Characterization and Quantification by LC-MS/MS of the Chemical Components of the Heating Products of the Flavonoids Extract in Pollen Typhae for Transformation Rule Exploration

The Traditional Chinese Medicine herbs Pollen Typhae and Pollen Typhae Carbonisatus have been used as a hemostatic medicine promoting blood clotting for thousands of years. In this study, a reliable, highly sensitive method based on LC-MS/MS has been developed for differentiation of the heating products of total flavonoids in Pollen Typhae (FPT-N). Twenty three peaks were detected and 18 peaks have been structurally identified by comparing retention times, high resolution mass spectrometry data, and fragment ions with those of the reference substances and/or literature data. Additionally, 15 compounds have been quantified by multiple reaction monitoring in the negative ionization mode. It was found that the contents of the characterized compounds differed greatly from each other in FPT-N samples. Among them, the content of huaicarbon B significantly increased at first, while it decreased after heating for 25 min, which could be considered as the characteristic component for distinguishing FPT-N. The present study provided an approach to rapidly distinguish the differences of FPT-N samples. In addition, the actively summarized characteristic fragmentation might help deducing the structure of unknown flavonols compounds. Furthermore, transformation rules of flavonoids during the heating process in carbonisatus development could contribute to hemostatic therapeutic component exploration.


Introduction
Pollen Typhae (PT), namely Puhuang, is a well-known traditional Chinese medicine used in China, Japan and other Asian countries for more than two thousand years [1]. Having both a hemostatic and anti-thrombotic action, it is widely used in the clinic to treat hemorrhagic diseases, especially for the bleeding caused by obstruction due to congealed blood, such as in menorrhagia caused by trauma, hysteromyoma or heavy bleeding due to retained placenta [1,2]. Flavonoid glycosides are responsible for the herb's multifaceted pharmacological bioactivities [3][4][5].
Carbonisatus is a process by putting Chinese medicines into a heating container to burn them to blacken or brown the surface, and even form a brown or pale brown interior. In the case of Pollen Typhae its product was widely used in the clinic to treat bleeding diseases [6]. Pollen Typhae Carbonisatus (PTC), which is the legal processed product of PT [1], has obviously enhanced anti-hemorrhagic activity, as suggested by pharmacological and clinical studies [7][8][9]. It was reported that total flavones were the main active components of PTC [10,11]. Comparative analysis between PT and PTC found that the changes in the flavonoid components were obvious [12], with a decreased content in total flavonoid and flavonoid glycosides [13], while content of flavonoid aglycones was greatly increased [12,14]. No studies however have revealed the active components that account for the enhanced anti-hemorrhagic activity.
In this paper, the hypothesis that the hemostatic therapeutic effect could be attributed to transformation of flavonoids in Pollen Typhae (FPT) during the heating process in carbonisatus development was proposed. Thus, it is necessary to characterize and quantify the simulated processing (heating) products of total flavonoids in PT (FPT-N).
The initial analytical work was based on HPLC with UV or diode array detection (DAD) [15,16]. However, their low selectivity and sensitivity cause a disadvantage in the analysis, which has led to the development of methods based on liquid chromatography-tandem mass spectroscopy (LC-MS/MS) [17,18]. According to the literature, LC-PDA-MS and high performance capillary electrophoresis (HPCE) have been developed to determine flavonoids in PT, but these assays mainly focused on only a few components [19,20]. Therefore, it is worthwhile to develop a new sensitive method for characterization and quantitation of the components. Recently, full-scanning high resolution mass analyzers, mainly represented by the time-of-flight (TOF) mass analyzers, have offered new challenges [21]. High resolution and accurate mass MS was used to quantify components with excellent accuracy and high reproducibility [22]. LC-MS/MS instruments operating in Multiple Reaction Monitoring (MRM) are widely used for targeted quantitation in triple quadrupole and hybrid triple quadrupole linear ion trap (QTRAP) systems due to their well-known selectivity and sensitivity [23], especially for the reliable separation of isobaric compounds. Thus, enabling reliable analysis of even very complex matrices represents the key benefit [24].
The aim of this study was to develop and validate a single analytical method employing LC-MS/MS for simultaneously characterization and quantification of FPT-N samples to reveal the preliminary transformation rules of flavonoids in PT during the heating process for carbonisatus development. To the best of our knowledge, no other published method covers such a number of flavonoid analytes in FPT-N samples. Additionally, the transformation rules of flavonoids in PT under heating are reported for the first time.

