Synchronized Survey Scan Approach Allows for Efficient Discrimination of Isomeric and Isobaric Compounds during LC-MS/MS Analyses

Liquid chromatography-mass spectrometry- (LC-MS-) based multiple reaction monitoring (MRM) methods have been used to detect and quantify metabolites for years. These approaches rely on the monitoring of various fragmentation pathways of multiple precursors and the subsequent corresponding product ions. However, MRM methods are incapable of confidently discriminating between isomeric and isobaric molecules and, as such, the development of methods capable of overcoming this challenge has become imperative. Due to increasing scanning rates of recent MS instruments, it is now possible to operate MS instruments both in the static and dynamic modes. One such method is known as synchronized survey scan (SSS), which is capable of acquiring a product ion scan (PIS) during MRM analysis. The current study shows, for the first time, the use of SSS-based PIS approach as a feasible identification feature of MRM. To achieve the above, five positional isomers of dicaffeoylquinic acids (diCQAs) were studied with the aid of SSS-based PIS method. Here, the MRM transitions were automatically optimized using a 3,5-diCQA isomer by monitoring fragmentation transitions common to all five isomers. Using the mixture of these isomers, fragmentation spectra of the five isomers achieved with SSS-based PIS were used to identify each isomer based on previously published hierarchical fragmentation keys. The optimized method was also used to detect and distinguish between diCQA components found in Bidens pilosa and their isobaric counterparts found in Moringa oleifera plants. Thus, the method was shown to distinguish (by differences in fragmentation patterns) between diCQA and their isobars, caffeoylquinic acid (CQA) glycosides. In conclusion, SSS allowed the detection and discrimination of isomeric and isobaric compounds in a single chromatographic run by producing a PIS spectrum, triggered in the automatic MS/MS synchronized survey scan mode.

From the above, it can be surmised that secondary metabolites are diverse and, in some cases, are unique to specific plant species and, as such, can be used as chemotaxonomic markers [15][16][17][18]. erefore, the unambiguous detection and identification of these metabolites using analytical techniques such as liquid chromatography-mass spectrometry (LC-MS) is of paramount importance. For targeted analysis, multiple reaction monitoring (MRM) using a triple quadrupole LC-MS/MS system can be employed to selectively screen for and detect and quantify metabolites of interest [19][20][21][22]. However, studies dedicated to providing specific fragmentation patterns that discriminate between isomeric and isobaric secondary metabolites during MRM analyses are limited [23,24]. As such, a novel method/approach referred to as synchronized survey scan (SSS) is proposed as a possible way of discriminating structurally similar metabolites during MRM analyses, especially when they produce similar fragmentation transitions. Using a SSS function, we have demonstrated that positional isomers of diCQAs and their isobaric compounds, CQA glycosides, can be distinguished in a single chromatographic run. Here, authentic standards and plant extracts of Bidens pilosa [25] and Moringa oleifera were employed, since these plant species are reported to, respectively, accumulate/produce these compounds. us, the overall aim of the current study was to use the SSS approach as an orthogonal identification component of the MRM method for efficient discrimination of structurally related plant metabolites.

Sample Preparation.
A 1 mg/mL solution for each diCQA positional isomer was prepared with 100% methanol. e solution (for each positional isomer) was diluted 10× with 100% methanol. Furthermore, equal amounts (e.g., 40 µL) were taken from each positional isomer sample to prepare a mixed sample (e.g., of a final volume of 200 µL). e samples (individuals and the mixture) were placed in amber vials and subjected to HPLC-PDA analyses.

Metabolite Extraction.
e dried leaves of B. pilosa and M. oleifera were pulverized, respectively, using a clean and dry quartz mortar and pestle. Extraction was conducted using an organic solvent-based extraction. e respective amounts of samples (0.2 g) were mixed with 2 mL of 50% aqueous methanol, and these extracts were placed (with the lids of the tubes closed to avoid evaporation) in a heating block at 60°C for 2 h. e samples were sonicated for 30 min using an ultrasonic bath and then centrifuged at 9740 ×g for 10 min at 4°C. e resulting supernatants were subjected to UHPLC-MS analyses.

