Bifunctional 3-Hydroxy-4-Pyridinones as Potential Selective Iron(III) Chelators: Solution Studies and Comparison with Other Metals of Biological and Environmental Relevance

The binding ability of five bifunctional 3-hydroxy-4-pyridinones towards Cu2+ and Fe3+ was studied by means of potentiometric and UV–Vis spectrophotometric measurements carried out at I = 0.15 mol L−1 in NaCl(aq), T = 298.15 K and 310.15 K. The data treatments allowed us to determine speciation schemes featured by metal-ligand species with different stoichiometry and stability, owing to the various functional groups present in the 3-hydroxy-4-pyridinones structures, which could potentially participate in the metal complexation, and in the Cu2+ and Fe3+ behaviour in aqueous solution. Furthermore, the sequestering ability and metal chelating affinity of the ligands were investigated by the determination of pL0.5 and pM parameters at different pH conditions. Finally, a comparison between the Cu2+ and Fe3+/3-hydroxy-4-pyridinones data herein presented with those already reported in the literature on the interaction of Zn2+ and Al3+ with the same ligands showed that, from the thermodynamic point of view, the 3-hydroxy-4-pyridinones are particularly selective towards Fe3+ and could therefore be considered promising iron-chelating agents, also avoiding the possibility of competition, and eventually the depletion, of essential metal cations of biological and environmental relevance, such as Cu2+ and Zn2+.


Introduction
Copper (Cu) and iron (Fe) are essential metals for plants, animals and humans, ensuring their normal biochemical and physiological functions [1,2]. In healthy situations, living organisms are provided with homeostatic mechanisms and buffers to keep normal metal concentration levels and to avoid anomalous phenomena, as metal decompartmentalization, release and mobilization [3].
In plants, Cu plays important roles in photosynthetic and respiratory electron transport processes, occurring in chloroplasts and mitochondria. It participates in the oxidative stress protection, acts as cofactor of many enzymes and plays key functions in the cell wall metabolism, namely for Fe-mobilization, oxidative phosphorylation and the biogenesis of the molybdenum cofactor [4]. In the human body, copper favours the normal development of the brain and nervous system and maintains a fair level of white blood cells. Cu is also necessary to keep the muscle tone and functions; it is involved in the formation of red blood cells and in the processes of absorption and transport of iron (Fe 3+ ) in the body. Furthermore, the generation of cellular energy in the form of ATP into the mitochondria depends on the participation of a copper-containing enzyme. as well as their chelating efficacy towards Fe 3+ with respect to the commercially available chelating agents [13,18].
Herein, pursuing our previous strategy, we present the results of a potentiometric and UV-Vis (Ultraviolet-Visible) spectrophotometric investigation on the interaction of five bifunctional 3-hydroxy-4-pyridinones ( Figure 1) with Cu 2+ and Fe 3+ , metal cations with a borderline and a hard character, respectively, that are carried out at I = 0.15 mol L −1 in NaCl (aq) , T = 298.15 K and 310.15 K. Furthermore, the obtained thermodynamic data are compared with those already reported in the literature on the binding ability and chelating affinity of the five ligands towards Zn 2+ [19] and Al 3+ [13], also featured by borderline and hard behaviour, respectively, at the same experimental conditions. The aim of this work was to evaluate whether the 3-hydroxy-4-pyridinones under study could be exploited as selective chelating agents for the treatment of Fe 3+ overload in humans or, alternatively, in environmental matrices. Another relevant issue worth investigating was to ascertain whether, from a thermodynamic point of view, along with an effective Fe 3+ -sequestration, a significant competition, and possibly depletion, of divalent metals of biological and environmental relevance such as Cu 2+ and Zn 2+ may occur, despite the different charge density, acid-base behaviour [20][21][22] and ionic radius [23]. and UV-Vis (Ultraviolet-Visible) spectrophotometric investigation on the interaction of five bifunctional 3-hydroxy-4-pyridinones ( Figure 1) with Cu 2+ and Fe 3+ , metal cations with a borderline and a hard character, respectively, that are carried out at I = 0.15 mol L −1 in NaCl(aq), T = 298.15 K and 310.15 K. Furthermore, the obtained thermodynamic data are compared with those already reported in the literature on the binding ability and chelating affinity of the five ligands towards Zn 2+ [19] and Al 3+ [13], also featured by borderline and hard behaviour, respectively, at the same experimental conditions. The aim of this work was to evaluate whether the 3-hydroxy-4-pyridinones under study could be exploited as selective chelating agents for the treatment of Fe 3+ overload in humans or, alternatively, in environmental matrices. Another relevant issue worth investigating was to ascertain whether, from a thermodynamic point of view, along with an effective Fe 3+sequestration, a significant competition, and possibly depletion, of divalent metals of biological and environmental relevance such as Cu 2+ and Zn 2+ may occur, despite the different charge density, acid-base behaviour [20][21][22] and ionic radius [23].
Along the text, in the tables and figures, the five 3-hydroxy-4-pyridinones under study will be indicated with the abbreviations: H2

Equilibria for the Formation of Metal-Ligand Species
The formation or stability constants of the metal-ligand species are expressed considering the following stepwise (1) and overall (2) equilibria: pM n+ + qL z-+ rH + = M p H r L q (pn+r-qz) The equilibrium constants, concentrations and ionic strengths are expressed in the molar (c, mol L −1 ) concentration scale.

