ESI-MS Study of the Interaction of Potential Oxidovanadium(IV) Drugs and Amavadin with Model Proteins

In this study, the binding to lysozyme (Lyz) of four important VIV compounds with antidiabetic and/or anticancer activity, [VIVO(pic)2(H2O)], [VIVO(ma)2], [VIVO(dhp)2], and [VIVO(acac)2], where pic–, ma–, dhp–, and acac– are picolinate, maltolate, 1,2-dimethyl-3-hydroxy-4(1H)-pyridinonate, and acetylacetonate anions, and of the vanadium-containing natural product amavadin ([VIV(hidpa)2]2–, with hidpa3–N-hydroxyimino-2,2′-diisopropionate) was investigated by ElectroSpray Ionization-Mass Spectrometry (ESI-MS). Moreover, the interaction of [VIVO(pic)2(H2O)], chosen as a representative VIVO2+ complex, was examined with two additional proteins, myoglobin (Mb) and ubiquitin (Ub), to compare the data. The examined vanadium concentration was in the range 15–150 μM, i.e., very close to that found under physiological conditions. With pic–, dhp–, and hidpa3–, the formation of adducts n[VIVOL2]–Lyz or n[VIVL2]–Lyz is favored, while with ma– and acac– the species n[VIVOL]–Lyz are detected, with n dependent on the experimental VIV/protein ratio. The behavior of the systems with [VIVO(pic)2(H2O)] and Mb or Ub is very similar to that of Lyz. The results suggested that under physiological conditions, the moiety cis-VIVOL2 (L = pic–, dhp–) is bound by only one accessible side-chain protein residue that can be Asp, Glu, or His, while VIVOL+ (L = ma–, acac–) can interact with the two equatorial and axial sites. If the VIV complex is thermodynamically stable and does not have available coordination positions, such as amavadin, the protein cannot interact with it through the formation of coordination bonds and, in such cases, noncovalent interactions are predicted. The formation of the adducts is dependent on the thermodynamic stability and geometry in aqueous solution of the VIVO2+ complex and affects the transport, uptake, and mechanism of action of potential V drugs.


■ INTRODUCTION
Vanadium compounds (VCs) show a wide variety of pharmacological actions, among which are antiviral, antiparasitic, and antituberculosis effects, even though the most studied and promising application in medicine could be the treatment of diabetes and some types of cancer. 1 The study of the biospeciation of potential vanadium drugs is of fundamental importance to the prediction of (i) the form in which they are transported to the target organs, (ii) which is the active species in the organism, and (iii) the mechanism of action and binding to the biological receptors. In this context, the interaction of antidiabetic and anticarcinogenic VCs with the main components of the blood, including proteins, has been widely investigated. 2,3 In particular, proteins play a central role in the biospeciation and biotransformation of a VC in the organism, for both their high affinity toward V and high concentration in biological fluids.
In addition to X-ray diffraction, alternative techniques to obtain information on the metal−protein adducts are desirable; among them, NMR, EPR, ESEEM, ENDOR, UV− vis, and CD spectroscopy have been applied to systems containing V III , V IV , or V V . 4 More recently, other methods, such as voltammetry and polarography, HPLC-ICP-MS, sizeexclusion chromatography, gel-electrophoresis, and MALDI-TOF were used in combination with the other techniques. 3o,5 All these tools present some limitations: for example, NMR can be used mainly with V V , and EPRthe spectroscopy most frequently used for paramagnetic vanadium(IV) compounds 6 provides information on the type of residues involved in the metal binding but less on the stoichiometry of the formed adducts. Moreover, the relatively high vanadium concentration necessary to ensure a good signal-to-noise ratio in NMR and EPR measurements exceeds by several orders of magnitude that found in healthy humans (0.2−15 nM 1e,7 ) and in the patients treated with V drugs (1−10 μM for inorganic salts 1b,3a,8 and 40−60 μM when complexes such as bis-(maltolato)oxidovanadium(IV) (BMOV) or bis(ethylmaltolato)oxidovanadium(IV) (BEOV) are administered to rats 3m,9 ). Therefore, alternative techniques that are more sensitive at the physiological concentrations are needed.
