Predictive Population Dynamics of Escherichia coli O157:H7 and Salmonella enterica on Plants: a Mechanistic Mathematical Model Based on Weather Parameters and Bacterial State.

Weather affects key aspects of bacterial behavior on plants but has not been extensively investigated as a tool to assess risk of crop contamination with human foodborne pathogens. A novel mechanistic model informed by weather factors and bacterial state was developed to predict population dynamics on leafy vegetables and tested against published data tracking Escherichia coli O157:H7 (EcO157) and Salmonella enterica populations on lettuce and cilantro plants. The model utilizes temperature, radiation, and dew point depression to characterize pathogen growth and decay rates. Additionally, the model incorporates the population level effect of bacterial physiological state dynamics in the phyllosphere in terms of the duration and frequency of specific weather parameters. The model accurately predicted EcO157 and S. enterica population sizes on lettuce and cilantro leaves in the laboratory under various conditions of temperature, relative humidity, light intensity, and cycles of leaf wetness and dryness. Importantly, the model successfully predicted EcO157 population dynamics on 4-week-old romaine lettuce plants under variable weather conditions in nearly all field trials. Prediction of initial EcO157 population decay rates after inoculation of 6-week-old romaine plants in the same field study was better than that of long-term survival. This suggests that future augmentation of the model should consider plant age and species morphology by including additional physical parameters. Our results highlight the potential of a comprehensive weather-based model in predicting contamination risk in the field. Such a modeling approach would additionally be valuable for timing field sampling in quality control to ensure the microbial safety of produce. IMPORTANCE Fruits and vegetables are important sources of foodborne disease. Novel approaches to improve the microbial safety of produce are greatly lacking. Given that bacterial behavior on plant surfaces is highly dependent on weather factors, risk assessment informed by meteorological data may be an effective tool to integrate into strategies to prevent crop contamination. A mathematical model was developed to predict the population trends of pathogenic E. coli and S. enterica, two major causal agents of foodborne disease associated with produce, on leaves. Our model is based on weather parameters and rates of switching between the active (growing) and inactive (nongrowing) bacterial state resulting from prevailing environmental conditions on leaf surfaces. We demonstrate that the model has the ability to accurately predict dynamics of enteric pathogens on leaves and, notably, sizes of populations of pathogenic E. coli over time after inoculation onto the leaves of young lettuce plants in the field.

Survival and predictive modeling of Listeria monocytogenes under simulated human gastric conditions in the presence of bovine milk products.

Listeria monocytogenes is an opportunistic foodborne pathogen which has been implicated in many outbreaks of foodborne diseases. This study evaluated the effects of gastric acidity and gastric digestion time of adults, L. monocytogenes strain and food type on the survival of L. monocytogenes under simulated stomach conditions of adults in in vitro gastric models with dynamic pH changes occurring throughout the exposure. Individual strains as well as a cocktail of L. monocytogenes, inoculated at 8 log CFU/mL in filtered bovine milk products, 0 % milk, 2 % milk, 2 % chocolate milk and 3.25 % milk, were introduced to the gastric models for 2 h. The survival of L. monocytogenes depended on a combination of factors, including gastric acidity and gastric digestion time of adults, L. monocytogenes strain, food type and recovery method (P < 0.05). The survival rates of L. monocytogenes inoculated in 2 % milk after a 2-h exposure to simulated gastric fluids with pH values of 1.5, 2.0 and 3.0 were 0.003 to 0.040 %, 22.7 to 43.4 % and 16.6 to 27.2 %, respectively. Fluid milk containing a higher milk fat content (3.25 % vs 0 % milk) protected L. monocytogenes from being inactivated when they were exposed to the human stomach model with a gastric acidity of pH 2.0. Compared to 0 % and 3.25 % milk, L. monocytogenes survived the best in 2 % chocolate milk, which appears to be due to the presence of milk fat (2 %) and the additional nutrients that are present in chocolate milk. A predictive mathematical model was developed that captured the population of the strains of L. monocytogenes under the in vitro conditions. This study advances our understanding of the behaviour of L. monocytogenes under various human gastric conditions and provides key parameters that can affect the survival of L. monocytogenes in the stomachs of adults. The mathematical models developed in this study can be used as a supplementary tool to help predict the survival of L. monocytogenes under similar scenarios and for relevant risk-assessment studies.

