Comparison between swabbing devices in order to analyse the microbial flora found on surfaces of community kitchens by classical microbiology and amplicon sequencing

Introduction
The work in community kitchens involves several manual steps bearing the risk of transmitting pathogenic microorganisms. This fact justifies accurate control of surfaces to prevent foodborne illnesses. Different sampling methods have been proposed to isolate bacteria from surfaces, while the ensuing analysis of the surfaces is usually performed by classical microbiology and molecular biology methods [1,2,3]. Studies found in the literature do compare different swabbing devices for surfaces but not in terms of 16S rRNA gene profiling [4]. This study aimed at comparing the efficiency of recovery of different swabbing devices in terms of culture methods and culture-independent 16S rRNA amplicon sequencing, the latter in order to characterise the present microbial flora.
Material and Methods
To mimic surfaces which can be found in community kitchens, sterile polypropylene cutting boards and stainless steel surfaces were spiked with 1mL of 2-fold dilution of chicken and with 1mL of several concentrations of Escherichia coli, Salmonella enterica ssp. enterica serovar Enteriditis and Bacillus cereus, respectively. Collection of the bacteria was performed with sterile moistened swabs, including commercial swabs (sponges HS10NB2G, Sponge-Sticks SSL100; 3M) and cotton and gauze pads. Enumeration was carried out by aerobic plate count on Plate Count Agar (Bio-Rad), Rapid’E.coli 2 (Bio-Rad), Rapid’Salmonella (Bio-Rad) and MYP (Oxoid) which were incubated for 24h at 37°C. The V1-V3 16S rRNA amplicon libraries obtained from total bacterial DNA extraction were sequenced on a Illumina MiSeq sequencer and the resulting sequences were analysed with mothur software. Statistical analyses were performed by 2-way ANOVA and Tukey’s multiple comparisons by using Prism 7.
Discussion
The perfect swab for kitchen analyses should recover the highest number of viable bacteria with a high population diversity from the surfaces. Classical culture method showed best recovery numbers for the three inoculated bacteria with Sponge-Sticks (no significant difference between inoculation dosis and recovered number of bacteria, p > 0.1234). The 16S rRNA amplicon sequencing allows to conclude that sponge samples were loaded with DNA from Microbacterium genus (from Neutralizing buffer). Furthermore, a high relative abundance of Bacillus genus was found in cotton pad, gauze pad and Sponge-Stick samples. Salmonella genus was detected in only low proportions, whereas Escherichia genus are problematic as their DNA can be contaminants of reagents used during library creation. However, differences in recovery or enumeration with each method must be considered, as they can induce estimation bias on the initial concentration or recovered CFU/mL. Finally, low amount of DNA in controls leads to the emergence of free DNA contaminants like Elizabethkingia population, which can be considered as a bias.

References
1. Ismaïl R. et al. (2013). Int. J. Environ. Res. Public Health, 10, 6169-6183.
2. Biranjia-Hurdoyal, S., & Latouche, M. C. (2016). Can J Infect Dis Med Microbiol., 2016.
3. Malorny B. et al. (2008). Appl Environ Microbiol, 74(5), 1299-1304.
4. Lahou, E., & Uyttendaele, M. (2014). Int. J. Environ. Res. Public Health, 11(1), 804-814.