Prediction of 3-hydroxypyridin-4-one (HPO) log K1 values for Fe(iii)

Article abstract of DOI:10.1039/c2dt31254a

As a means to aid in the design of 3-hydroxypyridin-4-ones (HPOs) intended for use as therapeutic Fe³⁺ chelating agents, a novel methodology has been developed using quantum mechanical (QM) calculations for predicting the iron binding affinities of the compounds (more specifically, their log K ₁ values). The reported/measured HPO log K₁ values were verified through their correlation with the corresponding sum of the compounds' ligating group pK_a values. Using a training set of eleven HPOs with known log K₁ values, reliable predictions are shown to be obtained with QM calculations using the B3LYP/6-31+G(d)/CPCM model chemistry (with Bondi radii, and water as solvent). With this methodology, the observed log K ₁ values for the training set compounds are closely matched by the predicted values, with the correlation between the observed and predicted values giving r² = 0.9. Predictions subsequently made by this method for a test set of 42 HPOs of known log K₁ values gave predicted values accurate to within ±0.32 log units. In order to further investigate the predictive power of the method, four novel HPOs were synthesised and their log K₁ values were determined experimentally. Comparison of these predicted log K₁ values against the measured values gave absolute deviations of 0.22 (13.87 vs. 14.09), 0.02 (14.31 vs. 14.29), 0.12 (14.62 vs. 14.50), and 0.13 (15.04 vs. 15.17). The prediction methodology reported here is the first to be provided for predicting the absolute log K₁ values of iron-chelating agents in the absence of pK_a values.

Full text of DOI:10.1039/c2dt31254a

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Dalton
Transactions
Cite this: Dalton Trans., 2012, 41, 10784
www.rsc.org/dalton
PAPER
Prediction of 3-hydroxypyridin-4-one (HPO) log K1 values for Fe(III)†
Yu-Lin Chen, Dave J. Barlow, Xiao-Le Kong, Yong-Min Ma and Robert C. Hider*
Received 11th June 2012, Accepted 17th July 2012
DOI: 10.1039/c2dt31254a
As a means to aid in the design of 3-hydroxypyridin-4-ones (HPOs) intended for use as therapeutic Fe3+
chelating agents, a novel methodology has been developed using quantum mechanical (QM) calculations
for predicting the iron binding affinities of the compounds (more specifically, their log K1 values). The
reported/measured HPO log K1 values were verified through their correlation with the corresponding sum
of the compounds’ ligating group pKa values. Using a training set of eleven HPOs with known log K1
values, reliable predictions are shown to be obtained with QM calculations using the B3LYP/6-31+G(d)/
CPCM model chemistry (with Bondi radii, and water as solvent). With this methodology, the observed
log K1 values for the training set compounds are closely matched by the predicted values, with the
correlation between the observed and predicted values giving r2 = 0.9. Predictions subsequently made by
this method for a test set of 42 HPOs of known log K1 values gave predicted values accurate to within
0.32 log units. In order to further investigate the predictive power of the method, four novel HPOs were
synthesised and their log K1 values were determined experimentally. Comparison of these predicted log
K1 values against the measured values gave absolute deviations of 0.22 (13.87 vs. 14.09), 0.02 (14.31 vs.
14.29), 0.12 (14.62 vs. 14.50), and 0.13 (15.04 vs. 15.17). The prediction methodology reported here is
the first to be provided for predicting the absolute log K1 values of iron-chelating agents in the absence of
pKa values.
Introduction
Iron is vital to all living organisms but becomes toxic when it
saturates the natural cellular buffering mechanisms, the excess
iron causing free radical formation via the Fenton reaction, and
thereby leading to oxidative stress.1 When there is a build-up of
iron in humans – for example, in repeatedly transfused patients
Fig. 1 3-Hydroxypyridin-4-one (HPO). Deferiprone, R1 = R2 = CH3
suffering from β-thalassaemia,2 or sickle cell anaemia3 – iron
and R5 = R6 = H.
chelation therapy becomes necessary. The most widely used
therapeutic chelator, desferioxamine (DFO)4 is a hexadentate
ligand with a very high affinity for Fe3+, but it suffers the major
excellent iron scavenging properties of Deferiprone, but which
drawback of being orally inactive. The more recently introduced
lack this compound’s relatively high metabolic instability and its
synthetic alternatives, the 3-hydroxypyridin-4-ones (HPOs,
undesirable effect, in a small number of patients, of lowering a
Fig. 1), have also been shown to be useful therapeutic chelators,
patient’s white blood cell count.10
exhibiting the necessary affinity and selectivity for Fe3+ along
The rational design of such molecules can be greatly facili-
with good oral activity.5 One such compound, Deferiprone6
tated by having accurate means to predict their affinity for Fe3+
(Fig. 1), has emerged as a critically important therapeutic, able
and H+ – the latter demanding an accurate means to predict the
to remove accumulated excess iron from the heart7,8 and mito-
pKa values for their iron co-ordinating moieties. In our attempts
chondria.9 However, the search continues for other such syn-
to develop such methodologies, we previously sought to explore
thetic chelators, with the aim to develop analogues that retain the
the accuracy and reliability of the pKa predictions that could be
achieved for a series of HPOs using a number of existing, low-
cost, Quantitative Structure Property Relationship (QSPR)11,12
and Quantum Mechanical (QM)13 methods. Using a training set
of 15 HPOs with known hydroxyl pKa values, reliable predic-
Institute of Pharmaceutical Science, King’s College London, Franklin-
tions were shown to be obtained with QM calculations using the
Wilkins Building, 150 Stamford Street, London, SE1 9NH, UK.
B3LYP/6-31+G(d)/CPCM model chemistry (with Pauling radii,
E-mail: robert.hider@kcl.ac.uk; Tel: +44 (0)20 7848 4882
and water as solvent).14 With this methodology, the observed
†Electronic supplementary information (ESI) available. See DOI:
10.1039/c2dt31254a
hydroxyl pKa values for the training set compounds closely
10784 | Dalton Trans., 2012, 41, 10784–10791
This journal is © The Royal Society of Chemistry 2012

View Article Online
matched the predicted pKa values (with r2 = 0.98) and the pre-
dictions subsequently made by this method for a test set of 48
HPOs of known hydroxyl pKa values gave predicted pKa values
accurate to within 0.2 log units.
In the studies reported here, our aim was to extend the model-
ling of these chelators towards prediction of their log K1 values
– hopefully, therefore, providing a rapid and reliable in-silico
method for predicting their affinities for Fe3+.
water molecules, with r2 = 0.90, |M| = 0.16 and |S| = 0.09
(Fig. 3). Adopting the same solvent model, B3LYP*/6-31+G(d)
(which has 10%, 15%, 25%, 30%, 35%, or 40% Hartree–Fock
exchange compared to 20% for B3LYP), CAM-B3LYP/6-31+G
(d) (a modified functional which is suitable for long orbital dis-
tance calculations), B3LYP/3-21G (a lower basis set), PM6
(a semi-empirical method), were also tested. However, no
superior outcomes were detected with respect to either compu-
ting accuracy or time.
