Bacterial community structure analysis on Listeria monocytogenes inoculated spinach leaves is affected by PCR based methods to exclude chloroplast co-amplification

Consumption of ready-to-eat leafy vegetables has increased in popularity due to their anticipated health benefits, but their consumption also poses a potential health risk in the form of foodborne pathogens. Factors determining growth of pathogens goes beyond plant species and cultivation practice and may include the phyllosphere bacteriome. This study investigated the bacteriome of spinach leaves, stored under EURL challenge conditions for 9 days after inoculation with L. monocytogenes using two methods of excluding chloroplast co-amplification (COMPETE, BLOCK) at the PCR step as well as a post-PCR chloroplast sequence filter option (CONTROL). While all three approaches allowed to charcterize a change of bacterial communities over time, the BLOCK (peptide nucleic acid, pPNA) approach resulted in greater diversity similarities to the CONTROL option. The COMPETE (competing primer) solution with a specifically designed primer to prevent chloroplast amplification had a strong underrepresentation of the phylum Planctomycetota and to a lesser extend underrepresentation of Verrucomicrobiota due to the inheritance of the selected primer region that allowed to deselect chloroplast co-amplification. However, the COMPETE approach achieved a 180-fold reduction in chloroplast co-amplification, while BLOCK only achieved a 40-fold reduction. Higher relative abundances of Pseudomonadaceae and lower numbers of Lactobacillales coincided with higher growth potential of L. monocytogenes from day 7–9, suggesting that particular phylogenetic groups may support or restrict growth of L. monocytogenes. While chloroplast co-amplification with spinach in the present study was relatively modest (<16.3 %), other leafy vegetables may require one of the demonstrated co-amplification prevention solutions. Although the COMPETE solution in the present study was linked to some amplification bias, the approach may be useful when otherwise co-amplification is very high and the demonstrated BLOCK approach with pPNA is insufficient