Quantification of chlorogenic acid, rosmarinic acid, and caffeic acid contents in selected Thai medicinal plants using RP-HPLC-DAD

The chlorogenic acid, rosmarinic acid, and caffeic acid contents in 100 selected plants were determined using reversed phase high performance liquid chromatography equipped with diode array detector. The optimum condition was 0.2% phosphoric acid in water (solvent A) and methanol (solvent B) as the mobile phase, which was set at 45% B for 20 minutes at a flow rate of 1.2 mL/min. The column temperature was maintained at 30 °C and the detection wavelength was 325 nm. Among 100 selected plants, 39.64% contained all 3 compounds, 40.54% contained 2 compounds, 14.41% contained only 1 compound, and 5.41% could not detect any of the 3 compounds. The highest contents of chlorogenic acid, rosmarinic acid, and caffeic acid were found in Lonicera japonica flowering buds, Melissa officinalis leaves, and Coffea canephora seeds at the concentration of 9.900 ± 0.004, 19.908 ± 0.171, and 1.233 ± 0.003 g/100 g of dried plant, respectively.


INTRODUCTION
Phenolic compounds or polyphenols, the secondary metabolites of plant, are one of the most abundant and extensively distributed groups of substances in the plant kingdom which appear in all plant organs. However, the polyphenolic profile of plants differs between varieties of the same species. For decades, polyphenols have interested many researchers for their antioxidant, antioxidative stress activities, and great abundance in food. The varieties of natural polyphenols range from simple molecules (such as phenolic acids) to highly polymerized compounds (such as tannins). Polyphenols occur primarily in a conjugated form with one or more sugar residues linked to hydroxyl groups, although direct linkages of the sugar unit to an aromatic carbon atom also exist (Bravo, 1998;Manach et al., 2004). Hydroxycinnamic acid, one of two major groups of phenolic acids, is usually found in plants. The hydroxycinnamic acid derivatives consist of a large group of simple phenolic acids, and are bountiful in fruits, seeds of fruits, vegetables, and cereals. In addition, they have been arranged into structural and functional constituents of plant cell walls and also as bioactive ingredients of diets. The derivatives of hydroxycinnamic acids are synthesized through the shikimate pathway in which phenylalanine and tyrosine are used as starting precursor molecules. The main hydroxycinnamic acid derivatives are ferulic acid, caffeic acid, p-coumaric acid, chlorogenic acid, sinapic acid, and rosmarinic acid (Lafay, Gil-Izquierdo, 2008;Manach et al., 2004;Teixeira et al., 2013). Caffeic acid ( Figure 1A) is one of the most common phenolic acids that biosynthesise by hydroxylation of p-coumaric acid and is more broadly present in several food sources such as berries, coffee drinks, and dietary supplements (Magnani et al., 2014). Chlorogenic acid ( Figure 1B) is an ester form of caffeic acid and quinic acid, which is widely distributed in the human diet with plants, fruits, and vegetables especially in coffee, apples, and pears (Upadhyay, Mohan Rao, 2013). Rosmarinic acid ( Figure 1C), an ester of caffeic acid and 3, 4-dihydroxyphenyllactic acid, is commonly found in species of the boraginaceae, lamiaceae, and in some ferns and hornworts (Petersen, Simmonds, 2003). High performance liquid chromatography (HPLC) is a primary method for the separation and analysis of chemical compounds in many fields such as agriculture, cosmetics, pharmaceutical industries, environments, and food. It is commonly used for qualitative and quantitative analyses of chemicals in herbal extracts. The identification of compounds depends on the retention time and light spectral characteristics of each chromatographic peak (Zeng et al., 2011).
T h e a i m o f t h i s s t u d y w a s t o e s t a b l i s h a RP-HPLC-DAD condition for analysis and provide the approximate quantification of chlorogenic acid, rosmarinic acid, and caffeic acid in 100 selected Thai medicinal plants.

Sample collection
A selection of 100 fresh plants was obtained by randomized collection from various places in Thailand and also purchased from local markets in Thailand based on chemotaxonomy. They were authenticated by Associate Professor Dr. Nijsiri Ruangrungsi. All plant materials were dried at 45 °C in a hot air oven, and voucher specimens were deposited at College of Public Health Sciences, Chulalongkorn University. After the removal of any foreign matter, crude drugs were grounded into coarse powders before use.

Sample extraction
Ten grams of each selected plant sample were exhaustively extracted with petroleum ether and followed by 95% ethanol using a Soxhlet apparatus. The ethanolic extract was filtered through filter-paper and evaporated to dryness under reduced pressure by a rotary evaporator. The extract yields were weighed, recorded, and stored at -20 °C to avoid the possibility of degradation of the active compounds.

Preparation of standard solutions
One milligram of each standard was dissolved in 1 mL of methanol. The solution was filtered through a 0.45 µm PTFE membrane syringe filter.

Preparation of sample solutions
Fifty milligrams of each extract were dissolved in 1 mL of methanol and diluted to appropriate concentrations for further RP-HPLC analysis. The solution was filtered through a 0.45 µm PTFE membrane syringe filter.

