Isomer detection on the basis of analyte adduct formation with the components of the mobile phase and tandem mass spectrometry

1 Introduction

Liquid chromatography–tandem mass spectrometry (LC–MS/MS) is one of the most widely used analytical techniques applied for different purposes (Bhaskara et al., 2011; Dziadosz et al., 2014a; Jain et al., 2015; Venugopal et al., in press; Wang et al., in press). However, an important limitation of LC–MS/MS is the detection of small molecules since they can produce very small ion fragments or do not produce any stable ion fragments at all (Dziadosz et al., 2014b). The traditional way to analyse this problematic substances with LC–MS/MS is based on the detection of a pseudo mass transition which in fact is a simple monitoring of the protonated or deprotonated drug (Gao et al., 2011; Jain et al., 2007; Kim et al., 2011; Matsuura et al., 2008; Shibata et al., 2012). Therefore, the advantage of liquid chromatography–tandem mass spectrometry based on real mass transitions (defined contrary to pseudo mass transitions) as a consequence of analyte structural change cannot be applied in drug identification/quantification.

The latest research demonstrated that the problems mentioned can be solved by the application of the analyte adduct formation process with the components of the mobile phase such as sodium acetate and acetic acid if appropriate buffers are used. Analyte adduct ions can be applied to generate bigger ion fragments, when the analyte produces very small ones or to generate real mass transitions in cases when the analyte does not produce any stable ion fragments at all (Dziadosz et al., 2014c, 2013). A recent paper demonstrated also that this analytical strategy can be applied to generate MS³ mass transitions of small molecules (Dziadosz, 2015a). As a consequence these problematic substances can be analysed with LC–MS³ with higher sensitivity.

In previous papers it was stated that the analyte adduct formation process is an interesting alternative especially for small drugs whose fragmentation is problematic (Dziadosz et al., 2014b, 2013). The application of analyte adduct formation opened a way to generate MS² and MS³ mass transitions for these small molecules (Dziadosz, 2015a; Dziadosz et al., 2014c). An effective validation/quantification could also be performed with LC–MS/MS on the basis of forensic guidelines. The application of this strategy for drugs whose fragmentation is not problematic was thought to be limited (in some methods if sensitivity problems occur) or even without analytical sense. However, the study of analyte adduct formation induced a consideration about the similarities/differences in isomer adduct formation and about possible applications if differences occur. Therefore, the aim of this paper was to investigate the analyte adduct formation/fragmentation of similar substances which can be characterised by the same molecular mass but slight differences in the structure. An analyte group that fulfils these requirements can be represented by α- (AHB), β- (BHB) and γ-hydroxybutyrate (GHB) (Fig. 1) with molecular masses (M) = 104 Da. These substances were used as model compounds for the appropriate investigation. AHB and BHB are substances which can interfere with GHB when a GHB analysis is performed for forensic/clinical purposes. GHB itself is an endogenous molecule; however, it gained public attention as a drug of abuse. Thus, these model compounds represent a real analytical challenge in the field mentioned. A validated method for GHB analysis in human serum on the basis of analyte adduct formation was already presented (Dziadosz, 2015a; Dziadosz et al., 2013).

Close

Figure 1

The chemical structure of AHB, BHB and GHB.

Export to PPT

Since the application of different analyte adduct ions can also improve method sensitivity in some applications, they have been used for a long time in LC–MS/MS analysis especially in cases, when the signal intensity of protonated/deprotonated molecules is not satisfactory. An example is the analysis of sirolimus or asiatic acid based on ammonium adduct ions (Kushnir et al., 2005; Nair et al., 2012; Vogeser et al., 2002). However, the advantage of adduct detection in minimising the problem of interfering substances was not investigated/described. Therefore, the study presented was focused on the possibility of using analyte adduct formation/fragmentation to improve mass transition specificity of problematic molecules for MSⁿ detection. Furthermore, the observation examined and described on the basis of different AHB/BHB/GHB signals in this paper (mass transition of: deprotonated analyte, deprotonated analyte adduct with sodium acetate, deprotonated analyte dimer adduct with sodium and deprotonated analyte adduct with two molecules of sodium acetate) could signal the possibility to use the analyte adduct formation/fragmentation strategy presented earlier not only for small drugs which are problematic in their detection when LC–MS/MS is applied but also for all other drugs with appropriate molecular structure for adduct formation and different problems with interfering matrix components.

