Journal of Pharmaceutical and Biomedical Analysis 168 (2019) 1–12 Contents lists available at ScienceDirect Journal of Pharmaceutical and Biomedical Analysis journal homepage: www.elsevier.com/locate/jpba Phytochemical profile of Syzygium formosum (Wall.) Masam leaves using HPLC–PDA–MS/MS and a simple HPLC–ELSD method for quality control Thi Phuong Duyen Vu, Trong Quan Khong, Thi Minh Nguyet Nguyen, Young Ho Kim, Jong Seong Kang ∗ College of Pharmacy, Chungnam National University, Daejeon 3414, South Korea a r t i c l e i n f o Article history: Received 29 November 2018 Received in revised form 23 January 2019 Accepted February 2019 Available online 11 February 2019 Keywords: Eugenia Formosa HPLC–ELSD LC–MS Optimized extraction Syzygium formosum a b s t r a c t Syzygium formosum (SF) leaves have been used for thousands of years in traditional Vietnamese medicine for the treatment of allergy or skin rash However, the role of the phytochemical profile of SF leaves on their activities is poorly understood Additionally, there is currently no quality control method for this herbal material, which is required by the pharmaceutical industry Therefore, this study aimed to investigate chemical profile of SF leaves using high–performance liquid chromatography–photodiode array–tandem mass spectrometry (HPLC–PDA–ESI-MS/MS) and establish a simple and efficient HPLC method for controlling major bioactive compounds The characterization of 28 components, including eleven flavonoids, thirteen triterpene acids, and four phenolic acids, was achieved on the basis of their maximum ultraviolet wavelength values and MS fragmentation pathways An HPLC–evaporate light scattering detector (HPLC–ELSD) method over 35 of analysis time for quality control of SF leaves was proposed Using response surface methodology based on a Box–Behnken design and Derringer’s desirability function, the optimal conditions for extracting the main bioactive compounds in SF leaves were determined The content of marker compounds within SF leaves decreased in the order asiatic acid > corosolic acid > betulinic acid > maslinic acid The developed HPLC–ELSD method is appropriate for quality control testing of SF leaves © 2019 Elsevier B.V All rights reserved Introduction Syzygium formosum (Wall.) Masam is an evergreen tree of the Myrtaceae family (also known under the synonym Eugenia formosa) that is widely distributed across Southeast Asia, India, Taiwan, and Brazil This multipurpose plant is ingested as food or tea and used in traditional Asian medicines In Vietnam, the leaves are typically used in traditional medicine for the treatment of skin rash, allergy, scabies, sore throat, bronchitis, and cystitis [1,2] S formosum (SF) leaf extract has pharmacological activities such as antibacterial activity, especially toward Staphylococcus aureus and Escherichia coli [3,4], anti-allergic, anti-inflammatory [5], and antioxidant activity [6,7], and Moloney murine leukemia virus inhibition [8] Despite of the medicinal applications of the plant, only few studies on the constituents of SF leaves have been reported For instance, ∗ Corresponding author E-mail address: kangjss@cnu.ac.kr (J.S Kang) https://doi.org/10.1016/j.jpba.2019.02.014 0731-7085/© 2019 Elsevier B.V All rights reserved several studies have reported that SF leaves contain flavonoids and triterpenoid acids [4–6] However, the detailed composition of SF leaves, which is very important for understanding their pharmacological mechanism, is largely unknown Furthermore, no method for the quality control of SF leaves has been reported to date Therefore, the objectives of our study were to investigate the phytochemicals in SF leaves and establish a simple and precise method for the quality control of SF leaves in the market Due to the great diversity of the constituents in herbal samples and unavailability of reference compounds, liquid chromatography-tandem mass spectrometry (LC–MS/MS) is currently considered as