Synthesis of the Ligands
The five 3-hydroxy-4-pyridinones ( Figure 1) have been synthesized and characterized in the neutral form (H r L 0 ), following procedures already reported in the literature [13].

Acid-Base Properties of Ligands and the Metal Cations
The 3-hydroxy-4-pyridinones under study are featured by different protonable groups highlighted in Figure 1 with dotted rectangles. The ligands' structure consists of a hydroxyl group as substituent on the N-heterocyclic ring, a -NH 2 and/or -COOH on the alkyl chain and a pyridinone nitrogen atom (proton provided by an excess of inorganic acid) [13], each of them with different acidity. The 3-hydroxy-4-pyridinones protonation constants have been already reported in the literature at I = 0.15 mol L −1 in NaCl (aq) , T = 298.15 and 310.15 K (Table S1) [13].
The hydrolytic constants of Cu 2+ and Fe 3+ have already been published [20][21][22]. In the case of Fe 3+ , the solubility product related to the formation of Fe(OH) 3 0 (s) sparingly soluble species has been also considered [21].

Metal-Ligand Studies
The elaboration of potentiometric and UV-Vis spectrophotometric data on the binding ability of the ligands towards Cu 2+ and Fe 3+ allowed us to determine various speciation schemes, based on the different acid-base properties of the 3,4-HPs in NaCl (aq) , the hydrolytic behaviour and the charge density of the metal cations. The best possible speciation models were selected on the basis of criteria such as the simplicity and probability of the model, the species formation percentages in the whole investigated pH, the statistical parameters (like the standard deviation on equilibrium constants and on the fitting values), the corresponding ratios between single variances compared with those from the accepted model. The high number of experiments carried out and experimental points collected allowed for the consideration of differences in variance between the accepted model and others to be significant.
In the case of the Cu 2+ and Fe 3+ /(3,4-HPs) interactions investigated with both of the mentioned analytical techniques, an average of the potentiometric and UV-Vis stability constants was calculated with the aim of describing the systems in a more complete way, considering a wide range of metal and ligand concentrations used, namely c~10 −3 mol L −1 and 10 −4-10 −5 mol L −1 , for potentiometric and UV-Vis spectrophotometric measurements, respectively.

Cu 2+ /(3,4-HPs) Systems
For each investigated Cu 2+ /(3,4-HPs) system, the treatment of potentiometric and UV-Vis spectrophotometric data recorded at I = 0.15 mol L −1 in NaCl (aq) , T = 298.15 K and pH ranges 2.0-10.0 and 2.0-11.0, respectively, allowed us to obtain speciation models featured by complex species with 1:1 stoichiometry (CuL (2-z) ) and different protonation degrees (CuH 2 L (4-z) , CuHL (3-z) ). The experimental formation constants obtained by each analytical techniques are in accordance with each other, and they are reported in Table 1. As can be observed in Table 1, at the mentioned experimental conditions a trend of the complexes' stability can be observed, based on the common species CuL (2-z) : Cu(L3) 0 (aq) > Cu(L4) 0 (aq) > Cu(L2) 0 (aq) > Cu(L5) + > Cu(L1) 0 (aq) . This trend could be explained considering that the stability of the Cu 2+ /(3,4-HPs) species may be favoured by the concomitant presence of the extra-functional groups in the 3,4-HP ligand molecules, namely -COOH, -NH 2 and -CHNH 2 COOH bearing groups in the alkyl chain bound to the N-heterocyclic ring, which, in some cases have also inserted an amide moiety (H 2 (L2), H 2 (L3)) ( Figure 1). Generally, the complexes with higher stability are those where the 3,4-HP ligands are extra-functionalized with α-amino-carboxylic groups (H 2 (L2), H 2 (L3), H 2 (L4)), probably due to their inherent chelating capacity [24]. The different length of the alkyl moiety is also another factor influencing the stability of the species; in fact, from the comparison between the data obtained for Cu(L2) 0 (aq) and Cu(L3) 0 (aq) species, which only differ in the ligand structures by an additional -CH 2 group present in the H 2 (L3) alkyl chain (Figure 1), a decrease of the formation constants with alkyl moiety length decreasing can be observed (Table 1). Furthermore, from the comparison among the Cu(L1) 0 (aq) and Cu(L5) + stability constants it can be observed that the ligand featured by the only amino group (H(L5)) in the alkyl chain forms Cu 2+ complexes with higher stability than the carboxylic-3-hydroxy-4-pyridinone (H 2 (L1)), a trend which is in accordance with data reported in the literature [24][25][26] on the interactions of alkylamines and carboxylic acids towards Cu 2+ , also following the Pearson's principle of "hard and soft acids and bases" theory (HSAB) for ligand-metal preferences [27][28][29].  A further comparison between the speciation of the different Cu 2+ /3-hydroxy-4pyridinone systems may be performed based on the distribution diagrams reported in Figure 2, for H 2 (L3), and Figure S1, for the other ligands. In the case of H 2 (L1) ( Figure S1a), the diagram shows that the formation of the CuH(L1) + and Cu(L1) 0 (aq) species reaches the 68% and 99% maximum percentages at pH~3.9 and pH~6.6, respectively. As regards the distribution of Cu 2+ /H 2 (L2) (Figure 2), H 2 (L3) ( Figure S1b) and H 2 (L4) ( Figure S1c) species, the metal-ligands complexation occurs up to pH~3.2-3.3 with the formation of CuH 2 L 2+ species exceeding the 52% formation. The CuHL + complex achieves more than the 86% formation at pH~5.0-5.1 for H 2 (L2) and H 2 (L4), pH~5.5 for H 2 (L3). The 1:1 stoichiometry complex starts to form at pH~4.0, 4.5 and 3.6 and reaches more than the 99% formation at pH~8.8, 9.1 and 8.2, for H 2 (L2) , H 2 (L3) and H 2 (L4), respectively. In the case of the Cu 2+ /H(L5) system ( Figure S1d), the formation of the Cu(L5)H 2+ and Cu(L5) + species reaches their maximum percentages at pH~5.4 and pH~9.6, respectively.

