The Electrochemical and Statistical Evaluation of Isolation of Mellitin and Apamin from Honey Bee (Apis Mellifera) Venom

Abstract

We present in this manuscript for the first time the electrochemical and statistical evaluation of FPLC isolation of mellitin and apamin from honey bee (Apis mellifera) venom. Venoms are extremely complex blends of diverse substances that target a myriad of receptors or ion channels. Therefore, toxins, isolated from venomous organisms can be a valuable tool with diverse biological applications. In this study we decided to optimize the purification of honey bee venom by using fast protein liquid chromatography, to obtain biologically active peptide - melittin (2846.46 Da). Due to a presence of other compounds with similar molecular weight (apamin 2027.34 Da), we optimized a differential pulse voltammetry method with adsorptive transfer technique (AdT DPV), utilizing Brdicka supporting electrolyte for measurements. Typical voltammograms - fingerprints for each substance were obtained and numerical projections of voltammograms were employed to propose an artificial neural network. Our suggested neural network can simply predict the content of each peptide in fraction with following performance: 100 % for training and 100 % for testing.We present in this manuscript for the first time the electrochemical and statistical evaluation of FPLC isolation of mellitin and apamin from honey bee (Apis mellifera) venom. Venoms are extremely complex blends of diverse substances that target a myriad of receptors or ion channels. Therefore, toxins, isolated from venomous organisms can be a valuable tool with diverse biological applications. In this study we decided to optimize the purification of honey bee venom by using fast protein liquid chromatography, to obtain biologically active peptide - melittin (2846.46 Da). Due to a presence of other compounds with similar molecular weight (apamin 2027.34 Da), we optimized a differential pulse voltammetry method with adsorptive transfer technique (AdT DPV), utilizing Brdicka supporting electrolyte for measurements. Typical voltammograms - fingerprints for each substance were obtained and numerical projections of voltammograms were employed to propose an artificial neural network. Our suggested neural network can simply predict the content of each peptide in fraction with following performance: 100 % for training and 100 % for testing.