
(t, 6H, J \ 7.9 Hz, CH CH CN), 2.40 (t, 6H, J \ 7.9 Hz,
2
2
CH CN), 9.42 (s, 1H, CHO). 13C NMR (300 MHz, DMSO-
d ): d 12.07, 26.94 (CH CH CN), 50.91 (Cq), 118.01 (CN),
2
6
2
2
does not alter the complexing ability of the ligand (compared
to the widely used TREN spacer) is very promising. The car-
boxylic function located outside of the coordination sphere
may allow the connection of any substituent. Studies are in
progress concerning (i) the grafting of a Ñuorescent probe and
(ii) the grafting of polyoxyethylene units allowing control of
the hydrophic/lipophilic balance without the help of the
charged and hydrophilic sulfonate groups commonly used to
confer water solubility.
200.88 (CHO).
4-(2-Cyanoethyl)-4-hydroxymethylheptane-1,7-dinitrile (3).
NaBH (1.5 g) was added to a solution of 2 (5 g, 24.6 mmol) in
4
distilled methanol (250 mL) under nitrogen at 0 ¡C within 30
min. The solution was stirred for an additional 2 h at room
temperature. Water (50 mL) was added and the resulting
mixture was cooled to 0 ¡C and then acidiÐed with concen-
trated HCl to pH 1. Methanol was evaporated and the
product was extracted with dichloromethane (3 ] 75 mL), the
combined extracts were dried with Na SO . The solvent was
Experimental
2
4
removed to yield 3 as a white crystalline material (4.6 g, 92%).
Mp 69 ¡C. IR (KBr) l(cm~1): 3462, 2248. 1H NMR (300
MHz, CDCl ): d 1.79 (t, 6H, J \ 7.8 Hz, CH CH CN), 2.44 (t,
Materials and equipment
3
2
2
Solvents were puriÐed by the usual techniques. All other com-
pounds were of reagent grade and were used without further
puriÐcation. Iron(III) stock solutions were prepared by dis-
solving appropriate amounts of ferric perchlorate hydrate
(Aldrich) in standardized HClO and NaClO solutions. The
6H, J \ 7.8 Hz, CH CN), 3.56 (d, 2H, J \ 3.6 Hz, CH OH).
2
2
13C NMR (300 MHz, CDCl ): d 11.85, 29.35 (CH CH CN),
39.59 (Cq), 65.02 (CH OH), 119.37 (CN). MS (DCI, NH
] isobutane): m/z 206 (M ] 1).
3
2
2
2
3
4
4
solutions were spectrophotometrically calibrated for ferric ion
by using a molar extinction coefficient of 4160 M~1 cm~1 at
240 nm.19 IR spectra were collected on a Nicolet Impact 400
spectrometer. UVÈvisible absorption spectra were recorded on
a PerkinÈElmer Lambda 2 spectrometer using 1.000 cm path
length quartz cells and connected to an IBM PC 340 micro-
computer. Mass spectra were recorded on a NERMAG 101 ¡C
mass spectrometer. Microanalysis were performed by the
Central Service of CNRS (Solaize, France). Melting points
were determined with a Buchi apparatus and are not cor-
rected. 1H and 13C NMR spectra were obtained in 5 mm
tubes at 25 ¡C with a Bruker AM 300 spectrometer. EPR
4-(tert-Butyldimethylsilanyloxymethyl)-4-(2-
cyanoethyl)-heptane-1,7-dinitrile (4). To a solution of 3 (2 g,
9.75 mmol) in freshly distilled dichloromethane (50 mL) under
nitrogen, were added tert-butyldimethylsilyl chloride (1.62 g,
10.72 mmol), 4-dimethylaminopyridine (1.19 g, 9.75 mmol)
and triethylamine (1.37 mL, 9.75 mmol). The solution was
heated to reÑux for 48 h, cooled, washed with brine, Ðltered
and dried with Na SO . Evaporation of the solvent gave 4 as
2
4
a white solid (2.55 g, 82%). Mp 72 ¡C. IR (KBr) l(cm~1):
2247. 1H NMR (250 MHz, CDCl ): d 0.04 (s, 6H, CH ), 0.87
3
3
(s, 9H, CH of But), 1.68 (t, 6H, J\7.8 Hz, CH CH CN), 2.31
3
2
2
(t, 6H, J \ 7.8 Hz, CH CN), 3.34 (s, 2H, CH O). 13C NMR
2
2
286
New J. Chem., 2000, 24, 281È288

View Article Online
(250 MHz, CDCl ): d [5.78 (CH ), 25.59 (CH of But),
(OCH ), 66.14 (CH O), 115.09 (CH), 122.61 (CH), 124.25 (CH),
126.79 (Cq), 147.29 (Cq), 152.45 (Cq), 165.19 (C2O). MS (FAB,
3
3
3
3
2
11.75, 29.38 (CH CH CN), 17.96 (CqSi), 39.80 (Cq), 65.77
2
2
(CH ), 119.31 (CN). MS (DCI, NH ] isobutane): 320
NBA matrix): m/z 711 [M ] 1]`. Anal. calcd. (found) for
2
3
(M ] 1).
