Active role of the buffer in the proton-coupled electron transfer of immobilized iron porphyrins

Depósito de investigación de la Universidad de Sevilla https://idus.us.es/ "This document is the Accepted Manuscript version of a Published Work that appeared in final form in Inorganic Chemistry, copyright © American Chemical Society after peer review and technical editing by the publisher. To access the final edited and published work see https://doi.org/10.1021/acs.inorgchem.0c01091” The Active Role of the Buffer in the Proton-Coupled Electron Transfer of Immobilized Iron Porphyrins Inmaculada Márquez, José Luis Olloqui-Sariego, Miguel Molero, Rafael Andreu, Emilio Roldán and Juan José Calvente* Departamento de Química Física, Universidad de Sevilla. C/Profesor García Conzález, 1. 41012 Sevilla. Spain *Corresponding author. e-mail: pac[email protected] 1 ABSTRACT Evaluation of the proton-coupled electron transfer thermodynamics of immobilized hemin is challenging due to the disparity of its electrochemical titration curves reported in literature. Deviations from the one-electron/one-proton transfer at circumneutral pHs have been commonly ascribed to either the formation of dimeric species or to the ionization of a second iron-bound water molecule. Herein, however, we report on non-idealities in the more acidic region, whose onset and extent vary with the nature and concentration of the commonly used phosphate and acetate buffers. It is shown that these deviations originate in the ligand-exchange binding between the oxidized aquo-hemin complex and the anionic components of the buffer, so that they are restricted to the pH interval where these forms coexist. A stepwise approach was developed to quantify unambiguously both the apparent and intrinsic binding equilibrium constants. The apparent binding equilibrium constant exhibits a peak-shaped pH dependence, whose maximum is located at ca. the midpoint between the pKa of the iron-bound water and the first pKa of the buffer, and its magnitude is greater for the phosphate than for the acetate buffer. But strikingly, the opposite trend was found for the magnitude of the intrinsic binding equilibrium constants determined from the apparent ones, due to the different relative locations of the phosphoric and acetic pKa values with respect to that of the oxidized aquo-hemin. To probe the role of the heme propionic residues, a similar study was carried out with a propionic-free iron porphyrin containing eight ethyl residues. These substituents decrease the acidity of the iron-bound water, strengthens the iron(III)-acetate binding, weakens the iron(III)-dihydrogen phosphate binding, and enables the binding between iron(III) and monohydrogen phosphate, that was hampered in hemin by the presence of the negatively charged propionate residues. Overall, this work provides a more complete speciation of immobilized iron porphyrins in acid 2 conditions than previously considered, showing the substitutional lability of the aqua ligand in the oxidized state of the iron center and the reluctance of its hydroxyl counterpart to anion exchange. Knowledge of these redoxand pH-dependent bindings with the buffer components is crucial for a rigorous quantification of the proton-coupled electron transfer and the electrocatalytic activity of iron porphyrins. 3 INTRODUCTION The study of the redox and acid/base properties of iron porphyrins is important to understand the influence of both metal oxidation and acid/base ionization states on the coordinative chemistry of the heme group. This information is also valuable to quantify the proton-coupled electron transfer (PCET)1–4 and the electrocatalytic response of heme-containing species,5–14 as well as to identify the molecular factors that control the redox potential of hemin15–20 and hemoproteins,21–27 and to rationalize the relationship between their conformational changes and the exchange of ionizable ligands.28–30 The pH dependence of the redox potential of immobilized hemin onto different types of electrodes has shown that its interfacial electron transfer is coupled to a protonation transfer for pH values above a critical threshold. A close inspection of the published results reveals a significant dispersion of this threshold within the 2.5 - 4.0 pH interval, and major differences in the shape of the voltammetric titration curve ( voltammetric midpoint potential vs. pH ). For instance, Ksenzhek et. al.31 reported a pH dependence of the redox potential of immobilized hemin at graphite electrodes that was composed of a limiting asymptotic value for pH < 4.65, followed by two linear segments (slope of ~-60 mV/pH), separated by an intermediate levelling off, for pH > 4.65 (up-triangles in Figure 1). Deviation of the titration curve with respect to the 1e-/1H+ predictions (solid line) was interpreted by invoking dimerization of hemin in the two redox states. Later on, Shigehara and Anson,32 in the context of a study of the electrocatalytic reduction of oxygen and hydrogen peroxide mediated by hemin, reported a titration curve that differs from the previous one by the extension of the asymptotic and linear segments, and by the presence of a wider intermediate plateau between pH 5.5 and 8.5 (down-triangles in Figure 1). In a later 4 study, Bianco et al.33 reported on a simpler titration curve, composed of the limiting acidic value followed by a linear dependence for pH > 3, that fulfills the predictions for a 1e-/1H+ transfer (circles in Figure 1). They ascribed the disparity of the previously reported titration curves to variations in the formation of dimers due to different experimental conditions. It should also be noted that a small intermediate plateau located at more acidic pHs (2 – 4) is observed in the titration curve reported more recently by De Groot and Koper34 for hemin immobilized at graphite electrodes (squares in Figure 1). On the other hand, Pilloud et al.35 reported on a levelling off of the redox potential for covalently-attached hemin at dimercaptoalkane-modified gold electrodes in the circumneutral pH region ( 7 – 8 ), that was interpreted as evidence of the formation of mono and bis(hydroxo) complexes of hemin in its monomeric or µ-peroxo dimeric states. The fact that this levelling off persisted in the esterified hemin derivative revealed that it Figure 1. Shift of the redox potential, with respect to its acidic asymptotic value, for hemin immobilized onto graphite electrodes as a function of the solution pH: (up triangles) Ksenzhek et al.,31 (down triangles) Shigehara et al.,32 (circles) Bianco et al.33 and (squares) De Groot et al.34 Solid lines correspond to theoretical predictions for 1e-/1H+ redox processes calculated from 1Ox m a,AH H E -(RT / nF )ln( (K / c ))     with the indicated Ox a,AH pK values. 