Synthesis of 3-hydroxypyridin-4-ones
Experimental
Synthesis of CP70
Titrimetry
3-Benzyloxy-6-methyl-pyran-4(1H)-one.
The stability constants of HPOs were determined using an auto-
mated titration system.15 The log K1 values were evaluated using
a 10 : 1 molar ratio of Ligand (L) : Metal (Fe3+). The titration
data were analysed using the pHab program16 and species plots
were calculated using the HYSS program.17
To a solution of 6-methyl-3-hydroxypyran-4(1H)-one (63 g,
0.5 mol) in methanol (500 mL) was added sodium hydroxide
(22 g, 0.55 mol) dissolved in water (50 mL), and the mixture
was heated to reflux. Benzyl chloride (70 g, 0.55 mol) was
added dropwise over 30 min, and the resulting mixture was
refluxed overnight. After removal of solvent by rotary evapo-
ration, the residue was mixed with water (200 mL) and extracted
with dichloromethane (3 × 150 mL). The combined extracts
were washed with 5% aqueous sodium hydroxide (2 × 200 mL)
followed by water (200 mL). The organic fraction was then dried
over anhydrous sodium sulphate, filtered and rotary evaporated
to yield an orange oil which solidified on cooling. Recrystallisa-
tion from diethyl ether affords colourless needles of the com-
Prediction of HPO log K1 values using QM calculations
By analogy with a related study,14 in order to simplify calcu-
lations and to decrease computing time, only two types of opti-
mised structures, namely deprotonated ligand (L−) and iron
(Fe3+) complex ([Fe3+L1]2+), in solvent models were computed.
The calculated free energy differences (ΔG*calculated) between the
two optimised structures of the starting 11 HPOs (Table 1) were
plotted against their experimentally determined log K1 values in
order to assess the correlation. The correlation coefficient (r) for
the regression between predicted and experimental values and
their mean and standard deviation of absolute deviations (|M|
and |S|) were used to investigate the performance of various
model chemistries. The different model chemistries at B3LYP/6-
31+G(d) level explored here vary very significantly according to
the solvent model employed18 (IEF-PCM, CPCM, SMD, I-PCM
or SCI-PCM) and the radii employed in the CPCM models
(Bonding, Pauling, UFF, UAHF, or UAKS). For [Fe3+L1]2+,
there are two constructed starting structures: one, with no
explicit water (Fig. 2(a)) and the other, with four explicit water
molecules octahedrally distributed around the Fe3+ (Fig. 2(b)).
All the optimised geometries (direct optimised in solvent
models; force constants calculated at each step; without sym-
metry constraints; the spin multiplicity = 6 for [Fe3+L1]2+) were
calculated using Gaussian 09,19 and the conformers generated
verified as corresponding to local minima on their potential
energy surfaces (with no imaginary frequencies existing in the
output files). Prior to running the Gaussian 09 optimisation, con-
formational searches at AM1 level were performed using
HyperChem Release 8,20 to obtain the global (or nearly global)
minimum energy conformers for all the molecules studied here.
(Any further calculations including, for example, Boltzmann
averaging over multiple low energy conformers were not per-
formed since the additional computing time required could not
be justified.)
1
pound (88.5 g, 82%). H NMR (CDCl3): δ 2.23 (s, 3H, 6-CH3),
5.07 (s, 2H, CH2Ph), 6.22 (s, 1H, 5-H), 7.3-7.41 (m, 5H, Ph),
7.46 (s, 1H, 2-H).
3-Benzyloxy-6-methyl-pyridin-4(1H)-one. To
a
solution of
3-benzyloxy-6-methyl-pyran-4(1H)-one (21.6 g, 100 mmol) in a
mixture of EtOH and water (200 mL; 1 : 1 v/v) was added
ammonia (100 mL). The mixture was heated at 70 °C overnight.
After evaporation to remove solvent, the crude product was
purified by recrystallisation using EtOH–diethyl ether to afford
white crystals of 3-benzyloxy-6-methyl-pyridin-4(1H)-one
(17.4 g, 81%). 1H NMR (DMSO-d6): δ 2.16 (s, 3H, 6-CH3),
4.97 (s, 2H, CH2Ph), 6.01 (s, 1H, 5-H (pyridinone)), 7.31–7.41
(m, 6H, 2-H (pyridinone) and Ar), 11.15 (br., 1H, NH). MS
(ESI): m/z, 216.2 [M + H]+.
3-Hydroxy-6-methyl-pyridin-4(1H)-one (CP70) hydrochloride. A
solution of 3-benzyloxy-6-methyl-pyridin-4(1H)-one (12.9 g,
60 mmol) in N,N-dimethylformamide (30 mL) was subjected to
hydrogenolysis in presence of 5% Pd/C (w/w) catalyst (0.5 g)
for 5 h. The catalyst was removed by filtration and the filtrate
acidified to pH 1 with concentrated hydrochloric acid. After
removal of the solvent in vacuo, the residue was purified by
recrystallisation from methanol–acetone to afford a white solid
(CP70) (8.6 g, 89%). 1H NMR (DMSO-d6): δ 2.49 (s, 3H,
6-CH3), 7.13 (s, 1H, 5-H), 8.01 (s, 1H, 6-H). MS (ESI): m/z,
126.1 [M + H]+. Anal.Calc. for C6H7NO2.HCl: C, 44.60; H,
4.99; N, 8.67; Cl, 21.94%. Found C, 44.52; H, 4.75; N, 8.53; Cl,
22.05%.