Fe 3+ /(3,4-HPs) Systems
The investigation on the binding ability of the H2(L1), H2(L2), H2(L4) and H(L5) ligands towards Fe 3+ was carried out by potentiometric titrations at I = 0.15 mol L −1 in NaCl(aq) and T = 298.15 K. In the case of H2(L2) and H(L5), UV-Vis experiments were performed also at the same ionic strength and T = 310.15 K. The data treatment allowed for the determination of FeHL (4-z) , FeL (3-z) and FeL2 (3-2z) species in the pH range 2.0-5.0, due to the formation of a red colour precipitate, attributable to the sparingly soluble Fe(OH)3 0 (s) species [21]. This limitation was overcome by spectrophotometric titrations performed at more diluted conditions, thus allowing to explore the measurements in a wider pH range (2.0-9.1).

Fe 3+ /(3,4-HPs) Systems
The investigation on the binding ability of the H2(L1), H2(L2), H2(L4) and H( ands towards Fe 3+ was carried out by potentiometric titrations at I = 0.15 mol L −1 in and T = 298.15 K. In the case of H2(L2) and H(L5), UV-Vis experiments were per also at the same ionic strength and T = 310.15 K. The data treatment allowed for th mination of FeHL (4-z) , FeL (3-z) and FeL2 (3-2z) species in the pH range 2.0-5.0, due to mation of a red colour precipitate, attributable to the sparingly soluble Fe(OH)3 0 (s) [21]. This limitation was overcome by spectrophotometric titrations performed a diluted conditions, thus allowing to explore the measurements in a wider pH ran 9.1).
The stability constants determined at the different experimental conditions ported in Table 2. The values obtained by the two analytical techniques are in qui

