The reference standards of naringin, hesperidin, aloe emodin, rhein, honokiol, magnolol, emodin, chrysophanol and 1, 8-dihydroxyanthraquinone, were purchased from the National Institute for the Control of Pharmaceutical and Biological Products of China. The purity of all reference standards was above 98%. Methanol and acetic acid were purchased from TEDIA (USA) and Chongqing Chemistry Co Ltd (China) respectively. Granules of Radix et Rhizoma Rhei, Cortex Magnoliae Officinalis, Fructus Aurantii Immaturus and Natrii Sulfas were purchased from Chengdu Green Pharmacy Co Ltd (China). All reagents used in the study were of HPLC grade. Ultra-pure water used for the mobile phase was prepared by an ultra-pure water purification system (Shanghai Anting Co Ltd, China).

Each DT granule sample (5 g) was ground to fine powder (80 meshes). One hundred milligrams of the fine powder was accurately weighed and dissolved in 50 ml of 65% (v/v) aqueous methanol solution. The internal standard (1, 8-dihydroxyanthraquinone, 1 ml, 5 mg/ml) was added to the solution which was then extracted in an ultrasonic water bath (70°C) for 120 min. The sample solution was centrifuged (× 3000 g) for 15 min and the supernatant was filtered through a 0.45 μm Millipore membrane filter (Millipore, USA).

All separations were performed on a Waters HPLC system equipped with a binary pump, auto-sampler and Waters 2487 dual λ absorbance detector (Waters, USA). A RP-C₁₈ HPLC column (150 × 4.6 mm, S-5 μm, 12 nm) and a guard column (GL Sciences, Japan) were used. Data were processed on a Waters Empower chromatographic workstation (Waters, USA).

The mobile phase was a mixture of solvent (A) 100% methanol and solvent (B) 0.2% aqueous acetic acid (pH 3.12, 1:500, v/v) (A-B: 20 min 36:64; 19 min 65:35; 21 min 70:30). The flow rate was 1.0 ml/min and injection volume was 10 μl. A ten min interval was given between sample injections. The column effluents were simultaneously monitored at 254 nm (for aloe emodin, rhein, chrysophanol and emodin) and 280 nm (for honokiol, magnolol, naringin and hesperidin). The column temperature was controlled at 26°C.

While it is difficult to separate naringin and hesperidin, honokiol and magnolol, aloe emodin, rhein, emodin and chrysophanol as they are structurally similar, separation was improved when 0.2% (1:500, v/v) aqueous acetic acid was added to the sample solution. A mixture of water and methanol was chosen for the separation as all eight markers dissolve in both water and methanol. The ratio 38:62 (v/v) of methanol and water was optimal for the separation of hesperidin and naringin, and 65:35 (v/v) for the separation of the other six markers, hence the above mentioned gradient elution program. The HPLC column of C₁₈ (150 × 4.6 mm) was chosen to ensure the run time to be within 60 min.

All calibration curves were plotted based on linear regression analysis of the concentrations of the eight markers (x) versus the peak area ratio (y). The concentration of the internal standard was 100 μg/ml for all markers. Each calibration curve was obtained with six concentrations in triplicates.

The standard solutions of the eight markers were diluted with 65% aqueous methanol to provide appropriate concentrations with the internal standard (100 μg/ml). The diluted solutions were injected three times. The quantity of each marker was determined according to the corresponding calibration curve. The limit of detection (LOD) for each marker was determined when the peak signal-to-noise ratio was at 3. The limit of quantification (LOQ) for each marker was determined when the RSD of the peak's quantity from the corresponding calibration curve was less than 20%.

Intra-day and inter-day variability was studied to evaluate the precision of the method. Three solutions (high, medium and low concentrations) of the eight markers with the internal standard (100 μg/ml) were prepared. The quantity of each component was calculated from the corresponding calibration curve. The relative standard deviation (RSD) was taken as the measure of precision. The inter-day reproducibility test was carried out on three different days.

Three known quantities of the eight markers were added to 50 ml of 65% aqueous methanol solution with 1 ml of the internal standard (5 mg/ml) and the powder of DT (400 mg). The resultant samples were extracted and analyzed as described above. The quantification of each marker was obtained according to the corresponding calibration curve.

Statistical analysis was performed with the PEMS3.1 for Windows (Sichuan University, China). The quantitative results were presented as mean with standard deviation if the distribution was normal and as maximum, median and minimum if the data distribution was not normal. Linear regression for the calibration curves was tested. Single factor ANOVA was performed on each component for their determination in the three batches of DT. The comparison of two independent samples was evaluated with the Student's t-test when the distribution was normal. Test result was considered to be statistically significant when its two-tailed P value was less than 0.05.

A chromatogram of standard naringin, hesperidin, aloe emodin, rhein, honokial, magnolol, emodin and chrysophanol is shown in Figure 2A. These standards are well resolved with relatively high sensitivities at 14.484, 16.429, 30.065, 37.355, 40.579, 47.373, 49.060 and 54.463 min. The internal standard is separated at 43.309 min. The peaks are well separated from each other (peak purity > 95%) and show a characteristic profile.

All eight calibration curves displayed good linear relationships under the present chromatographic conditions (Table 1). The LOD and LOQ were in the ranges of 0.21–0.43 ng/μl and 0.76–1.74 ng/μl respectively for all eight markers. The recovery tests of all eight markers were performed by adding the standards to the powder. The results indicated that the recovery for all eight markers were in the range of 95–104% (Table 2). The overall intra- and inter-day variability was less than 4.0% for all eight markers (Tables 3 and 4).

The method described above was subsequently applied to the simultaneous qualification of the eight markers in three batches of commercial DT samples. The data of the three batches of samples are highly consistent (Figure 2B and Table 5). The present method is feasible and reliable for simultaneous determination of many anthraquinone, flavone and neolignan compounds in complex Chinese medicine formulae such as DT.