The Liquid Chromatographic Determination of Chlorogenic and Caffeic Acids in Xu Duan (Dipsacus asperoides) Raw Herb

A validated analytical method is reported for the analysis of chlorogenic and caffeic acids in Xu Duan (Dipsacus asperoides) in the dried raw herb. The ground samples were extracted by ultrasonication in water and the extract was analysed by LC-PDA with identity confirmation by (+)ESI-MS/MS.AC18 columnwas usedwith a 0.1% aqueous formic acid :methanol gradientmobile phase. The analytes were quantified 325 nm. With the MS detector, using the chlorogenic acid precursor ion with m/z 354, ions withm/z 191, and 85 were produced. For caffeic acid the precursor ionwithm/z 181, ions withm/z 163, 135, and 89were produced.The amount of chlorogenic and caffeic acids in the raw herb was found to be 4.46 and 0.63mg/g, respectively, and the method LOD was 0.13 and 0.02mg/g, respectively.


Introduction
Dipsacus asperoides or commonly known as Xu Duan is a perennial plant distributed in mountainous regions of southwestern China and has been used in Chinese traditional folk medicine for its antibacterial and anti-inflammatory properties [1].The herb has a bitter, pungent, and a slightly warm flavour.The root has been used extensively for the treatment of liver and kidney dysfunctions.The herb is claimed to tonify the liver and kidneys, strengthen bones and sinews, prevent miscarriage, and help in the regeneration of flesh.Other uses for the herb include the treatment of trauma and fractures of bones and removal of blood stasis, thus relieving pain and swelling [2,3].
There is very little published literature on the HPLC (high performance liquid chromatography) analysis of Xu Duan (Dipsacus asperoides).The majority of published articles on this herb are on their biological activity or effect.The analysis of chlorogenic and caffeic acids in Xu Duan has only been reported in one paper [3].In that study the powdered herb was extracted with boiling ethanol and the extract analysed using a C18 column with a mobile phase of acetonitrile : water (20 : 80)-no wavelength of detection was stated, and no aspects of validation reported.
Chlorogenic and caffeic acids (structure shown in Figure 1) are hydroxycinnamic acids and are related in that chlorogenic acid is the ester of caffeic acid with the 3-hydroxyl group of quinic acid.
Both compounds are known to have antioxidant properties [3][4][5][6][7][8][9] and show an ability to inhibit platelet aggregation [2].Caffeic acid also has antiprostaglandin properties which are a property of conventional drugs used to treat dysmenorrhoea [10].This ability could explain its frequent inclusion in herbal formulations for the treatment of the gynaecological ailment dysmenorrhoea.
The lack of comprehensively validated HPLC methods for Xu Duan used in this study reinforced the need to develop and validate methods used for this herb.Analysis of Xu Duan had very little published literature.The validated method developed was used to quantify the analytes of interest in the raw herb sourced from different suppliers.The differences in concentration of the selected analytes in the different sourced herbs may be due to factors like growing conditions, time of harvest, postharvest treatment, storage conditions, and age of herb.This work reports a LC-PDA (liquid chromatographyphotodiode array detection) method for the determination of chlorogenic and caffeic acids in the Xu Duan raw herb, with MS identity confirmation of the analyte.

Preparation of Standard Solutions.
A mixed stock calibration solution containing an accurately known concentration of approximately 174 g/mL chlorogenic acid and 66 g/mL caffeic acid was prepared by making up 1.8358 mg and 0.6705 mg of chlorogenic acid and caffeic acid, respectively, to 10 mL with methanol.The standards were stored at −10 ∘ C and discarded if not used within 2 days as the peak area decreased by ≥2% after this time.

Preparation of the Raw Herb for Analysis.
Grind the dried Xu Duan (100 g) samples to pass through a 212 m sieve.Sieve the sample twice to ensure that it is homogenous.

Extraction Solvent and Method Optimisation.
The ground raw herb sample was extracted using six different solvents (with five replicates per solvent) to determine which was most efficient.The solvents used were methanol, water, aqueous methanol (50 : 50% v/v), absolute ethanol, ethyl acetate, and hexane.Accurately weighed quantities of the sample (0.950-1.050 g) were transferred into volumetric flasks (100 mL), 95.0 mL of the solvent added, and the mixture sonicated for 2 × 30 min, with the sample allowed to cool between sonications.The sample was made to volume with the extraction solvent and passed through a 0.45 m PVDF (polyvinylidene fluoride) syringe filter into a 2 mL autosampler vial for analysis.
The ground raw herb sample was extracted with the selected optimum solvent (which was found to be water) using three different extraction methods (with five replicates per method).The extraction methods to be used were sonication, reflux, and Soxhlet.
(a) Sonication Extraction.This method is the same as described in the extraction solvent optimisation, except that the optimum solvent (which was found to be water) was used.(b) Reflux Extraction.The sample was accurately weighed (0.950-1.050 g) into a round bottom flask (100 mL) and 90.0 mL water added before refluxing for 1 h.After cooling the solvent was decanted into a volumetric flask (100 mL) and made up to volume with water.(c) Soxhlet Extraction.A known quantity (0.950-1.050 g) of the sample was weighed into a 100 mL cellulose extraction thimble and placed in a Soxhlet extraction apparatus.90.0 mL water was placed in round bottom flask (100 mL) and the sample extracted for 3 h.After cooling, the solvent was quantitatively transferred to a volumetric flask (100 mL) and made up to volume for analysis.