Fe 3+ /(3,4-HPs) Systems
The investigation on the binding ability of the H 2 (L1), H 2 (L2), H 2 (L4) and H(L5) ligands towards Fe 3+ was carried out by potentiometric titrations at I = 0.15 mol L −1 in NaCl (aq) and T = 298.15 K. In the case of H 2 (L2) and H(L5), UV-Vis experiments were performed also at the same ionic strength and T = 310.15 K. The data treatment allowed for the determination of FeHL (4-z) , FeL (3-z) and FeL 2 (3-2z) species in the pH range 2.0-5.0, due to the formation of a red colour precipitate, attributable to the sparingly soluble Fe(OH) 3 0 (s) species [21]. This limitation was overcome by spectrophotometric titrations performed at more diluted conditions, thus allowing to explore the measurements in a wider pH range (2.0-9.1). The stability constants determined at the different experimental conditions are reported in Table 2. The values obtained by the two analytical techniques are in quite good agreement. Similarly to the Cu 2+ /(3,4-HPs) studies, in this case the data average was also calculated.
In the case of the H 2 (L2) and H(L5) ligands, the formation constants were also determined at I = 0.15 mol L −1 in NaCl (aq) and T = 310.15 K, as reported in Table 2: the obtained values increase with temperature.
A further deepening on the speciation of the different Fe 3+ /(3,4-HP) systems may be performed considering the distribution diagrams drawn from potentiometric data at I = 0.15 mol L −1 in NaCl (aq) and T = 298.15 K, as reported in Figure 4 for H 2 (L2) and in Figure S2 for H 2 (L1), H 2 (L4) and H(L5). and H(L5), respectively. As regards the FeL2 (3-2z) complex, it starts to form from pH ~ 3.2 with ligands as H2(L2) and H2(L4), from pH ~ 2.6 with H(L5). For the Fe 3+ /H2(L2) system ( Figures 5 and S3), a band with λmax = 568 nm is observed at pH ~ 2.0, followed by an intensity decrease at pH ~ 3.7. A first band hypsochromic shift In the case of H 2 (L1) (Figure S2a), the diagram shows the FeH(L1) 2+ and Fe(L1) + species reaching the maximum percentages of 93% and 19% at pH~2.8 and 3.5, respectively. Introducing the solubility product of Fe(OH) 0 3(s) [21] in the speciation model in the HySS programme [30], used to calculate the formation percentages and to represent the distribution diagrams, the formation of the sparingly soluble species should occur at pH~3.5, hindering the possible formation of the Fe(L1) 2 species which should only start to form at the mentioned pH value. However, since the precipitation was experimentally observed at pH~5.0, the formation of the Fe(L1) 2 species in the pH range 3.5-5.0 could be considered as a probable complex, and therefore it was reported in Table 2. Regarding the species distribution of Fe 3+ /H 2 (L2) (Figure 4), Fe 3+ /H 2 (L4) ( Figure  S2b) and Fe 3+ /H(L5) ( Figure S2c) species, the FeHL (4-z) species achieves 99% formation at pH~2.3-2.4. The 1:1 stoichiometry complex reaches 26%, 37% and 7% formation at pH~3.8, 3.7, 3.2 for H 2 (L2) , H 2 (L4) and H(L5) , respectively. As regards the FeL 2 (3-2z) complex, it starts to form from pH~3.2 with ligands as H 2 (L2) and H 2 (L4), from pH~2.6 with H(L5).  For the Fe 3+ /H2(L2) system ( Figures 5 and S3), a band with λmax = 568 nm is observed at pH ~ 2.0, followed by an intensity decrease at pH ~ 3.7. A first band hypsochromic shift (λmax = 510 nm) and an absorbance increase occurs at pH ~ 4.5. Then, a second blue shift and a band is observed with λmax = 460 nm from pH ~ 6.1-7.1, depending on the experimental conditions, up to the formation of precipitate, which hindered further investigations.  In the case of the Fe 3+ /H(L5) system, at metal/ligand stoichiometric conditions ( Figure  S4), the mentioned band with λmax = 568 nm at pH ~ 2.0, as well as its two hypsochromic shifted bands (λmax = 510 nm, 460 nm) at pH ~ 4.9 and 5.9-6.0, respectively, are observed. For cFe 3+ /cligand = 1/2 and cFe 3+ /cligand = 1/3 ( Figure 6), the first recorded band is featured by  The deconvolution of the UV-Vis spectrophotometric data allowed us to calculate the molar absorptivity (ε/L (mol −1 cm −1 )) values for each metal-ligand species. Graphical representations of the ε determined for the Fe 3+ /H2(L2) and Fe 3+ /H(L5) systems are reported in Figures S5 and 7, respectively, at I = 0.15 mol L −1 in NaCl(aq) and different temperatures. As a representative example, the calculated molar absorptivities for the For the Fe 3+ /H 2 (L2) system ( Figure 5 and Figure S3), a band with λ max = 568 nm is observed at pH~2.0, followed by an intensity decrease at pH~3.7. A first band hypsochromic shift (λ max = 510 nm) and an absorbance increase occurs at pH~4.5. Then, a second blue shift and a band is observed with λ max = 460 nm from pH~6.1-7.1, depending on the experimental conditions, up to the formation of precipitate, which hindered further investigations.
In the case of the Fe 3+ /H(L5) system, at metal/ligand stoichiometric conditions ( Figure S4), the mentioned band with λ max = 568 nm at pH~2.0, as well as its two hypsochromic shifted bands (λ max = 510 nm, 460 nm) at pH~4.9 and 5.9-6.0, respectively, are observed. For c Fe3+ /c ligand = 1/2 and c Fe3+ /c ligand = 1/3 (Figure 6), the first recorded band is featured by λ max = 536 nm at pH~2.0, with a blue shift occurring at pH~4.9 with a band at λ max = 515 nm, whilst the last blue shift corresponds to a band at λ max = 460 nm, similarly to the previous described spectra.
The deconvolution of the UV-Vis spectrophotometric data allowed us to calculate the molar absorptivity (ε/L (mol −1 cm −1 )) values for each metal-ligand species. Graphical representations of the ε determined for the Fe 3+ /H 2 (L2) and Fe 3+ /H(L5) systems are reported in Figure S5  nm, whilst the last blue shift corresponds to a band at λmax = 460 nm, similarly to the previous described spectra. The deconvolution of the UV-Vis spectrophotometric data allowed us to calculate the molar absorptivity (ε/L (mol −1 cm −1 )) values for each metal-ligand species. Graphical representations of the ε determined for the Fe 3+ /H2(L2) and Fe 3+ /H(L5) systems are reported in Figures S5 and 7, respectively, at I = 0.15 mol L −1 in NaCl(aq) and different temperatures. As a representative example, the calculated molar absorptivities for the