5 was not caused by ionization of the hemin propionic residues. This finding is consistent with the presence of an intermediate plateau (pH 6 – 8) in the titration curve of a propionate-free, water-soluble iron porphyrin reported recently by Costentin et al.,9 that was explained by a coupling between the electron transfer, the protonation/deprotonation of the iron-bound water, and the formation of µ-oxo dimers. A possible cause for the disparity of the titration curves reported so far for hemin immobilized at graphite electrodes, and that has been overlooked, is the interaction of hemin with the buffer components, among which are the ionized and non-ionized forms of the phosphoric, acetic, citric and boric acids. A detailed inspection of the buffer compositions used in previous studies reveals that the simplest pH dependence for the hemin redox potential was obtained by using a low buffer concentration (10 mM),33 whereas the more intricate dependence was obtained with 0.1 M buffer solutions containing phosphate, acetate or citrate anions depending on the pH region.32 There is some evidence indicating that interaction of hemin with either the acetate or phosphate buffer components may be plausible under certain experimental conditions. Thus, for instance, interaction between Fe(III) and propionate pertaining to different hemin molecules is crucial for the formation of β-hematin, and its natural analog hemozoin (commonly known as malaria pigment) at aquo-lipid interfaces.36–38 Due to the hydrophobicity of the graphite, a similar interaction may occur between immobilized hemin and freely-diffusing acetate ions. In this regard, recent quantum chemical calculations have shown the following sequence for the binding affinity of oxidized hemin with hydroxyl > acetate > water, so that the acetate anion is able to displace the iron-bound water molecule but not its ionized form (hydroxyl).39,40 Likewise, a binding between the haem containing horseradish peroxidase enzyme and acetate has been reported in previous 6 works,41–44 where acetate stacks parallel with the heme plane without an intervening water molecule between the iron center and the acetate molecule.44 On the other hand, experimental evidence for the interaction of aquo-ferric ions with the anionic forms of the phosphoric acid has been reported in acid media, and the stability constant of the resulting 2 24 FeH PO  and 4 FeHPO complexes have been determined.45,46 In a more recent work, the interaction of aquo-ferric ion with a phosphonic acid terminated self-assembled monolayer has been used to study its electrochemical response in the immobilized state.47 Herein, motivated by the disparity and non-idealities of the titration curves reported for hemin immobilized at electrodes, and suspecting that they might originate in its interaction with buffer components, we have carried out a thorough study of the effect of the solution pH and buffer composition on the voltammetric response of two iron porphyrins that differ in their  substituents. Particularly, the propionic acid-free iron octaethylporphyrin is used to gauge the role of the propionic acid ionization on the proton-coupled electron transfer of hemin. As part of this study, a two-step strategy has been developed to quantify both the apparent and intrinsic equilibrium constants for the binding of the two iron porphyrins with the buffer components. It is shown that only the oxidized aquo-heme form binds acetate, dihydrogen phosphate or monohydrogen phosphate, the latter binding being very sensitive to the presence of negative charge at the propionate residues in the porphyrin ring. Replacement of the hemin β substituents by the more hydrophobic ethyl group decreases the acidity of the iron-bound water, strengthens the iron(III)-acetate binding, weakens the iron(III)-dihydrogen phosphate binding, and enables the binding of the hemo iron(III) with monohydrogen phosphate that was hampered in the case of hemin by the presence of the negatively charged propionate residues. 7 EXPERIMENTAL SECTION a) Chemicals Ferriprotoporphyrin IX chloride (hemin), ferriprotoporphyrin IX hydroxide (hematin) and 2,3,7,8,12,13,17,18-Octaethyl-21H,23H-porphine iron (III) (abbreviated FeOEP) were purchased from Sigma-Aldrich and used without further purification (Figure 2). Solutions of these iron porphyrins for their immobilization onto the graphite electrode were prepared in dimethylsulfoxide (Sigma-Aldrich). Phosphoric acid, tri-sodium phosphate, acetic acid and sodium acetate were from Sigma-Aldrich. Sodium dihydrogen phosphate, disodium hydrogen phosphate, sodium hydroxide and sodium chloride were from Fluka. Hydrochloric acid and ethanol were from Merck. Buffer solutions were prepared from water purified with a Millipore Milli-Q system (resistivity of 18 M cm). These solutions contain the desired concentration of the buffer and 0.1 M of either hydrochloric acid or sodium chloride depending on the desired pH value. For the titration curves, the pH of the solution in the electrochemical cell was increased by adding aliquots of a concentrated sodium hydroxide aqueous solution. To Figure 2. Schematics of the iron porphyrins used in the present work. 14 where 1 , Ox a A H K , 2 , Ox a A H K and 1 , Rd a A H K , 2 , Rd a A H K are the acid dissociation equilibrium constants of the A1H and A2H groups in the oxidized and reduced forms of the electroactive species, respectively. Fit of eq 2 to the experimental titration curve requires that one of the acid groups only ionizes in the oxidized state. If such a group is chosen to be A1H, then 1 , Rd a A H H Kc  along the operative pH interval, and the fit provides 220 Rd a,A H pK . and two alternative estimates of the pKa values in the oxidized state, either 147 Ox a,A H pK . and Figure 6. Square and cubic schemes for the proton-coupled electron transfer (1e-/1H+) of an immobilized electroactive species that bears a) one (AH) or b) two (A1H, A2H) ionizable groups. Vertical and horizontal arrows stand for the individual electron and protonation/deprotonation steps, respectively. Solid arrows stand for the particular set of equilibria that has been considered to derive the voltammetric response in the Supporting Information. 