Chromatographic analysis of salicylic compounds in different species of the genus Salix.

The separation of nine phenol glycosides--salicin, salicortin, 2'-acetylsalicortin, populin, tremulacin, salidroside, triandrin, picein and helicin--by normal phase (NP), reversed phase (RP) HPLC techniques and a coupling of NP and RP monolithic silica columns was studied. Among the above nine compounds only five--salicin, populin, tremulacin, salidroside and triandrin--were resolved in an NP system with a mobile phase comprising hexane/isopropanol/methanol (87:12:1, v/v/v). Optimized separation was performed with two coupled monolithic silica columns of different polarity (bare silica and RP-18). The method was applied to verify the presence of salicylic compounds and other phenolic derivatives in the bark of six species from the genus Salix, namely S. purpurea, S. daphnoides clone 1095, S. alba clone 1100, S. triandra, S. viminalis, and S. herbacea. Gradient elution with a mobile phase composed of acetonitrile and water containing 0.05% of trifluoroacetic acid, with increasing acetonitrile concentration from 3% to 48%, was chosen as optimal. For the selective detection of the salicylic compounds, an evaporative light scattering detector was employed along with a UV detector. The differences in the composition of phenols in the different plant materials were confirmed. Additionally, it must be emphasized that for the first time the presence of 2'-acetylsalicortin was revealed in S. alba clone 1100. Furthermore, an SPE-HPLC method was developed for the rapid analysis of the salicin content, analyzed as free and total fraction, in willow barks. The determined concentrations of total salicin varied from 25.4 mg/g in S. alba clone 1100 to 96.47 mg/g in S. daphnoides clone 1095.

Evaluation of different warping methods for the analysis of CE profiles of urinary nucleosides.

Nowadays, numerous metabolite concentrations can readily be determined in a given biological sample by high-throughput analytical methods. However, such raw analytical data comprise noninformative components due to many disturbances normally occurring in the analyses of biological material. To eliminate those unwanted original analytical data components, advanced chemometric data preprocessing methods might be of help. Here, such methods are applied to electrophoretic nucleoside profiles in urine samples of cancer patients and healthy volunteers. In this study, three warping methods: dynamic time warping (DTW), correlation optimized warping (COW), and parametric time warping (PTW) were examined on two sets of electrophoretic data by means of quality of peaks alignment, time of preprocessing, and way of customization. The application of warping methods helped to limit shifting of peaks and enabled differentiation between whole electropherograms of healthy and cancer patients objectively by a principal component analysis (PCA). The evaluation of preprocessed data and raw data by PC analysis confirms differences between the applied warping tools and proves their suitability in metabonomic data interpretation.