Simultaneous Extraction Optimization and Analysis of Flavonoids from the Flowers of Tabernaemontana heyneana by High Performance Liquid Chromatography Coupled to Diode Array Detector and Electron Spray Ionization/Mass Spectrometry

Flavonoids are exploited as antioxidants, antimicrobial, antithrombogenic, antiviral, and antihypercholesterolemic agents. Normally, conventional extraction techniques like soxhlet or shake flask methods provide low yield of flavonoids with structural loss, and thereby, these techniques may be considered as inefficient. In this regard, an attempt was made to optimize the flavonoid extraction using orthogonal design of experiment and subsequent structural elucidation by high-performance liquid chromatography-diode array detector-electron spray ionization/mass spectrometry (HPLC-DAD-ESI/MS) techniques. The shake flask method of flavonoid extraction was observed to provide a yield of 1.2 ± 0.13 (mg/g tissue). With the two different solvents, namely, ethanol and ethyl acetate, tried for the extraction optimization of flavonoid, ethanol (80.1 mg/g tissue) has been proved better than ethyl acetate (20.5 mg/g tissue). The optimal conditions of the extraction of flavonoid were found to be 85°C, 3 hours with a material ratio of 1 : 20, 75% ethanol, and 1 cycle of extraction. About seven different phenolics like robinin, quercetin, rutin, sinapoyl-hexoside, dicaffeic acid, and two unknown compounds were identified for the first time in the flowers of T. heyneana. The study has also concluded that L16 orthogonal design of experiment is an effective method for the extraction of flavonoid than the shake flask method.


Introduction
Herbal or medicinal plant products, in various forms, have been used to treat different illness for many hundreds of years across the world. About 70-80% of the world population, particularly in the developing countries, rely on nonconventional medicine in their primary healthcare [1]. India has a rich �ora that is widely distributed throughout the country, and a large number of Indian medicinal plants are attributed with various pharmacological activities, because of diversi�ed class of phytochemicals, but still, the efficacy of these plants are yet to be scienti�cally documented [2]. In general, phytochemical constituents are essential for the survival and proper functioning of plants. ey provide protection against herbivores, microorganisms, and competitors, regulate growth (e.g., delaying seed germination until an appropriate time), and control pollination, fertilization, and rhizosphere environment [3]. e main secondary metabolite present in plants includes lignins, �avonoid, phenols, alkaloids, amino acid derivatives, organic acids, terpenoids, steroids, and sugar derivatives. Among different phytochemicals, �avonoid exerts a wide range of biochemical and pharmacological properties, including free radical scavenging, inhibitors of lipid peroxidation, antimicrobial, antiviral, antioxidant, antithrombogenic, hepatoprotectivity, and nephroprotectivity [4].  1  55  1  65  1 : 05  1  2  65  2  75  1 : 10  2  3  75  3  85  1 : 15  3  4  85  4  95 1 : 20 4 Plants of Apocynaceae family (Dogbane) are rich in alkaloids or glycosides, especially in seeds and latex. Some species are valuable sources of medicine, insecticides, �bers, and rubber [5]. is botanical family includes 4555 species, distributed in 415 genera [6], and the genus Tabernaemonana is included under this family that consists of shrubs or small trees. Tabernaemontana heyneana Wall. (syn. Ervatamia heyneana) is included in the oldest script Amarakosam or Namalingkanusasanum written by Amarasshimhan somewhere in between the �rst and sixth century AD [7]. It is known as kundalam paalai in Tamil, possesses curative properties against venereal diseases, gonorrhoea, respiratory problems, nervous disorders, diabetes, chronic bronchitis, rheumatism, cardiotonic ailments, and snake bite [8,9]. Preliminary phytochemical screening of the ethanolic extract of the roots of T. heyneana Wall. revealed the presence of alkaloids, sterols, triterpenoids, resins, and �avonoids [10]. Sathishkumar et al. [11] have proved the presence of quercetin and rutin related �avonoids in the leaves of T. heyneana. Reports are available for the therapeutic effect of �ower �uice (mixed along with coconut oil) against burning sensation of eyes and improved vision [12].
Extraction is a very important process for production of �avonoid concentrate from rich sources. e traditional extraction methods possess several limitations such as time consuming, laborious, low selectivity, and yield as well as utilization of large amount of organic solvents [13]. At present, there is a renewed interest in developing new processes based on the use of different variables like temperature, solvent modi�ers, and material ratio for the extraction of low molecular weight components that may be environmentally friendly and benign. Previous research documentation authenticates that temperature-assisted, enzyme-assisted, supercritical-�uid-based, and semibionic-based extractions are superior over conventional soxhlet-mediated extraction [14]. It is quite complex to predict the suitable experimental conditions for a given separation task, and therefore good experimental design becomes signi�cantly important. Such experiments are oen executed in the form of orthogonal arrays. Because of the cost efficiency, experimenters always consider as many factors as possible in a design with minimum number of runs in order to make the design saturated [15].
Although the roots and leaves have been reported to have �avonoids, no scienti�c information is available about the �avonoid content in �owers of Tabernaemontana heyneana. is paper reports about the development of extraction optimal process and high-performance liquid chromatography-diode array detector coupled with electron spray ionization/mass spectrometry (HPLC/DAD-ESI/MS) method for characterizing the chromatographic �ngerprinting of �avonoid and other possible polyphenolic compounds.