15 211 Ox a,A H pK . or 111 Ox a,A H pK . and 247 Ox a,A H pK . . According to previous studies,31–35 the A1H group can be ascribed to the iron-bound water, provided that water replaces the coordinating chloride of the hemin. In fact, the similarity of the titration curves measured for the hemin (initially coordinated to chloride) and hematin (initially coordinated to a hydroxide), depicted in the left panel of Figure 5, points to the exchange of chloride by a water molecule when the hemin-modified electrode contacts the aqueous solution. As for the second ionizable group A2H, Pilloud et al.35 ascribed it to the ionization of either the second iron-bound water molecule of the bis(aquo)hemin complex, or to the water molecule coordinating the iron of the hemin µ-peroxo dimer. However, this is unlikely hypothesis for the present system, as physisorption of hemin at graphite electrodes in a lying-down configuration involves the loss of an axially coordinating water molecule and the formation of dimers is not favored at the acid conditions where the intermediate plateau develops.64,65 Indeed the pKa values reported by Pilloud et al.35 for the second ionizable group are higher than those determined in the present study. To probe whether the second redox-dependent ionizable group can be identified with any of the two propionic acid residues of the porphyrin ring, the voltammetric titration curve of the propionic acid-free iron octaethylporphyrin (FeOEP) was determined. The similar shape of the FeOEP and hemin titration curves (Figure 5) reveals that their intermediate plateau is not related to the protonation/deprotonation of the propionate residues, although the FeOEP titration curve can be reproduced by eq 2 with 220 Rd a,A H pK . and two alternative estimates of the pKa values in the oxidized state, either 154 Ox a,A H pK . and 210 Ox a,A H pK . or 110 Ox a,A H pK . and 254 Ox a,A H pK . . 16 Then, we considered the possibility that the intermediate levelling off of the titration curve might be originated in the interaction between the hemin and some of the buffer components. Therefore we measured the titration curves for distinct phosphate buffer concentrations in the 10 – 200 mM range. As can be seen from Figure 7, a decrease of the phosphate concentration reduces the extent and depth of the intermediate plateau, so that it almost disappears for a 10 mM phosphate buffer concentration. For this low concentration, the titration curve approaches the truncated sigmoidal dependence predicted for a PCET process in which the acid group only ionizes in the oxidized state of the redox center. Interestingly, titration curves are insensitive to the phosphate concentration for pH > 6, and approach a common value at low enough pHs, so that a maximum deviation from the ideal PCET behavior is observed between pH 2.5 and 4.0. According to the acid/base speciation of the hemin and the phosphate buffer, two scenarios are consistent with the above findings, Figure 7. Effect of the phosphate (left) and acetate (right) buffer concentration on the voltammetric titration curve of immobilized hemin onto an edge pyrolytic graphite electrode, measured at 0.1 V s-1 and 25 ºC, with a starting solution containing the indicated buffer concentration and 0.1 M HCl, whose pH was increased by adding aliquots of a concentrated sodium hydroxide solution. Solid lines are theoretical fits to the binding-proton-coupled electron transfer process depicted in Figure 10b (eq 10). Dashed line represents the theoretical prediction in the absence of the binding event. The inset represents a magnification of the acidic portion of the titration curves. 17 namely i) binding of the aquo-hemin complex with the ionized forms of the buffer, or ii) binding of the hydroxo-hemin complex with the un-ionized form of the buffer. On the basis of previous theoretical39 and experimental66 studies, that showed a greater affinity of the oxidized hemin for anionic ligands than for their neutral counterparts, the first scenario is the most plausible. To further explore the role of the buffer components, and to search for reference conditions where their interaction with the hemin is absent, the voltammetric titration curve of hemin was also measured in the presence of sodium acetate buffer. As illustrated in Figure 7, replacement of phosphate by acetate results in simpler titration curves, with only a small reminiscence of the intermediate plateau for high acetate concentrations. As a matter of fact, the titration curve determined starting from the 10 mM acetate, 0.1 M HCl solution can be reproduced quantitatively with the theoretical expression for a single acid/base group that only ionizes in the oxidized state of the hemin: , ln 1 Ox a AH o m prot H K RT EE nF c         (3) 18 Fit of eq 3 to the titration curve of hemin in the 10 mM acetate buffer leads to ,=3.85 Ox a AH pK . This estimate is similar to the value determined by De Groot and Koper34 ( ,4.0 Ox a AH pK  ) after discarding the small intermediate plateau between pH 2 and 4.5. These results indicate that a low concentration of the acetate buffer is an appropriate choice to determine the acid/base properties of immobilized heme-containing molecules, due to its smaller propensity to interact with the heme group. We take advantage of this finding to probe the role of the propionic residues in the ionization of the iron-bound water molecule, by measuring the titration curve of the propionic acid-free FeOEP in 10 mM acetate, 0.1 M HCl buffer (Figure 8). Fit of eq 3 to experimental