Literature Data Comparison
From the best of our knowledge, no studies have been reported on the Cu 2+ /(3,4-HP) systems. Two papers have been published by Santos et al. [31,32] on the binding ability of the same H 2 (L1) and H 2 (L4) ligands ( Figure 1) towards Fe 3+ at I = 0.10 mol L −1 in KNO 3(aq) and T = 298.15 K. The authors determined a speciation scheme featured by FeH r L q (3+r-qz) (q, r = 1-3) species with different stoichiometry, including FeHL 2+ . This complex was also reported in the current work for the same two ligands (Table 2), and so a comparison between the experimental and literature data can be made. The stability constants determined by Santos et al. are logK 111 = 9.58 for H 2 (L1) [32] and logK 111 = 15.16 for H 2 (L4) [31] (Table S2). The value obtained for H 2 (L4) is in good accordance with the logK 111 = 15.21 (Table 2) presented in this paper at I = 0.15 mol L −1 in NaCl (aq) and T = 298.15 K. However, the value previously reported for H 2 (L1) is slightly higher than the value determined herein (logK 111 = 7.42, Table 2).
Some other comparisons could also be made considering metal-ligand investigations on compounds with similar structures and functional groups ( Figure S6) with respect to the 3,4-HP ligands under study. Nevertheless, some little differences in the ligand struc-tures, discrepancies between the experimental conditions and, in particular, the different approaches sometimes used by the authors for the data treatment (determination of ligands' acid-base properties, apparent neglect or very few information reported on the metals' hydrolytic behaviour), make it difficult to establish a direct comparison among the stability constants. However, an attempt of comparison could be performed, considering the logK 110 values reported in this paper for the ML (n-z) species and the data published in the literature for complexes with the same stoichiometry.
Overall, the generally much higher values found for the stability of the 1:1 metal complex with the ligands bearing a terminal α-amino-carboxylic group (H 2 (L2), H 2 (L3), H 2 (L4)) may be mainly attributed to the probable co-adjuvation of the main hydroxypyridinone (O,O) metal coordination by the (N,O) glycine type coordination, and also the inserted amide bond, which can further interfere in the length and rigidity of the linker between both main groups.

Sequestering Ability
The evaluation of the sequestering ability of the 3-hydroxy-4-pyridinones towards Cu 2+ and Fe 3+ can be performed by calculating the pL 0.5 empirical parameter which represents the total ligand concentration required for the 50% sequestration of a metal cation if present in trace amount in solution. The pL 0.5 can be described using a sigmoidal type Boltzmann equation, with asymptotes equal to 1 for pL→−∞ and 0 for pL → +∞ (Equation (3)): where x M is the mole fraction of metal cation complexed by the ligand, pL = −log c L and pL 0.5 = −log c L , if x M = 0.5. The evaluation of the sequestering ability is very important for detoxification, remediation of polluted systems and water treatment processes, requiring the use of a chelating agent with the aim of trying to optimize the working conditions. A more detailed description of the pL 0.5 determination, its importance and other possible applications is reported in the literature [40].
The study of the sequestering ability of the ligands towards Cu 2+ and Fe 3+ was performed at I = 0.15 mol L −1 in NaCl (aq) , T = 298.15 K and different pHs. In the case of Fe 3+ /H 2 (L2) and Fe 3+ /H(L5) systems, the pL 0.5 was also determined at the same ionic strength and T = 310.15 K.
From the analysis of the data reported in Table 3 and Figure S7a for the Cu 2+ /L2 system, it can be concluded that the sequestering ability increases with pH, probably due to the gradual ligand deprotonation, which favours the metal-ligand electrostatic interaction. At pH = 7.4 (physiological value), I = 0.15 mol L −1 in NaCl (aq) and T = 298.15 K, the pL 0.5 trend is: H 2 (L2) ≥ H 2 (L3) > H 2 (L4) > H(L5) > H 2 (L1) (Table 3, Figure S7b). As already observed for the stability constants, the sequestering ability is also influenced by the presence in the ligands structure of the -CO 2 H, -NH 2 , -CHNH 2 CO 2 H [24] and, possibly, the amidic moiety in the alkyl chain. In addition, the pL 0.5 value obtained for the amino-3-hydroxy-4-pyridinone (H(L5)) is slightly higher than the one calculated for H 2 (L1) (terminal -CO 2 H group), highlighting a better Cu 2+ sequestration by the ligand featured by the terminal -NH 2 group (H(L5)) with respect to the carboxylic one [24][25][26]. As regards the iron-containing systems, the formation of precipitate at pH~5.0 during the potentiometric measurements allowed us to evaluate the sequestering ability of the ligands in a quite narrow pH range. As can be observed in Figure 8, the pL 0.5 trend at pH = 4.0 is: H 2 (L2) (8.12) > H 2 (L4) (7.94) > H(L5) (6.77) > H 2 (L1) (5.14), confirming that, analogously to what was observed for the stability constants, the sequestration is mainly favoured by the presence in the ligand structures of the amide-amino-carboxylic, aminocarboxylic or amino moieties [24] in the alkyl chain bound to the N-heterocyclic ring. favoured by the presence in the ligand structures of the amide-amino-carboxylic, aminocarboxylic or amino moieties [24] in the alkyl chain bound to the N-heterocyclic ring.
Furthermore, since in the case of H2(L2) and H(L5) UV-Vis experiments were carried out in the pH range 2.0-9.1, for these ligands the pL0.5 values were also calculated at different pHs, T = 298.15 K and 310.15 K (Table S3), considering the spectrophotometric data ( Table 2).
The sequestering ability of H2(L2) and H(L5) towards Fe 3+ was found to increase with pH, possibly owing to the gradual ligand deprotonation with pH increasing. The pL0.5 values also increase with temperature, in accordance with the stability constants trend ( Table 2).

