Detection-confirmation-standardization-quantification: a novel method for the quantitative analysis of active components in traditional Chinese medicines

Materials and chemicals

Danshen was purchased from the Qiancao herbal wholesale company (Beijing). The standards of cryptotanshinone (1), tanshinone I (2), and Tanshinone IIA (3) (Figs. 1A–1C) were obtained from Traditional Chinese Medicine Solid Preparation National Engineering Research Center (Nangchang, Jiangxi Province, China), with a purity of ≥98%. Dimethyl fumarate (4) (Fig. 1D) with a purity of 99% was obtained from Alfa Aesar (Ward Hill, MA, USA). CDCl₃ (99.8% pure) was obtained from Cambridge Isotope Laboratories Inc. (Andover, MA, USA). Acetonitrile and methanol for HPLC were obtained from Fisher Scientific (Fair Lawn, NJ, USA). Formic acid (>98% purity) for HPLC was obtained from Sinopharm Chemical Reagent Co. Ltd. (Shanghai, China). All other chemicals were of analytical grade. Deionized water was obtained from Wahaha Company (Hangzhou, Zhejiang, China).

Identification of cryptotanshinone, tanshinone I, and tanshinone IIA in Danshen by HPLC-ESI-MS/MS

Sample preparation: Dried Danshen samples were powdered using a mill and sieved through a No. 24 mesh. Approximately 0.3 g of the powder sample was weighed and extracted by refluxing in 50 mL of methanol for 1 h. The weight loss was adjusted with methanol after the extraction. One mL of the sample solution was filtered through a 0.45 μm nylon filter into a HPLC amber sample vial for injection.

Data acquisition: An Agilent 1,100 series HPLC system (Agilent Technologies, Santa Clara, CA, USA) equipped with a quaternary solvent delivery system, an on-line degasser, an auto-sample injector, a column temperature controller, and a variable wavelength detector (VWD) was coupled to an Agilent 6,460 triple quadruple mass spectrometer (QQQ-MS) equipped with a dual electrospray ion source (ESI) (Agilent Technologies, Santa Clara, CA, USA) formed the HPLC-ESI-MS/MS system.

The samples were separated using a Diamonsil C18 column (4.6 × 250 mm, 5 μm, Dikma Technologies Inc., Foothill Ranch, CA, USA). The mobile phase was a mixture of acetonitrile (mobile phase A) and water containing 0.1% formic acid (mobile phase B). The gradient elution started at 60% A, followed by a linear increase to 80% A at 30 min, 90% A at 40 min, a linear decrease to 60% A at 41 min, and held constant at 55 min before the next injection was performed. The flow rate was 1.0 mL min⁻¹. The column temperature was maintained at 22 °C. The injection volume was 20 μL.

The mass spectrometric data were acquired using the positive electrospray mode. The full-scan mass spectrum was recorded over the range of m/z 250-325. N₂ was used as the sheath and auxiliary gas of the mass spectrometer. The capillary voltage for ESI spectra was 3.5 kV, and the capillary temperature was set at 300 °C. Ultra-high pure helium was used as the collision gas in the collision-induced dissociation (CID) experiments. The MS/MS product ion spectra were acquired through the CID of the peaks with the protonated molecular ion [M + H] ⁺ of each analyte. The collision energy for the target protonated molecular ions was set at 20 eV to obtain the appropriate fragment information.

Preparation of multicomponent reference standard

Preparation of total tanshinones: The total tanshinones were prepared using a method of preparing tanshinone extract as recorded in the National Commission of Chinese Pharmacopoeia (2020b) edition. Fifty gram of the Danshen powder was extracted by refluxing in 500 mL 95% ethanol for 2 h. The Danshen extract was evaporated under reduced pressure at 40 °C. The resulting solid was washed three times with hot water (80 °C) to remove the water-soluble components and obtain the total tanshinones.

Preparation of multicomponent reference standard (MCRS): the MCRS containing the three tanshinones was prepared using a CXTH LC-3000 semi-preparative HPLC series equipped with a binary solvent delivery system, a Rheodyne 7725i manual injection valve (five mL sample loop) and a UV-visible detector (Chuangxintongheng Co. Ltd., Beijing, China). The total tanshinones were dissolved in 30 mL of methanol and separated using a Thermo BDS Hypersil C18 prep-column (22.2 × 150 mm², 5 μm, Thermo Scientific, Waltham, MA, USA) eluting with methanol/water (77:23, v/v). The flow rate was 9.0 mL min⁻¹, and the detection wavelength was set at 200 nm. The injection volume was 1.0 mL. All the effluents containing cryptotanshinone, tanshinone I, and tanshinone IIA were collected and evaporated to dryness. A total of 210 mg of the MCRS was obtained from 30 mL of the total tanshinone solution.