Em vs. pH data of FeOEP leads to ,4.3 Ox a AH pK  , which is higher than the pKa value determined for hemin ,3.85 Ox a AH pK  ; revealing that replacement of the propionic groups by alkyl substituents in the porphyrin ring reduces the acidity of the iron-coordinating water molecule. This effect can be ascribed to an increase of the electron density in the porphyrin ring brought about by the electron Figure 8. pH dependence of a) the voltammetric midpoint potential and b) its shift with respect to the acidic asymptotic value of immobilized hemin (circles) and iron octaethylporphyrin (squares) onto an edge pyrolytic graphite electrode, measured at 0.1 V s-1 and 25 ºC, with a starting solution containing 0.01 M acetic acid and 0.1 M HCl, whose pH was increased by adding aliquots of a concentrated sodium hydroxide solution. Solid lines are theoretical fits of eq 3 to the experimental data. 19 donating inductive effect of the alkyl residues.67 Interestingly, the same effect is responsible for the more negative value of the FeOEP redox potential at low enough pHs with respect to that of hemin, as an increase of the electron density in the porphyrin ring stabilizes the ferric vs. the ferrous state. b) Quantifying the hemin-anion binding b.1) Binding curves It has been shown in the previous section that the PCET of immobilized hemin is coupled to anionic binding events in the pH interval between 2 and 6. For a precise determination of the binding equilibrium constants, we have measured the voltammetric response of the hemin-modified electrode for increasing buffer concentrations, while keeping the same pH value. In the case of the acetate buffer, the starting solution was 1 mM acetate and 0.1M of either HCl or NaCl, whose acetate concentration was progressively increased by adding appropriate volumes of a solution containing 0.5 M acetate, 0.1M HCl (or 0.1M NaCl) at the same pH. Figure 9 shows the dependence of the voltammetric midpoint potential on the logarithm of the acetate buffer concentration for eleven pHs in the 1.2 < pH < 6.0 interval. The insensitivity of Em to the acetate buffer concentration for pH < 2.5 reveals a negligible interaction of the aquo-hemin complex with the un-ionized form of the acetate buffer. For pH > 2.5, however, the negative-going shift of Em with the acetate buffer concentration reveals a stabilization of the oxidized form of the hemin with respect to the reduced form that is likely to be originated by the binding of aquo-hemin with the acetate anion. The extent of this negative-going shift of Em reaches a maximum at ca. pH 4.5, and almost disappears at pH 6, revealing a negligible binding of the hydroxo-hemin complex with the acetate anion. The observed deviation of Em toward more negative values 20 at low enough acetate buffer concentrations for pH > 5.5 results from a limitation of the acetate buffering capacity. 21 In the case of the phosphate buffer, to overcome its limited buffering capacity in the 3.5 < pH < 5.5 range, we have started the titration with a 1 mM phosphate, 10 mM acetate and 0.1 M HCl (or 0.1 M NaCl) solution at the desired pH, taking advantage of the negligible hemin-acetate binding at the 10 mM acetate concentration. Then, the phosphate concentration was progressively increased upon addition of appropriate volumes of a solution containing 0.5 M phosphate, 10 mM acetate and 0.1 M HCl (or 0.1 M NaCl) with Figure 9. Upper panel: Dependence of the voltammetric midpoint potential of immobilized hemin onto an edge pyrolytic graphite electrode on the acetate (left) or phosphate (right) buffer concentration measured at the indicated pH values with a scan rate of 0.1 V s -1 and 25 ºC. All solutions contain 0.1 M of either HCl or NaCl. The phosphate buffer solution also contains 10 mM acetic acid/acetate. Solid lines represent the theoretical fits of eq 5 to experimental data (symbols). The observed deviations at low enough buffer concentrations for the less acidic pHs result from a limitation of the buffer capacity. Lower panel: Dependence of the apparent binding equilibrium constant KOx·L on the solution pH for the binding of oxidized hemin with acetate (left) or phosphate (right) buffer components. Solid lines represent the theoretical fit of eq 7 to experimental data (symbols) as explained in the text. 22 the same pH as the initial solution. The corresponding binding curves (Figure 9) were similar to those obtained with the acetate buffer, except that now the negative-going shift of Em starts at more acidic pHs (as low as 1.2) and reaches its maximum extent at ca. pH 3.5, in agreement with the relative pKa values of the phosphoric (2.16) and acetic (4.76) acids,68 if hemin binding is assumed to involve the ionized form of the buffer only. A stepwise strategy was used to quantify the hemin-anion binding. First, an estimate of the apparent hemin-buffer binding equilibrium constant KOx·L for each solution pH was obtained from the buffer concentration dependence of Em. KOx·L refers to the global binding constant between the oxidized form of the hemin (Ox) and the freely-diffusing ligand (L), without taking into account explicitly their acid/base states. Then, from the pH dependence of KOx·L, an estimate of the intrinsic hemin-anion equilibrium constant 2 HO Ox A K was obtained by taking into account explicitly the acid/base states of both hemin and buffer. As shown later, 2 HO Ox A K refers to the ligand-exchange binding constant between the aquo-hemin complex (Ox·H2O) and the ionized form of the buffer (A). The KOx·L value for a given solution pH was estimated from the theoretical expression of Em associated with the square scheme depicted in Figure 10a, that combines electron transfer and binding equilibria without considering the acid/base speciation of the relevant species (eq S28): · / · 1 ln1 oOx L L m Ox Rd Rd L L Kc RT EE nF K c    (4) where / o Ox Rd E stands for the formal standard potential of the unbound redox couple (Ox/Rd), and KOx·L, KRd·L are the apparent binding equilibrium constants of the oxidized and reduced forms of the redox couple with the freely-diffusing ligand L, respectively, and cL is 23 the ligand bulk concentration. The negative-going