The UPLC–ESI–QqQLIT–MS/MS method for quantitative determination of phytochemicals in ethanolic extracts of different parts of eight Ficus species: Development and validation

ABSTRACT Ficus and validation of the ultra performance liquid chromatography–electrospray ionization hybrid triple quadrupole–linear ion trap–tandem mass spectrometry (UPLC–ESI–QqQLIT–MS/MS) method in a multiple-reaction monitoring (MRM) mode for the quantitative determination of 19 phytochemicals. The chromatographic separation of targeted phytochemicals was performed using the Waters ACQUITY UPLC BEH™ C18 column (1.7 μm, 2.1 mm × 50 mm) with 0.1% formic acid with water and acetonitrile as a mobile phase at a flow rate of 0.25 mL/min. The validation parameters showed the overall recoveries from 95.78−101.44% (RSD ≤ 3.25%), precision (intra-day: RSD ≤ 2.96%; inter-day: RSD ≤ 2.89%), linearity (R2 ≥ 0.9982), limit of detection (8.60 × 10–10−2.18 × 10–6 mg/mL), and the limit of quantitation (2.60 × 10–9–6.63 × 10–6 mg/mL) in the concentration range from 0.5 to 1000 × 10–6 mg/mL. This method was successfully applied in ethanolic extracts of different parts (fruits, leaves, and barks) of selected eight Ficus species. Quinic acid was predominant followed by rutin and chlorogenic acid among the studied nineteen phytochemicals. Ficus benjamina showed the maximum total content in fruits and leaves. The UPLC–ESI–QqQLIT–MS/MS method combined with principal component analysis (PCA) was successfully used for Ficus species discrimination on the basis of the contents of 15 compounds. The UPLC–ESI–QqQLIT–MS/MS method combined with PCA could be used for quality control.


Introduction
Genus Ficus (Moraceae) is a tropical, deciduous, evergreen tree and cultivated worldwide since ancient times due to its delicious, nutritive fruits and medicinal properties. [1] The fruit of the Ficus species is commonly known as " Fig." The annual worldwide production of raw and dried fig is about 1.1 million tonnes and 0.13 million tonnes, respectively, with the highest production in Asia followed by Africa, Europe, and America. [2] The Ficus species is also used in traditional medicine for treating a wide range of ailments of the central nervous system, endocrine system, gastrointestinal tract, reproductive system, respiratory system, and infectious disorders. [3,4] The crude extracts of the Ficus species have been reported to possess antioxidant, antidiabetic, antiinflammatory, antiulcerogenic, antifungal, antibacterial, anticancer, and hypotensive activities. [4][5][6][7][8][9] Drying method The fruits, leaves, and barks were washed with tap water followed by Milli-Q water and water removed by placing them under blotting papers. The cleaned samples were kept in zipped polyethylene bags containing silica gel of 6-20 mesh, which was replaced regularly until the samples were fully dried. The completely dried plant materials were ground by mechanical mill into fine powder.

Extraction
The powdered samples (10 g) were each placed in ethanol (200 mL) and sonicated (Bandelin SONOREX, Berlin) for 30 min and allowed to stand at room temperature (26-28°C). After 24 h, ethanol was filtered off using a Whatman number 1 filter paper. The filtrate was concentrated in a rotary evaporator (Buchi Rotavapor-R2, Flawil, Switzerland) at 40°C under reduced pressure . The residue was reextracted three times following the same procedure. Each extract was packed separately in glass vials and stored at −20°C.