The model chemistry which produced the best result was
found to be B3LYP/6-31+G(d)/CPCM (Bondi radii, water as
solvent) and [Fe3+L1]2+ structures are in the absence of explicit
This journal is © The Royal Society of Chemistry 2012
Dalton Trans., 2012, 41, 10784–10791 | 10785

View Article Online
Table 1 The structures, predicted (experimental) log K1 values and calculated free energies of 3-hydroxypyridin-4-ones (HPOs)a
HPOs
R1
R2
R5
R6
ΔG*calculated (G*calculated for [Fe3+L1]2+, L−) (Hartree)c
log K1
CP20b
CP21b
CP23b
CP60b
CP61b
CP68b
CP69b
CP90b
CP93b
CP94b
CP99b
CP360
CP369
CP374
CP502
CP509
CP513
CP514
YMF5
YMF7
YMF8
YMF14
YMF15
YMF16
YMF17
YMF21
YMF22
YMF24
YMF25
YMF26
YMF33
CP102
CP110
CP359
CP364
CP365
CP366
CP372
CP511
CP545
CP510
CP28
–CH3
–CH3
–CH3
–CH3
–H
–H
–CH3
–H
–H
–H
–H
–H
–H
–H
–H
–H
–H
–H
–H
–H
–H
–H
–H
–H
–H
–H
–H
–F
–H
–H
–CH3
–F
–F
–F
–F
–F
–H
–F
–F
–H
–H
–H
–H
–H
–H
–H
–H
–H
–H
–H
–H
–H
–H
–H
–H
–H
–H
–H
–H
–CH3
–F
–H
–H
–H
–H
−1263.2292 (−1740.0518, −476.8226)
−1263.2305 (−1779.3418, −516.1113)
−1263.2309 (−1818.6296, −555.3987)
−1263.2209 (−1700.7550, −437.5340)
−1263.2227 (−1740.0471, −476.8244)
−1263.2333 (−1779.3443, −516.1110)
−1263.2245 (−1740.0498, −476.8253)
−1263.2199 (−1661.4703, −398.2503)
−1263.2303 (−1779.3381, −516.1078)
−1263.2307 (−1818.6283, −555.3975)
−1263.2282 (−1740.0596, −476.8314)
−1263.2241 (−1854.5518, −591.3277)
−1263.2242 (−1893.8405, −630.6162)
−1263.2212 (−1933.1295, −669.9084)
−1263.2174 (−1948.0243, −684.8069)
−1263.2169 (−1948.0272, −684.8103)
−1263.2164 (−1987.3136, −724.0972)
−1263.2183 (−1987.3120, −724.0937)
−1263.2103 (−1760.7086, −497.4982)
−1263.2141 (−1799.9948, −536.7807)
−1263.2146 (−1839.2858, −576.0712)
−1263.2178 (−1839.2906, −576.0728)
−1263.2117 (−1799.9937, −536.7820)
−1263.2129 (−1839.2857, −576.0728)
−1263.2196 (−1839.2903, −576.0706)
−1263.2131 (−1800.0077, −536.7946)
−1263.2170 (−1800.0099, −536.7929)
−1263.2144 (−1878.5745, −615.3601)
−1263.2124 (−1878.5741, −615.3617)
−1263.2133 (−1878.5751, −615.3618)
−1263.2147 (−1917.8633, −654.6486)
−1263.2278 (−1893.8415, −630.6136)
−1263.2271 (−2007.2015, −743.9744)
−1263.2226 (−1854.5508, −591.3281)
−1263.2206 (−1815.2591, −552.0386)
−1263.2197 (−1893.8404, −630.6207)
−1263.2203 (−1969.0530, −705.8327)
−1263.2195 (−2008.3458, −745.1263)
−1263.2063 (−1908.7534, −645.5471)
−1263.2005 (−1983.9593, −720.7588)
−1263.2172 (−1908.7311, −645.5139)
−1263.2268 (−1700.7715, −437.5446)
−1263.2265 (−1967.9156, −704.6891)
−1263.2280 (−1854.5560, −591.3280)
−1263.2276 (−2046.4899, −783.2623)
−1263.2162 (−2200.0370, −936.8208)
−1263.2249 (−1893.8275, −630.6026)
−1263.2280 (−1933.1183, −669.8903)
−1263.2056 (−2030.3502, −767.1446)
−1263.2013 (−2105.5593, −842.3581)
−1263.2338 (−1779.3475, −516.1137)
−1263.2358 (−1958.7859, −695.5501)
−1263.2455 (−1859.5496, −596.3041)
−1263.2226 (−1700.7691, −437.5465)
−1263.2257 (−1933.1174, −669.8917)
−1263.2279 (−1972.4049, −709.1770)
−1263.2309 (−1740.0669, −476.8360)
14.79 (15.09)21
14.98 (14.82)21
15.04 (15.00)21
13.64 (13.57)21
13.88 (14.11)21
15.37 (15.22)21
14.14 (13.86)21
13.50 (13.73)21
14.96 (14.77)21
15.01 (15.12)21
14.65 (14.82)21
14.09 (14.39)22
14.10 (14.37)22
13.67 (14.41)22
13.14 (13.41)24
13.08 (12.85)24
13.00 (12.84)24
13.28 (12.96)24
12.15 (11.57)23
12.69 (11.95)23
12.75 (11.98)23
13.20 (13.14)23
12.35 (12.14)23
12.51 (12.69)23
13.46 (13.61)23
12.54 (12.50)23
13.09 (13.21)23
12.73 (12.20)23
12.44 (11.88)23
12.57 (11.76)23
12.77 (12.37)23
14.61 (14.69)d
14.50 (14.78)d
13.88 (14.26)d
13.59 (14.28)d
13.47 (14.15)d
13.55 (13.82)d
13.43 (14.29)d
11.59 (12.03)d
10.77 (11.32)d
13.12 (12.20)d
14.46 (14.49)d
14.42 (14.62)d
14.63 (14.61)d
14.57 (14.58)d
12.97 (13.02)d
14.19 (14.45)d
14.64 (15.00)d
11.49 (11.59)d
10.88 (10.98)d
15.44 (15.51)d
15.72 (15.06)d
17.08 (16.52)d
13.87 (14.09)d
14.31 (14.29)d
14.62 (14.50)d
15.04 (15.17)d
–C2H5
–CH(CH3)2
–CH3
–C2H5
–CH3
–CH3
–H
–CH3
–C2H5
–H
–CH3
–C2H5
–CH3
–CH3
–H
–CH3
–C2H5
–H
–CH3
–C2H5
–CH3
–CH3
–C2H5
–CH3
–H
–CH3
–CH3
–H
–H
–H
–C2H5
–C2H5
–C2H5
–H
–CH2OH
–CH2OH
–CHC2H5OH
–CONHCH3
–CON(CH3)2
–CONHC2H5
–CONHCH3
–H
–F
–F
–F
–H
–H
–CH3
–H
–CH3
–F
–H
–H
–F
–CH3
–CH3
–CH3
–CH3
–CH3
–CH3
–CH3
–H
–H
–H
–H
–H
–H
–H
–CH3
–H
–H
–H
–H
–H
–H
–H
–H
–C3H7
–C3H7
–CH(CH3)2
–C4H9
–C2H4OH
–C2H4COOH
–C2H5
–CH3
–C2H5
–C3H6OH
–C3H6OH
–H
–H
–CH3
–H
–C2H4COOH
–C2H4OH
–C3H6COOH
–C2H4OH
–CH3
–CH3
–CH3
–C2H5
–C2H5
–H
–H
–H
–H
–H
–H
–CH2OH
–CH2OH
–CHCH3OH
–CH2OH
–CHCH3OH
–CONHCH3
–CONHC2H4OH
–H
–CH3
–CH3
–CH3
–C2H5
–CH(C6H5)OH
–CH2OCH3
–CH(CH3)OCH3
–CH2(NC5H10
–CH2(NC5H10
–CH3
–F
–F
–H
–CH3
–H
–CONHCH3
–H
–H
–H
CP38
CP40
CP111
CP352
CP362
CP363
CP414
CP417
CP751
YMF3
YMF4
CP70
–H
–CH3
–CH3
–CH3
–CH3
–CH2OH
–H
–F
–F
–CH3
–CH3
–CH3
–H
)
)
–CH3
–CH3
–H
–H
–H
–C2H5
–CH3
–H
–H
–H
–H
–CH3
CP370
CP375
CP616
–CH2OCH3
–CHC2H5OCH3
–CH3
–H
a Positions of substituents R1, R2, R5 and R6 (Fig. 1). b The first 11 HPOs were used to test different model chemistries. c ΔG*calculated: the calculated
free energy differences between [Fe3+L1]2+ and L− using B3LYP/6-31+G(d)/CPCM (Bondi radii, water as solvent) (G*calculated: the calculated free
energy); 1 Hartree = 627.5095 kcal mol−1 d The experimental values were determined in this study.