shift of Em with the buffer concentration reveals that KOx·L > KRd·L, whereas the lack of a limiting asymptotic value of Em at high enough buffer concentrations is consistent with ·1 Rd L L Kc along the operative concentration range. This result agrees with the recently reported stronger binding of acetate to the ferric vs. the ferrous form of a related cationic iron porphyrin,66 and is a consequence of the different acid/base hardness of Fe(III) and Fe(II), that favors the binding of Fe(III) to hard oxygen ligands. The negligible binding of ferrous hemin to the Figure 10. a) Square schemes for the binding-coupled electron transfer of an immobilized electroactive species that binds the freely-diffusing ligand L (eq 4). b) Thermodynamic cycle for the binding proton-coupled electron transfer of an immobilized electroactive species whose oxidized form binds the ionized forms of a b.1) monoprotic or b.2) diprotic freely-diffusing acid (eqs 10 and S50). Solid arrows stand for the particular set of equilibria that has been considered to derive the voltammetric response in the Supporting Information. 30 Eqs 11 - 15 can be used to obtain a first estimate of the relevant thermodynamic parameters from the transitional pH values and the intermediate basin depth. We have found that eq 10 is able to reproduce quite satisfactory the hemin titration curves measured for distinct buffer concentrations (Figure 7) by using the parameter values determined in the previous section from the pH dependence of KOx·L (solid lines in Figure 9). Figure 13. Upper panel: General shape of the theoretical voltammetric titration curve associated with the binding-proton-coupled electron transfer depicted in Figure 10b, calculated from eq 10 with the following parameter values 2 ,7.0 Ox a H O pK  , 4.0 L a pK  and 24 1·10 HO Ox·A K . Dashed lines represent its dissection into the linear segments whose intercepts provide the values of the characteristic transition pHs. Lower panel: Theoretical titration curves computed from eq 10 for: (left) distinct L a pK values with 2 ,5.0 Ox a H O pK  and 23 510 HO Ox·A K· , and (right) distinct 2 HO Ox·A K values with 2.0 L a pK  , 2 ,5.0 Ox a H O pK  . In all cases 0.1 M L c , 255.5 M HO c and 0 o prot E . 31 c) Role of the heme propionic residues 32 To probe the role of the heme propionic residues in the Fe(III)-anion binding, the voltammetric response for the propionic-acid free FeOEP was measured for increasing concentrations of the acetate and phosphate buffers at distinct solution pHs. The corresponding binding curves (Figure 14) are qualitatively similar to those obtained for hemin, but significant differences are observed in the KOx·L values determined from them with eq 5. In the acetate buffer, FeOEP shows a greater value of (KOx·L)max than hemin, that 33 accordingly to eq 9 is partly due to the higher 2 , Ox a H O pK value of FeOEP. The fit of eq 7 to the experimental KOx·L vs. pH data, by using the pKa values of the acetic acid ( 4.76 L a pK  ) and FeOEP porphyrin ( 2 ,= 4.30 Ox a H O pK ), provides the following estimate of the intrinsic Fe(III)OEP-acetate binding equilibrium constant ( 24 ·1.5·10 HO Ox AcO K ), which is greater than the value determined for the hemin-acetate binding (1.1·104). Although small, the difference between these two values is statistically significant, as illustrated by comparing 34 the experimental KOx·L vs. pH data of FeOEP with the theoretical predictions calculated by using the value of the hemin-acetate binding constant (dotted green line in the lower left panel of Figure 14). The large mismatch observed between experiment and theory supports the significance of the reported difference between the acetate binding constants to FeOEP and hemin. The enhancement of the Fe(III)-carboxylate binding upon increasing the 35 hydrophobicity of the porphyrin ring might originate from the additional interaction between the methyl group of the acetate anion and the porphyrin ring. In the case of the phosphate buffer, FeOEP shows a smaller and more asymmetric KOx·L vs pH peak (Figure 14) than the one obtained for hemin (Figure 9), with a significant broadening for pH > 4.5. In fact, this peak can only be reproduced quantitatively with eq 7 Figure 14. Upper panel: Dependence of the voltammetric midpoint potential of immobilized iron octaethylporphyrin onto a graphite electrode on the acetate (left) or phosphate (right) buffer concentration measured at the indicated pH values with a scan rate of 0.1 V s -1 and 25 ºC. All solutions contain 0.1 M of either HCl or NaCl. The phosphate buffer solution also contains 10 mM acetic acid/acetate. Solid lines represent the theoretical fits of eq 5 to experimental data (symbols). The observed deviations at low enough buffer concentrations for the less acidic pHs result from a limitation of the buffer capacity. Lower panel: Dependence of the apparent binding equilibrium constant KOx·L on the solution pH for the binding of oxidized iron octaethylporphyrin with acetate (left) or phosphate (right) buffer components. Dashed and solid lines represent the theoretical fit of eq 7 and eq 16, respectively, to experimental data (symbols) as explained in the text. Dotted green lines represent the theoretical predictions calculated with the values of the hemin-acetate and hemin-dihydrogen phosphate binding constants. 