2 Materials and methods

An Applied Biosystems API 4000 QTrap tandem mass spectrometer operated in the negative electrospray ionisation mode and a Shimadzu UFLC Prominence System equipped with two solvent delivery units (LC-20AD), Communication Bus Module (CBM-20A), Autosampler (SIL-20AC HT), degasser (DGU-20 A3) and column oven (CTO-10 AS VP) were used in the study. Data acquisition was performed by the Analyst 1.5 software. Chemicals/solvents used were of analytical/LC–MS grade and purchased from the following: Merck (Darmstadt, Germany), J.T. Baker (Deventer, Netherlands) and Sigma–Aldrich (Taufkirchen, Germany). ESI-MS/MS was performed in the negative scan mode (−Q1 MS), negative enhanced product ion mode (−EPI) and negative multiple reaction monitoring mode (−MRM) with the following source/gas parameters: curtain gas (N₂) – 10 psi, collision gas (N₂) – high, temperature – 0, ion source gas 1 (N₂) – 22 psi, ion source gas 2 (N₂) – 0 psi, ion spray voltage – −4500 V. Additionally, the −Q1 MS mode used for the adduct identification experiments was performed with following declustering (DP) and entrance potential values (EP): −60 V and −10 V respectively and the −EPI mode used for the fragmentation experiments was performed with following DP, collision energy (CE) and collision energy spread (CES) values: −60 V, −20 V and 10 V. The −MRM mode was used to investigate the influence of CE on signal intensity at a given mass transition at DP = −60.

A 1 mg/mL stock solution of each substance investigated (AHB/BHB/GHB) was prepared in methanol and used for the preparation of infusion solutions. Appropriate dilutions were performed in methanol with 1 mmol L⁻¹/10 mmol L⁻¹ CH₃COONa/CH₃COOH. A continuous infusion at a flow rate of 10 μL/min of each diluted substance (1 and 10 μg/mL) made the adduct identification/fragmentation experiments possible. By the application of this strategy it was expected that the infusion experiments performed on the basis of these solutions will reveal that signal intensities of some m/z values increase in the higher concentrated drug solution. One of these m/z values would be related to the deprotonated analyte and other values with a comparable signal increase would be related to appropriate analyte adduct ions. The calculated differences of m/z values related to the analyte adduct ions and m/z values related to the deprotonated analyte would characterise the mass increment expected by the adduct formation process and would made the adduct identification possible.

The application of −Q1 MS experiments was aimed to be used to identify adduct ions formed under conditions applied and to compare the analyte adduct formation of substances with small differences in the structure and the same molecular mass (AHB, BHB, GHB). −EPI experiments were planned to be used to identify possible mass transitions of adducts characterised by the −Q1 experiments. Finally these mass transitions were planned to be investigated by −MRM experiments to answer the question if the analyte adduct formation process with the components of the mobile phase can be applied successfully to “separate” the GHB detection performed in the forensic/clinical toxicology on the basis of MS/MS systems from signals generated by other substances with the same molecular mass and similar structure. These problematic substances are relevant, since only slight differences in the structure in comparison with the analyte lead to similar retention times, when liquid chromatography–tandem mass spectrometry is used in the analysis. Therefore, they are a potential interference and can cause important separation problems especially when the analyses are performed by the application of reversed-phase chromatography.

