O-phosphorylethanolamine and O-phosphorylcholine chloride solutions were prepared by weighing and subsequent dissolution of the corresponding products (Sigma-Aldrich/Merck, Darmstadt, Germany). The purity of the ligands, determined by alkalimetric titration, was greater than 99%. Magnesium chloride solutions were prepared by weighing and dissolving the Fluka (Fluka/Honeywell, Charlotte, North Carolina, United States) product. These solutions were standardized using the EDTA (Ehylenediaminetetraacetic acid disodium salt, BioUltra, ≥99%) standard.

Sodium chloride solutions were obtained by weighing the salt (puriss., Sigma-Aldrich/Merck, Darmstadt, Germany), after drying at 110°C. The solutions of hydrochloric acid and sodium hydroxide were obtained by diluting the Fluka (Fluka/Honeywell, Charlotte, North Carolina, United States) vials. Subsequently, they were standardized by titrations using sodium carbonate (≥99.5%, Sigma-Aldrich/Merck, Darmstadt, Germany) and potassium acid phthalate (≥99.5%, Sigma-Aldrich Merck, Darmstadt, Germany), respectively. These salts were previously dried in an oven at 110°C.

Two distinct potentiometric systems were used for titrations. Each system has an identical configuration with a Metrohm model 809 Titrando potentiometer, an automatic dispenser Metrohm Dosino 800, and a Metrohm LL-Unitrode WOC combined glass electrode. A PC was connected to each potentiometric system to acquire experimental titration data by Metrohm TIAMO 2.2 software. Several parameters, such as the titrant delivery and e.m.f. stability, were controlled by this software. Estimated accuracies of these systems are ±0.15 mV and ±0.002 ml for e.m.f. and for titrant volumes, respectively.

For the ligand protonation, for each titration, volumes of the NaOH standard were added to 25 ml of the solution containing PEA or PPC at C_L = 5–10 mmol L⁻¹, 0.15 ≤ I/mol L⁻¹ ≤ 1 in NaCl at t = 25°C, and I = 0.15 mol L⁻¹ at t = 15, 37°C. For metal–ligand complexes, 25 ml of the solution containing Mg²⁺ and PEA or PPC at C_M = 1–4 mmol L⁻¹, C_L = 2–4 mmol L⁻¹, C_M/C_L = 0.33-2, and I = 0.15 mol L⁻¹ in NaCl was titrated by using the NaOH standard at t = 15, 25, and 37°C. All the solutions during the titrations were in glass jacket thermostated cells, under magnetic stirring and by bubbling pure N₂. Independent titrations of HCl with standard NaOH were performed to obtain the values of the standard electrode potential E⁰ and pK_w, under the same ionic strength and temperature conditions of the corresponding measurement.

A Varian NMR spectrometer 500 Mhz was used to process ¹H-NMR spectra. 1,4-dioxane was used as the internal reference (δ _CHdioxane = 3.70 ppm), and all chemical shifts refer to tetramethylsilane (TMS). All measurements were carried out using the presaturation technique to reduce the water signal, in 9:1 H₂O/D₂O solution at t = 25°C. The spectra containing ligands PEA or PPC (at C_L = 7 mmol L⁻¹) and NaCl (I = 0.15 mol L⁻¹) solutions were recorded in a pH range between 2 and 11. The spectra containing PEA or PPC and Mg²⁺ (C_M = 6 mmol L⁻¹, C_L = 7 mmol L⁻¹) and NaCl solutions were recorded in the same pH range of the free ligands.

All metal complexes were prepared, as published elsewhere (Chillè et al., 2020; Aiello et al., 2021a). Briefly, all ligands (1 or 2 mmol) were dissolved in 100 μl of water; the pH was adjusted to 8 with NaOH and then added to an aqueous solution (200 μL) of MgCl₂ (1 mmol). The resulting solutions were maintained under magnetic stirring, at room temperature for 2 h. MALDI mass spectrometry analysis was performed on a 1 μl portion of a premixed solution containing the reaction mixture and the matrix α-CHCA (0.3% in TFA), in a 2:10 (v:v) ratio.