Analysis of the pM Values
The study of the metal-chelating affinity of a ligand or the comparison between different ligands' behaviour towards one or more metal cations can be performed by means  (Table S3), considering the spectrophotometric data ( Table 2).
The sequestering ability of H 2 (L2) and H(L5) towards Fe 3+ was found to increase with pH, possibly owing to the gradual ligand deprotonation with pH increasing. The pL 0.5 values also increase with temperature, in accordance with the stability constants trend ( Table 2).

Analysis of the pM Values
The study of the metal-chelating affinity of a ligand or the comparison between different ligands' behaviour towards one or more metal cations can be performed by means of the pM parameter, with pM = −log [M] free (with M = Cu or Fe) for c Mn+ =1.0·10 −6 mol L −1 and c ligand = 1.0·10 −5 mol L −1 [41].
The pM values of all the Cu 2+ and Fe 3+ /(3,4-HPs) systems investigated in this paper were calculated at pH = 7.4 (physiological value). Furthermore, an attempt to compare the obtained data with those determined for ligands with similar molecular structures and functional groups ( Figure S5) was carried out using literature stability constants [22,[31][32][33][34][35][36][37][38][39]42], taking into account the already mentioned experimental and methodological differences used for the data treatment.
The analysis of the pCu values reported in Table 4 and in Figure 9a showed that, at physiological pH, the copper-chelating affinity by the ligands is favoured by the concomitant presence of -COOH, -NH 2 and amino-carboxylic groups in the 3,4-HP molecules. They follow the trend: H 2 (L3) > H 2 (L2) > H 2 (L4) > H 2 (L1) > H(L5). At the selected pH value, an inversion of pCu tendency can be observed for H 2 (L1) (terminal -COOH) and H(L5) (terminal -NH 2 ) with respect to the already mentioned stability constants and sequestration trend. This aspect could be explained considering that at pH = 7.4, the amino group present in H(L5) is still protonated while the carboxylic one in H 2 (L1) is already deprotonated, thus favouring the Cu 2+ /H 2 (L1) electrostatic interaction. At higher pH values, with the deprotonation of -NH 3 + to NH 2 in H(L5), the metal affinity increases with respect to H 2 (L1), and the pCu trend becomes analogous to that observed for the stability constants and pL 0 . 5 values. A comparison between the pCu data (Table 4, Figure 9b) determined for the The Fe 3+ chelating efficiency was evaluated at pH = 7.4 only for L2 and L5 ligands, since their interaction with the metal cation was also investigated by UV-Vis spectrophotometry, an analytical technique not used for the Fe 3+ /H2(L1) and Fe 3+ /H2(L4) systems. In fact, the UV-Vis studies were performed at lower component concentrations (c ~ 10 −4 mol L −1 ) than those used for the potentiometric ones (c ~ 10 −3 mol L −1 ), allowing us to investigate a wider pH range (2.0-9.1) without being stopped at pH ~ 5.0, as occurred for potentiometric titrations, owing to the formation of a precipitate possibly attributable to the formation of the Fe(OH)3(s) species [21].
Analysis of the pFe values, reported in Table 4 for the systems studied herein, showed that at physiological pH the iron-chelating affinity is favoured by the concomitant presence of extra-functional groups in the 3,4-HP ligand molecules, namely the amideamino-carboxylic moiety (H2(L2)), with respect to the simple terminal group -NH2 (H(L5)). The pFe data (Table 4) determined for Fe 3+ /H2(L2) and Fe 3+ /H(L5) systems were also compared with the values reported in the literature for ligands such as H2(L1) [32], H2(L4) [31], The Fe 3+ chelating efficiency was evaluated at pH = 7.4 only for L2 and L5 ligands, since their interaction with the metal cation was also investigated by UV-Vis spectrophotometry, an analytical technique not used for the Fe 3+ /H 2 (L1) and Fe 3+ /H 2 (L4) systems. In fact, the UV-Vis studies were performed at lower component concentrations (c~10 −4 mol L −1 ) than those used for the potentiometric ones (c~10 −3 mol L −1 ), allowing us to investigate a wider pH range (2.0-9.1) without being stopped at pH~5.0, as occurred for potentiometric titrations, owing to the formation of a precipitate possibly attributable to the formation of the Fe(OH) 3(s) species [21].