C
H
N O É 1.25CH Cl É 0.75MeOH: C, 57.16 (57.19); H,
38 51
3
10
2
2
6.57 (6.78); N, 4.83 (4.99)%.
4-(3-Aminopropyl)-4-(tert-butyldimethylsilanyloxymethyl)
heptane-1,7-diamine (5). To a solution of 4 (5 g, 15.65 mmol)
in 95% ethanol (60 mL) at 0 ¡C, were added NaOH (2.4 g, 60
mmol) and hydrazine 99% (6 mL). To this mixture Raney
nickel (700 mg, 12 mmol) was added in small portions within
2 h; the solution was stirred for an additional 1 h. The reac-
tion mixture was heated to reÑux for 45 min, Ðltered hot and
evaporated to give an oil. Addition and evaporation of
toluene were repeated until NaOH precipitated. Evaporation
of the Ðltrate gave 5 as an oil (4.77 g, 88%). IR (KBr) l(cm~1):
2,2,2-Tris[3-(2,3-dimethoxybenzamido)propyl]ethanal (8). A
solution of dimethyl sulfoxide (51 lL, 0.66 mmol) in dry
dichloromethane (150 lL) was added to a solution of oxalyl
chloride (30 kl, 0.33 mmol) in freshly distilled dichloromethane
(1 ml). The mixture was stirred for 5 min and then a solution
of 7 (0.21 g, 0.3 mmol) in dry dichloromethane (200 lL) was
added within 2 min; the solution was stirred for an additional
15 min. Triethylamine (420 lL, 3 mmol) was added and the
reaction mixture was allowed to warm to room temperature.
Water (20 mL) and dichloromethane (50 mL) were added and
the organic layer was washed with brine and dried with
3375, 1607, 1442. 1H NMR (300 MHz, DMSO-d ): d 0.01 (s,
6
6H, CH ), 0.85 (s, 9H, CH of But), 1.10 (m, 6H, CH ), 1.21 (m,
3
3
2
6H, CH ), 2.40 (t, 6H, J \ 6.8 Hz, CH NH ), 3.23 (s, 2H,
MgSO . The solvent was removed and the yellow residue was
2
2
6
2
4
CH O). 13C NMR (300 MHz, DMSO-d ): d [5.82 (CH ),
puriÐed by column chromatography (silica gel; 1% MeOH in
2
3
25.75 (CH of But), 27.13, 30.96 (CH ), 42.99 (CH NH ), 17.93
CH Cl ). 8 was obtained as a white foam (0.17 g, 82%). IR
3
2
2
2
2
2 2
(CqSi), 39.08 (Cq), 66.12 (CH O). MS (DCI, NH
(Ðlm from CHCl ) l(cm~1): 3380, 2939, 1721, 1651. 1H NMR
3
3
] isobutane): m/z 332 (M ] 1). Anal. calcd. (found) for
(300 MHz, CDCl ): d 1.48 (m, 6H, CH ), 1.59 (m, 6H, CH ),
3.34 (br q, 6H, CH ), 3.79 (br s, 18H, OCH ), 6.90 (dd, 3H,
3
2
2
C
H
N OSi É 0.25CH Cl : C, 58.79 (58.72); H, 12.10
17 41
3
2
2
2
3
(11.85); N, 11.67 (11.91)%.