.
10786 | Dalton Trans., 2012, 41, 10784–10791
This journal is © The Royal Society of Chemistry 2012

View Article Online
200 mL). The combined organic extracts were dried, filtered and
concentrated in vacuo to yield the crude product. Recrystallisa-
tion from isopropanol afforded straw coloured plates (70 g,
1
60.8%): H NMR (DMSO-d6): δ 2.25 (s, 3H, 6-CH3), 6.2 (s,
1H, 5-H), 6.3–7.35 (br., 1H, 3-OH), 7.7 (s, 1H, 2-H).
2-Hydroxymethyl-3-hydroxy-6-methyl-pyran-4(1H)-one. 6-Methyl-
3-hydroxypyran-4(1H)-one (68 g, 540 mmol) was added to an
aqueous solution of sodium hydroxide (23.7 g, 594 mmol) in
distilled water (500 mL) and stirred at room temperature for
5 min. 35% Formaldehyde solution (49 mL) was added dropwise
over 10 min and the solution allowed to stir for 12 h. Acidifica-
tion to pH 1 using concentrated hydrochloric acid and cooling to
3–5 °C for 12 h gave a crystalline deposit (65 g, 75%), 1H NMR
(DMSO-d6): δ 2.3 (s, 3H, 6-CH3), 4.5 (s, 2H, 2-CH2OH),
4.6–5.7 (br., 1H, 2-CH2OH), 6.25 (s, 1H, 5-H), 8.7–9.2 (br., 1H,
3-OH).
Fig. 2 [Fe3+L1]2+ Structures: (a) without explicit water molecules; (b)
with 4 explicit water molecules octahedrally distributed around Fe3+
.
2-Hydroxymethyl-3-benzyloxy-6-methyl-pyran-4(1H)-one. Sodium
hydroxide (17.7 g, 434 mmol, 1.1 equiv) dissolved in 37 mL dis-
tilled water was added to a solution of 2-hydroxymethyl-
3-hydroxy-6-methyl-pyran-4(1H)-one (63 g, 403 mmol, 1 equiv)
in methanol (367 mL) and the reaction mixture was heated to
reflux. Benzyl chloride (76 g, 403 mmol, 1 equiv) was added
dropwise over 30 min and then refluxed for 12 h. The reaction
mixture was concentrated in vacuo, the residue taken up into
dichloromethane (1 L) and the inorganic salts filtered off. The
dichloromethane layer was washed with 5% sodium hydroxide
solution (6 × 100 mL), water (600 mL), dried, filtered and con-
centrated in vacuo to yield the crude product as a yellow crystal-
line solid. Recrystallisation from dichloromethane–petroleum
ether 40 : 60 afforded a white crystalline solid (70 g, 71%):
1H NMR (CDCl3): δ 2.2 (s, 3H, 6-CH3), 2.6 (br., s, 1H,
2-CH2OH), 4.3 (br., s, 2H, 2-CH2OH), 5.2 (s, 2H, CH2Ph), 6.15
(s, 1H, 5-H (pyranone)), 7.4–7.6 (m, 5H, Ar).
Fig. 3 The experimental log K1 values versus the calculated free
energy differences (ΔG*calculated) for the starting 11 HPOs using B3LYP/
6-31+G(d)/CPCM (Bondi radii, water as solvent). The errors related to
the experimental log K1 determination for the typical HPOs is estimated
to be of the order of 0.14 (assuming errors on the determined pKa
values of 0.1).
3-Benzyloxy-1-ethyl-2-hydroxymethyl-6-methyl-1H-pyridin-4(1H)-
one. To a solution of 2-hydroxymethyl-3-benzyloxy-6-methyl-
pyran-4(1H)-one (68 g, 277 mmol, 1 equiv) in dichloromethane
(830 mL) was added 3,4-dihydro-2H-pyran (47 g, 554 mmol,
2 equiv) followed by p-toluenesulfonic acid monohydrate
(800 mg, cat.). After being stirred at room temperature for 3 h,
the reaction mixture was washed with 5% aqueous sodium car-
bonate (500 mL) followed by water (4 × 250 mL). The organic
fraction was then dried, filtered, and rotary evaporated to yield a
light yellow oil. This oil was dissolved in ethanol (280 mL)/
aqueous ethylamine (280 mL) and refluxed at 70 °C for 12 h.
After removal of the solvent by rotary evaporation, the residue
was re-dissolved in ethanol (200 mL) and 2 N hydrochloric acid
(140 mL) and refluxed for 4 h. The solvent was removed by
rotary evaporation prior to addition of water (140 mL) and
washing with diethyl ether (2 × 300 mL). Subsequent adjustment
of the aqueous fraction to pH 9 with 10 N sodium hydroxide
solution was followed by extraction into dichloromethane (4 ×
300 mL), the combined organic layers were dried, filtered, and
rotary evaporated to give a light brown solid. Crystallisation
from methanol–diethyl ether afforded the pure product (44 g,
Synthesis of CP370
Chlorokojic acid.
Kojic acid (200 g, 1.4 mol) was dissolved in thionyl chloride
(800 mL), stirred for 1 h, after which time a yellow crystalline
mass formed. The product was collected by filtration and washed
with petroleum ether and then recrystallised from water to give
cream coloured needles (170 g, 75.9%): 1H NMR (DMSO-d6) δ:
4.7 (s, 2H, 6-CH2Cl), 6.6 (s, 1H, 5-H), 8.1 (s, 1H, 2-H), 9.3
(br., s, 1H, 3-OH).