Over the past decade, the potentiality of mass spectrometry (MS) to study the behavior of metallodrugs in biological samples and characterize at the molecular level their interaction with biomolecules and potential targets, such as proteins, has been discussed. 10, 11 Among the instrumental techniques based on MS, ESI (ElectroSpray Ionization) and MALDI (Matrix-Assisted Laser Desorption/Ionization) are very powerful methods to ascertain metallodrug−protein binding. While MALDI generates mainly singly charged pseudomolecular ions, ESI gives singly or multiply charged metal−protein adducts (upon deprotonation or association with protons and alkali ions). 10c, 11 For example, ESI-MS has been used to explore the interaction of cisplatin and its derivatives to low molecular mass bioligands, such as DNA nucleic bases, amino acids, oligonucleotides, and peptides, and high molecular mass biomolecules, such as proteins; 10a,c,12 moreover, the study of the reactivity of organometallic complexes of Ru II and Au III with potential antitumoral applications is another example of its potentiality. 10b,13, 14 An important advantage over other instrumental techniques is that the metal concentrations for ESI-MS experiments are in the range 1−100 μM, therefore very close to those found under physiological conditions. Among others, lysozyme, ubiquitin, myoglobin, superoxido dismutases, and insulin have been extensively studied by ESI-MS as model proteins; 11 they have structural features which make them ideal compounds for MS experiments: a small or moderate size (with molecular masses ranging from 5500 to 33000 Da) and high stability in solution under physiological-like conditions, commercial availability with high purity, water-solubility, and easy ionizability. Further ESI-MS studies have been carried out with potential molecular targets, such as metallothioneins, glutathione-S-transferases, cytochrome c, and calmodulin. 11 Model studies with the abovementioned proteins can be very useful, since the proteins involved in the transport and mechanism of action of a metallodrug (albumin and transferrin, for example) are not available in a purified form and have often high molecular masses. 10c Until now, the ESI technique was applied to inert metal complexes of the second and third transition series in which proteins bind with a coordination bond. 11 However, it has been pointed out that such a technique can also be used to study metal species which form labile coordinative bonds with biomolecules (such as vanadium, copper, and other metals of the first transition row) or to systems where noncovalent binding occurs (i.e., coordinatively saturated, such as Pt IV complexes). 11,15 Despite the high number of articles published on Pt, Ru, and Au complexes with antitumoral activity, 11 very few papers have been devoted to potential V drugs: beyond the "classical" applications of MS in the determination of the molecular mass of synthetic V complexes, the number of other studies is rather limited. One example is the ICP-MS investigation of the biospeciation of some antidiabetic V IV O 2+ compounds in the blood serum, 16  Preparation of the Solutions and ESI-MS Measurements. The solutions were prepared dissolving in ultrapure water (LC-MS grade, Sigma-Aldrich) a weighted amount of the VC to obtain a metal ion concentration of (1.0−2.0) × 10 −3 M. They were subsequently diluted in ultrapure water and mixed with aliquots of a stock protein solution (500 μM) in order to have a metal-to-protein molar ratio of 3:1, 5:1, or 10:1 with a final protein concentration of 5 or 50 μM. In all the solutions, pH was in the range 5.0−6.0. Argon was bubbled through the solutions to ensure the absence of oxygen and avoid the oxidation of the V IV O 2+ ion. ESI-MS spectra were recorded immediately after the solution preparation.
Mass spectra in positive-and negative-ion mode were obtained on a Q Exactive Plus Hybrid Quadrupole-Orbitrap (Thermo Fisher Scientific) mass spectrometer. The solutions were infused at a flow rate of 5.00 μL/min into the ESI chamber. The spectra were recorded in the m/z range 50−750 for the binary and 300−4500 for the ternary systems at a resolution of 140 000 and accumulated for at least 5 min to increase the signal-to-noise ratio. Generally the experiments were run more than once; when measurements were repeated, indistinguishable results were obtained. Only one set of the reproducible data is reported here.
