The aerial parts of Schizonepeta tenuifolia were obtained at seven different harvesting times (listed in Table 1) from a Good Agriculture Practice (GAP)-based herbal garden which is located in Anguo, Hebei province. Herbs were then authenticated by Professor Qinan Wu, the botanist of the College of Pharmacy at Nanjing University of Chinese Medicine (Nanjing, China). Voucher specimens were also deposited at the herbarium at the above location. Plant materials were carefully cleaned and separated into two or three parts: leaf, stem and spike (if it existed), then a clevenger apparatus method was employed for the steam-distillation of volatile oils from separated leaf and spike parts. About 60 g of dried leaves or spikes were used for every sample.

Ethyl acetate (gradient grade), n-decane (GC grade), naphthalene (GC grade) were purchased from Sinopharm Chemical Reagent (Shanghai, China). Chemical standards of 3-octanone, β-myrcene, (+)-menthofuran, and were purchased from Sigma-Aldrich (Vienna, Austria); (−)-menthone and (+)-pulegone were purchased from National Institutes for Food and Drug Control (Beijing, China); 3-octen-1-ol, (−)-limonene and β-caryophyllene were purchased from Tokyo Chemical Industural Co., Ltd (Tokyo, Japan). Their structures are presented in Figure 1.

Volatile oil samples were extracted by water distillation for 4 hours from 60 g of ST leaf and spike, respectively, using a set of standard apparatus, according to the procedure described in the Pharmacopoeia of the Peoples’s Republic of China [19]. The volatile oils were isolated from water and dried over anhydrous sodium sulphate until the last traces of water were removed and then stored in the dark glass bottles at −20 °C. Fifty milligrams of the volatile oil sample were accurately weighed and transferred into a 10 mL volumetric flask, and 1 mL solutions of the two internal standards (containing 1.727 mg/mL of naphthalene and 0.252 mg/mL of n-decane in ethyl acetate) were transferred into the same flask, respectively, and made up to volume with ethyl acetate.

A mixed standard stock solution containing the eight reference compounds was prepared in ethyl acetate. The concentration of each compound in the stock solution was 284.64 μg/mL for 1-octen-3-ol (1), 53.24 μg/mL for 3-octanone (2), 100.12 μg/mL for β-myrcene (3), 802.80 μg/mL for (−)-limonene (4), 3.990 mg/mL for (−)-menthone (5), 24.42 μg/mL for (+)-menthofuran (6), 7.389 mg/mL for (+)-pulegone (7) and 157.28 μg/mL for β-caryophyllene (8), respectively. Working standard solutions were prepared by diluting the mixed standard stock solution with ethyl acetate to give six different concentrations within the ranges: (1), 5.7–227.7 μg/mL; (2), 1.1–42.6 μg/mL; (3), 2.0–80.1 μg/mL; (4), 16.1–642.2 μg/mL; (5), 79.8–3191.7 μg/mL; (6), 0.5–19.5 μg/mL; (7), 147.8–5911.0 μg/mL and (8), 3.1–125.8 μg/mL for calibration curves. 1.0 mL of double internal standards was also added into each solution of calibration.

All of the sample and standard solutions were filtered through a 0.45 μm membrane prior to injection. They were stored in a refrigerator at 4 °C before analysis.

1.0 μL of assayed solution was injected into Agilent 6890N (Agilent Technology, Austin, TX, USA) gas chromatograph (consisting of a 5975B mass selective detector and Agilent chemstation software package) equipped with an auto sampler (split ratio; 20:1). The analysis GC column was an HP-5MS (5% Phenyl Methyl Siloxane) capillary column (30 m × 0.25 mm ID, 0.25 μm film thickness). Helium (He) was the carrier gas at a flow rate of 1.0 mL/min. The GC temperature program was as follows: initial temperature was 50 °C, increased to 90 °C at a rate of 10 °C/min, at the temperature maintained for 5 min; then increased to 160 °C at a rate of 10 °C/min and held for 10 min. The injection temperature was 220 °C, transfer line temperature was 280 °C, and ion source temperature was 230 °C. The mass spectrometer was operated at 70 eV in the electron impact mode with selected ion monitoring (SIM) mode. The selected ion groups for the identification of eight compounds in SIM mode were listed in Table 2. The dwell time for each ion was set at 50 ms. n-decane and naphthalene were used as two internal standards (IS) for quantification. The GC-MS-SIM chromatograph of eight compounds was shown in Figure 2. The chemical identities of the resulting components were determined by matching their recorded mass spectra with the data bank mass spectra (NIST libraries) provided by the instrument software. The structures of the components were further confirmed by authentic standards analyzed under the conditions mentioned above.