Sample Preparation Optimization
To obtain efficient dissolution of FPT-N, 30%, 60%, 100% (v/v) methanol were explored as the extract solvent. It was found that the dissolution efficiency by 60% methanol was higher than with the other two solvents, therefore 60% methanol was chosen for dissolution. Because of the known beneficial effects (higher efficiency, shorter extraction time, less solvent usage) [25], ultrasound-assisted dissolution was considered. Furthermore, dissolution repetitions (1-3 times) and dissolution time (15,30 and 60 min) were also optimized using ultrasonication. The results showed that ultrasound dissolution twice with 60% methanol for 30 min each was comparatively simple and complete for FPT-N dissolution.

Characterization of Detectable Components in FPT-N
A total of 18 of 23 detected chromatographic peaks were identified in FPT-N. The result is listed in Table 1. Thirteen compounds were identified by the accurate mass information and retention times of targeted reference substances. The structural characterization of the other five compounds was based on the basis of the accurate mass, the registered mass spectra fragmentation patterns and literature data. By summarizing the fragmentation patterns of these reference markers, it can be concluded that simple

Comparison of the Chemical Compositions of FPT-N
FPT-N samples were comparatively analyzed using HPLC-MS/MS. The representative base peak chromatograms (BPC) of FPT-N are shown in Figure 2. A total of 23 compounds were detected. The peak areas of the 23 compounds are shown in Table 2. The values of the coefficient of variance (C.V.%) of the peak area of compounds were all greater than 24.23%, showing that the peak area of component in different FPT-N samples varied significantly, especially the compounds huaicarbon A and huaicarbon B. The detailed differences between the chemical compositions could also be easily obtained by analyzing the entire FPT-N chromatogram. Compared to FPT, it was interesting to find that the peak heights of the compounds quercetin-   There are some other resultant compounds such as quercetin, kaempferol, isorhamnetin and peak 17 that were generated by heating for 15 min. All the results above suggested that these FPT-N samples showed significant differences due to the 23 different characteristic peaks. Based on all the above results, we can speculate that the O-triglycosylated 3-O-substituted isorhamnetin (quercetin) is transformed into O-di-/O-monoglycosylated 3-O-substituted isorhamnetin (quercetin) or directly into the corresponding flavonol aglycone isorhamnetin (quercetin) by removal of the sugar moiety or into some other newly generated hemostatic components (huaicarbon A and huaicarbon B) upon heating. However, the structures of the compounds are destroyed, until they completely disappear, after extended heating times.

Method Validation
To evaluate the linearity, calibration standards of six concentration levels of each analyte were prepared and analyzed. The calibration curve was constructed by plotting the peak area vs. the concentrations. As shown in Table 3, all 13 target compounds showed good linear regression within the test range. The linear regression analysis results showed that the coefficient estimation of the standard curve was >0.9987. The data showed excellent reproducibility. The LOD were estimated to be 0.15-3.20 ng/mL. The LOQs were estimated to be 0.52-5.66 ng/mL. The results of repeatability, stability and precision for the determination of the analytes are shown in Table 4, which indicating all the relative standard deviation (RSD) values were within 5.0%. The recovery values of the 13 compounds as shown in Table 4 were within the range of 92.03%-103.98%. The above validation data showed that MRM was sensitive for quantitation of the analytes.

Quantification Analysis of the 15 Components in FPT-N Samples
The developed method was applied to quantitatively analyze the 15 ingredients in FPT-N. The quantitative results are summarized in Table 5. The contents determination of qercetin-3-O-neohesperidoside and kempferol-3-O-neohesperidoside were respectively based on rutin and kaempferol-3-rutinoside with the same precursor/product ions in MRM mode and on the complete chromatographic separation between each other. As shown in Table 6

Materials and Reagents
Pollen Typhae (10 kg) was collected from Jiangsu Province, China. It was identified by Chungen Wang (College of Pharmacy, Nanjing University of Chinese Medicine) as belonging to the dried pollen of Typha orientalis Presl. LC-MS grade methanol, acetonitrile and MS grade formic acid (98%) were purchased from Merck (Darmstadt, Germany). Deionized water was purified by a Milli-Q system (Millipore, Bedford, MA, USA). Analytical reference standards of typhaneoside, rutin, isorhamnetin-3-O-neohesperidoside, kaempferol-3-O-rutinoside, isorhamnetin-3-O-rutinoside, kaempferol-3-O-glucoside, isorhamnetin-3-O-glucoside, quercetin, huaicarbon A, naringenin, huaicarbon B, kaempferol and isorhamnetin with 98% purity were purchased from the National Institutes for Food and Drug Control (Beijing, China). Stock solutions (about 1.0 mg/mL) were prepared in methanol. Diluted mix standards were prepared with acetonitrile/water (18/82, v/v). For the validation purposes, working standard solutions mixture at about 1.0 μg/mL was prepared in volumetric flask from stock solutions of particular analytical standards. Table 3. Validation data including the regression equations, correlation coefficient (r), linear ranges, limit of detection (LOQ) and lower limit of quantitation (LOD) for the determination of the analytes.