Chromatographic conditions
The Shimadzu HPLC LC-20A system (Shimadzu, Japan) consists of a system controller (CMB-20A), two solvent delivery units (LC-20A), an on-line degassing unit (DGU-20A3), an auto-sample (SIL-20A), a column oven (CTO-20A), and a photo-diode array detector (SPD-M20A). System control and data analysis were processed with Shimadzu LC Solution software. The chromatographic separation was performed with an Inertsil ® ODS-3 5 µm C 18 column (4.6 X 250 mm) and coupled with a ReproSil ® -Pur ODS-3 C 18 guard column (4.0 X 10 mm). The samples were analyzed using 0.2% phosphoric acid in water, pH 1.46 (solvent A), and methanol (solvent B) as a mobile phase. The isocratic program was set at 45% B for 20 minutes at a flow rate of 1.2 mL/min. The mobile phase was filtered through 0.45 µm nylon membrane filters and degassed using an ultrasonic bath before analysis. The column temperature was maintained at 30 °C and the injection volume of standards and sample solutions was 5 µl. The wavelength was set at 325 nm for monitoring chromatographic profile. All measurement was done in triplicate.

System suitability
The retention factor, theoretical plate number, and tailing factor were evaluated for system suitability parameters. The system performance was analyzed for five replicates of standard solution.

Method validation
According to the ICH guideline (ICH, 2005), the calibration range, specificity, accuracy, repeatability, intermediate precision, limit of detection (LOD), limit of quantitation (LOQ), and robustness were validated for analytical method. The Lonicera japonica flowering bud ethanolic extract was found to contain all 3 compounds so it was used as a sample matrix to evaluate the validity of the analytical method.

Calibration range
The calibration range was performed by plotting peak areas obtained from RP-HPLC analysis versus concentrations of standard. The stock solutions of chlorogenic acid, rosmarinic acid, and caffeic acid were dissolved in methanol and diluted together to give concentrations of 16.67, 33.33, 50.00, 66.67, and 83.33 µg/ mL for evaluation of the calibration range. The calibration range of these standards was fitted by linear regression. The regression equation was calculated in the form of y = ax + b, where y is peak area and x is concentration.

Specificity
The specificity was evaluated by a peak purity test. The peak purity index of the analyte was processed by Shimadzu LC Solution software. It was determined by comparing all the spectra within the chromatographic peak to the reference spectrum at the peak apex.

Accuracy
The accuracy of each sample was tested by recovery method. Three different levels of standard solutions (10, 25, and 50 µg/mL) were spiked into the extract. The spiked and un-spiked samples were evaluated under the same condition in triplicate, then percent recoveries were calculated by comparing the measured amount of those standards with the amount added.

Precision
The precision was determined by repeatability (intra-day) and intermediate precision (inter-day) studies. The method was performed by analyzing three level concentrations of sample solution in triplicate on the same day for repeatability and in the five different days for intermediate precision. The precision was calculated in terms of percent relative standard deviation (% RSD) of compound content.

Limit of Detection (LOD) and Limit of Quantitation (LOQ)
LOD and LOQ were determined from the calibration range using the following formulae: where: σ = the residual standard deviation of the regression line; S = the slope of the regression line

Robustness
The robustness was determined for variations in flow rates (1.195, 1.200, and 1.205 mL/min), variations in column temperature (29, 30, and 31 °C), and variations in wavelength (322, 325 and 328 nm). The robustness was calculated in terms of percent relative standard deviation (% RSD) of retention time and peak area.

Optimization of chromatographic condition
The chromatographic condition optimization including mobile phase, gradient elution procedure, flow rate, column temperature, and wavelength detection were performed to provide a better separation of constituents. Numerous mobile phases and gradient programs were trialled using various proportions of different aqueous phases and organic modifiers. Formic acid, phosphoric acid, and acetic acid were usually employed to the aqueous phase to enhance the resolution, restrain the ionization, and reduce the peak tailing of compounds (Ma et al., 2011). The most suitable mobile phase that showed good resolution and symmetric peak shape were obtained using two parts as Solvent A (0.2% phosphoric acid in water) and Solvent B (methanol) with an isocratic program. The column temperature was held at 30 °C for the duration of analysis to improve the retention time precision. Hydroxycinnamic acids have the maximum wavelength during 270 -360 nm (Köseoglu, Kolak, 2017). The UV spectra of standard chlorogenic, rosmarinic, and caffeic acids were compared at varying wavelengths, and based on the data from the literatures. The optimal detection wavelength in this study was to be 325 nm (Haghi, Hatami, 2010;Shan et al., 2013).