Weather stressors correlate with Escherichia coli and Salmonella enterica persister formation rates in the phyllosphere: a mathematical modeling study.

Enteric pathogens can enter a persister state in which they survive exposure to antibiotics and physicochemical stresses. Subpopulations of such phenotypic dormant variants have been detected in vivo and in planta in the laboratory, but their formation in the natural environment remains largely unexplored. We applied a mathematical model predicting the switch rate to persister cell in the phyllosphere to identify weather-related stressors associated with E. coli and S. enterica persister formation on plants based on their population dynamics in published field studies from the USA and Spain. Model outputs accurately depicted the bi-phasic decay of bacterial population sizes measured in the lettuce and spinach phyllosphere in these studies. Predicted E. coli persister switch rate on leaves was positively and negatively correlated with solar radiation intensity and wind velocity, respectively. Likewise, predicted S. enterica persister switch rate correlated positively with solar radiation intensity; however, a negative correlation was observed with air temperature, relative humidity, and dew point, factors involved in water deposition onto the phylloplane. These findings suggest that specific environmental factors may enrich for dormant bacterial cells on plants. Our model quantifiably links persister cell subpopulations in the plant habitat with broader physical conditions, spanning processes at different granular scales.

Formation of Escherichia coli O157:H7 Persister Cells in the Lettuce Phyllosphere and Application of Differential Equation Models To Predict Their Prevalence on Lettuce Plants in the Field.

Escherichia coli O157:H7 (EcO157) infections have been recurrently associated with produce. The physiological state of EcO157 cells surviving the many stresses encountered on plants is poorly understood. EcO157 populations on plants in the field generally follow a biphasic decay in which small subpopulations survive over longer periods of time. We hypothesized that these subpopulations include persister cells, known as cells in a transient dormant state that arise through phenotypic variation in a clonal population. Using three experimental regimes (with growing, stationary at carrying capacity, and decaying populations), we measured the persister cell fractions in culturable EcO157 populations after inoculation onto lettuce plants in the laboratory. The greatest average persister cell fractions on the leaves within each regime were 0.015, 0.095, and 0.221%, respectively. The declining EcO157 populations on plants incubated under dry conditions showed the largest increase in the persister fraction (46.9-fold). Differential equation models were built to describe the average temporal dynamics of EcO157 normal and persister cell populations after inoculation onto plants maintained under low relative humidity, resulting in switch rates from a normal cell to a persister cell of 7.7 × 10-6 to 2.8 × 10-5 h-1 Applying our model equations from the decay regime, we estimated model parameters for four published field trials of EcO157 survival on lettuce and obtained switch rates similar to those obtained in our study. Hence, our model has relevance to the survival of this human pathogen on lettuce plants in the field. Given the low metabolic state of persister cells, which may protect them from sanitization treatments, these cells are important to consider in the microbial decontamination of produce.IMPORTANCE Despite causing outbreaks of foodborne illness linked to lettuce consumption, E. coli O157:H7 (EcO157) declines rapidly when applied onto plants in the field, and few cells survive over prolonged periods of time. We hypothesized that these cells are persisters, which are in a dormant state and which arise naturally in bacterial populations. When lettuce plants were inoculated with EcO157 in the laboratory, the greatest persister fraction in the population was observed during population decline on dry leaf surfaces. Using mathematical modeling, we calculated the switch rate from an EcO157 normal to persister cell on dry lettuce plants based on our laboratory data. The model was applied to published studies in which lettuce was inoculated with EcO157 in the field, and switch rates similar to those obtained in our study were obtained. Our results contribute important new knowledge about the physiology of this virulent pathogen on plants to be considered to enhance produce safety.