**2. Results**

#### *2.1. Identification of the Constituents in G. straminea by UPLC-Q Exactive Mass Spectrometer*

The no. 3 sample from Sichuan province (SC-M-03-r3b) was selected for identification by UPLC-Q mass assay as the sample showing the most peaks during 0–30 min. Forty-six peaks were detected in *G. straminea* from MS and MSn in negative and positive ion mode by a UPLC-Q exactive mass spectrometer (Figure 1). The mass accuracy for all assigned components was less than 5 ppm compared with the molecular formulas of the published compounds in *G. straminea*. Gentiopicroside, swertiamarin, 6- -*O*-β-d-glucopyranosyl-gentiopicroside, sweroside, loganic acide and loganin were identified by comparison with the retention time and mass fragmentation of reference standards. The SciFinder Scholar and the PubChem databases were searched for the spectral data of other compounds reported previously in the genus *Gentiana* and *G. straminea* to identify the constituents of the herb [14–20]. Forty-three of these were identified by comparing the retention times and mass spectrometry, which has already been summarized [21], including 20 iridoids, 16 secoiridoids, 8 flavonoids, 2 triterpenoids, 2 lignins, 2 alkaloids, and 2 saccharides (Table 1).

Among these compounds, gentiopicroside (a type of secoiridoid) and loganin (a type of iridoid) were officially listed in the Chinese Pharmacopoeia 2015 edition for quality control of the herb [22]. In positive ion mode, secoiridoids and iridoids all showed highly abundant proton and sodium ion adducts, but a relatively lower proportion of potassium, and they also showed highly formate and chlorine ion adducts in negative ion mode. Moreover, small peaks for [2M <sup>+</sup> Na]+, [2M <sup>+</sup> HCOOH <sup>−</sup> H]− (Supplementary Figures S1 and S2), could be observed in the spectra for positive and negative experiments. All of these mass signals were helpful in the identification of secoiridoid and iridoid components [23,24].

**Figure 1.** Total ion current chromatograms of substances in the extract of *G. straminea* (the no. 3 sample of Sichuan province) with positive and negative ion modes.

For the first time, a pair of isomers (loganin and secologanol) were identified by mass spectrometry based on their fragmentation pathway. Loganin (a type of iridoid), eluted at 9.95 min, showed fragment ions at *m*/*z* 413.14157 [M + Na]<sup>+</sup> (Figure 2a) and *m*/*z* 803.29346 [2M + Na]+, with the elemental composition of C17H26O10Na (calculated 413.14240) in positive ion mode. In MS2, the compound formed product-characterized ions at *m*/*z* 285.09409 with the neutral loss of C6H8O3 (Δ*m* = 128.04831 Da). It also showed ions at *m*/*z* 185.04211 Da as glucose residue adducts sodium ion, and ions at *m*/*z* 219.06264 Da indacted the compound with the neutral loss of a glucose and methanol. The proposed fragmentation pathway of the loganin is shown in Figure 3a.

**Figure 2.** The mass and MS<sup>2</sup> spectra of (**a**) loganin and (**b**) secologanol.


#### *Molecules* **2019**, *24*, 4478


**Table 1.** *Cont.*

Notably, secoiridoids have always produced fragmentation by Retro–Diels–Alder (RDA) cleavage of the aglycon moiety [24] and these fragmentations are quite different from that of iridoids. Secologanol (a type of secoiridoid), eluted at 12.35 min, showed fragment ions at *m*/*z* 413.141172 [M + Na]<sup>+</sup> (Figure 2b) and *m*/*z* 803.29358 [2M + Na]<sup>+</sup> (Supplementary Figure S1b), and after generated [M + Na <sup>−</sup> Glu]<sup>+</sup> at *m*/*z* 251.08833 Da, which was identical to the aglycone fragment corresponding to the neutral loss of a glucose unit (Δ*m* = 162 Da). The precursor ion produced characterized ions at *m*/*z* 181.04694 with an RDA cleavage reaction of the base skeleton and showed the neutral loss of C4H6O (Δ*m* = 70 Da). The proposed fragmentation pathway of the secologanol is shown in Figure 3b.

**Figure 3.** (**a**,**b**) The proposed fragmentation pathways of loganin and secologanol, respectively.

#### *2.2. Multivariate Analysis of the Global Metabolomics Data*

To globally evaluate the chemical consistency of *G. straminea* samples of different geographical origins, the UPLC-Q exactive mass datasets were subjected to partial least-squares discriminant analysis (PLS-DA) and orthogonal partial least squared discriminant analysis (OPLS-DA) to highlight differences among the *G. straminea* samples. As shown in Figure 4a, the 42 samples were roughly clustered into three groups by PLS-DA analysis. With OPLS-DA analysis, all of the samples were clearly categorized into three groups in 3D space (Figure 4b), 10 samples (green dots) from Gansu province were assigned to group I, 7 samples (blue dots) from Qinghai were assigned to group II, and 25 samples (red dots) from Sichuan province were assigned to group III.

**Figure 4.** (**a**) Partial least-squares discriminant analysis (PLS-DA) score plot of *G. straminea* from three geographical origins. Green dots: samples from Gansu province, blue dots: samples from Qinghai province, red dots: samples from Sichuan province. (**b**) Orthogonal partial least-squares discriminant analysis (OPLS-DA) in 3D score plot of *G. straminea* from three rgeographical origins. Green dots: samples from Gansu province, blue dots: samples from Qinghai province, red dots: samples from Sichuan province. (**c**) Loading plot of OPLS-DA analysis of *G. straminea*. compound a: gentiopicroside (*tR* 9.78 min, *m*/*z* 379.09982), compound b: vitexin (*tR* 14.54 min, *m*/*z* 411.12610), compound c: swertiamarin (*tR* 8.51 min, *m*/*z* 397.11060), compound d: gentiobiose (*tR* 1.33 min, *m*/*z* 365.10510), compound e: sweroside (*tR* 10.28 min, *m*/*z* 381.11551), compound f: 2-methoxyanofinic acide (*tR* 22.15 min, *m*/*z* 271.08786), compound g: loganic acide (*tR* 6.87 min, *m*/*z* 399.12595), and compound h: 1β, α,3α,24-tetrahydroxyursa-12,20 (30)-dien-28-oic acid (*tR* 25.53 min, *m*/*z* 503.33640).