J \ 8.1 Hz, J \ 1.6 Hz, ArH), 7.07 (t, 3H, J \ 8.0 Hz, ArH),
1
2
1
7.56 (dd, 3H, J \ 8.1 Hz, J \ 1.6 Hz, ArH), 7.94 (br t, 3H,
2
2,2,2-Tris[3-(2,3-dimethoxybenzamido)propyl]-tert-
NH), 9.37 (s, 1H, CHO). 13C NMR (300 MHz, CDCl ): d
3
butyldimethylsilanyloxyethane (6). 2,3-Dimethoxybenzoic acid
(0.91 g, 5 mmol) was dissolved in 10 mL of thionyl chloride.
The reaction mixture was stirred under nitrogen for 6 h. The
solution was evaporated to dryness and the residue treated
with 4 ] 40 mL of dry hexane. The solid residue was dissolved
in freshly distilled dichloromethane (50 mL) under nitrogen.
Compound 5 (0.5 g, 1.51 mmol), dissolved in a solution of
freshly distilled dichloromethane (50 mL) and triethylamine
(0.9 mL), was added to the solution of acid chloride within 30
min. The mixture was stirred overnight at room temperature,
24.22 (CH ), 32.06 (CH ), 40.01 (CH NH), 48.21 (Cq), 56.1
2
2
2
(OCH ), 61.42 (OCH ), 115.39 (CH), 122.79 (CH), 124.22 (CH),
3
3
126.83 (Cq), 147.48 (Cq), 152.41 (Cq), 165.38 (C2O amide),
201.05 (C2O aldehyde).
2,2,2-Tris[3-(2,3-dimethoxybenzamido)propyl]acetic
acid
(9). Compound 8 (158 mg, 0.22 mmol) was dissolved in THF
(3 mL) and water (3 mL). Sulfamic acid (H NSO H, 29 mg,
2
3
0.29 mmol) was added, the solution was stirred for 5 min and
then sodium chlorite (NaClO , 26 mg, 0.29 mmol) was added.
2
Ðltered, washed with brine and dried with MgSO and then
The reaction mixture was stirred for 1 h at room temperature.
4
evaporated to dryness. The white residue was puriÐed by
CH Cl (40 mL) and water (50 mL) were then added and the
2
2
column chromatography (silica gel; 1% MeOH in CH Cl ). 6
organic layer was washed with water and dried with MgSO .
2
2
4
was obtained as a white foamy material (0.72 g, 58%). IR (Ðlm
The solvent was removed and the white residue was puriÐed
from CHCl ) l(cm~1): 3380, 1648. 1H NMR (300 MHz,
by column chromatography (silica gel; 4% MeOH in
3
CDCl ): d 0.08 (s, 6H, CH ), 0.75 (s, 9H, CH of But), 1.24 (m,
CH Cl ). Evaporation of the eluant gave 9 as a white foam
3
3
3
2 2
6H, CH ), 1.45 (m, 6H, CH ), 3.25 (s, 2H, CH O), 3.34 (br q,
(145 mg, 90%). 1H NMR (300 MHz, CDCl ): d 1.48 (m, 6H,
2
2
2
3
6H, J \ 6.44 Hz, CH NH), 3.80 (br s, 18H, OCH ), 6.94 (dd,
CH ), 1.59 (m, 6H, CH ), 3.35 (br q, 6H, CH ), 3.79 (br s, 18H,
2
3
2
2
1
2
3H, J \ 8.0 Hz, J \ 1.6 Hz, ArH), 7.05 (t, 3H, J \ 8.3 Hz,
OCH ), 6.91 (dd, 3H, J \ 8.1 Hz, J \ 1.6 Hz, ArH), 7.01 (t,
1
2
3
2
1
ArH), 7.58 (dd, 3H, J \ 8.0 Hz, J \ 1.6 Hz, ArH), 7.90 (br t,
3H, J \ 8.0 Hz, ArH), 7.54 (dd, 3H, J \ 7.9 Hz, J \ 1.6 Hz,
1
2
2
3H, NH). 13C NMR (300 MHz, CDCl ): d [5.61 (CH ),
ArH), 7.97 (br t, 3H, NH). 13C NMR (300 MHz, CDCl ): d
3
3
2
3
18.11 (CqSi), 23.38 (CH ), 25.81 (CH of But), 31.63 (CH ),
24.33 (CH ), 32.1 (CH ), 40.22 (CH NH), 48.29 (Cq), 56.05
2
3
2
2
2
39.34 (Cq), 40.49 (CH NH), 56.02 (OCH ), 61.29 (OCH ),
(OCH ), 61.39 (OCH ), 115.37 (CH), 122.77 (CH), 124.99 (CH),
2
3
3
3
3
66.87 (CH ), 115.15 (CH), 122.73 (CH), 124.30 (CH), 126.91
(Cq), 147.36 (Cq), 152.51 (Cq), 165.21 (C2O). MS (FAB, NBA
126.72 (Cq), 147.52 (Cq), 152.55 (Cq), 165.41 (C2O amide),
2
177.09 (CO H). MS (FAB, NBA matrix): 725 [M ] 1]`.
