Improvement of methods for the detection of Gram-negative foodborne pathogens

Foodborne diseases are a major source of morbidity and mortality worldwide. In most cases, these diseases are caused by contaminated food products, but transmission can also subsequently occur via person to person contact. The ability to detect the pathogens is an important aspect in the verification of food safety. A major proportion of foodborne disease is caused by Gram-negative bacteria. In this thesis, the detection of Gram-negative foodborne pathogens is addressed by looking at the successive steps from enrichment to detection with Salmonella, Shiga toxin-producing E. coli and Cronobacter spp. as example pathogens. The detection of foodborne pathogens using microbiological culture media aiming at the resuscitation and growth of bacteria is still regarded as the gold standard and included in many reference methods. However, cultural methods are time and labour-intensive. Since an immediate response is required in case of contamination and during outbreaks there is a strong interest in methods that deliver information on the microbiological status of the product as quickly and reliable as possible. Rapid cultural methods and commercially available real-time PCR systems for the detection of Salmonella and STEC were compared with regards to their sensitivity and specificity. It was shown that most of the marketed systems are as reliable as the standard methods. However, false-positive results were obtained with real-time PCR systems for the detection of Salmonella. Rapid cultural methods that were based on procedures without the pre-enrichment step, reduced the time to detection but did show some ambiguous results with difficult matrices such as tea. Of the seven rapid tests for the detection of STEC, one did not detect relevant Stx subtypes. In order to be detected, pathogens need to multiply to reach a minimum threshold level. However, because they are often sublethally injured due to hostile processing and storage conditions, they first need to be resuscitated. For most pathogens, (Salmonella, STEC and Cronobacter spp.) the first step in the detection is an enrichment including resuscitation in a non-selective medium such as BPW. Modifications to BPW were compared with respect to their ability to promote growth of unstressed and stressed Gram-negative pathogens. The aim was to develop a medium that could be used for the enrichment of pathogens in horizontal methods using only one enrichment step. The resuscitation of stressed Cronobacter cells was improved in BPW supplemented with an additional iron source and sodium pyruvate along with low levels of compounds for the inhibition of Gram-positive bacteria. However, it was observed that BPW containing these supplements allowed for less resuscitation of STEC when compared to regular BPW. Based on these results it was concluded that the application of a one-broth enrichment in food products with a high number of competing bacteria is not recommended due to the overgrowth of the target bacteria. Limitations of the current method for the detection of STEC from sprouted seeds were noticed. Therefore, the growth of stressed STEC cells from different serotypes was assessed in media used for the enrichment of Enterobacteriaceae. In addition, the growth of STEC was examined in the enrichment of sprouts using different media and incubation temperatures. It was shown that the high level of competitors was inhibiting the detection of the target pathogen and that the similarity of target and competing bacteria prevents the design of a selective enrichment procedure. In order to get a better insight in the enrichment ecology, the microbiome of mungo bean sprouts was analysed using Illumina HiSeq sequencing prior to and during the enrichment in BPW and EE-broth at different temperatures. The majority of the sprout flora was composed of bacteria belonging to the phylum Proteobacteria. Enrichment in BPW increased the proportion of Firmicutes whereas the incubation in EE-broth enriched Proteobacteria. The results point out that with the application of a selective medium like EE-broth, growth of the competitive microflora that complicates the detection of STEC is promoted. It was shown that EE-broth also resulted in good growth of STEC however, the problematic situation of low maximum population densities of the target strain in the matrix is still present. The probability of detection is not only influenced by the natural flora of a food product, but also by the physiological state of the pathogen. The influence of stress on the lag time of single cells and the resulting probability of detection were determined for Cronobacter spp. in powdered infant formula. Lag time was calculated from optical density measurement data and different scenarios were modelled. Lag time was longest after acid stress and lag time increase coincided with increased lag time variability. The probability of detection, however, depended both on the sampling plan and on the duration of the lag phase. This thesis provides a critical evaluation of rapid methods and valuable new insights on enrichment procedures, the role of competitors in bacterial enrichment procedures and the limitations of selective agents. This information will be of great help to further improve microbiological methods and thereby contribute to more effective management of food safety.