Ultra-High Performance Liquid Chromatography Tandem Mass Spectrometry (UHPLC-MS/MS) Analysis.
Once prepared, samples were analyzed on a Shimadzu Nexera 8050 UHPLC (Kyoto, Japan) fitted with a Raptor biphenyl analytical column, with the column temperature set at 40°C. A binary solvent mixture consisting of MilliQ water made up of 0.1% formic acid (eluent A) and methanol made up of 0.1% formic acid (eluent B) at a constant flow rate of 0.4 mL/min was used to analyze 2 µL of the injected samples. For the gradient elution, the following conditions were used: isocratic 5% eluent B from 0 min to 1 min, linear 5%-20% eluent B from 1 min to 5 min, linear 20%-90% eluent B from 5 min to 40 min, and isocratic 90% eluent B from 40 min to 45 min. At the end of analysis, conditions were changed to the initial conditions (5% eluent B) from 45 min to 48 min and finally, the column was reequilibrated with isocratic 5% eluent B from 48 min to 52 min. e data were acquired using a UV detector set at 325 nm and 330 nm.
For the MS analysis, the chromatographic effluent was introduced to a MS source and ionized by electrospray (ESI). ESI conditions were as follows: the interface voltage was set at 3.0 kV (in the negative ESI mode), the source temperature was 300°C, nitrogen was used as the drying gas at the flow rate of 10.00 L/min, and as a nebulizing gas at a flow rate of 3.00 L/min. Argon was used as a collision gas with a pressure of ±230 kPa in the collision cell. Sensitive and qualitative analysis of isomeric and isobaric plant metabolites was achieved by developing a MRM and MRM-dependent product ion scan (PIS) method. e MRM transitions were developed or optimized using 3,5-diCQA as the sample of choice based on the work done by Clifford et al. [3]. According to Clifford and colleagues, 3,5-diCQA contains product ions (e.g., m/z 353 representing a caffeoylquinic acid moiety, m/z 179 representing a caffeic acid moiety, and m/z 191 representing a quinic acid moiety) characteristic of all diCQA isomers, as well as CQA glycosides [13,14]. e MRM transition parameters were automatically optimized to produce the transitions shown in Table 1. e dwell time for all the MRM transitions was 30 ms.
A synchronized survey scan (SSS) function was selected which automatically performed MS/MS analysis when the precursor threshold peak intensity exceeded 2,000,000. is resulted in a combined MRM and a MRM-dependent PIS, both of which were produced in a single analysis. For the PIS mode, ions were collected at a mass range 100-1000 Da with a continuous scan time of 1 sec, at a collision energy of 25 eV.

LC-MS/MS Method Optimization.
In this study, LC-MS analysis was used to sensitively and qualitatively analyze isomeric and isobaric plant metabolites by developing a MRM-dependent product ion scan (PIS) method. Samples (authentic standards and plant samples) containing positional isomers of dicaffeoylquinic acids (diCQAs) and caffeoylquinic acid (CQA) glycosides were analyzed under reverse-phase chromatographic conditions using a Raptor biphenyl column with methanol as part of the binary solvent mixture. To identify the respective diCQA positional isomers, a sample made up of a mixture of the positional isomers (authentic standards) was analyzed, and it produced a chromatogram showing well-resolved peaks representing the five isomers ( Figure 1; Table 2). e retention times (Rts) of the respective peaks were compared with the Rts of the individual (nonmixed) authentic standards, and the elution order under the abovementioned conditions was noted as 1,3-diCQA ( Figure 1, A), 3,4-diCQA ( Figure 1, B), 3,5-diCQA ( Figure 1, C), 1,5-diCQA ( Figure 1, D), and 4,5-diCQA (Figure 1, E). Furthermore, the respective product ion scan (PIS) spectra, triggered in the automatic MS/MS synchronized survey scan mode, were also referred to for the analysis of the fragmentation patterns of the diCQA positional isomers for further identification (Figure 2). Although the MRM transition parameters were automatically optimized using 3,5-diCQA, all five diCQA isomers were detected, since these compounds have been shown to share similar product ions [3,11].