3 Results and discussion

The −Q1 MS experiments performed on the basis of 1 and 10 μg/mL infusion solutions made the identification of following AHB (Fig. 2)/BHB/GHB signals possible: deprotonated analyte ([M – H]⁻, m/z = 103 Da), deprotonated analyte adduct with sodium acetate ([M – H + CH₃COONa]⁻, m/z = 185 Da), deprotonated analyte adduct with analyte ([2M – H]⁻, m/z = 207 Da), deprotonated analyte dimer adduct with sodium ([2M – 2H + Na]⁻, m/z = 229 Da) and deprotonated analyte adducts with n molecules (n = 1–10) of sodium acetate ([M – H + nCH₃COONa]⁻, m/z = 103 + n82 Da). An interesting observation made is the fact, that analyte adducts even with 10 molecules of sodium acetate could be identified during experiments performed (m/z = 923). However, in this work only analyte adducts with no more than two molecules of sodium acetate were investigated further (m/z = 267).

Close

Figure 2

Q1 mass spectrum registered in the m/z range from 40 to 500 for the 10 μg/mL AHB solution.

Export to PPT

The −Q1 MS experiments revealed that AHB, BHB and GHB form the same analyte adducts with the components of the mobile phase with comparable signal intensities under conditions applied. An example of the −Q1 mass spectrum registered for AHB is presented in Fig. 2.

The deprotonated analyte/analyte adduct fragmentation was investigated on the basis of −EPI experiments which revealed that the fragmentation of analyte adduct ions is comparable to the fragmentation of appropriate deprotonated analytes (Fig. 3). However, other fragments with bigger masses could also be identified.

Close

Figure 3

EPI mass spectra of AHB, BHB and GHB (all 10 μg/mL).

Export to PPT

GHB produces a mass fragment at m/z = 85 among other signals with lower intensity. Based on the small molecule of GHB its fragmentation can be defined as poor. Therefore, the mass transition 103/85 is applied widely together with the detection of very small ion fragments (m/z < 85) for LC–MS/MS analysis when −ESI is used in the traditional way without adduct formation (Dresen et al., 2007; Elian and Hackett, 2011; Lott et al., 2012; Sørensen and Hasselstrøm, 2012; Stout et al., 2010). Since the mass transition 103/85 enables to work with the best sensitivity it is used as target in MS/MS detection. However, this transition can be explained presumably by the loss of water and it is well known that such mass changes resulting from commonly lost small fragments tend to be non-specific (Kushnir et al., 2005). Other mass transitions based on the detection of even smaller problematic ion fragments are applied as qualifier. Although, in this way two mass transitions are used as required in forensic applications, GHB has to be seen as problematic molecule for traditional LC–MS/MS analyses (Peters et al., 2009). AHB and BHB are molecules which can interfere with GHB since their molecular weights are the same. Therefore an appropriate separation with different stationary phases is necessary when one of these substances is analysed with LC–MS/MS (Sørensen et al., 2013). In particular, the method selectivity between AHB and GHB has to be evaluated since BHB does not produce the 103/85 mass transition in −ESI.

Since it was stated above that the fragmentation of analyte adduct ions is comparable to the fragmentation of appropriate deprotonated analytes (−EPI experiments) it was interesting to evaluate whether the m/z = 85 ion fragment is generated efficient enough by adduct fragmentation. Because the MS/MS signal intensity is affected strongly by CE, signal intensities of appropriate AHB, BHB and GHB adduct ions identified on the basis of −Q1 MS experiments were monitored at different values of CE applied in −MRM experiments (Fig. 4). The results revealed that GHB adduct ions form stable ion fragments at m/z = 85. Therefore, signals with appropriate intensities can be expected when the mass transitions investigated are used for GHB detection in LC–MS/MS analysis. On the other hand, like expected BHB adduct ions do not form stable ion fragments (m/z = 85) since the deprotonated BHB does not form this ion fragment neither (Fig. 3). This fact can be pointed out by CE functions obtained (Fig. 4). Thus, it can be stated that BHB is not a potential interfering molecule when −ESI is used for GHB analysis with liquid chromatography–tandem mass spectrometry on the basis of m/z = 85 ion fragment detection.

Close

Figure 4

MRM experiments with CE ramping (signal intensity as a function of collision energy) – all 10 μg/mL.

