Author:

Keywords:

Salmonella, MOL-PCR, Luminex, Whole genome sequencing, Bioinformatics, Public Health

Abstract:

Characterisation of pathogens below the subspecies or serovar level, i.e. subtyping, is essential in public health for routine surveillance of pathogens, outbreak detection and timely confinement of an outbreak. An ideal subtyping method is rapid, highly discriminative, inexpensive and robust. The classical subtyping methods for the bacterium Salmonella, a major foodborne pathogen, are phage typing, multiple-locus variable-number of tandem repeats analysis (MLVA) and pulsed-field gel electrophoresis (PFGE). Each of these techniques has however its intrinsic disadvantages. In the main part of this PhD, a molecular alternative for the subjective phage typing method was developed. Salmonella enterica subsp. enterica serovar Typhimurium was taken as a case study, because this is the most frequent Salmonella serovar in Belgium. In addition, it was shown by a serial passage experiment that 3 of the 5 the MLVA loci in the S. Typhimurium scheme are unstable, which especially complicates outbreaks investigations. For the time-efficiency of the alternative subtyping assay, several molecular markers had to be combined in a single assay. Therefore, the Luminex technology, which allows multiplexing by microsphere suspension arrays, was implemented in the Scientific Institute of Public Health (WIV-ISP) during this PhD research. The multiplex oligonucleotide ligation-PCR (MOL-PCR) assay technique was selected for the novel method for subtyping of S. Typhimurium, since this is a sufficiently fast and relatively inexpensive technique. The thorough optimisation of the MOL-PCR assay indicated that the DNA isolation, the probe concentration, the amount of microspheres and the concentration of reporter dye are the parameters that have a major influence on the MOL-PCR performance. The final developed MOL-PCR assay for subtyping of I. Typhimurium and S. 1,4,[5],12:i:- combines 52 molecular markers, including prophage genes, amplified fragment length polymorphism (AFLP) elements, Salmonella genomic island 1 (SGI1), allantoinase gene allB, MLVA locus STTR10, antibiotic resistance genes, single nucleotide polymorphisms (SNPs) and phase 2 flagellar gene fljB. The method proved to have a discriminatory power equal to that of phage typing and could deliver results in less than 8 hours, making it suitable for outbreak investigations. Additionally, an R application was created that allows an objective interpretation of the data and summarises the results in a MOL-PCR profile. During this PhD, whole genome sequencing (WGS) has gained the attention of public health institutes as the ultimate universal subtyping technique for surveillance of pathogens and for outbreak investigations. As such, this technique was explored in the second part of this work. Four issues associated to pathogen surveillance with WGS were elaborated using data of 32 S. Typhimurium and S. 1,4,[5],12:i:- isolates. The first issue relates to the workflow to be followed. For a SNP-based workflow, we showed that the applied tools, the used reference genome and error correction have an impact on the resulting phylogenetic tree. SNP-based analysis of WGS data proved to have more discriminatory power than the classical subtyping methods, which is especially of interest for isolates with frequently occurring subtypes. The alternative gene-based workflow, also referred to as whole genome MLST, could not be explored, as currently there is no whole genome MLST scheme available for Salmonella. The second issue concerns the maximum number of SNPs between two isolates before they are regarded as distinct subtypes. This number was again highly dependent on the applied tools. Existing web-based tools were used to look into the third and fourth issue, namely the inference of phenotypic characteristics from WGS data and the link between historical subtyping data and WGS data, which showed to be possible, albeit very limited. As a case study to show the ‘universal’ applicability of WGS, data of S. Enteritidis isolates related to two outbreaks were used, as for each of these outbreaks both food and human isolates were available. The WGS analysis with existing user-friendly tools confirmed the initial outbreak investigation with classical subtyping methods. In addition, the WGS data analysis could discriminate between the source of the two outbreaks, which was not possible with the classical subtyping methods. To conclude, the developed MOL-PCR assay for S. Typhimurium and S. 1,4,[5],12:i:- is a multiplex, molecular subtyping method that has the same discriminatory power as phage typing. For surveillance and outbreak investigation, WGS can currently be applied to complement the existing subtyping methods for increased discrimination between isolates. Before WGS can be implemented as a replacement of the existing subtyping methods, further optimisation of the data analysis tools is however required.