Transposon Mutagenesis: Himarl vs. Tn5367 and Tn5370

The thickness and impermeability of the cell wall makes it difficult to introduce plasmid vec­tors into MAP. Therefore, the shuttle plasmid vectors described in Table 7.1 yield low transfor­mation efficiencies.

Two types of transposons have been exten­sively used to generate random mutant banks: (i) IS1096- and (ii) Himar1 -derived transpo­sons (Fig. 7.1). Tn5367 (GenBank Accession KM232614) was the first transposon widely used in mycobacteria including MAP (Bardarov et al., 199 7; Shin et al., 2006; Rathnaiah et al., 2014). This transposon carries a kanamycin-resistant drug marker (aph), and the IS1096 transposase (tnpA) and resolvase (tnpR). This transposon has the transposase within the transposed element (Bardarov et al., 199 7) resulting in a relatively high transposition frequency (~1.0 ? 10^-5), at least in Mycobacterium smegmatis (Cirillo et al., 1991). Tn5367 transposes by a cut and paste mechanism (McAdam et al., 2002). This element

Table 7.1. Previously reported genetic systems and strategies used in Mycobacterium avium subsp. paratuberculosis (MAP) mutagenesis and gene expression studies (1993-2008).

Continued

G. Rathnaiah eta∕.

Table 7.1. Continued

Fig.

may also insert into another gene creating a new mutation and either a wild-type or mutated gene, depending on precise or imprecise exci­sions, in the location of the original insertion. Thus, the corresponding MAP mutants may not be stable enough to conduct long-term in vivo studies where bacilli multiply to large num­bers. An improved IS1096-derived transposon, Tn5370 (GenBank Accession KM232615), was constructed by removing open reading frames (ORFs) dispensable for transposition (e.g. tnpR) (McAdam et al., 2002). Moreover, the trans- posase was engineered into the phasmid outside of the transposable element. Thus, Tn5370, as opposed to Tn5367, generates stable mutants upon transposition. This transposon has an ad­ditional advantage of possessing a hygromycin (hyg) marker outflanked by the resolution sites (res) of the γ-δ resolvase (Wiater and Grindley, 1988). This construction allows for the excision of the drug marker and the generation of un­marked transposon insertion mutants. However, none of these transposons inserts in a fully ran­dom manner in mycobacteria (McAdam et al., 2002; Shin et al., 2006).

Himar1-derived transposons have also been used to generate random mutant libraries in Mycobacterium tuberculosis and more recently in MAP (Sassetti et al., 2001; Scandurra et al., 2009; Rathnaiah etal., 2014; Wang etal., 2014). The Himar1 transposon (MycoMarT7; GenBank Accession AF411123) was engineered into the shuttle phasmid phAE87 (Table 7.1). This vector carries the highly active C9 Himar1 transposase outside the inverted repeats, thus yielding stable mutants.

Most of the MAP transposon mutant li­braries were made with Tn5367 (Harris et al., 1999; Cavaignac et al., 2000; Shin et al., 2006; Rathnaiah et al., 2014) and fewer libraries have been constructed using the Himar1-derived transposon MycoMarT7 (Scandurra et al., 2009; Wang et al., 2014; Rathnaiah et al., 2016). A re­cent study performed a comparative analysis of MycoMarT7 and Tn5367 transposon insertion sites in MAP (Rathnaiah et al., 2016). They corre­lated the analysis of transposon recognition sites with actual insertions obtained in mutagenesis experiments to determine potential transposition biases. The analysis indicated that the distribution of IS 1096 insertion sites within ORFs is sparser, with 710 out of a total 4350 (16%) potential cod­ing sequences containing no recognition sites. The corresponding mutants in those ORFs will not be observed in the mutant libraries irrespective of whether the gene is essential or non-essential in the conditions studied. In contrast, there are only 37 coding sequences not targeted by the Himar1 transposase. Thus, Tn5367 or Tn5370 may not be used for a comprehensive analysis of virulence determinants and gene essentiality in MAP. Most importantly this study also revealed transposition biases that are independent of the transposon site recognition sequences. Both transposons have loci-dependent biases, with Tn5367 being even more skewed. Most notably, insertions into cat­egories of loci with fewer recognition sites seem underrepresented compared with the expected numbers of total sites.

Himar1 transposons have been used in MAP to isolate 111 mutants attenuated in bovine monocyte-derived macrophages from a random screen of 2290 (Scandurra etal., 2009). Though a sensible genetic approach was used, the screening method is labour intensive to screen a representative mariner transposon library. More recently, a larger study was performed that col­lected 90,000 Himar1 mutants but identified in­sertions in approximately 12,000 out of 52,000 (~25% saturation density) recognition sites (Wang et al., 2014). The authors attributed this low number of recognition sites represented to bottle-necking effects during DNA sample prepa­ration or the sequencing stage. Loci-dependent transposition biases may also contribute to this effect. None the less, this Himar1 library was used to screen for mutants capable of condi­tional growth without mycobactin (Wang et al., 2016). Based on this property, they identified a potential mycobactin-independent iron uptake system on the MAP-specific genomic island LSPP15. Step-by-step procedures to generate a MAP transposon mutant library are detailed elsewhere (Bannantine et al., 2019).

7.5