2
matrix): m/z 825 [M ] 1]`. Anal. calcd. (found) for
Anal. calcd. (found) for C
H
N O É CH Cl : C, 58.20
38 49
3
11
2 2
C
H
N O Si É CH Cl : C, 59.35 (59.46); H, 7.40 (7.43); N,
(57.92); H, 6.42 (6.36); N, 5.28 (5.20)%.
44 65
3
10
2 2
4.57 (4.62)%.
2,2,2-Tris[3-(2,3-dihydroxybenzamido)propyl]acetic
acid
2,2,2-Tris[3-(2,3-dimethoxybenzamido)propyl]ethanol
(7).
(10). To a solution of 9 (0.2 g, 0.28 mmol), dissolved in freshly
Compound 6 (0.5 g, 0.61 mmol) was dissolved in THF (10 mL)
at 0 ¡C. Tetrabutylammonium Ñuoride (0.57 g, 1.8 mmol) was
slowly added and the green reaction mixture was heated at
reÑux for 2 h. The solution was evaporated and the resulting
residue was dissolved in dichloromethane, washed with brine
distilled dichloromethane and cooled to 0 ¡C under nitrogen,
was added a solution of BBr in dichloromethane (1 M, 5
3
mL). The yellow suspension was stirred overnight at room
temperature. Methanol was added and the solvents were
evaporated. Addition and evaporation of methanol were
repeated Ðve times. The residue was applied to a Sephadex
LH-20 column (50% MeOHÈCH Cl ). A white product,
and dried with MgSO . The orange-green residue was puri-
4
Ðed by column chromatography (silica gel; 3% MeOH in
2
2
CH Cl ). 7 was obtained as a white foam (0.38 g, 89%). IR
eluted as a single band, was obtained (138 mg, 78%). 1H
2
2
(Ðlm from CHCl ) l(cm~1): 3378, 2937, 1643. 1H NMR (300
NMR (300 MHz, DMSO-d ): d 1.45 (m, 12H, CH ), 3.23 (br
3
6
2
MHz, CDCl ): d 1.27 (m, 6H, CH ), 1.48 (m, 6H, CH ), 3.34
q, 6H, CH NH), 6.66 (t, 3H, J \ 8.01 Hz, ArH), 6.88 (dd, 3H,
3
2
2
2
2
(m, 8H, CH NH ] CH O), 3.79 (br s, 18H, OCH ), 6.92 (dd,
J \ 8.0 Hz, J \ 1.2 Hz, ArH), 7.24 (dd, 3H, J \ 8.0 Hz,
2
3
1
2
2
1
3H, J \ 8.0 Hz, J \ 1.6 Hz, ArH), 7.07 (t, 3H, J \ 8.0 Hz,
J \ 1.2 Hz, ArH), 8.79 (t , 3H, J \ 5.4 Hz, NH), 12.84 (br s,
1
2
ArH), 7.65 (dd, 3H, J \ 8.0 Hz, J \ 1.6 Hz, ArH), 7.96 (br t,
1H, CO H). 13C NMR (300 MHz, DMSO-d ): d 23.75 (CH ),
1
2
2
6
2
3H, NH). 13C NMR (300 MHz, CDCl ): d 23.20 (CH ), 31.08
31.49 (CH ), 39.37 (CH NH), 48.51 (Cq), 114.8 (Cq), 117.05
3
2
3
2 2
(CH), 117.88 (CH), 118.76 (CH), 146.20 (Cq), 149.75 (Cq),
(CH ), 39.34 (Cq), 40.37 (CH NH), 55.95 (OCH ), 61.23
2
2
New J. Chem., 2000, 24, 281È288
287

View Article Online
169.70 (C2O amide), 177.68 (CO H). MS (FAB, NBA matrix):
The measurements of cell density were made daily. It was
often hampered by the occurrence of condensation droplets
on the cover of the 24-wells plates. To minimize light scat-
tering during densitometry, the numerized images of the cul-
tures were recorded from beneath through a mirror. The
Bio-Rad video camera system Geldoc 1000 was used, with the
molecular analyst software for quantiÐcation. To record
images of the plates, the hood of the camera was rotated to a
horizontal position. A mirror was introduced through the
open door of the hood, at 45¡ across the horizontal light path,
held by a special stand, which also maintained the 24-well
culture plate horizontally on a glass plate and held the Geldoc
