Tian Run-tao, Xie Pei-shan, Liu He-ping
Chromap Institute of Herbal Medicine Research, Zhuhai, Guangdong, China.
J Chromatogr A. 2009 Mar 13;1216(11):2150-5. doi: 10.1016/j.chroma.2008.10.127. Epub 2008 Nov 28.
Chaihu (Bupleuri Radix), roots of Bupleurum chinense and B. scorzonerifolium, is an authentic Chinese Materia Medica in the Chinese Pharmacopoeia. Some other species such as the roots of B. falcatum, B.bicaule and B. marginatum var. stenophyllum similar to Chaihu can also be occasionally found in local raw herb markets. The quality of 33 lots of authenticated Chaihu samples vs. 31 lots of commercial samples was evaluated by both high-performance liquid chromatography-evaporative light scattering detector (HPLC-ELSD) and high-performance thin-layer chromatography (HPTLC) analyses of its principal bioactive components (saikosaponins). The pre-treated data acquired from both HPLC fingerprints and HPTLC fluorescent images were processed by chemometrics for similarity and pattern recognition, including Artificial Neural Networks (ANNs), k-nearest neighbor (k-NN) and an expert's panel. It was apparent that k-NN classifier exhibited good performance with sufficient flexibility for processing HPTLC fingerprint images which were otherwise not easily dealt with by other algorithms due to the shift of R(f) values and varying hue/saturation of the band colours between different TLC plates. These two chromatographic fingerprint methods can be considered complementary measure of quality control. The roots of Chaihu from different species of the genus Bupleurum could readily be distinguished from each other so that commercial samples can easily be classified. Chaihu collected from several major herbal distribution centers was found to belong to B. chinense with great variation in the content of its major saikosaponins.
柴胡(柴胡根),为伞形科植物柴胡和狭叶柴胡的根,是《中国药典》中的正品中药材。在当地生药市场偶尔也能发现其他一些类似柴胡的物种,如镰叶柴胡、双茎柴胡和窄竹叶柴胡的根。通过高效液相色谱 - 蒸发光散射检测器(HPLC - ELSD)和高效薄层色谱(HPTLC)对33批经鉴定的柴胡样品与31批市售样品的主要生物活性成分(柴胡皂苷)进行分析,评估其质量。从HPLC指纹图谱和HPTLC荧光图像获得的预处理数据通过化学计量学进行相似性和模式识别处理,包括人工神经网络(ANNs)、k近邻(k - NN)和专家小组。很明显,k - NN分类器表现出良好的性能,具有足够的灵活性来处理HPTLC指纹图像,由于不同TLC板之间R(f)值的偏移和条带颜色的色调/饱和度变化,其他算法难以处理这些图像。这两种色谱指纹方法可被视为质量控制的补充措施。柴胡属不同物种的柴胡根很容易相互区分,因此市售样品很容易分类。从几个主要草药分销中心收集的柴胡被发现属于柴胡,其主要柴胡皂苷的含量有很大差异。