6-Methyl-3-hydroxypyran-4(1H)-one. Chlorokojic acid (150 g,
0.94 mol) was added to 500 mL of distilled water and heated to
50 °C with stirring. Zinc dust (122 g, 1.88 mol) was added fol-
lowed by the dropwise addition of concentrated hydrochloric
acid (280 mL) over 1 h with vigorous stirring maintaining the
temperature between 70–80 °C. The reaction mixture was stirred
for a further 3 h at 70 °C. Excess zinc was removed by hot fil-
tration and the filtrate extracted with dichloromethane (10 ×
1
75%) as a creamy white crystalline solid. H NMR (DMSO-d6):
δ 1.21–1.24 (t, 3H, J = 7.15 Hz, CH3, N-Et), 2.35 (s, 3H,
6-CH3), 4.06–4.12 (q, 2H, J = 7.13 Hz, CH2, N-Et), 4.6 (s, 2H,
2-CH2OH), 5.05 (s, 2H, CH2Ph), 6.16 (s, 1H, 5-H), 7.4–7.6
(m, 5H, Ar).
This journal is © The Royal Society of Chemistry 2012
Dalton Trans., 2012, 41, 10784–10791 | 10787

View Article Online
3-Benzyloxy-1-ethyl-2-methoxymethyl-6-methyl-pyridin-4(1H)-
one. To DMSO (30 mL) was added 3-benzyloxy-1-ethyl-
2-hydroxymethyl-6-methyl-pyridin-4(1H)-one (1 g, 3.66 mmol,
1 equiv) and stirred, until completely dissolved. This solution
was added to KOH (411 mg, 7.32 M, 2 equiv) and stirred for
another 3 h. The above reaction vessel was sealed and iodo-
methane (2.6 g, 18.3 mmol, 5 equiv) was injected dropwise over
10 min and allowed to stir overnight. A reddish brown clear
mixture was obtained at the end of the period. The reaction
mixture was reduced under pressure and the contents taken up in
500 mL water.
The aqueous layer was extracted with DCM (5 × 100 mL), the
combined extracts dried, filtered and dried to obtain the crude
product as yellow oil. Column chromatography using MeOH–
CHCl3 = 1 : 9 yielded the pure compound as a white solid
(900 mg, 95%). 1H NMR (CDCl3): δ 1.23–1.26 (t, 3H, J = 7.15 Hz,
CH3), 2.35 (s, 3H, 6-CH3), 3.23 (s, 3H, OCH3), 4.00–4.05
(q, 2H, J = 7.13 Hz, CH2), 4.37 (s, 2H, CH2OCH3), 5.24 (s, 2H,
CH2Ph), 6.4 (s, 1H, 5-H (pyridinone)), 7.4–7.6 (m, 5H, Ar).
(CDCl3): δ 1.12 (t, 3H, J = 7.2 Hz, 2-CHCH2CH3), 1.7–2.3
(m, 2H, 2-CHCH2CH3), 2.45 (s, 3H, 6-CH3), 4.95 (t, 1H,
J = 6.0 Hz, 2-CHCH2CH3), 5.0–6.0 (br, 1H, 2-CHOH), 6.3
(s, 1H, 5-H).
2-(1′-Hydroxypropyl)-3-benzyloxy-6-methyl-pyran-4(1H)-one. To
a solution of 2-(1′-hydroxypropyl)-3-hydroxy-6-methyl-pyran-4
(1H)-one (7.36 g, 40 mmol) in methanol (50 mL) was added
sodium hydroxide (1.76 g, 44 mmol) dissolved in water (5 mL),
and the mixture was heated to reflux. Benzyl chloride (5.6 g,
44 mmol) was added dropwise over 30 min, and the resulting
mixture was refluxed overnight. After removal of solvent by
rotary evaporation, the residue was mixed with water (50 mL)
and extracted with dichloromethane (3 × 50 mL). The combined
extracts were washed with 5% aqueous sodium hydroxide
(2 × 50 mL) followed by water (50 mL). The organic fraction
was then dried over anhydrous sodium sulphate, filtered and
rotary evaporated to yield a crude product as a yellow crystalline
solid. Recrystallisation from dichloromethane–petroleum ether
40 : 60 affords colourless needles (8.9 g, 81%). 1H NMR
(CDCl3): δ 0.8 (t, 3H, J = 7.2 Hz, 2-CHCH2CH3), 1.2–1.9
(m, 2H, 2-CHCH2CH3), 2.2 (s, 3H, 6-CH3), 2.4 (br., s, 1H,
2-CHOH), 4.5 (t, J = 6.2 Hz, 1H, 2-CHCH2CH3), 5.08 (s, 2H,
CH2Ph), 6.04 (s, 1H, 5-H), 7.28 (s, 5H, Ph).
1-Ethyl-3-hydroxy-2-methoxymethyl-6-methyl-pyridin-4(1H)-one
(CP370). To
a solution of 3-benzyloxy-1-ethyl-2-methoxy-
methyl-6-methyl-pyridin-4(1H)-one (800 mg, 2.78 mmol) in
DCM (10 mL) in a sealed container was added 6 mL of 33%
HBr in acetic acid and left to stir for 2 h. At the end of the
period diethyl ether was added to the reaction mixture to precipi-
tate the product. The crude off-white solid was filtered, dried and
then recrystallised from methanol and diethyl ether to yield pure
white crystals (250 mg, 50%). 1H NMR (DMSO-d6): δ
1.35–1.39 (t, 3H, J = 7.15 Hz, CH3), 2.64 (s, 3H, 6-CH3), 3.34
(s, 3H, OCH3), 4.35–4.40 (q, 2H, J = 7.13 Hz, CH2), 4.73
(s, 2H, CH2OCH3), 7.08 (s, 1H, 5-H). MS (ESI): m/z, 198.1
[M + H]+. Anal.Calc. for C10H15NO3. HBr: C, 43.18; H, 5.80;
N, 5.04; Br, 28.73%. Found C, 43.11; H, 5.59; N, 4.9; Br,
28.78%.
2-(1′-Methoxypropyl)-3-benzyloxy-6-methyl-pyran-4(1H)-one. To
a suspension of sodium hydride (0.96 g, 40 mmol) in 50 mL dry
DMF was added 2-(1′-hydroxypropyl)-3-benzyloxy-6-methyl-
pyran-4(1H)-one (5.48 g, 20 mmol) followed by dropwise
addition of iodomethane (8.52 g, 60 mmol) at 0 °C under nitro-
gen. After stirring for 30 min at this temperature, the reaction
mixture was poured into ice cold water (100 mL) and extracted
with dichloromethane (3 × 50 mL). The combined organic frac-
tions were dried over anhydrous sodium sulphate, filtered and
concentrated in vacuo to yield the crude product (5.2 g, 90%) as
an orange oil which solidified on cooling. Recrystallisation from
CH2Cl2–petroleum ether 40 : 60 afforded the pure product as
1
a white crystalline solid. H NMR (CDCl3): δ 0.9 (t, 3H, J =
Synthesis of CP375
2-(1′-Hydroxypropyl)-3-hydroxy-6-methyl-pyran-4(1H)-one.
7.2 Hz, 2-CHCH2CH3), 1.2–1.8 (m, 2H, 2-CHCH2CH3), 2.34
(s, 3H, 6-CH3), 3.18 (s, 3H, OCH3), 4.3 (t, J = 6.2 Hz, 1H,
2-CHCH2CH3), 5.24 (s, 2H, CH2Ph), 6.2 (s, 1H, 5-H), 7.38 (s,
5H, Ph).