Plant
Material. e �owers were collected from the medicinal garden of Kumaraguru College of Technology, Coimbatore, India. e species was identi�ed and con�rmed at Botanical Survey of India (BSI), Southern Circle, Coimbatore, India (BSI/SC/5/23/06-07/Tech. 478). About 5 g of air-dried fresh �owers was dissolved in 50 mL of the solvent (methanol, ethanol, distilled water, chloroform, heptane, and acetone) and kept in an orbital shaker for overnight (shake �ask method). e residue was reextracted under the same conditions. e obtained extracts were �ltered with Whatman number 1 �lter paper, and the �ltrate was used for total �avonoid estimation.

Experimental Design of Extraction
Process. e main factors that affect the extraction of �avonoid like temperature, extraction time, material ratio (weight of the �owers (g): volume of the solvent (mL)), solvent modi�er (%), and the number of extraction cycles were investigated. e factors and the experimental design were slightly modi�ed according to the type of the solvent used. e optimum extraction conditions were determined by L 16 (4 5 ) orthogonal design of experiment using ethanol as extracting solvent (Table 1). For ethyl acetate-mediated optimization, L 9 (3 4 ) orthogonal design of experiment was adopted ( Table 2). A singlefactor analysis of variance (one-way ANOVA) was adopted to investigate the effect of each factor in the extraction of �avonoid. GraphPad Prism 5 trial version soware was used to carry out the statistical analysis.   [16] with slight modi�cations. To 0.1 mL of the �ower extract, distilled water was added to make the volume to 5 mL. To this added 0.3 mL 5% NaNO 2 , aer �ve minutes, added 3 mL of 10% AlCl 3 and mixed well. Six minutes later, 2 mL of 1 M NaOH was added and the absorbance was measured at 510 nm. Rutin was used as a standard for constructing the calibration curve.

2.�. Identi�cation of Flavonoid by in Layer Chromatography (TLC)
. e glass plates (20 × 20 cm) were coated with silica gel G 60 (0.2-0.3 mm thick and 32 g/60 mL distilled water) and were dried at room temperature. e dried plates were activated at 100 ∘ C for 30 minutes in an oven and brought to room temperature. About 20 L of optimal extract along with standard markers (rutin and quercetin) was spotted 1.5 cm far from the edge of the plate. ese glass plates were developed one dimensionally in an air tight chromatography chamber containing about 200 mL of mobile phase solvent mixture consists of ethyl acetate-ethanol-water (5 : 1 : 5, v/v/v). e developed plates were air dried and visualized under UV light at 365 nm aer the application of liquid ammonia for the identi�cation of �avonoid [17].

Isolation of Flavonoid by Preparative in Layer Chromatography (PTLC)
. e glass plates (20 × 20 cm) coated with silica gel G 60 (0.5-1.0 mm thick and 46 g/85 mL distilled water) were activated at 100 ∘ C for 30 minutes in an oven and brought to room temperature. e procedure was repeated as same as that of TLC [18].

FT IR Analysis
. e PTLC �ower eluates were mixed with 200 mg KBr (FT-IR grade) and pressed into a pellet. e sample pellet was placed into the sample holder, and FT-IR spectra were recorded in the range 4000-450 cm −1 in FT-IR spectroscopy (PERKIN ELMER FT-IR Spectrometer, USA).