Comparison between M n+ /(3,4-HPs) Systems
The data presented in the current paper for Cu 2+ and Fe 3+ /(3-hydroxy-4-pyridinones) systems (Tables 1 and 2) were compared with those already reported in the literature on the interaction of the five ligands with Zn 2+ [19] and Al 3+ [13] (Table S4) at I = 0.15 mol L −1 in NaCl (aq) and T = 298.15 K. The speciation models determined for the different systems display a common species, namely the ML (n-z) , which can be used as reference to evaluate and compare the binding ability of the ligands towards the metal cations. The data analysis showed that the logK 110 trend is: Fe 3+ > Al 3+ > Cu 2+ > Zn 2+ , meaning that the 3,4-HPs are featured by a much higher tendency to form very stable complex species with Fe 3+ , followed by Al 3+ , with respect to the M 2+ . In addition, considering the stability constants reported in the literature for the Zn 2+ [19] and Al 3+ /(H 2 (L2), H(L5)) systems [13], the pZn and pAl values were also calculated at physiological pH (Table 4) and compared with the analogous results presented in this paper for Cu 2+ and Fe 3+ . Analysing the data obtained for the different metal cations and the graphs in Figure 10 and Figure S8, we can conclude that, similarly to the logK 110 behaviour, the pM values follow the trend for the metal ions: Fe 3+ > Al 3+ > Cu 2+ > Zn 2+ . Thus, both of the 3-hydroxy-4-pyridinones present a higher chelating affinity towards Fe 3+ , and in a lesser extent also to Al 3+ , with respect to divalent metal cations. reported in the literature for the Zn 2+ [19] and Al 3+ /(H2(L2), H(L5)) systems [13], the pZn and pAl values were also calculated at physiological pH (Table 4) and compared with the analogous results presented in this paper for Cu 2+ and Fe 3+ . Analysing the data obtained for the different metal cations and the graphs in Figures 10 and S8, we can conclude that, similarly to the logK110 behaviour, the pM values follow the trend for the metal ions: Fe 3+ > Al 3+ > Cu 2+ > Zn 2+ . Thus, both of the 3-hydroxy-4-pyridinones present a higher chelating affinity towards Fe 3+ , and in a lesser extent also to Al 3+ , with respect to divalent metal cations. These trends could probably be justified taking into account the already mentioned "hard-soft acids and bases" theory (HSAB) [27][28][29], according to which a hard acid-hard base or a soft acid-soft-base interactions are kinetically and thermodynamically favoured if compared with hard-soft ones. On this basis, the affinity between hard metal cations (acids: Fe 3+ , Al 3+ ) and hard-base functional groups (bases: -OH, -COOH) is higher with respect to those with borderline acids like Cu 2+ and Zn 2+ . These trends could probably be justified taking into account the already mentioned "hard-soft acids and bases" theory (HSAB) [27][28][29], according to which a hard acid-hard base or a soft acid-soft-base interactions are kinetically and thermodynamically favoured if compared with hard-soft ones. On this basis, the affinity between hard metal cations (acids: Fe 3+ , Al 3+ ) and hard-base functional groups (bases: -OH, -COOH) is higher with respect to those with borderline acids like Cu 2+ and Zn 2+ .
In light of these considerations, it can be claimed that, from a thermodynamic point of view, most of the bifunctional 3,4-HP ligands studied herein are particularly selective towards Fe 3+ and could be considered promising iron-chelating agents, also avoiding the possibility of a significant competition, and eventually a depletion, of divalent metals with biological and environmental relevance, such as Cu 2+ and Zn 2+ .

Chemicals
Riedel-deHäen concentrated ampoules were used to prepare sodium hydroxide and hydrochloride solutions standardized against potassium hydrogen phthalate and sodium carbonate, respectively. NaOH solutions were preserved from atmospheric carbon dioxide by means of soda lime traps. CuCl 2 ·2H 2 O and FeCl 3 ·6H 2 O salts purchased by Fluka were weighed to prepare the metal solutions without further purification and standardized against EDTA standard solutions [43]; their purity was always ≥98%. The synthesis of the functionalized 3-hydroxy-4-pyridinones was already reported in the literature [13]. The ligand solutions were prepared by weighing the products in the neutral form (H r (L) 0 ) without any further purification. Their purity was checked by means of alkalimetric measurements and, for all the ligands, it was found to be ≥99.5%. The ionic medium aqueous solutions were prepared by weighing the pure Fluka NaCl salt, previously dried in an oven at T = 383.15 K for two hours. The reagents used to carry out the studies were of the best available purity. The preparation of the solutions was performed using analytical grade water (R = 18 MΩ cm −1 ) and grade A glassware.

Potentiometric Equipment and Procedure
The interactions of the five 3-hydroxy-4-pyridinones towards Cu 2+ and Fe 3+ were experimentally investigated using a Metrohm 809 Titrando and a potentiometer with a combined Thermo-Orion glass electrode (Ross type 8102) connected to an automatic burette. This apparatus was coupled to a personal computer, and automatic titrations were performed by means of the MetrohmTiAMO 1.2 software, useful for the control of titrant delivery, data acquisition and e.m.f. stability. The estimated accuracy, for e.m.f. and titrant volume readings, was ±0.15 mV and ±0.003 mL, respectively. The measurements were carried out in 25 mL thermostatted cells under magnetic stirring, and purified presaturated nitrogen was bubbled into the solutions for at least 5 min to exclude the presence of oxygen and carbon dioxide inside. For all the experiments, titrations of hydrochloric acid with standard NaOH solutions were carried out at the same temperature, ionic strength and ionic medium conditions with respect to those used for the systems under study, for refining the value of the electrode potential (E 0 ), the acidic junction potential (Ej = j a [H + ]) and the ionic product of water (K w ). The pH scale employed was the free scale and pH≡ −log[H + ], with [H + ] that is the free concentration of the proton. From sixty to one hundred data points were collected during each titration, depending on the possible formation of sparingly soluble species.