the easiest and simplest technique for phytochemical profiling This system couples the powerful separating ability of LC with the strong identification ability of MS [9] Therefore, this study utilized a high-performance liquid chromatography–photodiode array detector coupled with a tandem mass spectrometer (HPLC–PDA–ESI-MS/MS) to establish the chemical profile of SF leaves, from which 28 compounds were identified or tentatively characterized Thereafter, major bioactive constituents in SF leaves, four triterpenoids, were determined and T.P Duyen Vu, T Quan Khong, T.M Nguyet Nguyen, et al / Journal of Pharmaceutical and Biomedical Analysis 168 (2019) 1–12 utilized as marker compounds to develop an analytical method using HPLC equipped with an evaporative light scattering detector (HPLC–ELSD) Finally, a multiresponse optimization of the extraction method was carried out which is very helpful for pharmacological-testing researches and pharmaceutical companies that use SF leaves as raw material 0.1% (v/v) formic acid in water (solvent A) and 0.1% (v/v) formic acid in acetonitrile (solvent B) A linear gradient elution program from 5% to 100% of solution B over 130 with a flow rate of 0.5 mL/min was run for the separation By comparing the retention times, absorption spectra, and mass spectra to those of the standards and literature data, 28 compounds were identified or tentatively characterized Method and materials 2.4 HPLC–ELSD analysis 2.1 Chemicals and materials HPLC-grade acetonitrile, ethanol, and ethyl acetate were purchased from Burdick and Jackson (USA), and formic acid was obtained from Sigma-Aldrich (USA) Distilled water was prepared using a Milli-Q system (Millipore, USA) Gallic acid, catechin, isoquercitrin, quercitrin, maslinic acid, oleanolic acid, and betulinic acid were purchased from Sigma-Aldrich (USA) Methyl gallate, avicularin, quercetin, asiatic acid, corosolic acid, and ursolic acid were provided by the Natural Laboratory, College of Pharmacy, Chungnam National University (Daejeon, South Korea) The purities of all standard compounds exceeded 95% SF leaves collected from Hanoi, Vietnam, in September 2017 were identified by Dr Bui (Vietnam Academy of Science and Technology, Hanoi, Vietnam) The leaves were dried for 48 h at 45◦ in an oven The powdered sample was stored at 4◦ until analysis 2.2 Preparation of the samples and standard solutions For the profiling analysis of SF leaves, their bioactive compounds were enriched using liquid-liquid-extraction method In brief, 1.50 g of sample powder was extracted with 10 mL of ethanol (70%, v/v) by ultrasonication (45 kHz, 60 min) The solution was centrifuged for at 3000 rpm The supernatant was then concentrated in a rotary vacuum at 50◦ to yield the dry crude extract The obtained extract was suspended in 10.0 mL of water and subsequently partitioned three times with ethyl acetate (10.0 mL) The ethyl acetate layers were combined and evaporated to dryness in the rotary vacuum apparatus The residue was then dissolved in ethanol (2.0 mL) and filtered through a 0.22-m PVDF filter before LC–MS analysis To develop the HPLC–ELSD method, 1.0 g of sample powder was combined with 10 mL of ethanol (70%, v/v) and was extracted ultrasonically at room temperature for 60 The solution was centrifuged for at 3000 rpm and the supernatant was filtered through a 0.22-m PVDF filter before analysis Standard stock solutions were prepared by dissolving 1.0 mg of each standard in 1.0 mL of ethanol These solutions were then diluted 10-fold before LC–MS analysis For quantitative analysis, a mixed standard stock solution containing asiatic acid, maslinic acid, corosolic acid, and betulinic acid was prepared in ethanol The mixed standard stock solution was then diluted appropriately for validation analysis All solutions were stored in brown bottles at 4◦ until analysis 2.3 HPLC–PDA–MS/MS analysis An LCMS-8040 system (Shimadzu, Japan) with scan molecular mass from 100 to 1000 Da