Background

Transposon insertion mutagenesis coupled with next-generation sequencing (NGS) has been proven to be a powerful method to investigate gene function under a multitude of conditions (Chao et al., 2016; Price et al., 2018). In general, TIS analysis consists of the massive parallel sequencing of transposon insertion sites and statistical analysis of the abundance of the insertion events.

TIS-based screens can provide high-resolution maps of the fitness contributions of individual loci and domains based on the bacterial fitness of highly saturated transposon mutant libraries under a multitude of conditions (Chao et al., 2016). The insertion frequency at each locus or the relative abundance of corresponding mutants is generally inversely correlated with the locus’s contribution to fitness following the imposition of a selective pressure, such as that imposed by hosts and antibiotics (Chao et al., 2016). The principles of this methodology underlie a variety of related approaches, including TIS, transposon sequencing (TnSeq), insertion sequencing (INSeq), transposon-directed insertion site sequencing (TraDIS), high-throughput insertion tracking by deep sequencing (HITS), and transposon insertion site sequencing (TIS-Seq) (Gawronski et al., 2009; Langridge et al., 2009; Chao et al., 2016; Veeranagouda et al., 2017; Price et al., 2018). These approaches are revolutionizing microbiological studies by facilitating unprecedented and deep genome-wide explorations of a wide range of bacterial species.

With the aim to generate highly saturated mutant libraries for further analysis, the insertion events caused by transposon should be unbiased and the transposon will be delivered into recipient cells several times. However, this will lead to a large amount of nonsense mutation and the complexity of mutant libraries may cause bottleneck effect during the following screens. Compared with this, the defined mutant libraries containing mutants with a single insertion at a known genomic location will be more useful and facilitated (Fu et al., 2013; Abel et al., 2015). The first human pathogenic bacterium defined mutant library for Pseudomonas aeruginosa was constructed in 2003. Subsequently, many defined mutant libraries of human pathogens were also generated, such as a library of Francisella tularensis in 2007, Vibrio cholera in 2008, Burkholderia thailandensis in 2013, Chlamydia and Mycobacterium in 2015 (Jacobs et al., 2003; Gallagher et al., 2007 and 2013; Cameron et al., 2008; Kokes et al., 2015). With these defined mutant libraries, a myriad of important researches were completed, such as the identification of ciprofloxacin resistome from P. aeruginosa library, a novel T6SS regulator TsrA from V. cholera library, and the exploration of functional genomics with Desulfovibrio alaskensis G20 library (Breidenstein et al., 2008; Zheng et al., 2008 and 2010; Kuehl et al., 2014). Hence, such defined libraries may serve as valuable resources because they often consist of collections of insertion mutants in almost all non-essential loci for an organism of interest and have drawn more attention to the exploration of functional genomics of microbe.

In order to optimize TIS approach, we developed the DML-seq method which adopts the mixed defined mutant libraries for conditional essential fitness screenings (Figure 1) (Wei et al., 2019a and 2019b). Compared with massive mutant library in traditional TIS, the DML library has advantages as: 1) insertions within the 80% of the ORF to ensure the effective mutagenesis; 2) low complexity and few isogenes to ensure low bottleneck effects; 3) the mixture of an equal amount of individual mutants to avoid hotpots caused by screening; 4) can be directly used in the following experiments.

Compared to human pathogenic bacterium, the pathogenesis of marine pathogenic bacterium is poorly understood. In our model organism Edwardsiella piscicida, which causes Edwardsiellosis in farmed fish and leads to severe economic losses in aquaculture worldwide (Leung et al., 2019), DML-seq has been employed to identify conditional essential genes as well as those crucial for survival in different environments (Wei et al., 2019b), colonization and infection in host cells (Wei et al., 2019a). Briefly, we constructed MKGR, a derivative of Himar1 mariner transposon based on MAR2xT7 (Liberati et al., 2006), to generate a library of random transposon insertion mutants of E. piscicida EIB202 (Table 1). In total, 2,806 genes were disrupted out of the 3,599 predicted genes with an average of 5.04-fold coverage on E. piscicida chromosome. After that, 5 non-redundant subset libraries were created manually. The 1^st and 2^nd subset libraries contained a single mutant for each disrupted ORF in the parental library and they were paralleled for each other. The 3^rd subset library contains transcriptional fusion mutants to eliminate the location effect on transcription. The 4^th subset library is an intergenic library and the 5^th library was a composited library, which was generated by mixing the 1^st, 2^nd, 4^th subset libraries into a pool.

Here, we will focus on the specific steps for DML-seq because the excellent protocols deTAILing TIS analyses are reviewed and available elsewhere (Chao et al., 2016; Yamaichi et al., 2017). In principle, the procedure described here can be applied to conditional essential screens and facilitate the exploration of functional genomics of microbe.

Table 1. The statistic values of defined mutant library of E. piscicida EIB202

Figure 1. A defined transposon mutant library construction. A. pMKGR, a derivative of Himar1 mariner transposon, was constructed to build the defined transposon mutant library of E. piscicida EIB202. B. Work flow of a defined mutant library construction. E. coli SM10 is used as a donor for conjugation with recipient strain E. piscicida EIB202. Conjugation is processed between EIB202 and SM10 on LB plates for 3 h at 30 °C. Later, conjugators are separated on LB plates containing 20 µg/ml polymyxin and 15 µg/ml gentamicin. After 24 h of incubation at 30 °C, monoclonal will be picked into 96-well plates and incubated at 30 °C for about 16 h. Before storing at -80 °C, glycerol was added into every well and mix to make the final 20% glycerol concentration to preserve strains. C. Schematic of TAIL-PCR amplification used for determination of the Tn insertion loci on the genome. The pair of primer Sp1/AB2 is used for the first round of PCR and another pair of primer Sp2/ABS is used for the second round of PCR. Finally, primer Seq2 is used for sequencing of the amplified PCR products. D. The number of new disrupted genes increased linearly with new mutants and the number of genes hit approached a plateau.