UV-Vis Spectrophotometric Apparatus and Procedure
The UV-Vis spectrophotometric titrations were carried out using a Varian Cary 50 spectrophotometer presenting an optic fibre probe with a fixed 1-cm path length. This instrument was connected to a computer, and the recording of absorbance (A) signal vs. wavelength (λ / nm) was carried out by means of the Varian Cary WinUV software. At the same time, a Thermo-Orion combined glass electrode (Ross type 8102), linked to a potentiometer, was employed to collect potentiometric data. The NaOH titrant solution was delivered in a 25-mL titration cell using an automatic burette (Metrohm 665 model). The homogeneity of the solutions during the measurements was ensured using a magnetic stirrer. Nitrogen was bubbled in the solutions for at least 5 min before starting the experiments, also in this case, for excluding the presence of O 2(g) and CO 2(g) inside.

Computer Programmes
Appropriate computer programmes were employed for the treatment of experimental data from different analytical techniques. The non-linear least squares ESAB2M computer program [44] was used for the determination of the acid-base titrations parameters (E 0 , pK w , j a ) and the reagents' analytical concentration. The elaboration of potentiometric data was carried out by means of the BSTAC computer program [45], while the UV-Vis spectrophotometric ones were processed using the HypSpec 2008 [46]. The calculation of the M n+ /3-hydroxy-4-pyridinone species formation percentages and the representation of distribution diagrams was performed using the HySS program [30].

Conclusions
The binding ability of five bifunctional 3-hydroxy-4-pyridinones towards Cu 2+ and Fe 3+ was studied by means of potentiometric and UV-Vis spectrophotometric measurements carried out at I = 0.15 mol L −1 in NaCl (aq) and T = 298.15 K. The data treatments allowed us to determine the speciation schemes featured by metal-ligand species with different stoichiometry and stability, due to the various functional groups present in the 3-hydroxy-4-pyridinones structures, which could potentially participate in the metal complexation and in the Cu 2+ and Fe 3+ behaviour in an aqueous solution. The stability of metal-ligand species follows the trends: Cu(L3) 0 (aq) > Cu(L4) 0 (aq) > Cu(L2) 0 (aq) > Cu(L5) + > Cu(L1) 0 (aq) and: Fe(L4) + > Fe(L2) + > Fe(L5) 2+ > Fe(L1) + , respectively. They were favoured by the simultaneous presence of amino or amino-carboxylic bearing groups in the 3,4-HP ligands, and showed some dependence on the length and structure of the chains between the pyridinone ring and the extra-functional groups. The investigation of the sequestering ability and metal-chelating efficiency was carried out by the calculation of the pL 0 . 5 and pM parameters at different pHs and physiological value (pH = 7.4), respectively. Similarly to the complexation behaviour, the sequestration and Cu 2+ and Fe 3+ affinity by the ligands under study is affected by the presence in the whole 3-hydroxy-4-pyridinone molecules of terminal amino-carboxylic groups and amidic moiety in the alkyl chain or, at least, of the one single terminal group, as -NH 2 group (H(L5)), with respect to the carboxylic group (H 2 (L1)). In addition, the data presented in this paper for Cu 2+ and Fe 3+ /3-hydroxy-4-pyridinone systems were compared with those reported in the literature, for the interaction of the ligands with Al 3+ and Zn 2+ at I = 0.15 mol L −1 in NaCl (aq) and T = 298.15 K. The logK 110 and pM trend show a clear dependence on the metal ion (Fe 3+ > Al 3+ > Cu 2+ > Zn 2+ ), meaning that the 3-hydroxy-4-pyridinones display a higher stability and chelating affinity towards Fe 3+ and (in a lesser degree) also Al 3+ , with respect to divalent metal cations. In light of these considerations, it can be claimed that, from a thermodynamic point of view, the ligands are particularly selective towards Fe 3+ and could be considered promising iron-chelating agents, also avoiding the possibility of a significant competition, and eventually a depletion, of divalent metals with biological and environmental relevance, such as Zn 2+ and Cu 2+ .
Supplementary Materials: The following are available online, Table S1. Overall and stepwise protonation constants of the 3-hydroxy-4-pyridinones under study reported in the literature at I = 0.15 mol L −1 in NaCl (aq) and different temperatures; Table S2. Literature stability constants of Cu 2+ and Fe 3+ /ligand species reported at different temperatures, ionic strengths and ionic media in molar concentration scale; Table S3. pL 0.5 values of Fe 3+ /H 2 (L2) and H(L5) systems at different pHs and temperature, from UV-Vis data at I = 0.15 mol L −1 in NaCl (aq) ; Table S4. Literature stability constants of ZnL (2-z) and AlL (3-z)