目的 建立红花龙胆Gentiana rhodantha不同药用部位UV-Vis和UPLC指纹图谱,结合多变量分析研究红花龙胆根、茎、叶和花中化学成分的整体分布与质量分数变化。方法 Shim-pack XR-ODS III液相色谱柱(150 mm×2.0 mm,2.2 μm)。流动相为0.1%甲酸水溶液和乙腈,梯度洗脱,柱温40 ℃;检测波长242 nm,进样量0.3 μL;体积流量为1.00 mL/min。电喷雾离子源(ESI)正/负离子模式,多反应监测(MRM)模式,喷雾电压3.5 kV,脱溶剂管温度250 ℃,鞘气和干燥气体积流量分别为3.0 L/min和15.0 L/min,碰撞气体积流量0.15 mL/min;UV-Vis指纹图谱测定,检测波长200～500 nm,狭缝1.0 nm,采样间隔0.5 nm;采用偏最小二乘判别分析(PLS-DA),变量投影重要性准则(VIP)和系统聚类(CA)对不同产地药材进行分类研究。结果 UPLC和UV-Vis方法学考察显示,精密度、稳定性、重复性试验RSD均小于2.00%;马钱苷酸、芒果苷和当药苷加样回收率在97.89%～102.71%,RSD在1.09%～2.88%;11点平滑＋一阶导数预处理为最佳光谱预处理方法,PLS-DA模型R²_cal=0.931 5,RMSEE=0.302,R²_val=0.901,RMSEP=0.341,根、茎、叶和花UV-Vis光谱呈现指纹特性;UPLC指纹图谱相似度分析显示植株化学成分分别在叶部和花部较接近,根部和茎部较相似。不同产地红花龙胆根部UPLC指纹图相似性系数变化范围大,药材质量不稳定。马钱苷酸、芒果苷在叶片和花中量较高[(1.46±0.42)、(51.59±15.45)mg/g];当药苷在根中量较高 [(4.41±3.24)mg/g]。叶片总体呈现较高的化学成分量,马钱苷酸、芒果苷和当药苷量对不同药用部位的区分均有较大贡献。聚类分析显示,不同产地红花龙胆植株化学成分呈现一定地理特征,高海拔地区的药材和低海拔地区的药材成分整体差别较大,聚为不同类群。结论 UPLC和UV-Vis指纹图谱相结合可反映红花龙胆根、茎、叶和花中化学成分的整体分布与质量分数变化,研究结果为野生红花龙胆资源的质量评价提供方法和理论依据。

Objective To develope a method for UV-Vis and UPLC fingerprint on various medicinal parts of Gentiana rhodantha. The chemical compounds variation in roots, steams, leaves, and flowers were studied by using fingerprint data combined with multivariate analysis. Methods Using Shim-pack XR-ODS III liquid chromatographic column (150 mm × 2.0 mm, 2.2 μm) for gradient elution, mobile phase was water with 0.1% formic acid and acetonitrile, temperature was 40 ℃, detection wavelength was 242 nm, injection value was 0.3 μL, and flow rate was 1.00 mL/min. Electrospray ionization (ESI) source and MRM model, the interface voltage was set to 3.5 kV. Desolation line (DL) temperature was 250 ℃. Nebulizing gas and drying gas were nitrogen at a flow rate of 3.0 and 15.0 L/min, respectively. Collision gas was 0.15 mL/min. UV-Vis detection wavelength range at 200 - 500 nm, slit was 1.0 nm, step was 0.5 nm. Multivariate analysis methods including partial least squares discriminant analysis (PLS-DA), variable importance (VIP), and hierarchical cluster were used for this projection. Results The investigation of UPLC and UV-Vis showed RSD was lower than 2.00% for precision, repeatability, and stability. The recoveries of loganin acid, mangiferin, and sweroside were 97.89% - 102.71% and RSD was 1.09% - 2.88%. Pretreatment by 11 smooth + first order was the best data processing method for PLS-DA model (R²_cal = 0.9315, RMSEE = 0.302, R²_val = 0.901, and RMSEP = 0.341). UV-Vis spectra of roots, steams, leaves, and flowers had the significant fingerprint characteristic. Similarity analysis showed the chemical compounds in the leaves and flowers were similar and those in the stems and roots were similar too. Similarity index of root medicinal material has a widely range. The quality of the roots was not stable. The contents of loganic acid and mangiferin were the highest in the leaves and flowers [(1.46 ± 0.42) and (51.59 ± 15.45) mg/g]. The content of sweroside was the hightest in the roots [(4.41 ± 3.24) mg/g]. The total content of compounds of the leaves was higher than other medicinal part, the three compounds, loganic acid, mangiferin, and sweroside were very used for the discrimination of different parts of G. rhodantha. Cluster anslysis showed there were geography characteristics of the constituents. Sample collected from high altitude was higher than samples collected from lower altitude. And they were classified in different groups. Conclusion UPLC and UV-Vis fingerprint could describe the variation of chemical compounds in the roots, steams, leaves, and flowers. All the results could provide the evaluation method and basic theory of G. rhodantha resource.

国家自然科学基金资助项目(81260608);云南省自然科学基金资助项目(2013FD050,2013FZ150,2013FD066,2014FD068)