light source, upside 10 cm above the cultures. Full frame
images were captured under constant light conditions, that is
the white light being adjusted so that just a few saturated
pixels appeared in red in the empty optical Ðeld. Under these
2
m/z
641
[M ] 1]`.
Anal.
calcd.
(found)
for
C
H N O É 0.75CH Cl É 0.25MeOH: C, 55.70 (55.72); H,
32 37
3
11
2
2
5.85 (5.60); N, 6.19 (5.91)%.
Potentiometric experiments
All the measurements were made at 25 ¡C in deionized and
twice-distilled water. The ionic strength was kept constant at
0.1 M (sodium perchlorate, PROLABO, Puriss). Due to the
low solubility of the ligand in water, 2% MeOH was added.
Such a low ratio of MeOH does not change signiÐcantly the
solvent properties7 and thus the measured pH values were not
corrected. The potentiometric titrations were performed using
an automatic titrator system, DMS Titrino (Metrohm), with a
combined glass electrode (Metrohm Ðlled with NaCl saturated
solution) and connected to an IBM Aptiva microcomputer.
The electrodes were calibrated to read the pH according to
the classical method.The ligand and its iron(III) complex (0.5
mM) were titrated with standardized 0.02 M NaOH. Argon
was bubbled through the solutions to exclude CO and O .
light conditions,
a reproducible relationship was found
between the absorbance of wells in densitometric scanning
proÐles of recorded images and the amount of cells in the
wells. Images of culture plates without cells were used as base-
line reference. They showed an uneven light background. This
fact was compensated by the subtraction of an adapted base-
line. Measurements and baseline proÐles were drawn using the
rectangle tool of the image analysis software. Baselines were
drawn by joining the crest highlight points in each image
between the wells and then subtracted. This method was
found to be more accurate and reproducible than the use of 24
grids, or the elliptical objects tools proposed by the software.
2
2
Carbonate content was checked by GranÏs method. The titra-
tion data (106 points, pH range 6.0È10.6) were reÐned by the
nonlinear least squares Ðtting program SUPERQUAD9 to
determine the ligand deprotonation constants.
Spectrophotometric measurements
The temperature was maintained at 25 ¡C with a variable tem-
perature unit PerkinÈElmer PTP-1 and the acquisition was
made with the UV Winlab software (PerkinÈElmer). The ionic
strength has been kept constant at 0.1 M (sodium perchlorate
and pH bu†er when indicated). Methanol (2%) was added for
ligand titration and kinetic measurements. pH measurements
were made with a 713 Metrohm digital pH meter equipped
with a microelectrode. The titration data were analysed using
the nonlinear least squares Ðtting program LETAGROP
SPEFO.13,14 The calculations were performed using absorb-
ance values at 10 wavelengths and the residual square sums of
References
1
2
J. B. Neilands, J. Biol. Chem., 1995, 270, 26723.
G. S. Tilbrook and R. C. Hider, in Iron T ransport and Storage in
Microorganisms, Plants and Animals, ed. A. Sigel and H. Sigel, M.
Dekker, New York, 1998, ch. 18, p. 691.