was utilized for peak detection of compounds An electrospray ionization (ESI) with an interface voltage of –3.5 kV for the negative mode and 4.5 kV for the positive mode under a desolvation line temperature of 250◦ were used A production survey scan was performed in both positive and negative mode Other conditions included a heating block temperature of 400◦ , a nebulizing gas flow rate of L/min, and a drying gas flow rate of 15 L/min The separation was performed using a Hector C18 column (250 × 4.6 mm, m; RStech, Korea) The mobile phase contained The quantitative analysis was performed over 35 on a Sedex 55 ELSD combined HPLC-20AD series system The stationary and mobile phases were the same as for the LC–MS analysis Detector gain was set at Column temperature was maintained at 30◦ To optimize the chromatographic separation, the proportion of eluent, the flow rate of the mobile phase (0.3, 0.5, 0.7, and 1.0 mL/min), the drift tube temperature (50, 60, 70, and 80 ◦ ), and the pressure of the nebulizer gas (N2 ) (1.0, 1.5, and 2.0 bar) were studied 2.5 Method validation The developed HPLC–ELSD method was validated using the metrics of linearity, limit of detection (LOD), limit of quantitation (LOQ), precision, accuracy, and stability Linearity was verified using correlation of determination after analyzing five concentrations of mixed standard solution in triplicate The LOD and LOQ were estimated at signal-to-noise ratios of and 10, respectively Intraday precision and accuracy were calculated by analyzing three concentrations (low, medium, and high) of the mixed standard solution five times over a single day An analysis carried out over five consecutive days established the interday variability The stability of standard stock solutions was tested at 24 h after they were initially made and expressed by RSD 2.6 Optimization of the extraction method The extraction method to maximize the amount of active compounds is an important step in pharmacological testing and manufacturing In this study, preliminary experiments were conducted using the “one-factor-at-a-time” methodology to identify suitable levels of the studied factors All experiments were carried out using ultrasonication and were conducted in duplicate First, four extraction times (30, 60, 120 and 180 min) were investigated using ethanol 70% (v/v) as the solvent and a solid-to-solvent ratio of 1:10 Then, the solid-to-solvent ratio (1:5, 1:10, 1:15, and 1:20) was examined using ethanol 70% (v/v) as the solvent over 60 Finally, the organic solvent concentration (30, 50, 70, and 100%) for extracting samples was determined, using an extraction time of 60 and solid-to-solvent ratio of 1:10 The total content of the four analytes was used as the response to establish the suitable range between levels of each factor After determining the levels for the extraction factors, the effects of three factors (extraction time (X1 ), solvent-to-solid ratio (X2 ), and ethanol concentration (X3 )) were studied using a Box–Behnken design (BBD) and response surface methodology (RSM) Each factor was investigated at three coded levels All BBD experiments were conducted in duplicate The results were then analyzed by quadratic multiple regression The predicted responses were calculated as: Y = ˇo + ˇi Xi + ˇij Xi Xj + ˇii Xi2 (i = 1, 2, , k) where Y represents the response (concentration of analyte, mg/g) and ˇo , ˇi , ˇij , and ˇii are the constant coefficient, linear coefficient, interaction coefficient, and quadratic coefficient, respectively [10] Derringer’s desirability function was used to identify a compromise T.P Duyen Vu, T Quan Khong, T.M Nguyet Nguyen, et al / Journal of Pharmaceutical and Biomedical Analysis 168 (2019) 1–12 solution for the multiple response optimization First, a desirability function (di ) for each individual response was created according to the following equation: d= ⎧ if Y < ⎪ ⎪ ⎨ Y − L s ⎪ ⎪ ⎩ T −L if L ≤ Y ≤ T if Y > T Second, the overall