Sample Preparation
FPT was obtained first to prepare FPT-N. Prior to the extraction, Pollen Typhae were blended with 65% methanol (1:15, w/v) Section 2.1 says 60% for infiltrating 2 h, and then placed in boiling water bath (temp 95 °C) for 1 h. This step was repeated twice for complete extraction. The extracts were combined, filtered and evaporated with a rotary vacuum evaporator. The concentrated solution was subjected to macroporous resin column chromatography, and eluted with water/ethanol successively in a gradient (10:0 to 0:10). The fractions (eluted with 70% and 80% ethanol) were concentrated to dryness by evaporation to obtain the FPT with 85% purity. FPT-N was prepared using single component heating (TEMP = 200 °C) that simulated the carbonization process. FPT was heated in a muffle furnace for 5 min ( To support efficient dissolution of analytes, and to achieve the best recoveries of analytes from "dry" matrices (FPT-N in our case), ultrasound-assisted dissolution was utilized. fine homogenized FPT-N powder (10 mg) was weighed into screwed glass vial, and 10 mL methanol/water solvent (30/70, 60/40, 100/0, v/v) was added and left into an ultrasonic bath (Kunshan, Nanjing, China) for 15, 30, 60 min. This step was repeated 1-3 times for complete dissolution. The total extracts were transferred into a 50 mL volumetric flask and the volume was adjusted to the calibration mark by methanol. The supernatant solutions were filtered through a 0.22 μm syringe filter, and then were diluted in the same way as working standard solutions (about 1.0 μg/mL) for LC-MS/MS analysis. One mL of purified extract was removed into a vial prior to injection on LC-MS/MS system.

HPLC-MS/MS Conditions
UFLC XR Ultra high performance liquid chromatograph (Shimadzu, Kyoto, Japan) was used for the separation of target analytes and for the method development. SunFire TM C18 analytical column (100 mm × 2.1 mm, i.d., 3.5 μm particle size; Waters, Dublin, Ireland) coupled with a C18 pre-column (Waters, Milford, CT, USA) maintained at 35 °C was used for quantitation (I) and a Dubhe C18 column (250 mm × 4.6 mm, 5.0 mm; Hanbon, Nanjing, China) maintained was used for characterization (II). To optimize separation and detection conditions, two types of mobile phases were used. The mobile phase was a mixture of 0.

MRM Method Validation
Mass spectrometry was performed on 5500 QTRAP (AB SCIEX) equipped with ESI source operated in negative mode. The parameters were as follows: capillary temperature, 550 °C; curtain gas, 35; ion spray voltage, −4500 V; ion source gas 1 (Gas 1), 55 psi; ion source gas 2 (Gas 2), 55 psi. For each analyte the optimum conditions of MRM were determined in infusion mode are listed in Table 6. Each analyte was directly infused into the electrospray source via a 50 μm i.d. PEEK capillary at a flow rate of 10 μL/min. The quantification was done from the calibration curves obtained in MRM mode by using Multi Quant 2.1.1 software.
The utilized validation concept was based on the guidance [H] GPH5-1 (CFDA, Beijing, China). Linearity, limits of detection (LODs), repeatability, stability, precision and accuracy were carried out under the developed method mentioned above. The "intra-day" and the "inter-day" variances was intended to determine the precision. A recovery experiment of analyte at three spiking levels was performed to investigate the accuracy. A known amount of each of the 15 analytes at three concentrations of 75%, 100% and 125% was added to the corresponding FPT-N powder. Then the resultant samples were extracted and analyzed with the established method in triplets at each level. The estimations of the RSD were based on the ANOVA. The data obtained from the LC-MS/MS were extracted and processed in the same way to permit an unbiased comparison.

Conclusions
In this paper, a simple, reliable and highly sensitive LC-MS/MS method to characterize and quantify FPT-N samples has been developed and validated. A total of 23 peaks were detected and 18 peaks were identified. In addition, 15 compounds have been quantified. Comparing the chemical compositions in FPT-N, it was found that these FPT-N samples showed significant differences due to 23 different characteristic peaks, especially that due to the hemostatic compound huaicarbon B. Accordingly, in order to guarantee the hemostatic activity of PTC, the heating process should be controlled at 200 °C for 20-30 min. Our data demonstrate that LC-MS/MS may be a promising method for differentiating the original species from their processed products to discover the active compounds. All these results provide some useful references for the application of PT and its processed product PTC in practice.