36 in the 1.2 ≤ pH ≤ 4.5 range by using the following parameter values 2 ,4.30 Ox a H O pK  and 2.16 L a pK  and the optimum value of the fitting parameter 2 24 3 ·4.3x10 HO Ox H PO K (dashed line in the lower panel of Figure 14). We have found that the deviations for pH > 4.5 can be accounted for by including the binding of the iron center with the monohydrogen phosphate dianion in the proton-coupled binding scheme (Figure 11b). The following expression for the pH dependence of KOx·L has been derived in SI (eq S44 applied to the particular case of a ligand with two ionized states): 22 22 2 ,1 ,1 ,2 ·· 2 · ,1 ,1 ,2 , 2 1 L L L H O H O a a a Ox AH Ox A H O H O HH H Ox L L L L Ox a a a a H O H HH K K K KK c c c c c KK K K cK cc       (16) where ,1 L a K and ,2 L a K stand for the first and second acid dissociation equilibrium constants of the phosphate buffer, respectively, 2 , Ox a H O K is the acid dissociation equilibrium constant of the iron-bound water molecule, and 2 · HO Ox AH K and 2 · HO Ox A K are the intrinsic binding equilibrium constants for the binding between aquo-Fe(III)OEP and the first ( - 24 AH H PO ) and second ( 24 A HPO ) ionized forms of the buffer, respectively. For the particular case of the phosphate buffer, 22 24 ·· = H O H O Ox AH Ox H PO KK and 22 4 ·· H O H O Ox A Ox HPO KK . For the ,1 2.16 L a pK  and ,2 7.21 L a pK  values of the phosphoric acid68 and the typical 2 ,4.0 Ox a H O pK  value for immobilized iron porphyrins, eq 16 predicts the appearance of a shoulder in the peaked KOx·L vs. pH curve upon increasing the binding of the metal center with the second ionized form of the buffer, 2 · HO Ox A K , that results in a second peak for large enough values of 2 · HO Ox A K (Figure 15). The experimental pH dependence of KOx·L for FeOEP resembles the theoretical curves with the shoulder. In fact, it can be reproduced 37 quantitatively by using the following parameter values 2 ,4.30 Ox a H O pK  , ,1 2.16 L a pK  and ,2 7.21 L a pK  , and the optimum values of the fitting parameters 22 24 3 ·· = = 4.3·10 H O H O Ox H PO Ox AH KK and 22 4 5 ·· = =3.0·10 H O H O Ox HPO Ox A KK (solid line in the lower panel of Figure 14). The greater affinity of the positively charged Fe(III)OEP for the 24 HPO dianion than for the singly charged - 24 H PO reveals that electrostatic interactions play an important role in these bindings. This is consistent with a negligible 24 Fe(III)-HPO binding in the hemin, as the ionization of its propionic acid residues imparts an overall negative charge to hemin in the pH region of coexistence with 24 HPO . Theoretical and experimental estimates for the pKa value of the hemin propionic acid residues reported in literature are within the (4.3 – 6.7) range,39,69,70 so that a significant ionization of these residues occurs in the tail of the KOx·L vs. pH plot. In addition to the aforementioned electrostatic repulsion, a second factor Figure 15. Theoretical pH dependence of the apparent binding equilibrium constant KOx·L associated with the proton-coupled binding process depicted in Figure 11b, calculated from eq 16 with 240 Ox a,H O pK . , ,1 2.16 L a pK  , ,2 7.21 L a pK  , 255 5 HO Ox·AH K. and the indicated values of 2 HO Ox·A K . 38 contributing to the lack of the 24 Fe(III)-HPO binding in hemin is the low population of the aquo-ferric form that coexists with 24 HPO , as a consequence of the more acidic axial water molecule in hemin than in FeOEP. The insensitivity of the hemin-acetate and hemin-dihydrogen phosphate bindings to the ionization of the propionic acid residues is consistent with the smaller electrostatic repulsion between two singly charged species compared to that between a singly charged fragment and a dianion. Apart from its effect on the 24 Fe(III)-HPO binding, replacement of the propionic acid residues by ethyl groups weakens the - 24 Fe(III)-H PO binding, which contrasts with the strengthening Fe(III)-acetate binding. The difference between these values and those obtained for hemin is statistically significant, as illustrated with the dotted green lines in the lower panel of Figure 14, that represent the theoretical predictions calculated by using the values of the hemin-anion binding constants. If ionization of the iron-coordinated water molecule is formally described as a substitution reaction of the water ligand by the hydroxyl anion, the decrease of the - 24 Fe(III)-H PO binding constant in FeOEP is consistent with the higher pKa value of its axial water, as both effects originate from a more strongly bound water molecule to the iron center of FeOEP. Then, the strengthening of the Fe(III)-acetate binding can be ascribed to a greater contribution of the hydrophobic effects brought about by the presence of alkyl substituents in the porphyrin ring. Overall, these findings reveal that anion binding to the aquo-ferric form of iron porphyrins is the result of the interplay between electrostatic and hydrophobic factors, whose relative contribution depends on the charge and hydrophobicity of the metalloporphyrin and the anionic ligand. CONCLUSIONS 39 It has been found that the proton-coupled electron transfer of two immobilized iron porphyrins onto a graphite electrode is accompanied by binding events with the commonly used phosphate and acetate buffers. These bindings involve the aquo-oxidized form of the iron porphyrin and the ionized forms of the buffer, so that they only operate in the narrow pH window where these species coexist, giving rise to an incipient intermediate plateau in the corresponding voltammetric titration curve. A two-step strategy has been developed to quantify both the apparent and intrinsic binding equilibrium constants with the minimum number of fitting parameters in each step. It has been shown that the apparent binding equilibrium constant has contributions from the pKa values of the buffer and the iron-bound water molecule, so that it may lead to a misleading perception of the true binding affinity. The obtained results show that the ionized forms of the acetate and phosphate buffers are able to displace the iron(III)-bound water molecule but not its ionized form (hydroxyl anion), and that the aquo-oxidized hemin has a greater affinity for the acetate anion than for the monohydrogen phosphate, though their apparent binding equilibrium constant suggested otherwise. Replacement of the hemin  substituents by ethyl groups decreases the acidity of the iron-bound water molecule, strengthens the iron(III)-acetate binding, weakens the iron(III)-dihydrogen phosphate binding, and enables its binding with the monohydrogen phosphate dianion, following the sequence of binding affinities: monohydrogen phosphate >> acetate > dihydrogen phosphate. Overall, this work provides a more complete interfacial speciation of immobilized iron-porphyrins than previously considered, showing the richer ligand exchange chemistry of the iron-bound water molecule in the oxidized state of the iron with respect to its reduced state. Particularly, the unnoticed binding of iron porphyrins with the commonly used 46 (63) Laviron, E. Theoretical Study of a 1 Electron, 1 Proton Surface Electrochemical Reaction (Four-Member Square Scheme) When the Protonation Reactions Are at Equilibrium. J. Electroanal. Chem. Interfacial Electrochem. 