Quantitative determination of each tanshinones in multicomponent reference standard by QNMR

Generation of NMR calibration series: a series of volumetric solutions containing 20.50, 10.25, 6.83, 3.42 and 1.71 mg/mL of the MCRS and 2.96, 1.48, 0.74, 0.37 and 0.19 mg/mL of dimethyl fumarate in CDCl₃ were prepared and measured by NMR at 600 MHz.

NMR spectral acquisition and processing parameters: The spectra was acquired using a 14.1 T Bruker Avance 600 MHz NMR spectrometer equipped with a 5 mm broad band (BB) inverse detection probe tuned to detect ¹H resonances. The ¹H resonance frequency was 600.13 MHz. All the spectra were acquired at 298 K. A total of 64 scans of 32 K data points were acquired with a spectral width of 9615.4 Hz (16 ppm). A pre-acquisition delay of 6.5 μs, with an acquisition time of 1.7 s, recycle delay of 24 s, and flip angle of 90° were used. The chemical shift of all the peaks was referenced to the tetramethylsilane (TMS) resonance at 0 ppm. The spectra were Fourier transformed to afford a digital resolution in the frequency domain of 0.293 Hz/Point. The phase and baseline corrections of the spectra were carried out manually. Preliminary data processing was carried out using the Bruker software TOPSPIN 2.1.

The signals for the H-15 of cryptotanshinone (4.884 ppm, t, 2H), H-17 of tanshinone I (2.295 ppm, d, 3H), H-17 of tanshinone IIA (2.259 ppm, d, 3H), and olefinic H of dimethyl fumarate (6.864 ppm, s, 2H) were used to determine the contents of cryptotanshinone, tanshinone I, and tanshinone IIA in the MCRS.

The concentrations of the three tanshinones in the MCRS were calculated using the following general equation:

(1) $$\mathrm{C}\mathrm{x}=\frac{\mathrm{A}\mathrm{x}\times \mathrm{W}\mathrm{i}\times \mathrm{N}\mathrm{i}}{\mathrm{A}\mathrm{i}\times \mathrm{M}\mathrm{i}\times \mathrm{N}\mathrm{x}\times \mathrm{V}}$$where C_x (in mM) corresponds to the concentrations of the three individual tanshinones; A_x and A_i correspond to the peak areas of tanshinones and the internal standard. W_i corresponds to the mass of the internal standard (in mg); N_i and N_x correspond to the number of protons of the respective signals of the internal standard and tanshinones used for the quantitative analysis; M_i corresponds to the molecular weight (in Da) of the internal standard, and V (in L) corresponds to the volume of CDCl₃.

The contents of the three tanshinones in the MCRS were calculated using the following general equation:

(2) $$\mathrm{P}\mathrm{x}=\frac{\mathrm{C}\mathrm{x}\times \mathrm{M}\mathrm{x}\times \mathrm{V}}{\mathrm{W}\mathrm{m}}\times 100\mathrm{\%}$$where P_x (%) corresponds to the percentage of the three individual tanshinones in the MCRS; M_x corresponds to the molecular weight (in Da) of the three tanshinones, and W_m corresponds to the mass of the MCRS (in mg).

Quantitative analysis of the three tanshinones in the Danshen sample by HPLC using the MCRS as the reference standard

Sample preparation and HPLC conditions: The Danshen sample was prepared according to the procedure described according to the above steps.

Analytical HPLC was carried out using a Shimadzu LC-20AT series equipped with a quaternary solvent delivery system, an on-line degasser, and auto-sample injector, a column temperature controller, and an SPD-M20A diode-array detector (DAD) (Shimadzu corporation, Kyoto, Japan). Finally, the samples were separated using a Diamonsil C18 column (4.6 × 250 mm, 5 μm, Dikma Technologies Inc.). Notably, the HPLC conditions are compatible to the HPLC-ESI-MS experiments.

Construction of calibration curves: Appropriate amounts of the MCRS were weighed and dissolved into 100 mL of acetone. To construct the calibration curves, 2, 4, 8, 12, 16 and 20 μL of the solution were injected in triplicates. The calibration curves were constructed by plotting the peak areas versus the quantity of the three tanshinones in the MCRS.

Detection of the peaks corresponding to cryptotanshinone, tanshinone I, and tanshinone IIA from the TIC and HPLC chromatograms of Danshen

The initial investigation focused on determining the target components from the TIC and HPLC chromatograms of Danshen and establishing the optimized HPLC conditions suitable for the quantitative analysis of the active components in Danshen. A rough gradient elution was first performed, after which the peaks corresponding to cryptotanshinone, tanshinone I, and tanshinone IIA were searched with their protonated molecular ion [M + H]⁺ from the total ion chromatogram of Danshen (Wang et al., 2017). The results (Figs. 2A and 2C) indicated that the peak with the retention times of 6.4, 6.2 and 8.5 min corresponded to the m/z of 297, 277 and 295, respectively. Under these conditions, cryptotanshinone and tanshinone I were not adequately separated, even though the positions of cryptotanshinone and tanshinone I in the TIC and HPLC spectra of Danshen were successfully assigned. The result obtained by the optimization of gradient elution is shown in Figs. 2B and 2D; the peaks with the retention times of 17.2, 19.0 and 26.1 min corresponded to the m/z of 297, 277, and 295, respectively, in the TIC of LC-MS (Fig. 2B). Moreover, the three tanshinones were adequately separated from other compounds. This condition was adequate for the subsequent quantitative analysis of the three tanshinones in Danshen.