The experimental settings for the measurements of positive-ion mode spectra were: spray voltage, 2300 V; capillary temperature, 250°C ; sheath gas, 10 (arbitrary units); auxiliary gas, 3 (arbitrary units); sweep gas, 0 (arbitrary units); probe heater temperature, 50°C. The settings used for negative-ion mode spectra were: spray voltage, −1900 V; capillary temperature, 250°C; sheath gas, 20 (arbitrary units); auxiliary gas, 5 (arbitrary units); sweep gas, 0 (arbitrary units); probe heater temperature, 14°C. ESI-MS spectra were analyzed with Thermo Xcalibur 3.0.63 software (Thermo Fisher Scientific), and the average deconvoluted monoisotopic masses were obtained through the Xtract tool integrated in the software. Lysozyme represents one of the most frequently used models to study the metallodrug−protein interactions by mass spectrometry techniques. Lyz is a relatively small protein (with mass of ca. 14300 Da) formed of 129 amino acids with an enzymatic activity and a significant relevance in a number of host defense processes. 19 The specific structure of this enzyme  Inorganic Chemistry pubs.acs.org/IC Article relies on its stable three-dimensional conformation associated with the presence of four disulfide bonds between Cys residues alongside the peptide chain. It has only one histidine in its structure (His15), which could represent a general binding site for some transition metal complexes. 11 The positive-ion mode ESI-MS spectrum of lysozyme shows a series of well-resolved peaks with a charge distribution from z = 7 to z = 12, which fall in the m/z range of 1100−2100 (Figure 1). 20 To determine the exact mass of the protein, the spectrum was deconvoluted with the Xtract software, which allows an estimation of the mass of macromolecules from the mass-to-charge spectra consisting of several peaks for multiply charged ions. 21 The result is reported in Figure 2, where a central intense peak at 14305 Da can be noticed with a series of other signals due to the adducts containing Na + ions (23 Da) and/or H 2 O molecules (18 Da). Each peak is also split in several signals related to the isotopic distribution of the revealed species.
In order to interpret the spectra in the system containing [V IV O(pic) 2 (H 2 O)] and lysozyme, it must be considered that  Figure S1 of the Supporting Information. The spectrum shows the presence of two intense peaks at m/z = 124.04 and m/z = 146.02 attributed to the protonated ligand, [Hpic+H] + , and its sodium adduct, [Hpic+Na] + . It is also possible to observe the signals of the H + and Na + adducts of [V IV O(pic) 2 ], at m/z 311.99 and 333.98, respectively, confirming the presence in solution, under these experimental conditions, of the 1:2 V IV O−picolinato species; the equatorial H 2 O is not revealed, in agreement with the literature data suggesting that a weak monodentate ligand can be removed from the metal coordination sphere during the ionization process. 23,24 The comparison between the experimental and calculated isotopic pattern for the peak of the [V IV O-(pic) 2 +H] + ion is shown as an example in Figure S2; in  Inorganic Chemistry pubs.acs.org/IC Article particular, it must be observed the coincidence of the two expected peaks due to the natural abundance of the 13 C isotope (separated by m/z ∼ 1.00 for this adduct with charge +1) at the third decimal figure. In the negative-ion mode ESI spectrum, the signal of the complex in which the water molecule is deprotonated,  25 This oxidation process is probably favored by the similarity of the structure of cis-V IV O(OH) − and cis-V V O 2 + (Scheme S1). The species identified by ESI-MS spectrometry in the system with picolinate are listed in Table S1.