1,6-Dimethyl-2-(1′-methoxypropyl)-3-benzyloxypyridin-4(1H)-
one. To
a
solution of 2-(1′-methoxypropyl)-3-benzyloxy-
6-methyl-pyran-4(1H)-one (4.32 g, 15 mmol) in EtOH (10 mL)/
water (10 mL) was added 3.49 g (45 mmol) of 40% aqueous
methylamine followed by 2 N sodium hydroxide solution until
pH 13 was obtained. The reaction mixture was sealed in a thick-
walled glass tube and stirred at 70 °C overnight. After adjustment
to pH 1 with concentrated HCl, the solvent was removed by
rotary evaporation prior to addition of water (50 mL) and
washing with diethyl ether (3 × 50 mL). Subsequent adjustment
of the aqueous fraction to pH 7 with 10 N sodium hydroxide
solution was followed by extraction into dichloromethane (4 ×
50 mL). The combined organic fractions were dried over anhy-
drous sodium sulphate, filtered and concentrated in vacuo. The
residue was purified by column chromatography on silica gel
(eluent: 15% MeOH/85% CHCl3) to yield the title compound
6-Methyl-3-hydroxypyran-4(1H)-one (12.6 g, 100 mmol) was
added to 100 mL water and the pH of the solution was adjusted
to 10.5 using aqueous sodium hydroxide (10 M). The mixture
was stirred at room temperature for 5 min. Propionaldehyde
(8.7 g, 150 mmol) dissolved in 50 mL methanol was added
dropwise over 1 h and the solution allowed to stir at room temp-
erature for 2 days. After adjustment to pH 1 with concentrated
hydrochloric acid, the reaction mixture was evaporated to
dryness and the residue taken up into 300 mL of dichloro-
methane. The organic layer was washed with water (150 mL),
dried over anhydrous sodium sulphate, filtered and concentrated
to yield the crude product. Recrystallisation from toluene
1
as a yellow oil (1.7 g, 38%). H NMR (CDCl3): δ 0.9 (t, 3H,
J = 7.2 Hz, 2-CHCH2CH3), 1.1–1.9 (m, 2H, 2-CHCH2CH3), 2.3
(s, 3H, 6-CH3), 3.05 (s, 3H, OCH3), 3.65 (s, 3H, NCH3),
1
afforded the pure product as a white crystalline solid. H NMR
10788 | Dalton Trans., 2012, 41, 10784–10791
This journal is © The Royal Society of Chemistry 2012

View Article Online
4.65–5.0 (t, J = 5.4 Hz, 1H, 2-CHCH2CH3), 5.24 (s, 2H,
CH2Ph), 6.3 (s, 1H, 5-H), 7.1–7.6 (s, 5H, Ph).
1,6-Dimethyl-2-(1′-methoxypropyl)-3-hydroxypyridin-4(1H)-one
cooling. The crude product was recrystallised from ethanol–
acetone to afford colourless solid (2.58 g, 95%). 1H NMR
(DMSO-d6): 2.07 (s, 3H, CH3), 2.15 (s, 6H, N(CH3)2), 3.25 (s,
2H, NCH2), 5.03 (s, 2H, OCH2), 7.28–7.41 (m, 6H, 6-H and
Ph). MS (ES): m/z, 273.2 [M + H]+.
(CP375)
hydrochloride. 1,6-Dimethyl-2-(1′-methoxypropyl)-
3-benzyloxypyridin-4(1H)-one (1.65 g, 5.5 mmol) was dissolved
in MeOH (10 mL)/water (10 mL) and adjusted to pH 1 with con-
centrated HCl prior to hydrogenolysis for 4 h in the presence of
5% Pd/C catalyst (0.35 g). Filtration followed by rotary evapo-
ration gave the crude product as a white solid. Recrystallisation
from ethanol–diethyl ether gave the pure title compound as a
white crystalline solid (1.08 g, 79%). 1H NMR (DMSO-d6):
δ 0.9 (t, 3H, J = 7.2 Hz, 2-CHCH2CH3), 1.4–2.3 (m, 2H,
2-CHCH2CH3), 2.6 (s, 3H, 6-CH3), 3.28 (s, 3H, OCH3), 4.04 (s,
3H, NCH3), 5.15 (t, J = 5.8 Hz, 1H, 2-CHCH2CH3), 7.4 (s, 1H,
5-H). MS (ESI): m/z, 212.1 [M + H]+. Anal.Calc. for
C11H18NO3Cl: C, 53.33; H, 7.32; N, 5.65. Found: C, 53.30; H,
7.18; N, 5.56%.
2,5-Dimethyl-3-hydroxy-4(1H)-pyridinone (CP616). To a solu-
tion of 3-benzyloxy-5-(dimethylaminomethyl)-2-methyl-4(1H)-
pyridinone (1.36 g. 5 mmol) in anhydrous ethanol (30 mL) was
added cyclohexene (40 mL) and palladium hydroxide on carbon
(1 g). The mixture was refluxed for 4 days with further additions
of cyclohexene (40 mL) and palladium hydroxide (1.5 g) at
intervals and then cooled. The precipitate was removed by fil-
tration and resuspended in ethanol, filtered twice to recover
product and the combined filtrates were evaporated to dryness
under vacuum. The crude solid was recrystallised from acetone–
1
ethanol to afford white solid (87%). H NMR (DMSO-d6): 1.92
(s, 3H, 5-CH3), 2.18 (s, 3H, 2-CH3), 7.38 (s, 1H, 6-H).
MS (ES): m/z, 140.1 [M + H]+. Anal.Calc. for C7H9NO2: C,
60.42; H, 6.52; N, 10.07. Found: C, 60.38; H, 6.39; N, 10.01%.
Synthesis of CP616
2-Methyl-3-benzyloxypyran-4(1H)-one.
Results
Validation of experimental log K1 values
To a solution of maltol (63 g, 0.5 mol) in methanol (500 mL)
was added sodium hydroxide (22 g, 0.55 mol) dissolved in water
(50 mL), and the mixture was heated to reflux. Benzyl chloride
(70 g, 0.55 mol) was added dropwise over 30 min, and the
resulting mixture was refluxed overnight. After removal of
solvent by rotary evaporation, the residue was mixed with water
(200 mL) and extracted with dichloromethane (3 × 150 mL).