HPLC-DAD-ESI/MS
Analysis. e high-performance liquid chromatography-electrospray mass spectrometry (HPLC-MS) experiment was performed on THERMO Finnigan LCQ Advantage max ion trap mass spectrometer (USA) having connected Finnigan Surveyor HPLC system. e column was Waters ODS-2, 250 × 4.6 mm, and id., 5 m. e mobile phase A was made up of acetonitrile, while B was made of 0.1% formic acid (pH 4.0, adjusted with ammonium hydroxide) aqueous solution. e gradient elution was performed at 0.5 mL/min with an initial condition of 12% of mobile phase A and 88% of mobile phase B for 5 min. e mobile phase A was increased to 25%, and the elution was performed for 10 min, and then elution was performed for 10 min by a linear increase of mobile phase A to 60% and again for 5 min with an increase to 100% [19]. e eluates were monitored by PDA (multiwavelength) detector at 260 nm. About 20 L of the PTLC �ower eluates was introduced into the ESI source through Finnigan surveyor autosampler. e mass spectra were scanned in the range 150-750 m/z, and the maximum ion injection time was set 200 nS. Ion spray voltage was set at 5.3 KV and capillary voltage 34 V. e MS scan ran up to 60 min, and the data reductions were performed by Xcalibur 1.4 SRI.

Extraction of Flavonoid from the Flowers of T. heyneana
Using Shake Flask Method. e amount of total �avonoid present in the �owers was depicted in Figure 1. Among di�erent solvents utilized for the extraction of �avonoid, methanol was proved to be the best (1.2 ± 0.13 mg/g tissue), and heptane was observed as poor solvent (0.3 ± 0.11 mg/g tissue). Ethanol and acetone were found to be as moderate solvents in the extraction of �avonoid (0.8 ± 0.16 mg/g tissue). e standard calibration curve constructed using rutin has proved a signi�cant positive correlation ( 2 = 0.995). e �rst and foremost step in any analysis protocol is the extraction of target compound from the source material. e extraction of �avonoid can be in�uenced by various factors such as choice and polarity of the solvent, temperature, pressure, matrix type, solvent modi�cation, particle size, and extraction time [20]. Previous reports have explored and revealed that various solvents like methanol, ethanol, acetone, water, ethyl acetate, ether, and, to a limited level, dimethyl sulfoxide, propanol, butanol, and their consortium have been used for the extraction of phenolic compounds [20,21]. Our present investigation has well agreed with the previous reports explored about the extraction of �avonoid content in some fruits and vegetables by the adoption of alcoholic and organic solvents. Similarly, the results were also well corroborated with the previous investigations about the shake �ask method and seemed to be inferior in the recovery of total phenolics and �avonoid content [21,22].

Optimization of Flavonoid Extraction Using Orthogonal
Design of Experiment. e optimization of �avonoid extraction in the �owers using 65-95% aqueous ethanol has yielded about 80.1 mg/g tissue (66.8 fold increase compared to an initial yield about 1.2 mg/g tissue) and ethyl acetate (20.5 mg/g  Table 3). Similarly, the optimal process of �avonoid recovery using ethyl acetate was proved to be a temperature 80 ∘ C (A 3 ), material ratio of 1 : 20 (C 3 ), 2 times of extraction cycle (D 2 ), and 1 hr extraction duration (B 1 ), that is, Table 4). e results analyzed in the form of range analysis and one-way ANOVA have revealed that material ratio (signi�cant at 5%, ( ), and ethanol-mediated optimization) and temperature (signi�cant at 1%, ( 1), and ethyl-acetate-mediated optimization) were signi�cant variables for the extraction and recovery of �avonoid content. Overall analysis has proved that aqueous ethanol-mediated optimal process is the best one compared to that of ethyl acetate.
Normally, the conventional solvent-mediated extraction process may depend upon conductive and convective processes to induct thermal energy into the system, and thereby, create a prolonged extraction time and increase the risk of thermal decomposition of �avonoid. is may be overcome by certain experimental approaches like temperature-assisted extraction, pressurized hot water extraction, and central composite design (response surface methodology (RSM)) which has con�rmed that variables like temperature, ethanol concentration, extraction time, and material ratio were the ma�or in�uencing factors in the process [23].

