Investigation of fragmentation behaviors of steroidal drugs with Li⁺, Na⁺, K⁺ adducts by tandem mass spectrometry aided with computational analysis

2.1 Chemicals and reagents

Analytical grade and pure (>98%) solvents and chemicals were used in this study. Ultrapure.

filtered deionized water (with resistivity 18.1MΩcm at 25 °C) acquired through Barnstead MicroPure Water Purification System (Thermo Scientific, USA). Methanol was purchased from Merck KGaA, 64,271 (Darmstadt, Germany), formic acid from Daejung (Daejung Chemicals and Metals Co. Ltd, (Korea), acetonitrile from Merck KGaA, 64,271 (Darmstadt, Germany). Sodium chloride (NaCl) was purchased from BioM Laboratories (Cerritos, USA). Lithium chloride (LiCl) was purchased from Sigma-Aldrich, and Potassium chloride (KCl) was purchased from BioM Laboratories (Cerritos, USA). Pharmaceutical drug standards were obtained from the Drug Bank of Dr. Panjwani Center for Molecular Medicine and Drug Research (PCMD), International Centre for Chemical and Biological Sciences (ICCBS), University of Karachi, Karachi, Pakistan. Lithium chloride and sodium chloride were purchased from Samchun Pure Chemical (Pyeongtaek, Korea). Potassium chloride was obtained from Junsei Chemical (Tokyo, Japan).

2.2 Preparation of standard solution and salts solutions

One mg of each standard drug was weighed and dissolved in 1 mL of methanol. One mM solution of each salt was prepared in water at a concentration of 0.1 mM of each salt was prepared by dilution and added to 1000 ng/mL solution of each drug before being subjected to the mass spectrometer.

2.3 Mass spectrometry analysis

Tandem mass spectra of cationized steroidal drugs were obtained using Bruker amaZon speed Ion trap mass spectrometer (Bremen, Germany) in positive ion mode. Samples were injected into the ion trap mass spectrometer with a syringe pump at a flow rate of 5 μL/min; the electrospray voltage and capillary temperature were set to 4.0 kV and 200 °C respectively. Ion source parameters were as follows: capillary voltage at 4500 V (-3500 V for negative mode), endplate offset at 500 V, nebulizer gas 40.0 psi, drying gas at 8.0 L/min, and drying gas temperature at 200 °C. Mass spectra scan range was set at m/z 50 to 1000 while the number of spectral averages was set at 5. Fragmentation was performed under CID, and amplitude was optimized for each analyte to obtain the maximum abundance of fragment ions.

A Bruker maXis II HR-QTOF mass spectrometer (Bremen, Germany) was used to assign fragments and obtain accurate m/z values. Samples were injected into the QTOF mass spectrometer with a syringe pump at a flow rate of 5 μL/min; the electrospray voltage and capillary temperatures were 4.0 kV and 200 °C, respectively. Spectra of all samples were acquired over 200 scans and averaged.

2.4 Computational analysis

Standard quantum chemical calculations were carried out using density functional theory (DFT) with the Gaussian 09 suite of programs. Geometrical parameters of the structures were optimized using the B3LYP hybrid functional method with the 6–31 ++ (d)(p) basis set. No symmetry constraints were imposed in the optimization. Initial models for each structure were obtained by placing a metal ion at each putative adduct site, followed by geometry optimization. Zero-point energy corrections were applied to all candidate structures. The more energetically more favorable structures were used in this study. All calculations are done in the gas phase and at a temperature of 300 k.

3.1 Fragmentation patterns of drug adducts

Fragmentation pathways of the cationized drugs were investigated using low-energy collision dissociation ESI tandem mass spectrometry (ESI-MS/MS) experiments. Collision energy (CE) optimization was performed by selecting CE between 25 and 60 eV. Low CE (<25 eV) caused a decrease in the production of the characteristic fragments, while an increase in the CE (>60 eV) has not provided any further significant information related to characteristic ions that appeared in the MS/MS spectrum because during high collision energy so much noise is produced, and its main fragments could not be easily identified.