Residual Analyte following Extraction Using the Optimum Extraction Solvent and Method.
Using the optimal extraction method and solvent (sonication and water, resp.),Note: the flow rate was 1 mL/min.
the residual herb sample following the extraction described was reextracted three more times.The combined extracts were concentrated to 5.0 mL and analysed.This result gives an estimate of the extraction efficiency.

Extraction of the Raw Herb
Sample.Approximately 0.5 g (0.450-0.550 g) of the raw herb sample was weighed and transferred to a 100 mL volumetric flask.95.0 mL of water was added and the mixture was sonicated for 2 × 30 min.The sample solution was cooled to room temperature between extractions.After cooling the sample solution was made up to volume, before filtering through a 0.45 m PVDF membrane filter into a 2 mL autosampler vial for analysis.

Recovery Studies for the Raw Herb.
A mixed spiking solution was prepared by weighing 43.21 and 7.03 mg of chlorogenic and caffeic acids, respectively, into a 25 mL volumetric flask and was made to volume with methanol.For the 100% spike level, 250 L of the spiking solution was added to an accurately weighed quantity of the herb sample (∼1.0 g).
For the 50 and 200% spike levels, the spiking volumes were adjusted accordingly.Seven replicates for each spike level were carried out to give a total of twenty-one samples.

LC-PDA Analysis and ESI-MS/MS Conditions.
The gradient mobile phase composition for the LC analysis of chlorogenic and caffeic acids is summarised in Table 1.
Column oven temperature was set to 30 ∘ C and flow rate to 1.0 mL/min.With this gradient program the column head pressure varied from 1500 to 2300 psi.The injection volume was 6 L with a flush volume of 30 L.Quantitation for both analytes was carried out at 325 nm.
ESI-MS/MS was used for identity confirmation.The MS/MS conditions set as follows: nebulisation gas temperature 350 ∘ C, at 19 psi; needle voltage −4100 V at 15 A; shield voltage −300 V, capillary voltage −60 V, and MS detector at −1400 V for chlorogenic acid; nebulisation gas temperature 400 ∘ C, at 18 psi; needle voltage 4200 V at 11 A; shield voltage 300 V, capillary voltage 35 V, and MS detector at 1600 V for caffeic acid.Argon is used as the collision induced dissociation (CID) gas.The precursor ions for chlorogenic and caffeic acid were monitored at m/z 354.0 and m/z 181.0, respectively.The precursor ion for chlorogenic acid was subject to CID and The m/z of the daughter ions for the ESI-MS/MS analysis was selected by direct infusion of the pure standard solutions in methanol into the MS/MS.The daughter ions were chosen based on the area curves of each daughter as shown by the collision cell breakdown display of the operating software.Conditions were selected so as to produce two daughter ions for identity confirmation, which meets the standard set by the European Commission Directorate for Agriculture guidelines [11].

Results and Discussion
3.1.Chromatograms Obtained.Chromatograms obtained at 325 nm for chlorogenic acid and caffeic acid gave the best compromise between sensitivity and baseline noise.Figure 2 shows a representative HPLC-PDA chromatogram for a sample of Xu Duan raw herb extract.

Extraction Solvent and Method Optimisation. The results
for the extraction solvent optimisation showed water to be the most efficient extraction solvent followed by 50% aqueous methanol, methanol, ethanol, hexane, and ethyl acetate.It was observed that the sonication method was the most effective at extracting both analytes.The sonication method also has the advantage of being simpler to use and is amenable to batch processing.A study was carried out to determine the analytes remaining in the herb following extraction by the selected method.The percentage residual chlorogenic and caffeic acids left over was 0.227 ± 11.5%RSD and 0.036 ± 30.9%RSD (5 replicates each).The %RSD values are high due to the low concentrations of the residual analytes.From these findings it is evident that the selected method is able to recover ≥95.0% of the selected analytes in the raw herb sample.Three types of membrane filters were tested to    For the raw herb, the average recovery across the three spiking levels for chlorogenic acid and caffeic acid is 99.88% and 87.21%, respectively.The %RSD for each recovery level varied from 2.4 to 9.6% (average 4.7%).The calibration curves obtained for each analyte showed good linearity with correlations of determination ( 2 ) > 0.999 and the linear ranges for chlorogenic and caffeic acids were between 8.7-174.5 g mL −1 and 3.3-65.9g mL −1 , respectively.The limits of detection (LOD) for chlorogenic and caffeic acids are determined as three times the SD obtained from seven replicate extractions and analyses of the raw herb were 0.13 and 0.02 mg/g, respectively.The limits of quantitation (LOQ) for chlorogenic and caffeic acids are determined as ten times the SD obtained from seven replicate extractions and analyses of the raw herb were 0.43 and 0.05 mg/g, respectively.