Effect of Temperature in the Extraction of Flavonoid.
e contents of �avonoid gradually increased with a rise in the temperature in a range of 55 ∘ C to 85 ∘ C with a 10 ∘ C temperature interval (a slight decrease was observed between 55 ∘ C and 65 ∘ C) in ethanol-mediated optimization, and a similar pattern has been observed in ethyl-acetate-mediated optimal process (gradual and steep increase in �avonoid content from 60 ∘ C to 80 ∘ C) (Figure 2). Our present investigation was well accorded with the report documented by Sathishkumar et al. [11], on the extraction optimization of �avonoid in the leaves of T. heyneana Wall. According to Shi et al. [24], higher temperature would cause soening of the plant tissue, disrupting the interactions between phenolic compounds and protein or polysaccharides, and increasing the solubility of the phenolic compounds, which improves the rate of diffusion, thus giving a higher rate of extraction. On one hand, higher temperature can accelerate the solvent �ow and thus increase the �avonoid content, and on the other hand, it can decrease the �uid density and viscosity, an increase in solute spread ability that could be responsible for an increase in the solvating power because of the increase in the solute vapor pressure [25].

Effect of Material Ratio in the Extraction of Flavonoid.
e contents of raw �avonoid extracted from the �owers were maxima at 1 : 20 material ratio in both the solvents-mediated optimization. An entirely different contradictory result was documented by Sathishkumar et al. [11] about the optimal value of material ratio of the extraction of �avonoid from the leaves of T. heyneana, and this variation is probably because of the difference in the phytochemical distribution. Generally, when the solvent volume was increased, it can cause excessive swelling of the material by water and absorbing of the effective constituent [26]. A similar study performed by Xiao et al. [27] has showed that higher solvent volume may give lower yield which is totally inverse when compared with conventional extraction techniques. In contrast, an investigation made by Chen et al. [28] suggested that for a �xed amount of raw material, the more volume of solvent used, the more dilute effect in the solvent side. is gave a larger concentration difference between the interior of the plant cells and the external solvent, thus a faster extraction rate could be obtained. e basic mechanism is that the increasing ratio of solvent to raw material could decrease solution concentration difference inside and outside plant cells, which consequently prompted diffusion rate of solute particles and made more �avonoid molecules to enter the solution.

Effect of Ethanol in the Extraction of Flavonoid.
e contents of �avonoid extracted by 75% ethanol reached maxima, and further increase in ethanol concentration may not yield increased �avonoid content. �sually, addition of a small amount of a liquid modi�er can enhance significantly the extraction efficiency and, consequently, reduce the extraction time and improve the recovery of different types of natural products from plant materials [29]. Ethanol was selected as a right choice because it is environmentally benign, relatively safe to human health and interacts with the �avonoid probably through noncovalent interactions and promotes a rapid diffusion into the solution. Eventhough alone water can extract maximal �avonoid, more proteins and polysaccharides may be extracted along with it, and removal of water from the system is not cost effective, and therefore, aqueous ethanol can act as suitable extracting agent [24].

Effect of Extraction Time in the Extraction of Flavonoid.
e optimal time duration of �avonoid extraction was found to be 3 hrs for aqueous ethanol and 1 hr for ethyl acetate, respectively. e contents of �avonoid extracted for 3 hrs reached maxima, and prolonged extraction may not yield an increased content. ere is a kind of �uctuation (increase and decrease) in the �avonoid content extracted in function with time duration has been noticed in aqueous ethanol-mediated optimization. is is probably because of the synergistic effect of other parameters involved and degradation of �avonoid due to thermal-based oxidation. Increase in time also led to an increase in the adhesion of diffused particles (�avonoid) around the walls of the supporting material like glass or plastic tubes that may hinders the extraction process [30]. Similar reports made by [14,[31][32][33] have revealed that 2-3 hrs was the optimal extraction time and was well associated with the present investigation.

Effect of Number of Extraction Cycles in the Extraction
of Flavonoid. e optimum number of extraction cycles for �avonoid extraction was found to be 1 cycle for aqueous ethanol and 2 cycles for ethyl acetate, respectively. e effect of repeated and successive extractions of the residue is termed as extraction cycles. During each cycle the residue was taken back and reextracted using fresh solvent under the conditions mentioned in the design [27]. e contents of raw �avonoid decreased gradually with the number of extraction cycles for aqueous ethanol-mediated optimization, and a similar pattern with slight modi�cation has been observed for ethylacetate-mediated optimal process.