All MS and MS/MS experiments were performed, as published elsewhere (Aiello et al., 2018; Imbrogno et al., 2019). All experiments were conducted using a 5800 MALDI-TOF/TOF analyzer (AB-SCIEX), supplied with a neodymium–yttrium–aluminum–garnet laser, operating at 349 nm. MS spectra were obtained with a mass accuracy of 5 ppm, by collecting 4,000 laser shots, applying a laser pulse rate of 400 Hz. A total of 5,000 laser shots, at a pulse rate of 1000 Hz and 1 kV of collision energy, were collected and averaged for each MS/MS experiment. Δppm of the MS/MS experiments was 20 ppm. MS/MS experiments were achieved using ambient air as the collision gas (10^–6 Torr). Data Explorer (version 4.0) was used for handling all spectra.

The STACO and BSTAC programs were employed to process the experimental potentiometric data. With their use, the protonation constants of the ligands, the formation constants of the complexes, and the parameters of the acid–base titration (the standard potential E⁰, junction potential, and analytical concentration of the reagents) were obtained. The LIANA program was used in processing experimental results at various ionic strengths and temperatures to obtain the dependence of protonation and formation constants on ionic strength and temperature. More information about BSTAC, STACO, and LIANA can be found in the reference (De Stefano et al., 1997). The speciation diagrams and the percentages of complex species were obtained using the HySS program (Alderighi et al., 1999). HypNMR software was used to process the observed experimental signals, assuming a fast mutual exchange in the NMR time scale (Frassineti et al., 1995). With its use, the protonation constants of PEA and PPC, the formation constants of the complex species, and the individual chemical shifts of each species were calculated.

The protonation constants of the two ligands under study, PEA and PPC, necessary for the subsequent determination of the complexes with Mg²⁺, were determined. The protonation reactions as overall formation constants (β) and stepwise formation constants (K) are as follows, where the charges are omitted for simplicity: L + rH  =  LH r                                       β LH r , (1) LH r - 1   +  H  =  LH r                                       K LH r . (2) Protonation constant values obtained via potentiometric titrations under different temperature and ionic strength conditions are summarized in Table 1. The calculated values referred to PEA at I = 0.15 mol L⁻¹ and t = 25°C (logK ^LH = 10.141, logK ^LH2 = 5.590) are similar to those reported by Mohan et al., logK ^LH = 10.12, logK ^LH2 = 5.52 (at I = 0.2 mol L⁻¹, t = 25°C in KNO₃) (Mohan and Abbott, 1978a; Mohan and Abbott, 1978b). In a very recent study, the protonation constants of PEA at t = 20°C and I = 0.1 mol L⁻¹ in KNO₃ were proposed (logK ^LH = 10.41, logK ^LH2 = 5.70) (Gabryel-Skrodzka et al., 2021). It is not possible to make other comparisons at other temperatures or ionic strengths since in the literature, there are only data up to 0.2 mol L⁻¹ and t = 20 or 25°C (Datta and Grzybowski, 1959; Wozniak and Nowogrocki, 1979; May and Murray, 2001; Pettit and Powell, 2001; Martell et al., 2004). As far as we know, in the literature, there are no thermodynamic parameters on the protonation of PPC.

The species with Mg²⁺ were subsequently investigated. Both the protonation constants reported in this study and the hydrolysis constant of Mg²⁺, reported in the Supplementary Table S1 under various conditions, were considered. Potentiometric experimental titrations were carried out at different metal/ligand ratios and different concentrations, to select the most reliable speciation model and to obtain the formation constants of the complex species. These Mg²⁺(M)-ligand(L) formation constants are indicated as overall formation constants (β) or stepwise formation constants (K), based on the following reactions, where the charges are omitted for simplicity: M + L + rH = MLH r                                         β MLH r , (3) M + LH r = MLH r                                         K MLH r . (4) The choice of the speciation model that best reflects the system under study is made considering some requirements such as its simplicity, goodness of statistical parameters (standard and mean deviations referring to the fit), percentages of formation of complex species, variance ratio between the chosen model, and others (Filella and May, 2005).