desirability (D) was obtained using the following equation: D= m d1 d2 · · ·dm where m is number of responses in the optimization procedure [11] The experimental design and the calculation of predicted results were conducted using Design-Expert software (version 11; StatEase, USA) Results and discussion 3.1 Optimization of the HPLC conditions A full-scan method from 5% to 100% of solution B over 130 using the LC-PDA-MS/MS system was applied to detect compounds in SF leaves Peak separation was not achieved when watermethanol was used as the mobile phase Therefore, acetonitrile, which has a lower ultraviolet (UV) cut off and thus can be used to analyze compounds with low UV absorptions, was used as an alternative mobile phase Furthermore, due to its higher elution strength compared with methanol, the total analysis time when using acetonitrile as the mobile phase was reduced The column pressure also decreased To improve peaks shape and symmetry, 0.1% (v/v) of formic acid was added because of its compatibility with both the MS and the ELSD detector To develop a routine quality control method of SF leaves, four major triterpenoids, including asiatic acid, maslinic acid, corosolic acid, and betulinic acid, were selected as marker compounds Because of their poor UV chromophores, they can only be detected at very low wavelengths (e.g., 205 or 210 nm), when using HPLC-UV method Under such circumstances, analytical results are seriously affected by the buffer and mobile phase solvents that also absorb at these wavelengths, such as formic acid and methanol, with UV cutoffs at 210 and 205 nm, respectively [12], especially under gradient conditions At triterpenoid-detecting wavelengths, buffers and mobile phase solvents also absorb light, leading to high and fluctuating baseline signals and low detection sensitivity [13] Indeed, in an experimental work, Fang et al [14] proved that ELSD produced a lower quantitation error compared with UV in analyzing several specific standards using a single calibration curve For these reasons, a method for the routine quality control of SF leaves using triterpenes as marker compounds was developed based on ELSD as the optimal detector The mobile phase and stationary phase were the same as for the LC–MS analysis The drift tube temperature and gas flow rate were optimized to obtain good peak symmetries and efficiencies, as well as a low baseline The mobile phase flow rate and proportion of eluent were also investigated A drift tube temperature of 80 ◦ and a gas flow rate of 2.0 bar were the final optimized conditions The elution program was 75% B at 0–10 min, 75–100% B at 10–25 min, and 100% B at 25–35 with a flow rate of 0.7 mL/min (Fig 4) 3.2 Phytochemical profile of SF leaves by HPLC–PDA–MS/MS The components in the SF leaves were tentatively identified by comparing their elution order, UV spectra, and MS spectra with those in the available literature of compounds found in the same genus For further confirmation, several reference standards were processed and analyzed using LC–PDA–MS/MS The results were then compared with those of the identified compounds in the sample Table shows the characteristics of the identified compounds in the SF leaves In total, 28 compounds (eleven flavonoids, thirteen triterpene acids, and four phenolic acids) were detected within the 130-min elution period (Fig 1), most of which were previously reported in Syzygium species The chemical structures and the proposed fragmentation pathways of the 28 identified compounds are shown in Fig and Fig 3, respectively In the full-scan MS spectra, the deprotonated molecular ions [M−H]– and protonated molecular ions [M+H]+ had the most prominent mass-to-charge ratios for most 28 compounds (Figure S1) Therefore, these ions were chosen as the precursor ions for tentative identification of the constituents of SF leaves using the product ion scan mode Three benzoic acid derivatives corresponding to negative molecular ions at