The ESI-MS spectra on the system containing [V IV O-(pic) 2 (H 2 O)] and Lyz were recorded using two concentrations of the protein (5 and 50 μM) and metal-to-protein ratios (3:1 and 5:1). The results indicate that these two variables significantly influence the outcome of the experiments. In Figure 3, the raw spectra obtained with a protein concentration of 5 μM are represented. It can be noticed that, for each peak of the protein, other signals are present with higher m/z values, which correspond to the adducts n[V IV O(pic) 2 ]−Lyz. Since the arrangement of the two ligands is equatorial−equatorial and equatorial−axial, only one equatorial coordination site is available for the protein. The comparison of the raw spectra recorded in the system [V IV O(pic) 2 (H 2 O)]/Lyz with those obtained in the system with lysozyme only (cfr. Figures 3 and  1) suggests that the V IV O(pic) 2 binding does not alter the distribution of the protein charge states; this indicates that the conformation of lysozyme remains unchanged after the interaction with the metal moiety. 26 To determine the mass of these adducts, the ESI-MS spectra were deconvoluted, and the results are reported in Figure 4. The spectra are similar, and in addition to the peak of the free protein, the peaks at 14 616 and 14 926 Da (for the 3:1 ratio) and also at 15 238 Da  (Table 1). Therefore, an increase of the      Table 1). Concerning the interaction between [V IV O(acac) 2 ] with  25 The comparison between the experimental and calculated isotopic pattern confirms the presence of this species ( Figure S9).
The ESI-MS spectrum of the system [V IV (hidpa) 2 ] 2− /Lyz 5:1 was recorded ( Figure 9) and, together with the signal of the free protein, only a peak attributable to the adduct [V(hidpa) 2 ]−Lyz was observed (14706.8 Da), whose intensity increases with the vanadium concentration. Considering that several protonation states exist for the protein, it is not possible to determine if the oxidation state of vanadium in the adduct is +IV or +V. On the basis of the studies in the literature, one would expect that it is the V IV complex, because the oxidation to V V is not favored. 17b Interaction of [V IV O(pic) 2 (H 2 O)] with Myoglobin. Myoglobin (Mb) is a small protein, found in skeletal muscles and in the heart, where it stores molecular oxygen. It constitutes up to 5−10% of all the cytoplasmic proteins found in these muscle cells and consists of a polypeptide of 153 amino acid residues and a heme group. 33,34 Under physiological conditions, 70% of the Mb backbone is folded into eight alpha helices (named A−H), the folding leading to a tight globular structure with a cleft for the heme group at helices C, E, and F. 35 The ESI-MS spectrum of myoglobin dissolved in ultrapure water ( Figure S10) shows a series of peaks which correspond to different charge states (from +6 to +13). In the deconvoluted spectrum, two series of peaks can be recognized with an experimental mass of 16 951 and 17 566 Da, respectively, and a predominance of the second one. The mass difference (615 Da) suggests that they belong to the apo (without heme) and holo forms of the protein, with the latter maintaining the binding with the heme group. It is known that in myoglobin heme is bound to the protein chain by the covalent binding with the proximal histidine (His93) and that the stability of this coordinative bond may be heavily affected    Figure 10, it can be noticed that the most intense peak at 17 566 Da is revealed with a series of other signals due to the adducts with a certain number of Na + ions (23 Da). The presence of the free heme group was confirmed by the detection of a signal at m/z 616.2 in the raw spectrum of the protein. In Figure S11, a comparison between the experimental and calculated isotopic patterns attributable to [Fe III heme] + is shown; in agreement with the results in the literature, the oxidation state of iron in the heme group released from Mb is +III. 37 The interaction of [V IV O(pic) 2 (H 2 O)] with myoglobin was studied by ESI-MS with the same ratio and concentration used for the previous measurements with lysozyme. In the raw spectrum, the signals at m/z values higher than the peaks of free protein can be attributed to  (Table 2). This indicates that, under these experimental conditions, only two or three vanadium species are bound to the protein.