Export to PPT

AHB is a potential interfering drug for GHB analysis since the mass transition 103/85 can also be generated (Fig. 3). Since the AHB CE function of this mass transition is comparable to the CE function of GHB, the AHB interference cannot be eliminated by the application of a different CE value. When CE functions of appropriate AHB and GHB adducts are compared it can be observed that all GHB adducts investigated form stable ion fragments at m/z = 85. This tendency cannot be confirmed by AHB adduct ion fragmentation since only for the 185/85 mass transition a CE function can be supposed (Fig. 4). Other signal intensities registered for AHB adducts (mass transitions: 229/85 and 267/85) on the basis of CE ramping can be defined as background noise comparable with BHB adduct signals (Fig. 4). This finding is very interesting since it suggests that by the application of GHB adduct ion fragmentation an important problem concerning the AHB interference (−ESI) can be solved not only by the application of an appropriate separation column but also by the way of analyte detection.

The advantage of analyte adduct formation/fragmentation in solving the problem of AHB interference in GHB analysis with LC–MS/MS is presented in Fig. 5. For the analysis presented two solutions, one with 10 μg/mL AHB and another with 10 μg/mL GHB, were prepared in methanol with 1 mmol L⁻¹/10 mmol L⁻¹ CH₃COONa/CH₃COOH. Chromatographic separation was performed with a Luna 5 μm C18 (2) 100 A, 150 mm × 2 mm column (Phenomenex, Aschaffenburg, Germany) and the elution with a mobile phase consisting of 10% A (H₂O/methanol = 95/5, v/v) and 90% B (H₂O/methanol = 3/97, v/v), both with 10 mM ammonium acetate and 0.1% acetic acid. The injection volume was 10 μL and the flow program used was as follows: starting with 0.05 mL/min, ramping to 0.4 mL/min from 0.0 to 1.5 min, holding 0.4 mL/min from 1.5 to 2.2 min, reducing to 0.05 mL/min from 2.2 to 2.5 min, and holding 0.05 mL/min from 2.5 to 3.0 min. The MS/MS was operated in the −MRM mode with following conditions (Q1 mass/Q2 mass = 267/85): time = 100 ms, DP = −60, EP = −10, CE = −25, cell exit potential (CXP) = −5, curtain gas (N₂) – 25 psi, collision gas (N₂) – medium, ion source gas 1 and 2 (N₂) – 80 psi, temperature – 400 °C, ion spray voltage – −4500 V. The conditions chosen did not guarantee a successful AHB/GHB separation; therefore, an AHB interference occurs when the traditional GHB MS/MS detection is applied (data not shown). However, the results achieved demonstrated that the AHB interference can be minimised on the basis of the analyte adduct formation/fragmentation process since only background noise could be observed for AHB at the retention time expected (Fig. 5). Thus, this strategy can be a real advantage in GHB analysis in biological material after appropriate method optimisation. Since the analyte adduct formation is a promising alternative for drug analysis, the study presented is an important continuation of previous papers (Dziadosz, 2016a, 2016b, 2015a, 2015b; Dziadosz et al., 2014b, 2014c, 2013).

Close

Figure 5

Chromatogram of a 10 μg/mL AHB and GHB solution in methanol with 1 mmol L⁻¹/10 mmol L⁻¹ CH₃COONa/CH₃COOH registered at m/z = 267/85.

Export to PPT

4 Conclusion

AHB, BHB and GHB produce similar adduct ions in the negative electrospray mode. BHB is not a potential interfering substance for GHB analysis since the ion fragment at m/z = 85 is not produced neither by BHB ion nor by BHB adduct ion fragmentation. AHB is a potential interfering drug for GHB analysis (−ESI) when traditional mass transitions are used. The interference of AHB can be minimised when analyte adduct ion fragmentation is used since appropriate fragmentations to m/z = 85 do not produce ion fragments with efficiency required at different values of collision energy.

The analyte adduct formation process can be a real benefit when interfering substances with similar retention times do not produce adduct ions or produce adduct ions but with a different fragmentation pattern like presented in this paper. Therefore, this process can be considered for application not only for small drugs which are problematic in their detection when liquid chromatography–tandem mass spectrometry is applied but also for all other drugs with appropriate molecular structure for adduct formation and different problems with interfering matrix components.