1980, 109, 57–67. (64) De Villiers, K. A.; Kaschula, C. H.; Egan, T. J.; Marques, H. M. Speciation and Structure of Ferriprotoporphyrin IX in Aqueous Solution: Spectroscopic and Diffusion Measurements Demonstrate Dimerization, but Not μ-Oxo Dimer Formation. J. Biol. Inorg. Chem. 2007, 12, 101–117. (65) Asher, C.; De Villiers, K. A.; Egan, T. J. Speciation of Ferriprotoporphyrin IX in Aqueous and Mixed Aqueous Solution Is Controlled by Solvent Identity, pH, and Salt Concentration. Inorg. Chem. 2009, 48, 7994–8003. (66) Martin, D. J.; Mercado, B. Q.; Mayer, J. M. Combining Scaling Relationships Overcomes Rate versus Overpotential Trade-Offs in O2 Molecular Electrocatalysis. Sci. Adv. 2020, 6, eaaz3318. (67) Jeon, S.; Bruice, T. C. Redox Chemistry of Water-Soluble Iron, Manganese, and Chromium Metalloporphyrins and Acid-Base Behavior of Their Lyate Axial Ligands in Aqueous Solution: Influence of Electronic Effects. Inorg. Chem. 1992, 31, 4843– 4848. (68) CRC Handbook of Chemistry and Physics, 97th ed.; Haynes, V. M., Ed.; CRC Press: Boca Raton, FL, 2017. (69) Savitskii, A. P.; Vorob’eva, E. V; Berezin, I. V; Ugarova, N. N. Acid-Base Properties of Protoporphyrin IX; Its Dimethyl Ester and Heme Solubilized on Surfactant Micelles: Spectrophotometric and Fluorometric Titration. J. Colloid Interface Sci. 1981, 84, 175–181. (70) Crespo, M. P.; Tilley, L.; Klonis, N. Solution Behavior of Hematin under Acidic Conditions and Implications for Its Interactions with Chloroquine. J. Biol. Inorg. Chem. 2010, 15, 1009–1022. 47 For Table of Contents Only Synopsis: Evidence of a pHand redox-dependent interaction of immobilized iron porphyrins with the commonly used acetate and phosphate buffers has been uncovered. This interaction involves the ligand-exchange binding between the oxidized aquo-complex of the metalloporphyrin and the anionic forms of the buffer. Recognition of these binding events is crucial for a rigorous quantification of the ubiquitous proton-coupled electron transfer in iron-porphyrins. 48 Supporting Information The Active Role of the Buffer in the Proton-Coupled Electron Transfer of Immobilized Iron Porphyrins Inmaculada Márquez, José Luis Olloqui-Sariego, Miguel Molero, Rafael Andreu, Emilio Roldán and Juan José Calvente* Departamento de Química Física, Universidad de Sevilla. C/Profesor García Conzález, 1. 41012 Sevilla. Spain 49 a) Voltammetry for a Proton-Coupled Electron Transfer with Two Ionizable Groups Let us consider the following cubic scheme to derive the voltammetric response of an immobilized redox couple with two ionizable groups. Because of thermodynamic restrictions in the different cycles, we only need to explicit seven out of the twelve thermodynamic parameters. The chosen parameters are shown on the solid arrows in Figure S1. o prot E stands the formal standard potential of the redox couple in its fully protonated state (HA1·Ox·A2H / HA1·Rd·A2H). 2 1 () , Ox A H a A H K and 2 1 () , Rd A H a A H K are the acid dissociation equilibrium constants of the A1H group in the oxidized and reduced forms of the electroactive species when the other ionizable group is in its protonated state (A2H). 1 2 () , Ox A H a A H K 1 2 () , Rd A H a A H K , 1 2 () , Ox A a A H K and 1 2 () , Rd A a A H K are the acid dissociation equilibrium constants of the A2H group in the oxidized and reduced forms of the redox Figure S1. Cubic scheme for the proton-coupled electron transfer of an immobilized electroactive species with one redox center and two ionizable groups. Vertical and horizontal arrows stand for the individual electron and protonation/deprotonation steps, respectively. Solid arrows represent the particular set of equilibria that has been used to derive the voltammetric response. o prot E 1 2 Ox( AH ) , A aH K HA1·Ox·A2H A1·Ox·A2HA1·Ox·A2 HA1·Ox·A2 1 2 AH Ox() , A a K 2 1 Ox( AH) , A aH K HA1·Rd·A2HHA1·Rd·A2 A1·Rd·A2 A1·Rd·A2H 1 2 Rd( AH ) , A aH K 2 1 Rd( AH) , A aH K 1 2 AH Rd() , A a K 50 couple when the first ionizable group is in its protonated (A1H) and ionized (A1) states, respectively. By applying the equilibrium condition to the electron transfer and protonation/deprotonation steps, it is obtained: 12 12 ·· ·· exp ( ) HA Ox A H o prot HA Rd A H nF EE RT          (S17) 1 2 1 2 22 11 1 2 1 2 · · · · ( ) ( ) ,, · · · · A Ox A H A Rd A H Ox A H Rd A H HH a A H a A H HA Ox A H HA Rd A H cc KK     (S18) 1 2 1 2 11 22 1 2 1 2 · · · · ( ) ( ) ,, · · · · HA Ox A HA Rd A Ox A H Rd A H HH a A H a A H HA Ox A H HA Rd A H cc KK     (S19) 1 2 1 2 11 22 1 2 1 2 · · · · ( ) ( ) ,, · · · · A Ox A A Rd A Ox A Rd A HH a A H a A H A Ox A H A Rd A H cc KK     (S20) where n is the number of exchanged electrons in the electron transfer step, F is the faradaic constant, R is the universal gas constant, T is the absolute temperature, j  stands for the surface concentration of the j species and H c for the proton bulk concentration. To complete the definition of the mathematical problem the redox mass balance condition is considered: 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 · · · · · · · · · · · · · · · · T rdox HA Ox A H A Ox A H HA Ox A A Ox A HA Rd A H A Rd A H HA Rd A A Rd A           (S21) where T rdox  stands for the total surface concentration of the electroactive species. By combining eqs S1 – S5, the following expressions are obtained for the partitioning of the electroactive species among their redox and acid/base states: 1 2 1 2 · · · · TT rdox rdox HA Ox A H HA Rd A H R O R O             (S22) 51 22 11 1 2 1 2 ( ) ( ) ,, · · · · Ox A H Rd A H TT a A H a A H rdox rdox A Ox A H A Rd A H R O R O HH KK cc              (S23) 11 22 1 2 1 2 ( ) ( ) ,, · · · · Ox A H Rd A H TT a A H a A H rdox rdox HA Ox A HA Rd A R O R O HH KK cc              (S24) 2 1 2 1 1 2 1 2 1 2 1 2 ( ) ( ) ( ) ( ) , , , , · · · · 22 Ox A H Ox A Rd A H Rd A TT a A H a A H a A H a A H rdox rdox A Ox A A Rd A R O R O HH K K K K cc              (S25) where O  and R  are defined