If the spectra are recorded with a protein concentration of 50 μM (Figure 11), the number of V IV O(pic) 2 fragments bound to the protein raises up to 4 when the 3:1 ratio is used and up to 5 with a 5:1 ratio, and a general increase of the relative abundance of the metal−protein adducts is revealed. It can be noticed that at 50 μM the peaks of {[V IV O(pic)] + n[V IV O(pic) 2 ]}−Mb are also observed, even though with an intensity lower than that of [V IV O(pic) 2 ]−Mb (cfr. Figures 11  and S13).

Interaction of [V IV O(pic) 2 (H 2 O)] with Ubiquitin.
Ubiquitin is a small regulatory protein constituted of a polypeptide chain of 76 amino acid residues with a molecular weight of 8.5 kDa. It contains a limited number of potential binding sites for metals, among them the N-terminal methionine (Met1), one histidine residue (His68), and some carboxylate O-donor groups. 10c Ub is involved in the post-transductional signal called ubiquination, which has a regulatory role in the cellular processes. Besides its functions in biological systems, ubiquitin is used as a model protein in mass spectrometry since it is commercially available with high purity, well characterized, and allows to obtain accurate data. 10c The interaction with some metal compounds with antitumoral activity (such as cisplatin and Ru-based complexes) was already studied by MS. 12c, 38 The ESI-MS spectrum of the free protein dissolved in ultrapure water at a concentration of 5 μM ( Figure S14) was recorded in this study and shows a series of peaks corresponding to the different charge states of ubiquitin (from +5 to +10). The deconvoluted spectrum reported in Figure 12 shows that the mass of the protein is 8564.6 Da, in accordance with the literature data. 12c,38 In addition to the major peak, a series of other signals due to the adducts with ubiquitous ions, such as Na + (23 Da) and K + (39 Da), are present.
With ubiquitin, the isotopic pattern of its charged states was simulated, allowing us to determine the exact formula. As an example, in Figure 13 the comparison of the experimental and calculated isotopic pattern for the peak at m/z 952.63 with z = 9 is reported. The empirical formula of the neutral protein is C 378 H 630 N 105 O 118 S, in agreement with the previous studies. 39 The ESI-MS spectra were recorded on the system [V IV O(pic) 2 (H 2 O)]/Ub with different protein concentrations (5 and 50 μM) and metal-to-protein molar ratios (3:1, 5:1, and 10:1). When the signals in the deconvoluted spectra are analyzed, it can be noticed that, in comparison with the free protein (with peak centered at 8564.6 Da, see Figure 12 To confirm the formation of the adducts, the simulation of the experimental isotopic pattern was carried out for the species formed by V IV O(pic) 2 with Ub obtaining an excellent agreement with the experimental data ( Figure S15) Finally, in the spectra recorded with an ubiquitin concentration of 50 μM, the maximum number of fragments bound to the protein increases up to 2 with the ratio 3:1 and up to 3 with the ratio 5:1 ( Figure S16). Moreover, at a ratio 5:1 also the adduct {[V IV O(pic)] + [V IV O(pic) 2 ]}−Ub, not observed with the same ratio when Ub concentration is 5 μM, is detected ( Figure S16). Therefore, according to the data in the literature, 36a it can be concluded that with increasing the concentration of the species, the VC−protein interactions  Table 2.
Rationalization of the ESI-MS Data. Analyzing the data presented in the previous sections, it can be observed that depending on the molar ratio, vanadium concentration, and type of V IV complexthe number and identity of the adducts revealed change significantly.