Chlorogenic acid, rosmarinic acid, and caffeic acid quantification
The 100 selected plants were edible vegetables, fruits, and herbal plants in Thailand. The plant samples were exhaustively extracted with petroleum ether and followed by 95% ethanol using a Soxhlet apparatus. The percent yields of crude extracts were shown in Table I.
A quantitative analysis of chlorogenic acid, rosmarinic acid, and caffeic acid in selected plants was performed by RP-HPLC analysis. The standard markers to quantify in this study are chlorogenic acid, rosmarinic acid, and caffeic acid which are hydroxycinnamic acid derivatives. Hydroxycinnamic acid derivatives, a subgroup of phenylpropanoids, are synthesised by the shikimate pathway where the starter precursor molecules are phenylalanine and tyrosine. Chlorogenic acid, rosmarinic acid, and caffeic acid in extracts were identified by comparing the retention time and UV spectrum of each peak with a reference of standard compounds (Figure 2). The contents of chlorogenic acid, rosmarinic acid, and caffeic acid in the 100 selected plants were shown in Table I. The results of RP-HPLC analysis demonstrated that the distribution of these 3 phenolic compounds varied in many samples. Among 100 selected plants, 39.64% contained all 3 compounds, 40.54% contained 2 compounds, 14.41% contained only 1 compound, and 5.41% could not detect these 3 compounds. Lonicera japonica flowering buds were found to be the richest source of chlorogenic acid content at 9.90 g/100 g of dried crude drug, and Melissa officinalis leaves showed the most rosmarinic acid content at 19.91 g/100 g of dried crude drug. The most caffeic acid content was found in Coffea canephora seeds at 1.23 g/100 g of dried crude drug. Chlorogenic acid was found in many families and is the main active constituent in L. japonica flowering bud (Chaowuttikul, Palanuvej, Ruangrungsi, 2017). It is also the main phenolic compound in coffee (Coffea spp.) that supported this study (Ayelign, Sabally, 2013). Rosmarinic acid was mostly found in the Labiatae family, relating to a previous report of high rosmarinic acid content in plants of this family, especially in Mentha spicata, Salvia officicalis, and Melissa officinalis (Shekarchi et al., 2012).

System suitability
The retention factor, theoretical plate number, and tailing factor were found to be 4.30 ± 0.01, 2745.17 ± 158.17, and 1.027 ± 0.07, respectively (Table II). These parameters confirmed that the condition is appropriate for analysis according to the FDA criteria.

Method validation
The analytical method validation is the process that confirms precise, accurate, and reliable quantitative data. According to the ICH guideline, calibration range, specificity, accuracy, repeatability, intermediate precision, limit of detection, limit of quantitation, and robustness should be validated for analytical analysis.
Standard chlorogenic acid, rosmarinic acid, and caffeic acid at 5 concentrations were investigated for linearity by RP-HPLC method. The calibration curves of standard compounds were linear in the range of 16.67 -83.33 µg/mL. The regression equation of chlorogenic acid, rosmarinic acid, and caffeic acid were y = 2874.5x + 813.03, y = 2833.8x -1858.3, and y = 5202.2x + 673.32, respectively (Figures 3 -5). The linearity showed good correlation (R 2 ≥ 0.999). An analytical technique is acceptable when the correlation of method (R 2 ) value achieved is 0.99 or better.
The specificity was evaluated by peak purity test and confirmed that analyte chromatographic peak is not attributable with another compound. This test is based on the absorbance spectrum, which is detected by diode array detectors. If all of the individual spectra recorded during the elution of a peak are identical, even if detected at any periods of a peak, the peak is considered pure (Hansen, Pedersen-Bjergaard, Rasmussen, 2011). An identical peak resulted in a peak purity index of 100% or peak purity index of 1.0, indicating that all spectra are similar. The results showed the peak purity index of the three compounds was more than 0.999 (Figures 6-8), thus no impurity was detected in these peaks.
The accuracy was evaluated by the recovery method.    LOD and LOQ analysis were calculated by the residual standard deviation of a regression line and the slope of the calibration curve. The LOD of chlorogenic acid, rosmarinic acid, and caffeic acid that is taken as the lowest concentration of analyte in a sample that could be detected was 1.64, 2.22, and 0.65 µg/mL, respectively.
The LOQ of chlorogenic acid, rosmarinic acid, and caffeic acid that is taken as the lowest concentration of analyte in a sample that could be accurately quantitated was 4.97, 6.72, and 1.97 µg/mL, respectively.
The robustness of sample and standard compounds was determined during the analysis of the RP-HPLC method when the flow rate of the mobile phase varied from 1.195-1.205 mL/min, the column temperature varied from 29-31 °C, and the wavelength varied from 322-328 nm. The results demonstrated no differences (%RSD <4) in the area of the curve and retention time as shown in Tables IV and V. However, the method validation in this study used L. japonica flowering bud ethanolic extract as a sample matrix which might not represent all of the plant samples. It was recommended that further quantification of chlorogenic acid, rosmarinic acid, and caffeic acid in each plant material extract as stated in this study should be verified for each sample matrix.
The RP-HPLC analysis in this study demonstrated the contents of 3 phenolic compounds in selected plants that could be useful as a chemical marker for quality control of plant material. The interesting plants with special reference to these markers could be further investigated for their biological activities involving hydroxycinnamic acid derivatives.