The CID-MS/MS of both the precursor protonated molecular ion at m/z 431.2418 and lithiated adduct at m/z 437.2518 of the budesonide steroid created seven product ions at m/z 413.2318, 395.2207, 341.1747, 323.1638, 305.1529, 263.1425, 293.1543 and 365.1931, 419.2402, 349.1979, 331.2088, 251.1614, 209.1507, 143.1042, whereas the precursor sodiated and potassiated adduct at m/z 453.2241 and m/z 469.1975 created seven product ions at m/z 397.2024, 383.1855, 379.1914, 365.1751, 355.1913, 337.1801, 319.1689, 163.1130 and 395.1252, 383.1621, 337.1207, 321.1089, 263.0688. The loss of the H₂O and C₄H₈O₂ are common for both sodiated and lithiated adducts as compared to potassiated adducts while other fragments are different for each adduct. The differentiations are obtained in the breaking of C-C bonds between each adduct. The hemiacetal group attached at C16 and C17 is broken in all adducts. The fragments BU1 and BU3 were only observed for lithiated adducts, and BU2 was for sodiated adduct while the BU4 protonated and lithiated adducts. BU5 was observed for protonated, lithiated, and potassiated adducts. The CID-MS/MS of both the precursor protonated molecular ion at m/z 415.2151 and lithiated adduct at m/z 421.2186 of the eplerenone steroid created nine product ions at m/z 397.2024, 383.1855, 379.1914, 365.1751, 355.1913, 337.1801, 319.1689, 163.1130 and seven at m/z 377.1921, 403.2086, 385.1986, 359.1835, 349.1975, 343.1874, 331.1898 respectively, whereas the precursor sodiated and potassiated adduct at m/z 437.1929 and m/z 453.1669 created six product ions at m/z 407.1818, 419.1820, 377.1719, 355.1911, 303.1356, 169.0617 and three product ions at m/z 435.1471, 341.1154, 379.1309, 323.1049 respectively. The loss of H₂O and 2H₂O is common for each adduct. The C-C fragments EP1 and EP2 were obtained only in the spectrum of potassiated adduct, while the EP2 and EP5 were only for sodiated adduct and EP4 only for lithiated adduct. The EP6 product ion was observed in the spectrum of all adducts. The CID MS/MS spectrum of fluoxymesterone showed seven fragments in the protonated molecule, six in lithiated adduct, six in sodiated adduct, and four in potassiated adduct. The loss of water (FL5) and/or hydrofluoric acid was observed in the spectrum of the protonated molecule, and all adducts. The product ion FL2 was observed for protonated and lithiated adducts, while FL3 and FL4 were unique fragments in the potassiated spectrum and the FL1 was observed in the sodiated adducts. The CID-MS/MS of the protonated oxandrolone at m/z 307.2258 afforded three major product ions respectively at m/z 289.2162, 271.2062, 247.1693 and the lithium adduct with m/z created four fragments at m/z 295.2244, 269.2087, 243.1931, 173.1153 while the sodiated and potassiated adducts at m/z 329.2091 and 345.1832 respectively created four and two product ions at m/z 311.1982, 285.1825, 259.1669 and 327.1721, 261.1251 respectively. The loss of water molecule (OX1) is common for all metallic adducts. The OX2, OX4, and OX5 product ions were observed for both protonated molecule and lithiated adduct, while the OX3 was observed in the sodiated adduct. The OX6 product ion was observed for protonated and all metal adducts. Both protonated molecule and a lithiated adduct of dienogest created five fragments, three for sodiated adduct, and only one product ion for potassiated adduct. The loss of H₂O (DI7) was common in all adducts and the loss of HCN (DI6) was observed in sodiated and lithiated adducts. The C-C breaking fragment DI4 was only observed in the potassiated adducts while DI2 was only observed for sodiated. DI3 was observed in lithiated adducts and DI5 for both, protonated and lithiated adducts. Finasteride showed two product ions in protonated molecule spectrum, four in lithiated adduct, five in sodiated adduct, and three in potassiated adduct. The loss of water (FI7) was observed only in the protonated ones, while FI1 product ion was observed only