The obtained results, in terms of formation constants of Mg²⁺-PEA and Mg²⁺-PPC species at I = 0.15 mol L⁻¹ in NaCl and t = 15, 25, and 37°C, are reported in Table 1. The speciation models include three species for the Mg²⁺-PEA system, namely, MLH₂, MLH, and ML and two species for the Mg²⁺-PPC system, namely, MLH and ML. Mass spectrometry measurements will also highlight the formation of ML₂ species. Despite the excess ligand employed in the experimental potentiometric conditions (M:L = 1:3), the formation percentage of ML₂ species was negligible for both ligands. Therefore, this species was not considered in the speciation models. The speciation diagrams of the systems containing PEA, PPC, Mg²⁺-PEA, and Mg²⁺-PPC are shown in Figures 2A–D. Under physiological conditions (pH = 7.4, t = 37°C, and I = 0.15 mol L⁻¹), considering PEA at C_L = 5 mmol L⁻¹, formation percentages of L, LH, and LH₂ species are 0.3, 97.4, and 2.2, respectively. Under the same conditions, considering PPC, formation percentages of L and LH species are 98.3 and 1.7, respectively (Figure 2B). In the presence of Mg²⁺, in the PEA system, MLH species achieves a significant formation percentage of 18.6 (Figure 2C); in the PPC system, both MLH and ML species achieve significant formation percentages equal to 15.3 and 23.7, respectively (Figure 2D). More in detail, Figure 2C—referring to the Mg²⁺-PEA system—shows that in the acid pH range, the MLH₂ species is formed, reaching percentages of up to 10%. The main species is MLH with 20% in the pH range 6.5–9.0. ML species is formed at pH > 9. With regard to the Mg²⁺-PPC system, shown in Figure 2D, the observed complex species are much higher than those of PEA under the same conditions. MLH species exceeds 40% at pH = 2–4. ML species reaches almost 40% at pH = 7–10. In the literature, fairly close to our results of formation constants of Mg-PEA species at I = 0.15 mol L⁻¹ and t = 25°C were reported by Hendrickson et al., logK ^ML = 2.20 and logK ^MLH = 1.48 (at I = 0.1 mol L⁻¹ in (C₃H₇)₄NI) and t = 20°C) (Hendrickson and Fullington, 1965). Other values were obtained by Mohan et al., logK ^ML = 1.56 and logK ^MLH = 1.17 (at I = 0.2 mol L⁻¹ in KNO₃ and t = 25°C) (Mohan and Abbott, 1978a; Mohan and Abbott, 1978b) and by Osterberg, logK ^ML = 1.70 and logK ^MLH = 1.23 (at I = 0.15 mol L⁻¹ in KCl and t = 25°C) (Osterberg, 1960). It was not possible to compare the results reported in this study on Mg²⁺-PPC species with those in the literature since as far as we know, the speciation patterns and the formation constants were not reported up to now.

¹H-NMR titrations were also carried out at I = 0.15 mol L⁻¹ and t = 25°C both for the determination of the protonation constants of the ligands under study as well as of the complexes with Mg²⁺, as already reported on other systems (Cardiano et al., 2011a; Cardiano et al., 2017). The Chemical shift and pattern of protons of the PEA solution are shown in Supplementary Figure S1 at different pH values. Some noticeable signals can be identified for the two types of PEA protons: a doublet triplet (td) assignable to the CH₂ protons in position 1 showing Δδ = 0.28 ppm from pH 1.67 to pH 11.00 and a triplet (t) assignable to the CH₂ protons in position 2 showing Δδ = 0.43 ppm from pH 1.67 to pH 11.00. The solutions of PEA with Mg²⁺ were analyzed in the same pH range by the NMR technique, and the proton spectra are shown in Figure 3. The data collected indicate that the chemical shift values have the same trend as the values recorded for the free ligand, as evidenced in the graph showing small differences (Supplementary Figure S2). Therefore, it can be assumed that the interaction between the metal and ligand occurs from the phosphoric moiety or by the electrostatic interaction of the negative oxygen atoms and the magnesium cation.