Preparation of calibration solutions and crude extract samples
The standards (quinic acid, protocatechuic acid, chlorogenic acid, mangiferin, catechin, epicatechin, rutin, hispidine, 7-hydroxycoumaric acid, ferulic acid, vanillic acid, luteolin, quercetin, apigenin, kaempferol, chrysin, betulinic acid, ursolic acid, and oleanolic acid) and the dried crude extracts (leaves, fruits, and barks) were separately weighed accurately to be about 1 mg each. Each standard and crude extract were dissolved in methanol (about 1 mL) according to weight for the preparation of 1 mg/mL stock solutions. Now, these stock solutions were sonicated for 30 min (Bandelin SONOREX, Berlin) and filtered through a 0.22 μm syringe filter (Millex-GV, PVDF, Merck Millipore, Darmstadt, Germany). The stock solutions of the standards were further diluted with methanol to prepare a series of working concentrations (0.5, 1.0, 2.0, 2.5, 5.0, 10.0, 25.0, 50.0, 100, 125, 200, 250, 500, 1000) × 10 -3 mg/mL for linearity and calibration. All stock solutions were stored at -20°C until LC-MS analysis. A triple-quadrupole linear ion trap mass spectrometer (API 4000 QTRAP TM MS/MS system from AB Sciex, Concord, ON, Canada) equipped with an electrospray (Turbo VTM) ion source was set up in a negative ionization mode. Source dependent parameters such as ion spray voltage, turbo spray temperature (TEM), nebulizer gas (GS 1), heater gas (GS 2), and curtain (CUR) gas were set at 5500 V, 550°C, 50 psi, 50 psi, and 20 psi, respectively. The collision-activated dissociation gas was set to be medium and the interface heater was on. Nitrogen was used in GS 1, GS 2, and CUR. Simultaneous quantitative analysis was performed using the MRM mode and its conditions were optimized for each compound during infusion. For full scan ESI-MS and MS/MS, the spectra scan range was m/z = 100-1000 at unit resolution. Optimization of the mass spectrometric conditions was carried out by infusing 50 × 10 -6 mg/mL solutions of the analytes dissolved in methanol at 10 μL/ min using a Harvard "22" syringe pump (Harvard Apparatus, South Natick, MA, USA). The transitions and optimized compound-specific dependent MRM parameters (declustering potential (DP), entrance potential (EP), collision energy (CE), and cell exit potential (CEP)) for each analyte are shown in Table 2.

Statistical analyses
The contents of the bioactive compounds obtained from three repeats (n = 3) of all samples were used for PCA on STATISTICA software, Windows version 7.0 (Stat Soft, Inc., USA).

Optimization of UPLC conditions
UPLC conditions were optimized to obtain a satisfactory separation within a short run time. Different mobile phases, such as water/methanol, water/acetonitrile, 0.1% formic acid in water/ methanol, and 0.1% formic acid in water/acetonitrile, at different flow rates (0.1, 0.15, 0.20, 0.25, 0.30, 0.35, and 0.4 mL/min) and different column temperatures (25,30,40, and 50°C) were tested. Similarly, different concentrations of formic acid (0.1, 0.2, 0.3, and 0.4% v/v) were also tested to improve the peak shape and restrain peak tailing. Satisfactory chromatographic separation was achieved with 0.1% formic acid in water and acetonitrile as the mobile phase at a flow rate of 0.25 mL/min at 25°C using ACQUITY UPLC BEH™ C18 column (1.7 μm, 2.1 mm × 50 mm) within a run time of 10 min.

Optimization of MS conditions
Each targeted analyte was infused into a mass spectrometer, and the mass spectra were studied in the negative ionization mode.

Method validation
The linearity, limits of detection (LOD), limits of quantitation (LOQ), precision (inter-day and intraday precision), stability, and accuracy were validated according to the guidelines of the International Conference on Harmonization (ICH). [32] The 2.50 × 10 -4 mg/mL stock solution was used in method development and validation.