Peak Purity Check and MS Identity Confirmation of
Analytes.The UV spectrum of the standard and sample peaks for chlorogenic and caffeic acids match closely as shown in Figure 3.At the quantitation wavelength of 325 nm there is almost exact overlap.
Identity confirmation is achieved by observing the presence of the correct m/z ions and the intensity ratios of those ions obtained for the sample and standard peaks.The daughter ions of chlorogenic acid (precursor [M-H] − m/z 354, daughter ions m/z 191 and 85) and the daughter ions of caffeic acid (precursor [M+H] + m/z 181, daughter ions m/z 163, 135 and 89) are summarised in Table 3.The proposed fragmentation patterns are shown in Figures 4 and 5.The differences in intensities for the standard and the sample peaks are within the acceptable tolerances set [11].
MS identity confirmation is thus achieved by observing the presence of the correct mass for the [M-H] − ion for  chlorogenic acid and [M-H] + ion for caffeic acid, the presence of the same m/z daughter ions between sample and standard peaks, and their relative intensity ratios.Furthermore the observed m/z fragments can be reasonably explained as shown by the proposed fragmentation pattern for the two acids.

Comparison of Concentration of Chlorogenic and Caffeic
Acids in Xu Duan Raw Herb Obtained from Different Suppliers Determined by HPLC-PDA.The chlorogenic acid and caffeic acid concentrations in the raw herbs are presented in Table 4.
The concentration of chlorogenic acid in the raw herbs varies significantly, with R-I and R-IV having marginally higher levels compared to the other two samples.The caffeic acid concentrations are quite low in all the samples with the sample used for the validation having the highest concentration of caffeic acid.There also does not seem to be a consistent ratio between the concentrations of the two acids.The observed variations could be due to many factors including different growing, harvesting, and storage conditions and the age of the herb.

Conclusion
A HPLC-PDA method has been developed for the quantitation of chlorogenic and caffeic acids in the Xu Duan raw

Figure 2 :
Figure 2: A HPLC-PDA chromatogram for the extract of the Xu Duan raw herb ( = 325 nm, chlorogenic acid at 11.9 min, and caffeic acid at 14.3 min).

Figure 3 :
Figure 3: Comparison of UV spectra of the standard and sample peaks for chlorogenic acid (a) and caffeic acid (b) peaks.

Figure 4 :
Figure 4: Proposed fragmentation pattern for chlorogenic acid showing the formation of the m/z 191 daughter ion from the [M-H] − = 353 precursor ion.

Figure 5 :
Figure 5: Proposed fragmentation pattern for caffeic acid showing the formation of the m/z 163, 135, and 89 daughter ions from the [M+H] + = 181 precursor ion.

Table 1 :
Mobile phase composition for the HPLC analysis of Xu Duan.

Table 2 :
Recoveries for chlorogenic acid and caffeic acid in the Xu Duan.
a Mean recovery ± %RSD calculated from 7 replicate extractions and analyses.

Table 3 :
Comparison of intensities of the qualifier ion(s) for chlorogenic acid (m/z 85) and caffeic acid (m/z 89, 135) obtained for the standard and raw herb.

Table 4 :
Concentration of chlorogenic acid and caffeic acid in several Xu Duan raw herbs determined by HPLC-PDA.theamount of analyte retained by the filter when filtering the prepared raw herb extract.It could be observed that the PVDF filter retained the least percentage of analyte.And had twice the least analyte retention efficiency of the PTFE and nylon filters.Thus PVDF filters were used to filter each sample prior to analysis.3.3.Method Validation.Sample R-I was used for the validation studies.The concentration of analytes in the unspiked sample analysed by the method was 33.2 mg/g and 1.8 mg/g for chlorogenic acid and caffeic acid, respectively.The 50, 100, and 200% spike recovery results for the Xu Duan raw herb are presented in Table2.For the raw herb the recoveries range from 80.52 to 104.56% (average 93.54%). determine