The combined extracts were washed with 5% aqueous sodium
hydroxide (2 × 200 mL) followed by water (200 mL). The
organic fraction was then dried over anhydrous sodium sulphate,
filtered and rotary evaporated to yield an orange oil which soli-
dified on cooling. Recrystallisation from diethyl ether gave white
In order to obtain a large group of HPOs with reliable experi-
mentally determined log K1 values, a plot of log K1 value versus
the corresponding sum of pKa values (hydroxyl pKa values plus
carbonyl pKa values) was utilised. From the existing
literature,21–25 41 HPOs were selected for initial analysis. Thirty
one of these HPOs were found to give the expected linear corre-
lation26 between their log K1 values and the corresponding sum
of their pKa values, r2 = 0.95 (Fig. 4). However, 9 HPOs were
found to have significantly higher log K1 values than predicted
from the linear relationship, while 1 HPO was found to possess a
significantly lower log K1 value (Fig. 4). After determining the
affinity constants using a 10 : 1 molar ratio of L : Fe3+ for the
10 HPO outliers (Table 1; Table S1 in ESI,† including former
values) and 16 previously uncharacterised HPOs (Table 1), 4 of
which were specifically synthesised for this study, 57 HPOs in
total were found to give the expected linear relationship between
1
needles (87.5 g, 81%). H-NMR (CDCl3): 2.12 (s, 3H, CH3),
5.11 (s, 2H, CH2), 6.25 (d, J = 6 Hz, 1H, 5-H), 7.28 (s, 5H, Ph),
7.47 (d, J = 6 Hz, 1H, 6-H).
2-Methyl-3-benzyloxypyridin-4(1H)-one. To
a
solution of
2-methyl-3-benzyloxypyran-4(1H)-one (13.8 g, 64 mmol) in
ethanol (25 mL) was added ammonia solution (50 mL) and
refluxed overnight. The solvent was removed under reduced
pressure, then taken into water and adjusted to pH 1 with con-
centrated hydrochloric acid. The aqueous mixture was washed
with ethyl acetate (3×) and the pH was adjusted to pH 10 with
sodium hydroxide (2 M). The aqueous phase was extracted with
chloroform (3×), dried over anhydrous sodium sulfate, filtered,
and evaporated under reduced pressure. Recrystallisation from
methanol–diethyl ether gave brown cubic crystals (75%). 1H
NMR (CDCl3) δ 2.15 (3H, s, CH3), 5.03 (2H, s, CH2), 6.35 (1H,
d, J = 6.9 Hz, 5-H), 7.25–7.31 (5H, m, Ph), 7.39 (1H, d,
J = 6.9 Hz, 6-H); MS (ES): m/z, 215.1 [M + H]+.
3-Benzyloxy-5-(dimethylaminomethyl)-2-methyl-4(1H)-pyridi-
none. To a mixture of dimethylamine (1.80 g, 40 mmol) and
40% formaldehyde (1.5 g, 20 mmol) in ethanol (100 mL)
was added 3-benzyloxy-2-methyl-4(1H)-pyridinone (2.15 g,
10 mmol). The mixture was refluxed for 24 h and followed by
rotary evaporation to afford an oil, which was crystallised after
Fig. 4 log K1 values versus sum of pKa values for 41 HPOsa, extracted
from the existing literature.21–25 ■ CP102, CP110, CP359, CP364,
CP365, CP366, CP372, CP511, CP545. ▲ CP510; the structures of the
41 HPOs are presented in the first 41 rows of Table 1.
a
This journal is © The Royal Society of Chemistry 2012
Dalton Trans., 2012, 41, 10784–10791 | 10789

View Article Online
their log K1 values and the corresponding sum of pKa values,
r2 = 0.95 (Fig. 5).
of explicit water molecules, was found to produce a relatively
reliable prediction of log K1 values for an initial series of eleven
HPOs. The log K1 values for an additional 42 HPOs were calcu-
lated using the same model chemistry to develop the final log K1
prediction model – Predicted log K1 = −140.18ΔG*calculated
−177063.32 (r2 = 0.90, |M| = 0.32 and |S| = 0.25, Fig. 7; values
in Table 1). In an attempt to further examine the predictive
power of this model, four novel HPOs, CP70, CP370, CP375
and CP616, were synthesised. The predicted values of all four
compounds were observed to be close to their experimental
values, indicating reliable predictions (Fig. 7). The absolute
deviations between the predicted and experimental values are
0.22 for CP70 (13.87 vs. 14.09), 0.02 for CP370 (14.31 vs.
14.29), 0.12 for CP375 (14.62 vs. 14.50), and 0.13 for CP616
(15.04 vs. 15.17).
Experimental log K1 determination
The UV/Vis spectra and a species plot of CP511 with a
10 : 1 molar ratio of L : Fe3+ are presented in Fig. 6. Two isosbes-
tic points occur in the spectral series resulting from titration with
Fe3+; one at λ 600 nm, corresponding to the formation of
[Fe3+(CP511)2]+ and the other at λ 504 nm, corresponding to the
formation of [Fe3+(CP511)3]. The speciation plot demonstrates
that [Fe3+(CP511)1]2+ dominates in solution from pH 0.5 to
pH 1; [Fe3+(CP511)2]+ dominates in solution from pH 1.5 to
pH 2 and [Fe3+(CP511)3] dominates in solution from pH 3
through to pH 10. The log K1 value (12.03) was determined
from curve fitting analysis of the spectra. Members of entire
HPO series have similar UV/Vis spectra for their [Fe3+L1]2+,
[Fe3+L2]+ and [Fe3+L3] species. The log K1 values of 26 HPOs,
4 of which were specifically synthesised for this study, were
determined using this methodology (Table 1).
Discussion
The methodology developed for prediction of HPO log K1
values is impressive (with the observed and predicted values
strongly correlated, with r2 of 0.9; Fig. 7). In comparison, the
errors associated with the experimental determination of log
K1 for the typical HPOs is estimated to be of the order of
Development of a log K1 prediction model
As discussed above, B3LYP/6-31+G(d)/CPCM (Bondi radii,
water as solvent), with [Fe3+L1]2+ structures being in the absence
Fig. 7 Developed log K1 prediction model. ■ CP70, CP370, CP375
and CP616; ▲ CP510 with the largest absolute deviation (0.92). Omis-
sion of the CP510 data point changes only the third decimal place in r2
and results in an insignificant change of the fitted line, and thus an
insignificant change of the predicted log K1 values.
Fig. 5 log K1 values versus sum of pKa values for 57 HPOs. The struc-
tures of the 57 HPOs are presented in Table 1.
Fig. 6 The UV/Vis spectra and species plot of CP511-iron(III) for log K1 determination when [CP511] = 445.4 μM and [Fe3+] = 44.2 μM.
10790 | Dalton Trans., 2012, 41, 10784–10791 This journal is © The Royal Society of Chemistry 2012

View Article Online
3 W. Hagar and E. Vichinsky, Br. J. Haematol., 2008, 141, 346–356.
0.14 (assuming errors on the determined pKa values of 0.1).
However, for some HPOs which are associated with experi-
mentally difficult log K1 determinations, the error can exceed 1.0
(Table S1 in ESI†). The plot of log K1 versus the corresponding
sum of pKa values (Fig. 5) was utilised to evaluate this exper-
imental error. The utilisation of the sum of pKa values rather
than the single carbonyl or hydroxyl pKa value was adopted
because both of the pKa values are associated with functional
groups which coordinate Fe3+.