The chemical shift and pattern of protons of the PPC solution are shown in Supplementary Figure S3 at different pH values. Some noticeable chemical shifts can be identified for the three types of PPC protons: a broad multiplet (m) assignable to the CH₂ protons in position 1 showing Δδ = 0.14 ppm from pH 1.90 to pH 10.00, a triplet (t) assignable to the CH₂ protons in position 2 showing Δδ = 0.13 ppm from pH 1.90 to pH 10.00, and a singlet of the nine methyl protons at 3.16 ppm, which does not have appreciable variations, in the pH range studied. The solutions of PPC with Mg²⁺ were analyzed by the NMR technique, and the proton spectra are shown in Supplementary Figure S4. Also, for PPC solutions, the chemical shift values of the free ligand and the ligand with magnesium are similar, evidencing small differences, as shown in the graph (Supplementary Figure S5). A similar assumption on PEA and PPC could also interact through the phosphate group oxygen, and this would justify the lack of chemical shift of the protons on the aliphatic chain of the ligand in the presence of magnesium ions. The results listed in Table 2, together with those obtained by potentiometry, were obtained from the processing of the measured chemical shifts. These results confirmed that the comparison between the results obtained with the two techniques, under the same experimental conditions, shows an excellent agreement both for the speciation model determined by potentiometry and for the values of the formation constants of the complexes, especially for PPC. Figure 4 highlights the excellent agreement between the experimental and the calculated chemical shift values over all pH ranges considered and therefore in the formation areas of the different species.

One of the aims of this investigation is to be able to use the acquired thermodynamic information making simulations under real fluid conditions, such as biological ones. Just as an example, the composition of two biological fluids was considered with the aim of evaluating the significance of the complex species under study for the purposes of a characterization of the fluid itself. Among the biological fluids, the cerebrospinal fluid (CSF) and the extracellular fluid in the brain intracellular space were considered. The CSF is formed in the brain. It is an aqueous solution containing higher contents of magnesium, sodium, and chloride and lower concentrations of potassium, calcium, bicarbonate, and phosphate, with respect to the plasma in humans (Artru, 2010). The latter in the brain plays a role in multiple key functions, including non-synaptic neurotransmission. The extracellular space makes up about 15% of the total brain volume and is filled with an extracellular fluid whose electrolyte composition differs enough from that of the cerebrospinal fluid (Heinemann et al., 2009).