Linearity, LOD, and LOQ
The calibration curves were constructed using more than five data points within the measured concentration range from 0.5 to 1000 × 10 -6 mg/mL using a weight (1/x 2 ) factor by the least-squares linear regression. The linear equation y = ax+b was used in the construction of the calibration curve, where "y" is the peak area ratio, "x" is the concentration of the analyte, "a" is the slope of the curve, and "b" is the intercept. The results showed linearity with correlation coefficients (R 2 ) from 0.9982 to 1.0000 within the test ranges. The LODs and LOQs were calculated with a signal-to-noise (S/N) = 3 and 10, respectively, as the criteria, and the calculated values were found to be 8.60 × 10 -10 −2.18 × 10 -6 mg/mL (LOD) and 2.60 × 10 -9 -6.63 × 10 -6 mg/mL (LOQ) ( Table 3).

Precision, stability, and recovery
Intra-day and inter-day experiments selected for the precision of the developed method was investigated by three replicates during a single day and by duplicating the experiments on six consecutive days. The percentage relative standard deviations (%RSD) of the peak area for all compounds ranged from 0.25 to 2.96% and 0.17 to 2.89%, respectively. Similarly, the stability of sample solutions stored at room temperature was investigated by replicate injections of the sample solution at 0, 2, 4, 8, 12, and 24 h. The %RSD values of the stability ranged from 0.45 to 3.45% (Table 3).
To evaluate the accuracy of this method, a recovery test was applied by the standard addition method. The mixed standard solutions with three different spike levels (low (1.00 × 10 -4 mg/ mL), middle (2.00 × 10 -4 mg/mL), and high (3.00 × 10 -4 mg/mL) were added into a sample and analyzed using the above method in triplicate experiments. The recovery was calculated by the formula: recovery = (a-b)/c × 100%, where "a" is the detected amount, "b" is the original amount, and "c" is the spiked amount. All analytes showed recoveries in the range of 95.78 −101.44% (RSD ≤ 3.25%). The results showed that the developed UPLC-ESI-QqQ LIT -MS/MS method was simple, reliable, and reproducible ( Table 3).

Fruits
Among fruits, the highest content of quinic acid was detected in F. virens (48.283 mg/g) followed by F. benjamina (43.067 mg/g) and F. lyrata (27.271 mg/g). Similarly, a high quantity of rutin was found in fruits of F. palmata subsp. virgata (72.337 mg/g) and F. benjamina (63.147 mg/g) followed by F. hispida (8.601 mg/g) and F. virens (8.020 mg/g). A significant content of chlorogenic acid was detected in F. benjamina (5.63 mg/g) followed by F. lyrata (1.38 mg/g). The fruits of F. benjamina showed the highest total content of all compounds (114.797 mg/g) followed by F. palmata subsp. virgata (77.139 mg/g) and F. virens (58.921 mg/g).

Principal component analysis
PCA is a chemometric method which discriminates variables and display differences among the samples. The PCA was run on the basis of the contents of 19 targeted bioactive compounds obtained   Fig. 4(A)). The loading factor of the first factor indicated a correlation among chlorogenic acid (0.848), catechin (0.827), and quercetin (0.719), whereas factor 2 has a correlation among quinic acid (0.840), oleanolic (0.762) acid, and kaempferol (−0.760). As a result, F. virens, F. lyrata, and F. racemosa were closer to each other, whereas F. hispida and F. palmata subsp. virgata were located in another quadrate. F. benghalensis was located in the borderline while F. benjamina and F. religiosa were far apart from other species. The PCA plot (Fig. 5(A)) of edible species showed that F. hispida, F. racemosa, F. benghalensis, and F. benjamina were far apart from each other. The non-edible species F. lyrata and F. virens were similar, hence located together, and F. palmata subsp. virgata and F. religiosa were in different quadrates (Fig. 5(B)). In leaves, PC1 and PC2 explained 41.91% and 23.09% variation which together explained 65.00%. Catechin (0.980), epicatechin (0.928), betulinic acid (0.798), and oleanolic acid (0.856) have a correlation in factor 1, whereas quinic acid (−0.756) and chlorogenic acid (−0.830) have a correlation  in factor 2. The PCA plot showed F. virens, F. racemosa, and F. palmata subsp. virgata located in the same quadrate. Similarly, F. hispida and F. religiosa were located in the same quadrate, whereas F. lyrata and F. benjamina were far apart from each other (Fig. 4(B)). The edible species F. virens, F. hispida, and F. racemosa were further analyzed in PCA which showed all edible species located falling far apart from each other (Fig. 5(C)). Similarly, PCA was further run on leaves of non-edible species (F. benghalensis, F. palmata subsp. virgata, F. lyrata, F. religiosa, and F. benjamina), and the results showed that F. benghalensis and F. palmata subsp. virgata were closer to each other while others remained far apart (Fig. 5(D)).

