Propidium monoazide (PMA) qPCR assay compared to the plate count method for quantifying the growth of Salmonella enterica serotypes in vacuum-packaged turkey breast combined with a mathematical modeling approach

Abstract

This study compares the plate count (PC) and the Propidium Monoazide-quantitative Polymerase Chain Reaction (PMA-qPCR) methods to assess the growth of a cocktail of three serotypes of Salmonella enterica (Heidelberg, Typhimurium, and Enteritidis) in cooked, sliced, and vacuum-packaged turkey breast (STB) under isothermal storage temperatures (8 °C–20 °C), using predictive models. Standard curves were developed for PMA-qPCR, demonstrating high efficiency (101%) and sensitivity, with quantification limits ranging from 1 to 2 log₁₀ CFU/g for all temperatures studied. Comparative analysis revealed a significant correlation (R² = 0.99; 95% CI) between the PC and PMA-qPCR methods; however, the agreement analysis indicated a mean difference (Bias) of −0.11 log₁₀ CFU/g (p < 0.05), suggesting underestimation by the PC method. This indicates the presence of stressed or viable but nonculturable (VBNC) cells, detectable by PMA-qPCR but not by PC. The Baranyi and Roberts model showed a good ability to describe the behavior of S. enterica cocktail in STB for PC and PMA-qPCR data under all isothermal conditions. The exponential secondary model more accurately represented the temperature dependence of the maximum specific growth rate compared to the Ratkowsky square root model, with R² values ≥ 0.984 and RMSE values ≤ 0.011 for both methods. These results suggest that combining PMA-qPCR with predictive modeling allows for a more accurate prediction of S. enterica growth, compared to PC method. In the event of cold chain disruptions of meat products, the use of PMA-qPCR method allow the quantification of VBNC cells, that can still pose a health risk to consumers, especially in ready-to-eat products.