The complexes that they form with Mg ²⁺ can be relevant and not negligible for the purposes of a characterization of the fluid itself. One example regards the calculation of the formation percentages of Mg²⁺-PEA and -PPC species under CSF conditions (C_Na = 141 mmol L⁻¹, C_K = 2.9 mmol L⁻¹; C_Ca = 1.25 μmol L⁻¹, C_Mg = 1.2 μmol L⁻¹, C_Cl = 124 mmol L⁻¹, C_HCO3 = 21 mmol L⁻¹, C_PO4 = 0.15 mmol L⁻¹, C_PEA = 1.70 μmol L⁻¹, C_PPC = 1.70 μmol L⁻¹, t = 37°C, and I = 0.15 mol L⁻¹) (Artru, 2010). All formation constants of the species taken into account in these simulations are listed in Supplementary Table S3. Under these conditions, none of the species containing PEA and PPC reaches significant formation percentages. Calculated percentages of Mg²⁺-PEA and -PPC species in CSF conditions at pH = 7.4, t = 37°C, and I = 0.15 mol L⁻¹ are shown in Figure 5B. On the contrary, considering the conditions of the brain intracellular space, where PEA and PPC concentrations are higher than in the CSF and plasma (C_Na = 155 mmol L⁻¹, C_K = 3.0 mmol L⁻¹; C_Ca = 1.6 mmol L⁻¹, C_Mg = 1.2 mmol L⁻¹, C_Cl = 135 mmol L⁻¹, C_HCO3 = 21 mmol L⁻¹, C_PO4 = 1.0 mmol L⁻¹, C_PEA = 0.59 mmol L⁻¹, C_PPC = 0.59 mmol L⁻¹, t = 37°C, I = 0.15 mol L⁻¹ (Klein et al., 1993; Eaton and Pooler, 2009; Barrett et al., 2013), the formation percentages, especially of the Mg²⁺-PPC species, increase significantly. More in detail, free magnesium reaches 67%, MgCl 13%, MgPPC 9.6%, and MgPEAH 5%. Calculated percentages of Mg²⁺-PEA and -PPC species, under the conditions of the extracellular fluid in the brain intracellular space, at pH = 7.4, t = 37°C, and I = 0.15 mol L⁻¹, are shown in Figure 5A. This result shows that the species MgPPC reaches a percentage not negligible, albeit not high. The availability of reliable thermodynamic constants makes simulation possible under the conditions of real fluids.

The dependence of protonation constants and formation constants on ionic strength was determined processing experimental measurements performed at different ionic strengths by considering the following Debye–Huckel equation, widely used in the 0 ≤ I ≤ 1 mol L⁻¹ range (Cardiano et al., 2018b; Chillè et al., 2018): logβ = log β 0  - 0 .51  ⋅ z* I 1 + 1 .5 I +   C   I, (5) where β is the stability constant at a given ionic strength, β⁰ is the stability constant at infinite dilution, z* = Σ(charge)² _reactants − Σ(charge)² _products, and C is an empirical parameter. Protonation constants at infinite dilution and C parameter values calculated by Eq. 5 for PEA and PPC species at t = 25°C in NaCl are listed in Table 3.

For the dependence on the temperature, the van’t Hoff equation was used, as for other systems (Cordaro et al., 2019; Giuffrè et al., 2020): log β T   =  log β θ + Δ H (   1 / θ − 1 / T ) Rln 10 , (6) where logβ^T is the stability constant at a given ionic strength and temperature (in Kelvin), logβ^θ is the value at the reference temperature (T = 298 K), and ΔH ⁰ is the formation enthalpy change expressed in kJ mol⁻¹ at T = 298.15 K and R = 8.314472 J K⁻¹ mol⁻¹.

The values of thermodynamic parameters, as formation enthalpy, entropy, and free energy changes, of the Mg²⁺-PEA, and -PPC species are reported in Table 4. Formation thermodynamic parameters referring to reaction (4) are shown as a bar plot in Figure 6, to highlight the prevalent contribution of entropy or enthalpy to free energy. As known for electrostatic interactions, the entropic term, related to the orientation disorder of the solvation water molecules, constitutes the main contribution to the free energy change. This behavior was found for most of the species, except for ML one formed by the PEA ligand.

The sequestering capacity of a ligand in a solution correlates with the tendency of the ligand to form complexes with a given metal cation. The higher the stability of the complex species, the lower is the concentration of the free metal cation in the solution. To evaluate the sequestering ability of a ligand toward a specific metal cation, all equilibria in which both the ligand and the metal cation participate are considered, as metal ion hydrolysis, ligand protonation, and weak interactions with the ionic medium. The pL_0.5 empirical parameter, i.e., the co-logarithm of the ligand concentration which sequesters 50% of the metal cation in traces, was proposed. Traces of metal cations were considered as they represent the conditions of concentration with which many of them are generally present in natural fluids. The sequestering capacity of a ligand toward a metal cation can be evaluated by the following Boltzmann-type equation with asymptotes 0 for pL → 0, 1 for pL→∞ (Cardiano et al., 2011b; Cardiano et al., 2018a): χ   = 1 1 + 10 ( pL – pL 0.5 ) , (7) where χ is the sum of the molar fractions of the metal–ligand species, and pL is the co-logarithm of the total ligand concentration. The sequestering ability strictly depends on pH, temperature, and ionic strength.