Discussion
In the present study, the UPLC-ESI-QqQ LIT -MS/MS method in the MRM mode was used to identify and quantify 19 bioactive compounds simultaneously in a single run from fruits, leaves, and barks of eight Ficus species namely F. religiosa L., F. benghalensis L., F. palmata subsp. virgata (Roxb.) Browicz, F. virens Aiton, F. lyrata Warb., F. benjamina L., F. hispida L. f., and F. racemosa L (Table 4). Hispidine, 7-hydroxycoumaric acid, and chrysin were detected in very small amounts while quinic acid and chlorogenic acid were found in high quantity in all the plant parts of all the studied species. A significant content of chlorogenic acid was detected in fruits (0.032-5.63 mg/g), leaves (0.053-4.857 mg/g), and barks (0.064-14.182 mg/g) as compared to the previously reported quantity in F. carica. [11] Catechin was also observed in higher quantity ranges of 0.041-0.262, 0.011-1.977, and 0.039-10.137 mg/g in fruits, leaves, and barks, respectively, than the previous report. [11] Epicatechin and rutin were also detected in high quantity ( Table 4). Many of the selected Ficus species have shown a high content of bioactive compounds than the other published reports on Ficus species. The HPLC method reported by Taskeen et al. [25] showed a very high amount of flavonoids in F. bhengalensis as compared to F. religiosa. The UPLC-ESI-QqQ LIT -MS/MS method in the MRM mode is rapid, sensitive, precise, efficient, and reproducible as compared to HPLC and LC-MS methods reported earlier. [3,6,8,9,[19][20][21][22][23] According to Yin-Xian et al., [7] the antioxidant property of crude extracts of Ficus species is due to the total phenolic and flavonoid contents. Therefore, F. benjamina fruits and leaves, which are edible, can be highly valuable for health due to the high total content of phenolics and flavonoids. Similarly, barks of F. benghalensis and F. virens can be utilized for medicinal purposes in comparison to other tested species on the basis of the high content of selected phytochemicals. PCA was successfully used to discriminate among the Ficus species on the basis of their chemical content. Figure 5. PC1 vs. PC2 plot showing discrimination between edible and non-edible fruits and leaves on the basis of 15 bioactive compounds (quinic acid, protocatechuic acid, chlorogenic acid, mangiferin, catechin, epicatechin, rutin, hispidine, 7-hydroxycoumaric acid, ferulic acid, vanillic acid, luteolin, quercetin, apigenin, kaempferol, chrysin, betulinic acid, ursolic acid, and oleanolic acid).

Conclusion
A rapid and sensitive UPLC-ESI-QqQ LIT -MS/MS method was developed and validated as per ICH guidelines for the simultaneous quantification of 19 bioactive compounds in a single run from different parts of eight Ficus species. It is a high throughput and more advance method for the identification and quantitation of chemical constituents. The results indicated that the total content of all the 19 compounds was high in F. virens barks followed by leaves and fruits of F. benjamina. All these selected Ficus species as well as edible and non-edible fruits and leaves were successfully discriminated using PCA. The quantitative results also indicated the importance of plant parts. The present method can be used for the quantitative analysis of these compounds and quality control of Ficus species.