The 10 HPO outliers in Fig. 4 can be categorised into 4 main
subclasses, namely 1-hydrogen-(2 or 6)-amido-, 2-hydroxy-
methyl-, 1-hydroxy(ethyl or propyl)- and 1-carboxyethyl-HPOs.
All these HPOs possess hydrogen bonding donors and acceptors
on the aromatic ring substituents and this may result in a distur-
bance of the first layer of solvation molecules surrounding each
iron (Fe3+) complex. This in turn will influence the magnitudes
of their stability constants; for instance, with CP511, which has
strong intramolecular hydrogen bonding,24 there is an appreci-
able decrease in log K1 values (Fig. S1 in ESI†) when increased
molar ratios of L : Fe3+ are employed in the titration. This effect
is probably associated with the partial formation of a μ-oxo
bridge species (Fe3+–O–Fe3+L−1) at relatively high Fe3+ molar
concentrations, thus disturbing the UV/Vis spectra. For this
reason, the 10 : 1 molar ratio of L : Fe3+ was adopted to minimise
the formation of μ-oxo bridges during the experimental determi-
nation of log K1 values in this study.
A knowledge of pKa values can be neglected using the de-
veloped prediction methodology for HPO log K1 values and this is
particularly useful to predict HPOs possessing at least three pKa
values. For the experimental log K1 determination of HPOs with
more than two pKa values, special attention is required to iden-
tify which two pKa values correspond to the iron-coordinating
oxygens. If the two pKa values are incorrectly assigned, there
will be an appreciable error in the associated experimental
log K1 value.
In summary, the methodology developed in this study is the
first quantum mechanical approach to predict the absolute log K1
value of iron-chelating agents in the absence of pKa values. The
plot of log K1 values versus the corresponding sum of pKa
values (of co-ordinating atoms) has proved to be extremely
useful in order to investigate the accuracy of experimentally
determined log K1 values. The stability constants for HPOs with
substituents possessing hydrogen bonding donors and acceptors
should be determined using a 10 : 1 molar ratio of L : Fe3+.
4 H. Keberle, Ann. N. Y. Acad. Sci., 1964, 119, 758–768.
5 T. Zhou, Y. Ma, X. Kong and R. C. Hider, Dalton Trans., 2012, 41,
6371–6389.
6 E. R. Huehns, J. B. Porter and R. C. Hider, Hemoglobin, 1988, 12,
593–600.
7 C. Borgna-Pignatti, M. D. Cappellini, P. De Stefano, G. C. Del Vecchio,
G. L. Forni, M. R. Gamberini, R. Ghilardi, A. Piga, M. A. Romeo,
H. Zhao and A. Cnaan, Blood, 2006, 107, 3733–3737.
8 A. Pepe, A. Meloni, M. Capra, P. Cianciulli, L. Prossomariti,
C. Malaventura, M. C. Putti, A. Lippi, M. A. Romeo, M. G. Bisconte,
A. Filosa, V. Caruso, A. Quarta, L. Pitrolo, M. Missere, M. Midiri,
G. Rossi, V. Positano, M. Lombardi and A. Maggio, Haematologica,
2010, 96, 41–47.
9 O. Kakhlon, H. Manning, W. Breuer, N. Melamed-Book, C. Lu,
G. Cortopassi, A. Munnich and Z. I. Cabantchik, Blood, 2008, 112,
5219–5227.
10 A. R. Cohen, R. Galanello, A. Piga, V. De Sanctis and F. Tricta, Blood,
2003, 102, 1583–1587.
11 ACD/pKa DB 12.0, http://www.acdlabs.com/ (assessed February, 2011).
12 Marvin 5.4.0.1, http://www.chemaxon.com/ (assessed February, 2011).
13 J. Ho and M. Coote, Theor. Chem. Acc.: Theory, Computation, and
Modeling (Theoretica Chimica Acta), 2010, 125, 3–21.
14 Y. L. Chen, D. J. Barlow, X. L. Kong, Y. M. Ma and R. C. Hider, Dalton
Trans., 2012, 41, 6549–6557.
15 X. L. Kong, T. Zhou, H. Neubert, Z. D. Liu and R. C. Hider, J. Med.
Chem., 2006, 49, 3028–3031.
16 P. Gans, A. Sabatini and A. Vacca, Ann. Chim. (Rome, Italy), 1999, 89,
45–49.
17 L. Alderighi, P. Gans, A. Ienco, D. Peters, A. Sabatini and A. Vacca,
Coord. Chem. Rev., 1999, 184, 311–318.
18 J. B. Foresman and Æ. Frisch, Exploring Chemistry with Electronic
Structure Methods, Gaussian, Pittsburgh, PA, 2nd edn, 1996.
19 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci,
G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian,
A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada,
M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida,
T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven,
J. Montgomery Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd,
E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi,
J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar,
J. Tomasi, M. Cossi, N. Rega, N. J. Millam, M. Klene, J. E. Knox,
J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts,
R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli,
J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski,
G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich,
A. D. Daniels, Ö. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski
and D. J. Fox, GAUSSIAN 09 (Revision A.02 SMP), Gaussian, Inc.,
Wallingford CT, 2009.
20 HyperChem(TM) Professional 8.0, http://www.hyper.com/ (assessed Feb-
ruary, 2011), Hypercube, Inc., 1115 NW 4th Street, Gainesville, Florida
32601, USA.
21 X. Kong, PhD thesis, King’s College London, University of London,
2008.
22 Z. D. Liu, H. H. Khodr, D. Y. Liu, S. L. Lu and R. C. Hider, J. Med.
Chem., 1999, 42, 4814–4823.
23 Y. Ma, S. Roy, X. Kong, Y. Chen, D. Liu and R. C. Hider, J. Med.
Chem., 2012, 55, 2185–2195.
24 S. Piyamongkol, Y. M. Ma, X. L. Kong, Z. D. Liu, M. D. Aytemir,
D. van der Helm and R. C. Hider, Chem.–Eur. J., 2010, 16, 6374–6381.
25 B. L. Rai, L. S. Dekhordi, H. Khodr, Y. Jin, Z. Liu and R. C. Hider,
J. Med. Chem., 1998, 41, 3347–3359.
References
1 O. Weinreb, T. Amit, S. Mandel, L. Kupershmidt and M. B. Youdim,
Antioxid. Redox Signaling, 2010, 13, 919–949.
2 E. J. Neufeld, Blood, 2006, 107, 3436–3441.
26 R. C. Hider and A. D. Hall, Prog. Med. Chem., 1991, 28, 41–173.
This journal is © The Royal Society of Chemistry 2012
Dalton Trans., 2012, 41, 10784–10791 | 10791