Analysis of Flavonoid by in Layer and Preparative in
Layer Chromatography Techniques. TLC plates developed and sprayed with liquid ammonia revealed the presence of �avonoid glycosides (tentative rutin related compounds, bright yellow-brown color), �avonols (tentative quercetinrelated compounds, bright yellow color), tentative phenolic acids (blue color), and certain other unknown phenolic compounds. e values of rutin and quercetin were found to be 0.96 and 0.94, respectively. e values of tentative rutin-and quercetin-related compounds were found to be 0.97 and 0.91, respectively. in layer chromatography (TLC) has frequently been used for the separation and the quantitative or semiquantitative analysis of natural compounds. TLC has some advantages such as rapidity, easiness, and cheapness.
An observation made by [16,18,34] regarding the identi�cation of �avonoids and phenolic acids under far UV light has been well documented and agreed with the present results. Similarly, studies carried out by [35] have substantiated that the calculated values of �avonoid and phenolic acids in any species may not be unique and purely depend upon the composition of the mobile phase used in the TLC. e overall results have well proved the presence of �avonols, �avonol glycosides, and phenolic acids in the �owers of T. heyneana. Likewise, based upon the PTLC technique, about two strong and one medium spots that have been tentatively predicted as rutin, quercetin, and phenolic acid-related compounds were successfully eluated from �owers ( Figure 3). All the present investigation was well agreed by the research report documented early [11].

FT IR Analysis.
A review report by [36] has explained that a spectral range between 1185 cm −1 and 965 cm −1 showed a strong indication of C-O-C and C-OH vibrations, and absorption band between 900 cm −1 and 500 cm −1 indicates the presence of sugar moiety (glycosylation), and a spectral range at 1645 cm −1 proved the presence of dienes (C=C). Similarly, it has documented that a spectral band at 3400 cm −1 showed a broad and strong indication of presence of -OH group, exclusively phenols [37]. All these previous documentations have well correlated with the present results proving the presence of phenolic compounds that includes �avonoid and phenolic acids in �owers (Figure 4).  100  95  90  85  80  75  70  65  60  55  50  45  40  35  30  25  20  15 (Figure 6(a)). All the present investigation was well correlated with the results documented early [38]. Similarly, initial research work carried out by Sakushima et al. [39] by fast atom bombardment method has revealed the fragmentation mode of several �avonoid glycosides. From their studies, robinin was found to possess the parent ion at m/z 741 and the fragment ion m/z 433/287. In our present investigation, the parent ion as [M − H] + was  A study made by Dubber et al. [40] has successfully explained the fragmentation pattern of rutin obtained from Ginkgo biloba. In this mechanism, rutin (m/z 609.12) gave rise to fragment ion at m/z 301 (aglycone portion of rutin, quercetin) with the corresponding loss of the rutinose unit (rhamnose plus glucose moieties) and subsequent retrocyclization of the C-ring (between bonds 1 and 2) leading to the A − fragment with m/z 179. A most alike fragmentation of rutin has been observed in the present studies; that is, parent ion m/z 610 was fragmented at m/z 460 with the loss of −150 u ((M + H -150), rhamnosyl moiety), and further fragmentation with the loss of −157 u ([M + H -150 -157], glucose moiety) yields quercetin at m/z 303 (Figure 8(a)). Similarly, the study carried out by Ferreres et al. [41] has explained the fragment pattern of sinapoyl-hexoside. e deprotonated parent ion m/z 385 has been fragmented as m/z 223 with the loss of −162 u (hexosyl radical) and at m/z 205 with the loss of −180 u (hexosyl radical + water). e current results have revealed the presence of sinapoyl-hexoside at parent ion m/z 383 that was fragmented as m/z 224 with the loss of −159 u (hexosyl radical) were in well concordance with the above said report (Figure 8(b)). e liquid chromatography (LC) of PTLC �ower eluate 3 has recorded a single peak G at retention time ( ) of 2.50. e mass spectra (MS) have recorded the presence of quercetin (RA-55%) at parent ion m/z 303 (Figure 9). Several research reports on the LC-MS analysis of quercetin have documented as m/z 303. e overall results have proved that the �owers of T. heyneana possess appreciable quantity of �avonoids and phenolic acids.

Conclusion
In summary, the �avonoid extraction process was successfully optimized by aqueous ethanol-mediated L 16 orthogonal design of experiment, and the components were characterized by HPLC-DAD-ESI/MS analysis. e optimal conditions were found to be 85 ∘ C, 3 hrs extraction duration, 75% ethanol concentration, a material ratio 1 : 20, and 1 time of extraction cycle. e HPLC-DAD-ESI/MS analysis performed in the PTLC �ower eluates has identi�ed compounds like rutin, quercetin, robinin, sinapoyl-hexoside, and two unknown phenolics.