Background: There has been increased interest in the study of molecular survival mechanisms expressed by foodborne pathogens present on food surfaces. Determining genomic responses of these pathogens to antimicrobials is of particular interest since this helps to understand antimicrobial effects at the molecular level. Assessment of bacterial gene expression by transcriptomic analysis in response to these antimicrobials would aid prediction of the phenotypic behavior of the bacteria in the presence of antimicrobials. However, before transcriptional profiling approaches can be implemented routinely, it is important to develop an optimal method to consistently recover pathogens from the food surface and ensure optimal quality RNA so that the corresponding gene expression analysis represents the current response of the organism. Another consideration is to confirm that there is no interference from the "background" food or meat matrix that could mask the bacterial response.
Findings: Our study involved developing a food model system using chicken breast meat inoculated with mid-log Salmonella cells. First, we tested the optimum number of Salmonella cells required on the poultry meat in order to extract high quality RNA. This was analyzed by inoculating 10-fold dilutions of Salmonella on the chicken samples followed by RNA extraction. Secondly, we tested the effect of two different bacterial cell recovery solutions namely 0.1% peptone water and RNAprotect (Qiagen Inc.) on the RNA yield and purity. In addition, we compared the efficiency of sonication and bead beater methods to break the cells for RNA extraction. To check chicken nucleic acid interference on downstream Salmonella microarray experiments both chicken and Salmonella cDNA labeled with different fluorescent dyes were mixed together and hybridized on a single Salmonella array. Results of this experiment did not show any cross-hybridization signal from the chicken nucleic acids. In addition, we demonstrated the application of this method in a meat model transcriptional profiling experiment by studying the transcriptomic response of Salmonella inoculated on chicken meat and exposed to d-limonene. We successfully applied our method in this experiment to recover the bacterial cells from the meat matrix and to extract the RNA. We obtained high yield and pure RNA. Subsequently, the RNA was used for downstream transcriptional profiling studies using microarrays and over 600 differentially regulated genes were identified.
Conclusions: Our result showed that 8 log cfu/g of Salmonella is ideal to obtain optimal RNA amount and purity. Our results demonstrated that RNAprotect yielded higher RNA amounts (approximately 10 to 30 fold) when compared to 0.1% peptone water. The differences between the RNAprotect and 0.1% peptone samples were significant at a p-value of 0.03 for the bead beater method and 0.0005 for the sonication method, respectively. The microarray experiment demonstrated that the chicken samples do not interfere with the hybridization of Salmonella cDNA on the array slide. Hence, the background chicken RNA will not interfere with the microarray analysis when poultry meat models are used. Finally, we successfully demonstrated the application of the poultry meat model proposed in this study by conducting transcriptional profiling analysis of Salmonella inoculated on the poultry. Results of this study proved that this method has the potential to be employed in other meat model studies.