m/z 169, 153, and 183 and product ions at m/z 125, 109, and 125 by losing carbonyl or methyl carbonyl groups were assigned as gallic acid (peak 1), protocatechuic acid (peak 2), and methyl gallate (peak 4), respectively Peak 3, which showed a negative molecular ion at m/z 289, positive molecular ion at m/z 291, and fragments at m/z 109 and 137, was deduced to be catechin or epicatechin By comparison with standards, this peak was confirmed to be catechin From peak to peak 13, in MS2 , all product ions exhibited fragments at m/z 319, 303, or 287, which corresponded to protonated myricetin, quercetin, and kaempferol aglycones, respectively Therefore, all of these compounds were identified as myricetin, quercetin, or kaempferol glycosides Peaks 5, 6, and 12 showed positive molecular ions at m/z 481, 465, and 507, which then lost a glucosyl, rhamnosyl, and acetyl-rhamnosyl moiety, respectively, to produce the same protonated myricetin aglycone Therefore, these peaks, were assigned as myricetin glucoside, myricetin rhamnoside, and myricetin acetyl-rhamnoside, respectively, which have been previously reported from Syzygium formosum [6] and Syzygium cumini [15] Similarly, the protonated ion of peak lost a glucosyl/galactosyl moiety, peak 8, 9, and 10 lost a pentosyl group, and peak 11 lost a rhamnosyl unit, to yield the same quercetin aglycone Based on the comparison with standard compounds, these peaks were identified as isoquercitrin, quercetin pentoside (pyran/furan ring), avicularin, and quercitrin, respectively Similarly, peak 13 was attributed to kaempferol pentoside (arabinoside or xyloside) Peak 14, which presented a [M+H]+ at m/z 303 with fragments at m/z 153 and 137, was assigned as quercetin, and was confirmed by comparison with standard Peak 15, which showed a [M−H]– at m/z 343 and a [M+H]+ at m/z 345 that yielded fragments at m/z 315 and 287 due to the loss of one and two methoxy groups, respectively, was tentatively identified as trimethoxyellagic acid, a compound previously isolated from Syzygium aromaticum [16] During negative ion electrospray mass spectrometry, peaks 16, 17, and 18 all formed [M−H]– and [M + HCOO]– ions at m/z 487 and 533, respectively Furthermore, during fragmentation, their deprotonated molecular ions lost a formic acid molecule and a water molecule to produce fragment at m/z 423 These compounds were all determined as triterpene acids with an empirical formula of C30 H48 O5 By comparison with the standard, peak 18 was confirmed to be asiatic acid which had been previously isolated from Syzygium formosum [4], while the two remain peaks were likely positional isomers of asiatic acid Similarly, peaks 23, 24, and 25 presented ions [M+H]+ , [M–H2 O+H]+ , [M–HCO2 H–H2 O+H]+ , [M−H]– and [M + HCOO]– at m/z 473, 455, 409, 471, and 517, respectively By comparison with literature data, these peaks were attributed to isomers of maslinic acid or corosolic acid, which share the same molecular weight of 472 Da Using authentic standards and the fragment pathways, peaks 24 and 25 were confirmed as Peak no RT (min) max (nm) Parent ion (m/z) Product ions (m/z) Molecular formula Identification 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 13.69 20.20 24.89 26.02 31.08 34.44 35.11 36.84 38.00 38.33 38.82 40.98 41.99 51.24 65.28 70.99 73.49 75.49 81.79 82.23 93.82 94.36 98.61 101.11 102.57 121.30 124.14 124.68 213, 270 259, 292 278 213, 271 255, 356 258, 349 249, 357 254, 354 254, 354 255, 355 253, 348 256, 347 257, 348 250, 370 247, 372 204 204 204 204 204 311, 204 311, 204 204 204 204 204 204 204 169 [M–H]– 153 [M–H]– 289 [M–H]– 183 [M–H]– 481 [M+H]+ 465 [M+H]+ 465 [M+H]+ 435 [M+H]+ 435 [M+H]+ 435 [M+H]+ 449 [M+H]+ 507 [M+H]+ 419 [M+H]+ 303 [M+H]+ 345 [M+H]+ 487 [M–H]– 487 [M–H]– 487 [M–H]– 485 [M–H]– 485 [M–H]– 633 [M–H]– 635 [M+H]+ 633 [M–H]– 635 [M+H]+ 473 [M+H]+ 473 [M+H]+ 473 [M+H]+ 455 [M–H]– 457 [M+H]+ 457 [M+H]+ 125 [M–COOH]–, 79, 97 109 [M–COOH]– 109, 137 125 [M–COOCH3 ]– 319 [M–glucose+H]+ , 153 319 [M–rhamnose+H]+ , 153 303 [M–glucose+H]+ , 153 303 [M–xylose+H]+ , 153 303 [M–xylose+H]+ , 153 303 [M–arabinose+H]+ , 153 303 [M–rhamnose+H]+ 319 [M–acetyl rhamnose+H]+ , 153 287 [M–pentose+H]+ , 153, 121 153, 137 315 [M–OCH3 ]+ , 287 [M–2OCH3 ]+ 487 [M–H]–, 423 [M–HCO2 H–H2 O–H]– 487 [M–H]–, 423 [M–HCO2 H–H2 O–H]– 487 [M–H]–, 423 [M–HCO2 H–H2 O–H]– 485 [M–H]–, 439 [M–HCO2 H–H]– 485 [M–H]–, 439 [M–HCO2 H–H]– 633 [M–H]–, 145 [coumaroyl–H]– 147 [coumaroyl+H]+ 633 [M–H]–, 145 [coumaroyl–H]– 147 [coumaroyl+H]+ n.d 391 [M–HCO2 H–2H2 O+H]+ 391 [M–HCO2 H–2H2 O+H]+ 455 [M–H]– 411 [M–HCO2 H+H]+ , 439 [M–H2 O+H]+ 411 [M–HCO2 H+H]+ , 439 [M–H2 O+H]+ C7 H6 O5 C7 H6 O4 C15 H14 O6 C8 H8 O5 C21 H20 O13 C21 H20 O12 C21 H20 O12 C20 H18 O11 C20 H18 O11 C20 H18 O11 C21 H20 O11 C23 H22 O13 C20 H18 O10 C15 H10 O7 C17 H12 O8 C30 H48 O5 C30 H48 O5 C30 H48 O5 C30 H46 O5 C30 H46 O5 C39 H54 O7 C39 H54 O7 C30 H48 O4 C30 H48 O4 C30 H48 O4 C30 H48 O3 C30 H48 O3 C30 H48 O3 Gallic acida Protocatechuic acidb Catechina,b Methyl gallatea,b Myricetin glucosideb Myricetin rhamnoside Quercetin glucoside (isoquercitrin)a Quercetin pentosideb Quercetin pentosideb Quercetin arabinofuranoside (avicularin)a,b Quercetin rhamnoside (quercitrin)a,b Myricetin acetyl rhamnosideb Kaempferol pentosideb Quercetina Trimethoxyellagic acidb Positional isomer of asiatic acidb Positional isomer of asiatic acidb Asiatic acida Dehydrogenated isomer of asiatic acidb Dehydrogenated isomer of asiatic acidb Positional isomer of coumaroyl asiatic acidb Positional isomer of coumaroyl asiatic acidb Isomer of maslinic acid/corosolic acid Maslinic acida Corosolic acida Betulinic acida Oleanolic acida Ursolic acida n.d.: not detected a These compounds were compared with authentic standards b These compounds are the first to be reported in Syzygium formosum leaves T.P Duyen Vu, T Quan Khong, T.M Nguyet Nguyen, et al / Journal of Pharmaceutical and Biomedical Analysis 168 (2019) 1–12 Table List of compounds detected in Syzygium formosum leaves by HPLC-PDA-ESI-MS/MS T.P Duyen Vu, T Quan Khong, T.M Nguyet Nguyen, et al / Journal of Pharmaceutical and Biomedical Analysis 168 (2019) 1–12 Fig HPLC–PDA chromatogram at 220 nm and total ion chromatograms (TIC) in the multiple reaction monitoring (MRM) mode of detected compounds in Syzygium formosum leaves (ethyl acetate fraction) (Compounds 1, 3, 4, 7, 10, 11, 14, 18, 24, 25, 26, 27, and 28 were compared with standards) 6 T.P Duyen Vu, T Quan Khong, T.M Nguyet Nguyen, et al / Journal of Pharmaceutical and Biomedical Analysis 168 (2019) 1–12 Fig Structures of compounds detected in Syzygium formosum leaves T.P Duyen Vu, T Quan Khong, T.M Nguyet Nguyen, et al / Journal of Pharmaceutical and Biomedical Analysis 168 (2019) 1–12 maslinic acid and corosolic acid, respectively Peak 23 was likely an other isomer of maslinic acid or corosolic acid In the same manner, peaks 26, 27, and 28 were identified as betulinic acid, oleanolic acid, and ursolic acid Both protonated molecular ions of oleanolic acid and ursolic acid fragmented at the carboxylic position and lost a water unit to produce fragments at m/z 411 and 439, respectively Betulinic acid, a pentacyclic lupane-type triterpene, otherwise, was too stable to produce product ions during fragmentation, which resulted in the unchanged fragment with a precursor ion at m/z 455 Peaks 19 and 20 presented molecular ions [M+H]+ at m/z 487 and [M−H]– at m/z 485 with carboxylmoiety-loss fragment [M-H-HCO2 H]− at m/z 439 These compounds corresponded to triterpene acids with the common empirical formula C30 H46 O5 Compared with above triterpene acids reported for SF leaves, compounds 19 and 20 had lower molecular weights than asiatic acid (2 Da) Hence, they were likely dehydrogenated isomers of asiatic acid Peaks 21 and 22 presented ions [M+H]+ at m/z 635 and [M–coumaroyloxy−OH]+ at m/z 455 in the positive scan mode and ions [M−H]– at m/z 633 and [M–coumaroyloxy–H]– at m/z 471 