1.10.1 Transposon mutagenesis experimental design
The availability of whole genome sequences, developments in bioinformatics and high throughput sequencing has led to the revival of transposon-based mutagenesis studies, to determine gene fitness/essentiality (Barquist et al., 2013a, van Opijnen and Camilli, 2013). Transposon mutagenesis involves the parallel production of a large number of individual transposon insertion mutants. An insertion in an essential or fitness-associated ORF will render the attenuated mutant nonviable or decrease its abundance in the pool. Comparison of input and output libraries will thereby allow identification of genes needed for fitness or survival of the organism in the test condition. The transposons most popularly used in bacterial random genome wide mutagenesis are Tn5 (originated from bacteria) and Mariner transposon derivatives
24
(Barquist et al., 2013a, Choi and Kim, 2009). The transposon mutants are negatively selected, after which, the genomic locus of the inserted transposons are identified using DNA hybridization, microarrays [e.g. transposon site hybridization (TraSH), microarray tracking of transposon mutants (MATT) and designer arrays for defined mutant analysis (DeADMAn)] (Mazurkiewicz et al., 2006) as well as transposon insertion sequencing [e.g. transposon directed insertion site sequencing (TraDIS), insertion sequencing (INSeq), high throughput insertion tracking by deep sequencing (HITS) and transposon sequencing (Tn-Seq)] (van Opijnen and Camilli, 2013). In this project, we used transposon insertion sequencing (Tn-Seq) to analyse the data from the mutant library, based on the workflow illustrated in Figure 1.4. Many factors affect the data generated from Tn-Seq including the experimental design and downstream analysis. Experimental design includes consideration of the choice of transposon, the density of the library, the loss of mutants due to stochastic selection as well as the introduction of bias during amplification of transposon-chromosome junctions (Chao et al., 2016).
25
Figure 1-4 Transposon insertion sequencing workflow
The workflow involved in the transposon insertion sequencing procedure involves the following steps. The transposon insertion mutant library is first created and pooled. The library is then grown in test conditions; selective (condition B) and non-selective (condition A). The viable mutants under each condition are recovered and the transposon junction in both pools are attached to sequencing adapters and amplified. The amplified product is then sequenced and mapped against the reference genome to identify the unique insertion sites. Reproduced from Chao et al. (2016).
26
1.10.2 Gene essentiality/gene fitness studies in N. meningococci
The term “essential gene” or “essentiality” is differently defined in the reviews on genomic studies and includes terms such as “operationally” (Judson and Mekalanos, 2000), “required” (Barquist et al., 2013a), “essential for survival” and “essential for fitness” (Gerdes et al., 2006). Nevertheless, there is general agreement that the essential genes/gene fitness is specific to a particular growth condition, thus it is crucial to always define how the “essentiality” was assessed. Even though the combination of transposon mutagenesis and high throughput sequencing is a powerful tool, not all the essential genes can be easily identified. In part, this is dependent on the density of the insertion in a specific gene, as well as the possibility of multiple redundant, or partially overlapping pathways of the biological process of interest (Barquist et al., 2013a, Judson and Mekalanos, 2000).
In N. meningitidis, transposon-based mutagenesis has been used to determine genes involved in competence (Sun et al., 2005), maltose catabolism (Pelicic et al., 2000), complement-mediated lysis (Geoffroy et al., 2003), systemic infections in infant rats (Sun et al., 2000), resistance to heme iron and other hydrophobic agents (Rasmussen et al., 2005) and in vitro and in sera growth (Mendum et al., 2011). When we started this research in 2014, two studies on meningococcal colonization was reported; Jamet et al. (2013) and Exley et al. (2009), but the libraries they used were relatively small, which would result in incomplete coverage. Exley et al. (2009) screened 576 mutants to identify genes involved in colonization of human nasopharyngeal tissue using an organ culture model (at 6 and 18 hours exposure), whilst Jamet et. al. screened 600 mutants for adhesion (at 18 hours exposure) to
27
T48 human colonic carcinoma derived epithelial cells and pharynx carcinoma derived FaDu epithelial cells. The findings from the two groups were quite variable. Exley et. al. identified 8 mutants that were impaired for colonization of human nasopharyngeal tissue, of which five were related to surface molecules: a gene predicted to encode a TonB receptor and two novel factors (related to a bacteriocin processing protein in Haemophilus influenzae and outer membrane protein OmpU). Jamet et al. (2013) identified some metabolic genes as described in 1.7., of which narP and estD appeared to be involved in adaptation to hypoxic conditions and stress resistance. One of the limitation of the two studies, was the use of low saturated mutant libraries for screening. Thus, my research using an extensive mutant library (~14,500 mutants) created by Mendum et al. (2011) would provide a more comprehensive genome wide screen to identify important genes not only in adhesion, but also invasion and traversal across the epithelial cells.
Whilst this research, Capel et al. (2016) described a similar approach in identifying genes involved in colonization of epithelial and endothelial cells. Using a mariner based transposon mutagenesis, they assessed the gene fitness during growth and colonisation using a microfluidic flow device with a 99.99% genome saturation library. A total of 288 genes and 33 intergenic regions containing small noncoding RNA candidates were found to be important in cell colonization. Genes involved in type IV pili biogenesis as well carbohydrate and amino acid metabolism were among the genes shown to be important.
This study uses the transposon library created by Mendum et al. (2011) and the work involved in its construction and the results obtained will be described. These authors
28
constructed a relatively large meningococcal transposon library, of approximately 14,500 mutants based on Tn5 insertions in N. meningitidis strain L91543. The fitness effects of the genes were assessed using transposon site hybridization (TraSH) when grown in complete media, minimal media (to identify metabolic requirements) and complement inactivated serum (to determine growth requirements in serum). The transposon data was used to interrogate a genome-scale metabolic model that the group had constructed, Nmb_iTM560. Twenty six percent of the genome, i.e. 585 genes was identified as essential when grown in Columbia blood agar with 6% defibrinated horse blood, the complete media. These genes were mostly associated with transcription, protein synthesis, protein export and modification, DNA replication, fatty acid biosynthesis and cell wall biosynthesis (Mendum et al., 2011). In addition, a few metabolic pathways were also predicted to be essential for growth in serum, including genes associated with synthesis of aromatic amino acid, purines, folic acid and panthothenate as well as genes required for iron acquisition. In addition, the essentiality screen also identified amino acid transporters and lactate permease.