in the potassiated spectrum, FI2 in sodiated adduct, and FI3 in the lithiated adduct. The FI2 and FI3 were obtained in the spectrum of both lithiated and sodiated adducts. The protonated molecule of etonogestrel showed three product ions, lithiated and sodiated adducts produced four product ions and potassiated adduct produced six ions. The product ion (ET6) indicates water loss observed in protonated, lithiated and sodiated adducts. Loss of acetylene (ET5) was observed for all adducts. The ET4 and ET3 were observed only in the spectrum of potassiated adduct, while the ET1 product ion observed for sodiated adduct and ET2 were only observed and lithiated adduct. The protonated ion of norethisterone showed three fragments, lithiated and sodiated adducts showed five product ions, and potassiated adduct produced two ions. The water loss (NO5) is observed for protonated, sodiated, and lithiated adducts. NO1 and NO4 product ions were obtained only in the spectrum of the sodiated adduct. The product ion NO3 was also observed for both protonated molecule and potassiated adduct. The CID-MS/MS of both the precursor protonated molecular ion at m/z 385.2371 and lithiated adduct at m/z 391.2459 of the estradiol dipropionate steroid created five and four product ions respectively while the sodiated and potassiated at m/z 407.2183 and 423.1949 created three and two product ions respectively. The losses of the carboxylic group (ED2) and aldehyde (ED3) groups were observed for all the adducts. The product ion ED1 was observed for both protonated and lithiated adducts, while ED4 was only observed in sodiated adducts. Cortisone showed six major product ions in the spectrum of the protonated adduct, three in lithiated adduct, four in sodiated adduct, and two in the potassiated adduct spectrum. The loss of water (CO5) is common for each molecular ion peak, and the loss of CH₂O molecule (CO4) is observed in the spectra of lithiated and sodiated adducts. The product ions CO3 and CO6 were observed for protonated molecule, sodiated and potassiated adducts, respectively. Norethisterone acetate showed four fragments in the spectrum of all protonated, lithiated, sodiated, and potassiated adducts. The product ion NA3, the loss of acetic acid, was observed in the protonated molecule and all adducts. The product ion NA2 and NA5 were observed in the spectrum of both sodiated and potassiated adducts. MS/MS spectrum of the protonated molecule of prednisolone showed four product ions, lithiated adducts showed six product ions, and both sodiated and potassiated adducts showed three product ions. The loss of water (PD4) and C₂H₄O₂ loss were observed in the lithiated and sodiated adducts. The product ion PD3 was observed only in the lithiated adduct. The protonated, lithiated adduct, sodiated adduct, and potassiated adduct of prednisolone acetate showed five, nine, seven, and four product ions respectively. The loss of water (PA6) is observed in the spectrum of all adducts. The product ions PA4 and PA5 were observed in the spectrum of both lithiated and sodiated adducts. The product ion PA1 was obtained only in the sodiated adduct spectrum, the PA2 in the lithiated adduct, the PA3 in the protonated molecule, and the PA7 in the potassiated adduct. The CID-MS/MS of both the precursor protonated molecular ion at m/z 317.2471 and lithiated adduct at m/z 323.2564 of the pregnenolone steroid created four product ions while the sodiated and potassiated at m/z 339.2283 and 355.2030 created three and two product ions respectively. The loss of water (PG4) is common for all adducts. The product ion PG1 is observed only in the spectrum of potassiated adduct, product ion PG2 in the protonated, spectrum, and product ion PG3 in the sodiated spectrum. The tandem mass spectral data of all drugs are presented in Table 1. The losses, product ions, and their elemental composition of all adducts of all steroids are presented in Fig. 1 and Supplementary Scheme 1–14.