in the negative scan mode Furthermore, during fragmentation, the protonated molecular ion and deprotonated molecular ion produced fragments at m/z 147 and 145, which corresponded to protonated coumaroyl and deprotonated coumaroyl, respectively Therefore, these compounds were suggested as esters of coumaric acid with the empirical formula C39 H54 O7 , a combination of coumaric acid C9 H8 O3 and a triterpene acid C30 H48 O5 Comparing with triterpene acids reported for SF leaves, C30 H48 O5 was likely isomer of asiatic acid Therefore, peak 21 and 22 were tentatively assigned as isomers of coumaroyl asiatic acid This was also confirmed by the maximum UV wavelength (max ) observed at 311 nm of peak 21 and 22 due to the coumaric functionality [17] This is the first time coumaroyl triterpene acids were reported from Syzygium genus The fragmentation pathway is shown in Fig In brief, among 28 identified compounds, 17 compounds were the first to be reported in SF leaves Additionally, 13 compounds were verified by comparing their retention times, mass spectra, and UV spectra with authentic standards (Fig 1) Next, we validated the HPLC–ELSD method for quality control of SF leaves using four triterpenoid standards as markers 3.3 Validation of the HPLC–ELSD method Four major bioactive components, i.e., asiatic acid, maslinic acid, corosolic acid, and betulinic acid, were used as marker compounds to develop an analytical method, which were detected in SF leaves using HPLC–ELSD The bioactivities of these four constituents have been reported in several pharmacological studies Asiatic acid demonstrates analgesic and anti-inflammatory effects in fulminant hepatitis and acute lung injury models [18–20] Inhibitory effects of asiatic acid against fungi and bacteria, including S aureus, E coli, Enterococcus faecalis, and Candida albicans, have also been reported [21,22] Anti-inflammatory and antibacteria activities of maslinic acid, corosolic acid, betulinic acid have also been reported recently [23–27] Among them, corosolic acid was the most active against S aureus, while maslinic acid was the most active against E faecalis [24] Remarkably, all four compounds exhibit strong anti- Fig Proposed fragmentation pathways of compounds identified in Syzygium formosum leaves 8 T.P Duyen Vu, T Quan Khong, T.M Nguyet Nguyen, et al / Journal of Pharmaceutical and Biomedical Analysis 168 (2019) 1–12 Fig (Continued) inflammatory effects, explaining the traditional use of SF leaves for the treatment of allergy-related diseases in Vietnam Furthermore, as the most predominant components in SF leaves, these components are potentially important marker compounds for the quality control of SF leaves and related products on the market Quality control of herbal samples using multiple pharmacological active compounds is widely practiced, which explains the selection of these components as marker compounds for developing the analytical method The proposed HPLC–ELSD method for quality control of SF leaves was validated using the metrics of linearity, LOD, LOQ, intraday/interday precision, accuracy and stability (Table 2) The calibration curves were linear with r2 values of at least 0.9990 over the investigated ranges of 0.3–0.9 mg/mL for asiatic acid, 0.067–0.2 mg/mL for maslinic acid, and 0.133–0.4 mg/mL for corosolic acid and betulinic acid The LODs and LOQs of the four marker compounds were less than 0.02 and 0.05 mg/mL, respectively, which proved that the developed method enabled the sensitive qualitative and quantitative analysis of all four analytes The accuracies for the four marker compounds varied from 100.3% to 103.7% within a single day and 99.5% to 107.6% over sequential days Meanwhile, the precisions ranged from 1.1% to 4.3%, which satisfied the Method Validation Guidelines of the Korean Ministry of Food and Drug Safety [28] The standard stock solutions of all four marker compounds were also stable over 24 h as indicated by a low RSD value (