MRM and MRM-Dependent Product Ion Scan Analysis of Plant Extracts.
After optimization, Bidens pilosa and Moringa oleifera plant extracts were analyzed under the abovementioned conditions (Section 3.1). When the MRM transition parameters, which were automatically optimized using 3,5-diCQA, were used to analyze extracts of B. pilosa and M. oleifera plant samples, several peaks were detected. Briefly, three peaks were detected in B. pilosa, and only two peaks were detected in M. oleifera. When compared, the two peaks detected from M. oleifera samples showed an earlier elution profile (Rt � 4.89 and 7.32) and were thus observed to be more hydrophilic than the three peaks detected in B. pilosa samples, which showed a later elution profile (Rt � 21.24, 22.05, and 24.46). ese observations are summarized in Table 2. Such chromatographic observations are important in scenarios whereby the analyzed plant sample contains both diCQAs and CQA glycosides, thus circumstances whereby the plant extract contains both the isomeric and isobaric compounds [11,13,14,26]. e Rts of the peaks (Rt � 21.29, 22.05, and 24.61) found in B. pilosa samples were compared with those observed with the diCQA authentic standard samples, and the respective peaks were identified as 3,4-diCQA (Figure 3(a)), 3,5-diCQA (Figure 3(b)), and 4,5-diCQA (Figure 3(c)) ( Figure 3; Table 2), since the fragmentation pattern of these peaks was identical to the fragmentation pattern observed in Figures 2(b), 2(c), and 2(e), respectively. For the peak representing 3,4-diCQA, the observed fragment ions were (Figure 3 Table 2). e peak representing 3,5-diCQA showed fragment ions (Figure 3 It is worth noting that in the absence of authentic standards, the ion at m/z 173 is diagnostic of HCA derivatives acylated at position 4 on the quinic acid (e.g., 3,4-diCQA and 4,5-diCQA), as it has been noted in the work done by Clifford et al. [3,10]. However, to distinguish between 3,4-diCQA and 4,5-diCQA an ion at m/z 335 is noteworthy, as its presence signifies the formation of a lactone group between a carboxylic group at position 1 and a hydroxyl group at position 5 on the quinic acid [27]. us, position 5 on the quinic acid need not be acylated to allow the formation of the lactone group as shown elsewhere [27]. erefore, 3,4-diCQA and 4,5-diCQA can be distinguished based on the presence of the ion at m/z 335 on the MS 2 spectra representing 3,4-diCQA [10]. e two peaks (Rt � 4.87 and 7.30) detected in the M. oleifera plant extracts were tentatively characterized as 3-caffeoylquinic acid (CQA) glycoside and 4-CQA glycoside ( Table 2) [26]. e peak (Rt � 4.87) representing 3-CQA glycoside showed fragment ions (Figure 3 [13,14]. In this study, both ions were secondary ions and not base peaks, and the intensity of both peaks could have been influenced by the differences in the MS conditions/parameters [28]. us, due to the low abundance/absence of the ion at m/z 323 on the MS 2 spectra (Figures 3(d) and 3(e)), the CQA glycosides were      Table 2).

Synchronized Survey Scan (SSS).
From the above, it is apparent that in the absence of the triggered product ion scan (PIS), the optimized MRM method would have detected the diCQA and the CQA glycosides as these compounds share similar transitions. However, the method would fail to differentiate between the structurally similar compounds and thus lead to misidentification. us, the novel approach, SSS, allowed the detection and differentiation of all five diCQA positional isomers ( Figure 2) as well as two CQA glycoside positional isomers (Figures 3(d) and 3 (e)) in a single chromatographic analysis. Furthermore, the method allowed the discrimination of isobaric compounds such as diCQA and CQA glycosides. Here, an ion at m/z 341 was found to be present on the PIS MS/MS spectra of CQA glycosides. is ion represents caffeoyl glycoside [11] and, as seen from our results, it was only observed on the PIS spectra but not the MRM spectra. is is an indication that SSS dependent PIS allows other diagnostic ions to be used for the differentiation of closely related molecules, which otherwise would be impossible if only MRM transitions are relied upon.
us, SSS produces an automatically performed MS/MS analysis which allows the simultaneous identification of these isomeric and isobaric compounds, without any ambiguity.

Conclusion
For the first time, we have successfully demonstrated synchronized survey scan (SSS) to be an efficient approach to detect and discriminate isomeric and isobaric plant metabolites. Briefly, this approach allowed the discrimination and identification of (1) diCQA positional isomers, (2) CQA glycosides positional isomers, and (3) isobaric compounds, diCQAs and CQA glycosides, in a single chromatographic run, albeit the developed MRM transition parameters were automatically optimized using the positional isomer 3,5-diCQA. Our results show that this method can be further applied in any method where isomers are expected. Furthermore, in phytochemistry, where identification is the key, SSS is expected to add a novel orthogonal feature during identification.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.