In order to identify the most significant discriminatory features between these regions that could act as potential barcodes, an extended statistical analysis was used to provide loading score plots of OPLS-DA (Figure 4c). In this plot, according to the importance of discriminating geographical characteristics, the size and color of these points have been highlighted, as seen in Figure 4c. The eight characteristic compounds were identified as: gentiopicroside (*tR* 9.78 min, *m*/*z* 379.09982); vitexin (*tR* 14.54 min, *m*/*z* 411.12610); swertiamarin (*tR* 8.51 min, *m*/*z* 397.11060); gentiobiose (*tR* 1.33 min, *m*/*z* 365.10510); sweroside (*tR* 10.28 min, *m*/*z* 381.11551); 2-methoxyanofinic acid (*tR* 22.15 min, *m*/*z* 271.08786); 1β,2α,3α,24-tetrahydroxyursa-12,20(30)-dien-28-oic acid (*tR* 25.53 min, *m*/*z* 503.33640); and loganic acid (*tR* 6.87 min, *m*/*z* 399.12595).

In the loading score plot of OPLS-DA, it is also clearly shown that samples from Gansu province are characterized by a high content of gentiopicroside, vitexin, and loganic acid, while samples from the Sichuan habitat location had a higher relative concentration of swertiamarin, and the populations of the Qinghai province were characterized by high contents of gentiobiose, sweroside, 2-methoxyanofinic acid, and 1β,2α,3α,24-tetrahydroxyursa-12,20(30)-dien-28-oic acid.

#### *2.3. Anti-Inflammatory E*ff*ect of Characterize Components*

With the aid of multivariate statistical analysis, gentiopicroside was confirmed as the most characteristic marker to distinguish the geographical origin of *G. straminea*. Anti-inflammatory activity is directly associated with therapeutic effects on arthritis, thus, to further evaluate the anti-inflammatory pharmacological function of characteristic markers, the inhibiting activities of nitric oxide production were evaluated in the macrophage cell line RAW 264.7 [25].

The cytotoxicity of gentiopicroside and LPS in RAW 264.7 cells were examined using CCK8 assay. As shown in Figure 5a, no significant difference in the viability of RAW 264.7 cells were observed among groups, suggesting that the concentrations of gentiopicroside and LPS used in the present study did not show any significant cytotoxic effects on RAW 264.7 cells. The inhibition effect of NO production induced by LPS in the macrophage-derived RAW 264.7 cells of the compound was assayed. It was found that the level of NO gradually decreased in a concentration-dependent manner in gentiopicroside. At a concentration of 100 μM, the compound significantly inhibited NO generation (83.76 ± 0.57%), with IC50 values of 44.2 ± 6.4 μM (Figure 5b).

In addition to gentiopicroside, nitric oxide production was also suppressed by loganic acid, swertiamarin, and vitexin. These compounds possessed the most potent inhibitory activity against NO production with IC50 values of 23.13 ± 5.4, 13.65 ± 7.1, and 15.71 ± 6.20 μM, respectively. All the results showed that gentiopicroside, loganic acid, swertiamarin, and vitexin were able to effectively inhibit NO production induced by LPS (1 μg/mL) in a dose-dependent manner in RAW 264.7 cells (Supplementary Figure S3). Specifically, the inhibiting effect against NO production of sweroside was also measured. However, the compound did not show inhibiting activities on LPS-induced NO production in RAW 264.7 macrophages, as has been reported [26], because the cells were incubated with sweroside for only 24 h in the experiment.

**Figure 5.** Effects of gentiopicroside on lipopolysaccharide (LPS)-induced NO production in RAW 264.7 cells. (**a**) RAW 264.7 cells were exposed to different concentrations of gentiopicroside (0, 5, 10, 20, 40, 80, and 100 μM) with or without LPS (1 μg/mL) for 24 h. Cell viability was determined by using the CCK-8 method. (**b**) RAW 264.7 cells were incubated with gentiopicroside (0, 5, 10, 20, 40, 80, and 100 μM) with stimulated by LPS (1 μg/mL) for 24 h. Extracellular levels of NO in culture media were measured using commercial Griess reagent. Data were folds of control and expressed as the mean ± SEM of six independent experiments. ### *p* < 0.001 compared with the control group. \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001, compared with the LPS alone.

#### **3. Conclusions**

In the present research, a selective and specificity approach was established to illustrate the chemical composition of 42 samples in *G. straminea* with a UPLC-Q exactive mass spectrometer, and an overall chemical profile of the herb was obtained. The significant differences in metabolite compositions between three geographical origins have been identified with multivariate analyses. The anti-inflammation effects of biomarkers on LPS-induced NO production in RAW264 macrophages

were examined. The results suggested that samples from Gansu province have a higher content of gentiopicroside and loganic acid, and showed better anti-inflammatory effects than others. From the legal point of view [27], the result also confirmed that samples of Gansu province have better quality than other samples. Altogether, this finding is crucial in realizing the discrimination of the botanical origin of *G. straminea*, and evaluating the herb quality.
