Improved assessments of milk microbiota composition via sample preparation and DNA extraction methods

Although bacterial detection by 16S rRNA gene amplicon DNA sequencing is now a widely-applied technique, standardized methods for sample preparation and DNA extraction are needed to ensure accuracy, reproducibility, and scalability for automation. To develop these methods for bovine milk, we assembled a bacterial cell mock community (BCMC) containing bacterial species commonly found in milk. Variations to the following methods were examined: BCMC enumeration method (colony enumeration or microscopy), sample volume (200 l to 30 ml), sample storage condition (frozen in PBS or 25% glycerol or exposure to freeze-thaw cycles), cell lysis method (bead-beating, vortex, enzymatic), and DNA extraction procedure (MagMAX Total, MagMAX CORE, and MagMAX Ultra 2.0, with and without either Proteinase K or RNase A). Cell enumeration by microscopy was shown to be more accurate for quantification of the BCMC contents. We found that least 10 mL ([≥] 104 cells in high quality milk) is needed for reproducible bacterial detection by 16S rRNA gene amplicon DNA sequencing, whereas variations in storage conditions caused minor differences in the BCMC. For DNA extraction and purification, a mild lysis step (bead-beating for 10 s at 4 m/s or vortexing at 1800 rpm for 10 s) paired with the MagMAX Total kit and Proteinase K digestion provided the most accurate representation of the BCMC. These approaches were confirmed to provide similar results in commercial milk samples. Overall, variations in cell lysis methods conferred the greatest changes to milk microbiota composition and therefore should be optimized for sample type.