An important difference between the four systems containing V IV O 2+ complexes and lysozyme is that for Hpic and Hdhp the formation of adducts with the 1:2 complex, The results are summarized in Table 3. As it can be noticed, at 15 μM, the formation of V IV OL 2 species is favored with Hpic and Hdhp, while with maltolate and acetylacetonate, V IV OL + is the predominant complex in solution. Therefore, in the first     (Table 3). In the overall examination of the data, it must be taken into account that the metal ion can be sequestered (that is completely removed from the ligand L − ) when the protein affinity for V IV O 2+ is very strong: a clear example is transferrin, which in the apo form has two iron binding sites which can accommodate hard metal ions like the oxidovanadium(IV) ion; 2a,d,3b,e,j,m metal ion abstraction could alter the species distribution of V IV O 2+ and L − . In contrast, when strong sites are not available (as in the case of the proteins examined in this work), the preferential interaction is with V IV OL + or V IV OL 2 moieties and, for this reason, the data obtained by ESI-MS can be used to evaluate the interaction of the metal species with proteins and stability of the adducts. The validity of the discussion can be tested analyzing the low m/z regions of the raw ESI-MS spectra. This is done for the system V IV O(ma) 2 /Lyz and is shown in Figure 15   In the other systems, a similar situation is observed. Notably, the ratio between the signal intensity of the 1:2 complexes to that of the 1:1 ones increases from acac − to pic − and to dhp − , in line with what is expected on the basis of the thermodynamic data (Table 3). The regions with a low m/z for these systems are reported in Figure S21 Pharmacological and Biological Implications. The data obtained allow us to advance some hypotheses on the metal−protein adducts formed by V IV O 2+ complexes with concentrations used in this study (between 15 and 150 μM, which is close to that found under the physiological conditions 3m ). The compounds can be divided into two classes: the first class comprises VCs formed by strong ligands (for example, picolinate and 1,2-dimethyl-3-hydroxy-4(1H)pyridinonate) which give adducts with a composition of In general, for Hpic and Hdhp adducts, the protein binds with an accessible protein residue indicated with an X in Scheme 2 in the fourth equatorial position. Specifically, the structure of the adduct between cis-V IV O(pic) 2 and Lyz has been determined by X-ray crystallography and computational methods, and the Asp52 residue of lysozyme was found to replace the equatorial water ligand of V IV O(pic) 2 (H 2 O) forming a distorted octahedral complex. 3h,42 Glu and His are other residues able to bind V in a monodentate manner. Particularly, accessible residues for the coordination to a cis-V IV OL 2 moiety are Glu35 with Lyz, 42 Glu16, Glu18, Asp21, and His68 for Ub, 43 while for Mb they are mainly histidines (His81, His113, His116, His199). 44  With ma − and acac − , the adducts formed have the general formula n[V IV OL]−protein. A considerable number of adducts are possible and should be considered. Since there are two free cis equatorial positions (plus the axial site) in the V IV OL + moiety, the protein binding mode could be mono-, bi-, or tridentate (X, Y, plus a further axial Z, Scheme 2). The bidentate or tridentate binding mode of the protein depends on the protein conformation and on the eventual presence of two or three neighboring residues able to occupy contemporaneously three facial positions of the octahedral vanadium geometry. This eventuality is rare in small proteins such as lysozyme but could be more common in larger proteins such as transferrin or albumin. The candidate residues able to coordinate the V center are Asp, Glu, and His, with the possible assistance of Ser, Thr, Tyr, Cys, or backbone-CO if they are in the right position. With the acac − ligand, adducts If the VC is very stable, such as amavadin, and no coordination sites in the complex are available, only noncovalent interactions could occur (Scheme 2). The strength of these interactions depends on the number of polar groups in the VC able to form hydrogen bonds, van der Waals, or hydrophobic contacts with the exposed polar groups on the protein surface. 45 These studies convincingly show that, if a VC reaches the blood, a significant amount of metal−protein adducts exist and the type of interaction can influence its fate in the organism and favor the cellular uptake and binding to the cellular receptor(s). For example, the adducts with transferrin, with the binding of V IV O 2+ to the active sites left free by Fe 3+ or to the accessible surface donors, can promote a conformational change which would lead to recognition by the cellular receptors in the endocytosis. 46 Moreover, the binding to the cellular targets can lead to pharmacological activity, as in the inhibition of phosphatases, to which the vanadium antidiabetic activity is attributed. 47 In this context, the results obtained with the ESI-MS technique could be considered in the design of new potential vanadium therapeutics. First of all, ESI-MS allows the study of biospeciation at the metal concentrations found in the patients treated with potential drugs. Second, the redox reactions in the biological media containing proteins of the three oxidation states, V III , V IV O, and V V O 2 , both in the free form and in the generated adducts, can be followed with the aim to determine the most stable state. Third, ESI-MS can suggest if the metal species exists in the free form or bound to the protein, and which equilibriums in solution are established and which adducts are formed (see Figure 15). Depending on the molecular receptor(s), the features of the organic ligand L − could be modulated to favor formation at physiological concentrations of moieties VOL + , that can form stable species with proteins through the binding of two amino acid donors, or VOL 2 , thatin contrastyields moderately stable adducts if it is in the cis arrangement or unstable if the interaction occurs in the axial position (Scheme 2).