In order to evaluate the sequestering capacity of PEA and PPC toward Mg²⁺, pL_0.5 values at different temperatures were calculated (Supplementary Table S4). Figures 7A,B show the comparison between the sequestering capacity of PEA and PPC under physiological conditions (pH = 7.4, and I = 0.15 mol L⁻¹). The plot confirms that under these conditions, PPC shows a slightly higher sequestering capacity than PEA toward Mg²⁺.

Mass spectrometry measurements were used to elucidate the mechanism of PPC and PEA interactions. In turn, this aspect is very useful to understand some biological phenomena such as the mechanism transport through membranes but also for evaluating the possible use of these compounds in some application fields. Phosphorylcholine-based biomaterials are well studied due to their biocompatibility and being used in many clinical applications (Matsuura et al., 2016; Goda and Miyahara, 2018). The performance of these biomaterials can be affected by electrolytes. In fact, the zwitterionic nature of the phosphorylcholine groups may be changed into a cationic system introducing divalent cations that strongly interact with the phosphate group. Although it required the knowledge of the sequestering ability, the coordination mode of the natural ligand in local microenvironments helps to understand the potential activity of the biomaterials, and speciation studies are seldom reported (Gabryel-Skrodzka et al., 2021).

Mass spectrometry (MS) techniques are generally used for the highly sensitive analysis of metal ion complexes. The positive ion mode is usually the polarity for metal complex analysis by mass spectrometry (Aiello et al., 2017). Here, this selection was strengthened using ligands containing a quaternary nitrogen with a fixed positive charge. Therefore, the matrix-assisted laser desorption mass spectrometry (MALDI-TOF/TOF-MS) platform in the positive ion mode was adopted to study Mg²⁺-L systems (L = PPC or PEA). This platform offers particular advantages in investigating biological systems (Aiello et al., 2021b). The most important peculiarities of MALDI MS-based methods rely in the rapid and sensitive detection of analytes (Aiello et al., 2020; Salvatore et al., 2020) and in obtaining the molecular profiling of complex mixtures. Finally, the structures of low molecular weight organic and organometallic compounds can be analyzed and determined by tandem mass spectrometry (MS/MS) experiments (Falcone et al., 2013). The interaction of Mg²⁺ with PPC and PEA was explored by MALDI using α–CHCA as the matrix. Full-scan positive ion MS of Mg²⁺-L systems (L = PPC, or PEA) displayed ion species indicating information on ML_n (n: 1, 2) species. The elemental composition of detected ML_n species, combined with the observed gas-phase fragmentation pathways, was used to identify the coordination sites and to ascribe the most probable structures of complexes. The isotope ratio patterns observed for all the Mg²⁺/L complexes matched with those obtained from theoretical calculations, suggesting both ligands acting as a bidentate. The simplest systems, represented by free ligands, will first be discussed (Figure 8A). The MS/MS spectrum of [PPCH]⁺ at 1 kV shows complementary fragment ion pairs of the m/z values 99/86 ([H₄O₄P]⁺/[C₅H₁₂N]⁺), 81/104 ([H₂O₃P]⁺/[C₅H₁₄NO]+), and 60/125 ([C₃H₁₀N]⁺/[C₂H₆O₄P]⁺) as the most abundant fragment ions. PPC is an aliphatic ester of phosphoric acid; consequently, it preferentially forms the fragment of m/z 99 rather than the phosphate marker ion of m/z 81. Meanwhile, an intramolecular H transfer, involving the alkyl backbone of the PPC molecule, promotes the release of trimethylamine and the formation of the m/z 125 ([C₂H₆O₄P]⁺). MS and MS/MS ion fragments of ligands are listed in Table 5.

