Wavelength Selection for Near-Infrared Spectroscopic Analysis of Teicoplanin

Abstract

Teicoplanin (TCP) is a multiple drug-resistant lipoglycopeptide antibiotic produced by fermenting Actinoplanes teichomyceticus . In this study, a mixture of TCP with the Tris-HCl buffer (TCP-Tris- HCl) was used to simulate TCP fermentation broth. The reagent-free, rapid and simultaneous quantitative analysis models for TCP and Tris in the TCP-Tris-HCl mixtures were established by near-infrared (NIR) spectroscopy. The equidistant combination partial least squares (EC-PLS) method and the equivalent model sets were proposed, the simplest equivalent model with the smallest number of wavelengths were further selected. The initial wavelength, number of wavelengths, number of wavelength gaps, number of PLS factors were 1520 nm, 28, 5, 5 for TCP and 1084 nm, 13, 6, 4 for Tris, respectively. Compared with the optimal EC-PLS models, the simplest equivalent models adopted fewer wavelengths. Thus, the redundant wavelengths were removed, the models were further simplifie d. The root-mean-square errors (SEP) and correlation coefficients (R P ) for prediction were 0.043 mg·mL -1 and 0.9998 for TCP, and 0.222 mg·mL -1 and 0.9989 for Tris, respectively. The results indicate that NIR method can be applied to highly accurate quantitative analysis for TCP and provide valuable references for further application to TCP fermentation broth.