Close

Fig. 1

The proposed C-C bond fragments of all drugs in each adduct.

Export to PPT

Close

Fig. 1

The proposed C-C bond fragments of all drugs in each adduct.

Export to PPT

Close

Fig. 1

The proposed C-C bond fragments of all drugs in each adduct.

Export to PPT

3.2 Effects of the alkali-metal ions cationization on structures and fragmentation, and computational studies

We have investigated the structural characteristics of selected alkali metal cations and their distinct fragmentation patterns. Different calculation methods were used to predict the position of the proton or the alkali metal ions in the complexes and reasonable fragmentation. To locate the position of the proton and alkali metal in all steroidal drugs, we have carried out a series of DFT calculations using the Gaussian 09 program (Frisch et al., 2016) suit at the B3LYP level of theory with the 6–31 ++ G(d) basis set. The minimum energy conformation of the neutral molecules was first optimized, and every possible site was then individually protonated and metalized after optimization. There are different possible protonation metalized sites, the oxygen’s, nitrogen groups, and it was found that all drugs showed minimum energy for protonation at carbonyl oxygen atom. In contrast, the metals showed binding with more than two electronegative atoms in all molecules. For example, in oxandrolone, it was found that the carbonyl oxygen (O-1) displayed minimum energy for protonation, while the metal showed binding with (O-1) and (O-2) with minimum energy. In the dienogest, etonogestrel, and pregnenolone metal cation interacts with metal showed binding with carbonyl oxygen (O-1) with minimum energy. In finasteride, the carbonyl oxygen (O-1) showed minimum energy for protonation. In contrast, the metal showed binding with (O-1) and (N-1) with minimum energy. In estradiol dipropionate, the protonation occurs at carbonyl oxygen atoms of the COO group while the metals showed minimum energy to attach with the ester group at C3 of ring A. In cortisone, the carbonyl oxygen (O-1) showed minimum energy for protonation while the metal showed binding with (O-3), (O-4), and (O-5) at C-17, C-20, and C-21 atoms, respectively. In prednisolone and prednisolone acetate, the carbonyl oxygen (O-1) showed minimum energy for protonation while the metal showed binding with (O-1), (O-3), and (O-6) with minimum energy (see Supplementary Scheme-15).

Geometries of the optimized structures, mechanisms of key fragments, and possible routes of all drug’s adducts were proposed. The binding energy of metal with molecules decreases as size increases from Li⁺, Na⁺, and K⁺. The fragmentation patterns observed for these Li⁺, Na⁺, and K⁺ ions are different. The number of fragments decreases as the metal size increases, suggesting the difference in binding position and strength of interaction of the molecule with metal cations. The proton can be readily bound to electronegative groups or atoms. However, for the molecules cationized by the metal ions, the positions of the metal are different.

Moreover, the bond lengths of protonated and metallated, and non-cationized forms were investigated. A comparison of bond lengths between all forms showed considerable bond elongation in the vicinity of the protonation and metallated site, thereby leading to the characteristic fragments. The lengths of dissociable bonds are the highest in lithium among the alkali metal complexes of all compounds. Lithium-ion's ion–dipole interactions induce the elongation of bonds more than other metals. The lithium adducts of all drugs have more fragments than other adducts due to the greater lengths of dissociable bonds (Table 3) and lower binding energy (Table 2). The highest ion–dipole interactions of lithium induce the elongation of the C–O bonds of the hydroxyl group and other C-C bonds. Elimination of the water molecule has been observed in the MS2 spectrum of protonated, lithiated, and sodiated ions in every drug with OH. In the potassiated ion, dehydration was not observed for some drugs. It is concluded that the formation of the dehydrated fragment is prevented by the strong ion–dipole interactions between the metal cation and the hydroxyl groups. In potassiated adducts, the metal bounds strongly with the functional –OH group and highly stable adduct.