by: 2 1 2 1 1 2 1 2 ( ) ( ) ( ) ( ) , , , , 2 1 Ox A H Ox A H Ox A H Ox A a A H a A H a A H a A H O HH K K K K cc       (S26) 2 1 2 1 1 2 1 2 ( ) ( ) ( ) ( ) , , , , 2 1 Rd A H Rd A H Rd A H Rd A a A H a A H a A H a A H R HH K K K K cc       (S27) The voltammetric current i is given by: 22 T rd d n F Av iRT d    (S28) where v is the potential scan rate and T rd  is the total surface concentration of the reduced forms of the redox couple, which is given by: 1 2 1 2 1 2 1 2 · · · · · · · · T rd HA Rd A H A Rd A H HA Rd A A Rd A       (S29) By substituting eqs S6 – S9 into eq S13 it is obtained: 1 ( ) T Trdox rd OR      (S30) By differentiating eq S14 with respect to  , and substituting the resulting expression into eq S12, the following expression is obtained for the voltammetric current:   22 2 () 1 ( ) T rdox O R OR n F Av iRT         (S31) 52 The corresponding expression for the voltammetric midpoint potential Em, equal to the peak potential, can be obtained by differentiating eq S15 with respect to  , and equating the resulting expression to zero, so that: 2 1 2 1 1 2 1 2 2 1 2 1 1 2 1 2 ( ) ( ) ( ) ( ) 2, , , , ( ) ( ) ( ) ( ) 2, , , , () ln () Ox A H Ox AH Ox A H Ox A a A H a A H a A H a A H oHH m prot Rd A H Rd A H Rd A H Rd A a A H a A H a A H a A H HH c K K c K K RT EE nF c K K c K K          (S32) For the particular case of independent ionization of the two acid groups (namely, 22 1 1 1 ( ) ( ) , , , Ox A H Ox A Ox a AH a A H a AH K K K , 11 2 2 2 ( ) ( ) , , , Ox A H Ox A Ox a A H a A H a A H K K K , 22 1 1 1 ( ) ( ) , , , Rd A H Rd A Rd a AH a A H a A H K K K and 11 2 2 2 ( ) ( ) , , , Rd A H Rd A Rd a A H a A H a A H K K K ), eq S16 reduces to: 1 2 1 2 1 2 1 2 2, , , , 2, , , , () ln () Ox Ox Ox Ox a A H a A H a A H a A H oHH m prot Rd Rd Rd Rd a A H a A H a A H a A H HH c K K c K K RT EE nF c K K c K K              (S33) so that for this particular case the pH dependence of Em is determined by four acid dissociation equilibrium constants. On the other hand, for the particular case of an electroactive species with only one ionizable group, eq S16 reduces to: , , ln Ox a AH oH m prot Rd a AH H cK RT EE nF c K      (S34) b) Voltammetry for a Ligand Binding-Coupled Electron Transfer Let us consider the square scheme depicted in Figure S2 to derive the voltammetric response of an immobilized redox couple that binds to a freely-diffusing ligand (L), without taking into account explicitly the acid/base speciation of the relevant species. o Ox/ Rd E and o Ox·L/ Rd·L E are the standard (formal) potentials of the unbound (Ox/Rd) and bound (Ox·L/Rd·L) redox couple, respectively, and KOx·L, KRd·L are the apparent binding 53 equilibrium constants of the oxidized and reduced forms of the redox couple with the freely-diffusing ligand, respectively. However, these parameters are not independent of each other since the following relationship is fulfilled in the cycle: · · / · / · ln oo Ox L Ox L Rd L Ox Rd Rd L K RT EE nF K    (S35) Thus, only three out of the four thermodynamic parameters are required to define the system. By applying the equilibrium condition to the electron transfer and binding steps depicted with solid arrows, it is obtained: / exp ( ) oOx Ox Rd Rd nF EE RT          (S36) ·· ·· Ox L Rd L Ox L Rd L Ox L Rd L KK cc    (S37) where L c stands for the ligand bulk concentration and the remaining parameters have been previously defined. By combining eqs S20 and S21 with the redox mass balance condition: Figure S2. Square scheme for the binding-coupled electron transfer of an immobilized electroactive species whose two redox forms bind to a freely-diffusing ligand L. Solid arrows represent the particular set of equilibria that has been used to derive its voltammetric response. / o Ox Rd E Ox·L K Rd·L Ox Rd Rd·L K Ox·L · / · o Ox L Rd L E 54 ·· T rdox Ox Ox L Rd Rd L       (S38) it is obtained: TT rdox rdox Ox Rd R O R O             (S39) · · · · TT rdox rdox Ox L Ox L L Rd L Rd L L R O R O K c K c             (S40) where O  and R  are defined as: ·· 1 1 O Ox L L R Rd L L K c K c      (S41) The expression for the voltammetric current can be obtained from eq S12, taking into account that T rd  is given by: ·1 ( ) T Trdox rd Rd Rd L OR           (S42) Thus, differentiating eq S26 with respect to  and substituting the resulting expression into eq S12, it is obtained:   22 2 () 1 ( ) T rdox O R OR n F Av iRT         (S43) Then, differentiating eq S27 with respect to  and equating the resulting expression to zero, the following expression is obtained for the voltammetric midpoint potential: · / · 1 ln 1 oOx L L m Ox Rd Rd L L Kc RT EE nF K c      (S44) c) Binding Between Immobilized and Freely-Diffusing Ionizable Species To quantify the intrinsic binding equilibrium constants between ionizable species from the corresponding overall apparent binding equilibrium constant, determined without 55 taking into account the ionization processes, we use the square scheme depicted in Figure S3 that combines binding and protonation/deprotonation steps of two ionizable species. For the sake generality, it is considered that the target molecule (Ox) possesses an ionizable group (H2O/OH) and that the ligand exists in four acid/base states (AH3/AH2/AH/A). It is also considered that binding implies the displacement of the host ionizable group by one of the ligand ionized forms. Because of thermodynamic restrictions in the corresponding cycles, only seven parameters are required to define the system at a thermodynamical level, namely: the acid dissociation equilibrium constants of both the target molecule ( 2 , Ox a H O K ) and the ligand ( ,1 L a K , ,2 L a K , ,3 L a K ), and the intrinsic binding Figure S3. Square scheme for the ionization and binding of an immobilized species and a freely-diffusing triprotic acid ligand. Ligand binding involves the displacement of a coordinating water molecule of the host. Solid arrows stand for the particular set of equilibria that has been used to derive the distribution of species among their acid/base and free/bound states. H+ 2 2AH· O x H O K Ox·OH2Ox·OH Ox·AH2 Ox·AH Ox·A 2Ax· OH H O K 2O Ox·A H K 2 x a,HO O K H2OAH2 H2OAH H2OA HO AH HO AH2 HO A a,2 KL AH3 a,1 KL a,3 KL AH2 AH A