However, all these considerations assume that the VCs remain intact after administration and do not address absorption, distribution, metabolism, and excretion (ADME) and, particularly, the events that occur before the drug enters the bloodstream. The possibility and likelihood that the complex survives the biological media and the uptake mechanisms is a question that is not addressed in this study. If the VC is administered orally, it will need to travel through the digestive tract to first reach the acidic pH of the stomach When more than one adduct is formed in aqueous solution, it is indicated with minor or major species. In round parentheses, the less probable binding of an axial residue Z is denoted. b The charges of the ligands and V=O ion (2+) are omitted for clarity.
Inorganic Chemistry pubs.acs.org/IC Article before continuing traveling into the intestine and being absorbed into the blood. If the VCs are delivered intravenously, then they reach the blood immediately, however the presence of other blood components and salts may decrease their concentration and, hence, reduce the chance of adduct formation. Alternative modes of administration exist, and the question of how much compound reaches the blood intact is not known, particularly if the complexes will need to pass any membranes in the process. Unfortunately, studies done after administration of the VCs to animals rarely have investigated the blood content of vanadium, and if this amount was determined, no speciation studies have been carried out. We are therefore left conducting studies on model systems similar to those described in this work. ) 3 ] − may occur, with the possible further oxidation to V V inorganic anions. Two important findings are: (i) the number of metal fragments bound to proteins increases with the VC concentration and (ii) the oxidation to V V is possible in the binary systems V IV O 2+ −L, but it is prevented or significantly slowed down after the formation of the adducts VC−protein. In fact, in the studied systems, there is no evidence of oxidation of V IV to V V . The results could allow the prediction of the fate of an administered V IV OL 2 potential drug. If the complex is stable, the proteins can coordinate the V IV OL 2 moiety only in a monodentate manner, remembering that the interaction and its strength are related to the geometry assumed in solution, cisoctahedral with a free equatorial position (moderately strong binding) or square pyramidal with an available axial site (weak coordinative or noncovalent interaction). If the complex has an intermediate stability, the proteins can interact with the VOL + fragment in a bi-or tridentate fashion, generating stable adducts with the contemporaneous binding of two amino acid side-chains in two adjacent equatorial positions. Finally, when the V IV complex is very stable without available coordination sites, as happens for amavadin, the proteins are not able to interact covalently and only weak noncovalent contacts are predicted. Since the formed adducts have different thermodynamic stability, it is clear that, depending on the properties of the ligand L, a pharmacologically active V IV O 2+ complex would reach the target organs in different forms, each of them being characterized by a different cellular uptake and biological activity. Thus, in relation to the pharmacological activity, not only the solubility and hydrophobicity of the VCs are important but also their ability to interact with proteins, which determines the adducts existing under physiological conditions. Overall, the results further support the fact that the biological properties of VCs must be related with their biospeciation in the organism and that a mixture of species (intact VC, adducts with proteins and/or low molecular mass bioligand, inorganic and hydrolytic ions, this latter eventually generated by redox reactions) could contribute to the pharmacological action, 48 with the active species depending both on the thermodynamic stability and structural requirement of the administered VC and, eventually, on its redox behavior.
Tables with identified species in the binary systems (Tables S1−S4)