The dilute solution of the reference compounds was further diluted to a series of concentrations with ethyl acetate to assess the limits of detection (LOD) and quantification (LOQ). The LOD and LOQ were determined as signal-to-noise (S/N) ratios of 3 and 10, respectively. The intra- and inter-day precision was determined by analyzing calibration samples during a single day and on three consecutive days, respectively. To confirm the repeatability, five different working solutions were analyzed. The R.S.D. was taken in order to measure precision and reproducibility. The recovery test was used to evaluate the accuracy of this method. In the test, reference compounds were added to the L3 sample, and then analyzed as described above. The average recoveries were estimated by the formula: recovery (%) = (amount found − original amount)/amount added × 100%, and R.S.D. (%) = (S.D./mean) × 100%. For determination of the eight compounds, a calibration curve with double internal standards for each marker was constructed and tested for linearity. Their regression equations were calculated in the form of Y = AX + B, where Y and X (μg/mL) were peak area ratio of each component to the internal standard and compound concentration injected, respectively.

GC-MS-SIM Chromatogram of volatile oils from ST leaves and spikes, were shown in Figure 2. All the main components were separated completely in 35 min, and the eight compounds for quantification were identified on the basis of comparison of their mass spectra with NIST05 database through MSD ChemStation D.05.01, and with mass spectra of standard compounds, respectively. The results were listed in Table 2, and the structures of the identified compounds were shown in Figure 1.

The proposed chromatographic method was validated to determine the linearity, LOD, LOQ, intraday and interday precisions, and accuracy. All calibration curves showed good linearity (r² > 0.9990) within relatively wide concentration ranges, and the overall LODs and LOQs were less than 12.5 and 41.7 μg/mL, respectively (Table 3). The intra- and interday variations, repeatability, and stability RSD values of the eight compounds were all <2.37% and were shown in Table 4. The overall recoveries fell between 97.50 and 105.45% for the eight reference compounds, with RSDs of <2.76% (Table 5), which indicated that the established method was accurate enough for the determination of the eight compounds in the ST volatile oil samples.

The established GC-MS-SIM method was then subsequently applied to the simultaneous determination of the eight compounds in the volatile oil samples collected at different harvesting times. The assay results of eight compounds in volatile oils were summarized in Table 6. (+)-Pulegone and (−)-menthone, two p-menthane monoterpenes (also found in other aromatic plants in the Lamiaceae, probably due to the same biosynthetic pathway [20]) were determined as being the predominant constituents of the ST volatile oil, accounting for 85%~95% of total oils. Harvesting period (around flowering time) was a key stage for the biosynthesis of menthone in S. tenuifolia, in which pulegone was rapidly converted into menthone in leaves under regulation of (+)-pulegone reductase. Meanwhile, such bioconversion occurred much slower in spikes. (+)-pulegone was also found to be the single dominant constituent of spikes.

In Table 6, the content of (−)-menthone, an important bio-active constituent, increased from the growing stage (in late August) to the stage of fruitage (in early October) in ST leaves. However, it decreased from the spikes’ formation stage (in early September) to the stage of fruitage in ST spikes. The content of (−)-limonene, another bio-active constituent, did not vary much with harvesting time in ST leaves or spikes. The content of (+)-pulegone, a main toxic constituent, obviously decreased from the growing stage to the stage of fruitage in ST leaves, but it showed a slightly increasing trend from the spikes’ formation stage to the stage of fruitage in ST spikes. The content of (+)-menthofuran, another toxic compound, had virtually the same variation trends in ST leaves and spikes. The contents of other four compounds for quantification remained relative stable during the different harvesting times.

All of the results demonstrated that the optimum harvesting time of ST leaves was possibly in the stage of fruitage (in early October), and the optimum harvesting time of ST spikes was possibly in the spikes’ formation stage (in early September). However, the content of (+)-pulegone, only one dominant constituent of ST spikes, showed a very slightly increasing trend during the harvesting times. Therefore, harvesting time is possibly not a key factor for influencing the quality of ST spikes.