Direct MS analysis of solutions containing both Mg²⁺ and PPC at a final M:L ratio of 1:1, pH 8 as stated by speciation experiments, showed the formation of the ions of m/z 206 ([MgPPC]⁺, [C₅H₁₃MgNO₄P]⁺) and m/z 389 ([Mg(PPC)₂]⁺, and [C₁₀H₂₇MgN₂O₈P₂]⁺). As shown in Figure 8B and Table 5, the product-ion spectrum of the [MgPPC]⁺ (m/z 206) contains prominent ions of m/z 147 ([C₂H₄MgO₄P]⁺) and m/z 91 ([C₂H₄O₂P]⁺). The pathways proposed for formation of these two ions involve an initial loss of trimethylamine ([MgL-59]), followed by an additional loss of Mg(OH)₂ (58 Da). Meanwhile, the observed daughter ion of m/z 191 ([C₄H₁₀MgNO₄P]^•+), 177 ([C₃H₈MgNO₄P]^•+), and 163 ([C₂H₆MgNO₄P]^•+) arises from direct and consecutive loss of methyl groups matching typical fragmentation of the positively-charged head group. The gas-phase behavior of [Mg(PEA)H]⁺ matches that observed for Mg/PPC systems (Figure 8C and Table 5). The observed fragmentation patterns of both Mg systems enable us to design the most probable molecular structures, where the coordination occurs at the phosphate group generating a four-membered cycle.

Despite the experimental potentiometric conditions, the significant formation of MgL₂ species was not found; to complete this study, a solution containing Mg²⁺:L (L = PPC or PEA), at a final ratio of 1:2, was analyzed by MALDI MS/MS spectrometry at different times, over 2 h. The experiments conducted allowed a better detection (% intensity of the total ion current) of [MgL₂]⁺ ions. The MALDI MS/MS spectrum of the [Mg(PPC)₂]⁺ (m/z 389 and [C₁₀H₂₇MgN₂O₈P₂]⁺ Figure 9A and Table 5) revealed the formation of the cations [C₅H₁₃MgNO₄P]⁺, [H₃MgO₇P₂]⁺, [C₂H₆O₄P]⁺, [C₅H₁₄NO]⁺, [C₅H₁₂N]⁺, [C₄H₁₀N]⁺, and [C₃H₈N]⁺, resulting from the initial loss of the ligand, followed by the additional fragmentations of [MgPPC]⁺. Furthermore, the direct loss of trimethylamine [Mg(PPC)₂-59] (m/z 330), followed by the additional loss of trimethylamine [Mg(PPC)-118]⁺(m/z 270) and/or ethylene (28) [Mg(PPC)₂-(145)]⁺ from the parent ion [Mg(PPC)₂]⁺, confirmed that the phosphate group to be involved in the coordination with Mg²⁺. Interestingly, the loss of one ligand molecule from the [M(PPC)₂]⁺ species was observed. These data indicated that coordination of Mg²⁺ with PPC causes weakening of specific bonds which break upon collision. The gas-phase behavior of [Mg(PEA)₂]⁺ does not match that observed for Mg/PPC systems. For this species, the direct loss aminoethanol ([Mg(PEA)₂-60], m/244) followed by the additional loss of the methylamine ([Mg(PEA)₂-93]⁺, m/212) dominates the gas-phase fragmentation (Figure 9B and Table 5). The formation of [Mg(PEA)-17]⁺ suggests that coordination of Mg(II) with PEA is more rugged than PPC under MS/MS conditions. The ML and ML₂ complexes of Mg²⁺ with PEA features coordination modes which were very similar to those observed for the species containing PPC. For all species detected, mass spectra suggested a common structure in which metal is coordinated to the phosphate group of the ligand frame.