Confirmation of the peak assignments by ESI-MS/MS product ion and UV spectra

The ESI-MS/MS product ion spectra were acquired through the CID of the peaks obtained with the protonated molecular ion [M + H]⁺ of the three tanshinones. The proposed MS fragmentation pathway of the compounds with the protonated molecular ions of m/z 297, 277 and 295 were summarized in Fig. 3, consistenting with the MS fragmentation pathway of cryptotanshinone, tanshinone I, and tanshinone IIA as studied by Wang et al. (2017). The UV spectra shown in Fig. 4 were obtained directly from the HPLC experiments. The UV spectra of the three peaks matched with those of cryptotanshinone, tanshinone I, and tanshinone IIA, as reported in the current literature (Huang et al., 2018).

Preparation of the MCRS containing mainly cryptotanshinone, tanshinone I, and tanshinone IIA

The MCRS containing mainly cryptotanshinone, tanshinone I, and tanshinone IIA was prepared by the preparative HPLC of the total tanshinone to perform a quantitative analysis of the three components, without individual reference standards. The chromatograms of the total tanshinone and MCRS were shown in Fig. 5A. Peaks 1, 2, and 3 were assigned to cryptotanshinone, tanshinone I, and tanshinone IIA, respectively, by collecting the main peaks in the preparative HPLC separately and then analyzing them by analytical HPLC.

Determination of the contents of cryptotanshinone, tanshinone I, and tanshinone IIA in the MCRS by NMR

The contents of the three tanshinones in the resulting MCRS were determined directly by ¹H NMR analyses. The ¹H NMR spectrum of the MCRS mixed with the internal standard, dimethyl fumarate, was shown in Fig. 6. The main signals were assigned to the three tanshinones according to the references, excluding the signals belonging to dimethyl fumarate (Mei et al., 2019; Zeng et al., 2017; Wu, Jiang & Wu, 2015). The composition of the MCRS was confirmed again by analyzing the signals observed in the ¹H NMR spectrum. To perform an accurate quantitative analysis, the data processing technique was used which named the Lorentz deconvolution function in the Bruker software TOPSPIN 2.1 . The linearity of signals were fitted by the Lorentz method. Next, the deconvolution was performed, and the peak areas were automatically generated. Regarding the quantitative analysis of the mixtures by NMR using the data processing, the accuracy was determined by the likelihood of lineshape fitting. The results showed that the contents of crytotanshinone, tanshinone I, and tanshinone IIA were at 1.635 g/100 g, 0.718 g/100 g, 1.953 g/100 g in Danshen sample, respectively, by MCRS method.

Under the preparation of MCRS of Danshen, the selection of the signals in the NMR spectra for the quantitative analysis of the three tanshinones and the validity of the NMR method were performed. The analytical method of NMR was achieved a high separation efficiency which was shown in Figs. 7A–7O. The NMR spectrum is very reproducible with good reproducibility and target three compounds high separation efficiency.

Determination of the contents of the three tanshinones in the Danshen sample by HPLC using the MCRS as the reference standard and comparison of the results with those obtained using individual reference standards

The calibration curve of the MCRS was constructed to investigate the validity of the NMR method. The calibration curves were constructed by plotting the concentrations of the three tanshinones versus that of dimethyl fumarate. Next, the linear regression lines were calculated. The calibration graph demonstrating the linearity of the NMR response with increasing concentrations of the three tanshinones was shown in Fig. 8.

Approximately 5 mg of the MCRS was dissolved into 100 mL of acetone to construct the calibration curve of the three tanshinones, corresponding to ~1.5, 0.5 and 1.7 mg of individual cryptotanshinone, tanshinone I, and tanshinone IIA, respectively. The contents of the three tanshinones in the Danshen sample were determined according to the calibration curves based on individual standards. The content of cryptotanshinone was determined at 1.635 g/100 g and 1.627 g/100 g, respectively, using MCRS and standard quantitation. Tanshinone I was detected at 0.718 g/100 g and 0.727 g/100 g, respectively, based on the two methods. Tanshinone IIA was measured at 1.953 g/100 g and 1.886 g/100 g, respectively. The mean contents of the three compounds obtained by the two methods were tested for no significance difference at the P ≤ 0.05 level. The results were compared with those obtained using individual reference standards. The deviations of the three tanshinones determined by the two methods were minimal (Fig. 9), and this could be attributed to the maximum likelihood lineshape fitting of their signals in the NMR spectrum of the MCRS.