3
4
C. Caris, P. Baret, C. Beguin, G. Serratrice, J.-L. Pierre and J.-P.
Laulhere, Biochem. J., 1995, 312, 879.
A. Shanzer and J. Libman, in Iron T ransport and Storage in
Microorganisms, Plants and Animals, ed. A. Sigel and H. Sigel, M.
Dekker, New York, 1998, ch. 8, p. 329.
the Ðt were 10~3. In order to determine b , a competition
110
5
A. M. AlbrechtÈGary and A. L. Crumbliss, in Iron T ransport and
Storage in Microorganisms, Plants and Animals, ed. A. Sigel and
H. Sigel, M. Dekker, New York, 1998, ch. 7, p. 239.
F. Thomas, P. Baret, D. Imbert, J.-L. Pierre and G. Serratrice,
Bioorg. Med. Chem. L ett., 1999, 9, 3035.
with EDTA was carried out for one week in the dark at 25 ¡C
(HEPES bu†er). Kinetic data were saved in ASCII format;
from these, Biokine ASCII Ðles were built. These data were
treated by using the BIOKINE 3.02 software (BIO LOGIC
Co., Claix, France).
6
7
8
9
S. J. Rodgers, C. W. Lee, C. Y. Ng and K. N. Raymond, Inorg.
Chem., 1987, 26, 1622.
F. Thomas, C. Beguin, J.-L. Pierre and G. Serratrice, Inorg. Chim.
Acta, 1999, 291, 148.
P. Gans, A. Sabatini and A. Vacca, J. Chem. Soc., Dalton T rans.,
1985, 1195.
Cell culture and growth measurement
Plant cells were grown in 24-well sterile cell culture plates
(Nunc); each independent axenic culture contained 1.5 ml of
culture medium. This was a modiÐed Murashige and Skoog
(MS) medium:20 for 1000 ml of medium 4.3 g macro and
micro elements powder (provided by Duchefa, catalog number
M0221, Haarlem Netherlands), 10 ml H KPO (20 g l~1), 0.5
10 T. M. Garrett, P. W. Miller and K. N. Raymond, Inorg. Chem.,
1989, 28, 128.
11 L. D. Loomis and K. N. Raymond, Inorg. Chem., 1991, 30, 906.
12 Z. Hou, T. D. P. Stack, C. J. Sunderland and K. N. Raymond,
Inorg. Chim. Acta, 1997, 263, 341.
2
4
13 L. G. Sillen and B. Warsquist, Ark. Kemi, 1969, 31, 377.
14 M. Meloun, J. Havel and E. Hogfeldt, Computation of Solution
Equilibria, Ellis Horwood, Chichester, 1988.
15 R. M. Smith and A. E. Martell, Critical Stability Constants,
Plenum Press, New York, 1989, vol. VI.
ml Kinetin (0.1 g l~1), 0.5 ml of (2,4-dichlorophenoxy) acetic
acid (0.2 g l~1); 1 ml of vitamin solution (Duchefa M 0409)
and 30 g saccharose. The complex Fe(III)ÈCacCAM was Ðrst
prepared at low pH then raised to pH 5.8 with NaOH and a
tris-maleate bu†er. Wells were inoculated at an absorbance of
0.003 (D1/12) with early stationary phase Arabidopsis thaliana
(var. Columbia) cell suspension. For axeny, plates were sealed
with Magic ScotchTM tape. 24-Well plates were agitated at 220
rpm at 25 ¡C in a New Brunswik Innova 4230 refrigerated
incubator shaker. Light was supplied 18 h a day by two Sylva-
nia Growlux 15-watt bulbs in the photosynthesis accessory.
16 A. Ringbom, Complexation in Analytical Chemistry, Interscience,
New York, 1963.
17 W. R. Harris, K. N. Raymond and F. L. Weitl, J. Am. Chem. Soc.,
1981, 103, 2667.
18 H. Boukhalfa, Ph.D. thesis, Universite de Grenoble, 1996.
19 R. Bastian, R. Weberling and F. Palilla, Anal. Chem., 1956, 28,
459.
20 T. Murashige and F. Skoog, Physiol. Plant., 1962, 15, 473.
288
New J. Chem., 2000, 24, 281È288