Generic and rapid determination of low molecular weight organic chemical contaminants in protein powder by using ultra-performance liquid chromatography-tandem mass spectrometry

Abstract

A rapid, simple, and generic analytical method that could simultaneously determine 291 undesirable low molecular weight chemical contaminants from different drug families in protein powder, such as veterinary drugs and pesticides, etc, had been developed. This method comprised the extraction with acetonitrile-dimethyl sulfoxide (DMSO), clean-up through dispersive solid phase extraction (D-SPE) and low temperature filtration, and analysis by ultra-performance liquid chromatography coupled with electrospray ionization tandem mass spectrometry at multiple-reaction monitoring mode. Acetonitrile-DMSO was more generic than acetonitrile or methanol for the extraction of large-scale organic chemical contaminants with different polarities in protein powder. Most interferences in the extract were eliminated by the combination of D-SPE and low temperature filtration, which simultaneously provided satisfactory recoveries of both hydrophobic and hydrophilic analytes. In particular, besides the purification function, the sorbent of D-SPE also played an important role in grinding samples to improve extraction efficiency during homogenization. This streamlined approach allowed the processes of extraction and the main purification were carried out in one-step, and dramatically reduced sample preparation turnaround times and solvent consumption. For quantification, matrix-fortified calibration curves showed competent linearity for most of the target compounds with linear regression coefficients (r) higher than 0.9900, except for two analytes. The limits of quantification ranged from 0.1 μg/kg to 50 μg/kg, which was usually sufficient to verify the compliance of products with legal tolerances. The average recoveries for spiked protein powder ranged from 65.6% to 142.2% with associated RSD values between 0.5% and 28.5%. For over 90% of the analytes, the recoveries were between 70% and 120% with RSD values in the range of 1%–15%. Applying this method in routine monitoring programs would drastically reduce both effort and time.