Throwing a spotlight on genomic dark matter: The power and potential of transposon-insertion sequencing

Abstract

Linking genotype to phenotype is a central goal in biology. In the microbiological field, transposon mutagenesis is a technique that has been widely used since the 1970s to facilitate this connection. The development of modern 'omics approaches and next-generation sequencing have allowed high-throughput association between genes and their putative function. In 2009, four different variations in modern transposon-insertion sequencing (TIS) approaches were published, being referred to as transposon-directed insertion-site sequencing (TraDIS), transposon sequencing (Tn-seq), insertion sequencing (INSeq), and high-throughput insertion tracking by deep sequencing (HITS). These approaches exploit a similar concept to allow estimation of the essentiality or contribution to fitness of each gene in any bacterial genome. The main rationale is to perform a comparative analysis of the abundance of specific transposon mutants under one or more selective conditions. The approaches themselves only vary in the transposon used for mutagenesis, and in the methodology used for sequencing library preparation. In this review, we discuss how TIS approaches have been used to facilitate a major shift in our fundamental understanding of bacterial biology in a range of areas. We focus on several aspects including pathogenesis, biofilm development, polymicrobial interactions in various ecosystems, and antimicrobial resistance. These studies have provided new insight into bacterial physiology and revealed predicted functions for hundreds of genes previously representing genomic "dark matter." We also discuss how TIS approaches have been used to understand complex bacterial systems and interactions and how future developments of TIS could continue to accelerate and enrich our understanding of bacterial biology.

Keywords: Tn-seq; TraDIS; bacterial genomics; dark matter; high-throughput screening; linking genotype-phenotype; novel genes; transposon mutagenesis.

Figure 1

Overview of TIS library sequencing. A high-density transposon mutant pool is generated via transformation of a transposon into cells, where it is then inserted at random into the target genome. The pool of mutants is harvested, and the genomic DNA is extracted. The DNA is then fragmented, either by shearing or enzymatic treatment (depending upon the kind of TIS being performed), and adapters ligated at the 5′ and 3′ end of the fragment. A PCR step is performed with transposon and adapter-specific oligos to enrich for transposon-containing fragments. Sequencing is then performed outwards from the transposon, with reads being mapped back to the reference genome to determine the location of all sites of transposon insertion.

Figure 2

Examples of using TIS to increase biological understanding.A, Using TIS to identify P. aeruginosa T6SS immunity genes (tsi) and adjacent toxin genes (tse) (44). Two high-density TIS libraries were generated in a T6SS active or inactive background. Each pool of mutants was separately incubated at high-density to promote cell contact-dependent T6SS killing via delivery of toxins into cells lacking the cognate immunity. This delivery only occurs in a T6SS active background which, results in cell death and absence of the tsi mutant in the library pool. Transposon insertions (red lines in graphs) are tolerated in the cognate immunity gene (tsi) only in the T6SS inactive background. B, identification of genes involved in S. aureus eDNA release in biofilms (78). A TIS library of S. aureus was used to grow biofilm. Extracellular DNA (eDNA) was separated from the genomic DNA of biofilm-grown cells and sites of transposon insertion in each pool identified. Genes involved in eDNA release are absent in the eDNA pool, and thus transposon insertions are not observed in these genes (e.g. gene b in the right-hand graph has no transposon insertions, as represented by the red lines in the graph) but present in all genes in the genomic DNA pool. C, assigning phenotypes for genes in diverse bacteria under many conditions using RB-Tnseq (113). For all 32 bacteria in the study a random-barcode Tn-seq (RB-Tnseq) library was generated, whereby a unique barcode (colored box) is associated with the transposon (black box) and randomly inserted into the genome, which creates a unique barcode for each transposon insertion site. Here shown are unique barcode-transposons in 3 genes, gene X, Y, and Z. The input library is sequenced to identify the gene that each unique barcode transposon is inserted into. The pooled RB-Tnseq library can then be inoculated into multiple selective conditions. In the Price et al. (113) study there were at least 173 conditions assayed for each bacterium. After growth a PCR step is performed followed by sequencing to determine the barcode abundance, which can then be used to determine the fitness of each gene under the conditions assayed. This can include significant numbers of “dark matter” genes as was the case in this study by Price et al. (113).

Figure 3

Building upon TIS outputs to go from a predicted function of “dark matter” genes to a comprehensive biological understanding. TIS approaches can be conducted in any number of assays, from in vitro to in vivo, and with single or multi-species communities. All these experiments can provide predictions about the function of genes in the “dark matter” category. Integration of these data with data generated in other approaches is essential to obtain true biological understanding. For example, to combine TIS outputs with outputs from 'omics datasets (e.g. transcriptomics, proteomics, and/or metabolomics) that were conducted under the same conditions; to validate the predictions using genome scale metabolic models from TIS outputs and experimentally validate these models; to use artificial intelligence (AI) and machine learning (ML) approaches to make structure-function predictions (e.g. AlphaFold3), to mine TIS datasets and make connections between TIS outputs for similar phenotypes (e.g. AMR, biofilms development, infection models) for different bacteria, and for future and retrospective TIS datasets.