Protecting the safety and quality of live oysters through the integration of predictive microbiology in cold supply chains

Vibrio parahaemolyticus is a bacterial species indigenous to marine environments and can accumulate in oysters. Some V. parahaemolyticus strains are pathogenic and seafoodborne outbreaks are observed worldwide. This pathogen can reach infectious levels in oysters if post-harvest temperatures are not properly controlled. The aim of this thesis was to support oyster supply chain management by developing predictive microbiological tools to improve the safety and quality of oysters in the market. A predictive model was produced by injecting Pacific oysters (Crassostrea gigas) harvested in Tasmania with a cocktail of pathogenic and non-pathogenic V. parahaemolyticus strains, and measuring population changes over time at static storage temperatures from 4 to 30¬¨‚à´C. In parallel, the total viable bacteria count (TVC) was measured. The V. parahaemolyticus and TVC growth models were then evaluated with Pacific and Sydney Rock oysters (Saccostrea glomerata) harvested in New South Wales containing natural populations of V. parahaemolyticus. Oysters were stored at static temperatures from 15 to 28¬¨‚à´C, and Vibrio parahaemolyticus and TVC viability were measured. In Pacific oysters, TVC growth was observed at all tested temperatures while V. parahaemolyticus growth was observed only at 23 and 28¬¨‚à´C. In Sydney Rock oysters, TVC growth was observed only at 24¬¨‚à´C and V. parahaemolyticus did not grow at any storage temperature tested. These interesting findings potentially indicate that Sydney Rock oysters have enhanced anti-bacterial defences compared to Pacific oysters, and that commercial temperature controls to manage V. parahaemolyticus growth can be different. Consistently higher growth rates of V. parahaemolyticus and TVC were observed in Tasmanian versus New South Wales oysters and may have been caused by different factors. They include variations in levels of background competitive flora, different growth rates among V. parahaemolyticus strains, and/or changes in the natural bacterial community structure influenced by conditions at the harvest site or during shipment to the laboratory. Nevertheless, the overall performance of the model was fail-safe‚ÄövÑvp for predicting growth of V. parahaemolyticus in Pacific oysters and would be a preferred public health tool. The V. parahaemolyticus and TVC predictive models for Pacific oysters were integrated in an Excel¬¨vÜ software tool. The model allows users to input time-temperature profiles and analyse the effects of dynamic storage temperatures normally found in oyster supply chains on bacterial growth. The tool was evaluated in five different simulated oyster supply chains (refrigerated and non-refrigerated). Observed and predicted V. parahaemolyticus and TVC growth rates were compared and a model over-estimation mean of 2.30 for V. parahaemolyticus and 2.40 for TVC were observed as determined by the bias factor index. Reasons for over-estimations are likely the same as those for model validation experiments. Uncertainty and variability are associated with oyster supply chains. Therefore, a stochastic model which encompassed the operations from oyster farm to the consumer was built using ModelRisk¬¨vÜ risk analysis software. This case study generated probabilistic distributions and the percentage of oysters containing V. parahaemolyticus and TVC during each operation of the supply chain. The results were used for an objective evaluation of the influence of short and long supply chains during summer and winter seasons. The stochastic model can help the oyster industry evaluate the performance of oyster cold chains, and potentially enable real-time decisions if coupled with suitable traceability systems. It can also provide risk managers with valuable information about V. parahaemolyticus exposure levels. Finally, in order to better understand microbial changes in oysters during distribution and storage, the dynamics of microbial communities in Pacific oysters was determined using 16S rRNA-based terminal restriction length polymorphism and clone library analyses. Significant differences in bacterial community composition were observed, and the predominant bacteria were identified for fresh and stored oysters at different temperatures. High microbial diversity in oysters was observed, with up to 73 different genera-related identified clones among all samples. The results identified Psychrilyobacter spp. as a potential spoilage indicator for future shelf-life studies, and Polynucleobacter and a bacterial group related to Alkaliflexus as possible indicators for storage temperature control in Pacific oysters. In future studies, quantitative correlations of the identified species and the freshness of oysters should be explored to determine whether the predominant microbes identified represent significant specific spoilage organisms‚ÄövÑvp, and to determine if they are antagonistic to human bacterial pathogens that are found in oysters.