No single gene acts by itself, instead the genome is organized into an intricate network of interacting components to ensure the organism mounts an appropriate response to its environment. A genetic interaction network (GIN) represents a global view of these relationships and, for instance, can depict a cell as a functional wiring diagram. Thereby GINs are key to develop an integrated understanding of all processes in a cell or organism. A genetic interaction is defined as a combination of mutations that have an unexpected phenotype with respect to the effect of the individual perturbations. For instance, two mutations that have little effect by themselves when combined may be lethal (a negative interaction) or two mutations that have a negative effect individually may have no effect when combined (a positive interaction). For model systems including yeast, tools such as synthetic genetic array analyses exist that allows for sampling of double gene knockouts on a genome-wide scale. This approach has enabled sampling of >23 million interactions and has resulted in the most detailed genetic interaction network to date consisting of ~900,000 genetic interactions. In contrast, an easily implementable approach for bacteria that can map genome-wide genetic interactions in high-throughput is lacking. In this proposal we solve this challenge by developing pooled and dual-guided CRISPRi (p&dgCRISPRi) in the bacterial pathogen Streptococcus pneumoniae. As a proof-of-principle we developed a relatively small version of p&dgCRISPRi. To enable this, we designed a cloning strategy aimed at combining two single guide RNAs (gRNAs) into a single genome targeting all pairwise combinations of a set of 105 genes in S. pneumoniae. Thereby ~5000 pairwise interactions were screened in a pool, resulting in ~500 negative interactions and ~200 positive interactions.
In Aim 1, we scale-up the approach and generate saturated libraries totaling ~1.2 million interactions. We first evaluate 10 gRNAs for each open reading frame (ORF) in the genome, and select two efficient ones. These gRNAs are than used to generate over 1.2 million pooled S. pneumoniae CRISPRi strains where each bacterium expresses 2 gRNAs. Each gRNA-pair is linked to two random barcodes, and the change in frequency of these barcodes in the population, which is determined by Illumina sequencing, is used to calculate their effect on fitness.
In Aim 2, we build the first genome-wide genetic interaction network for S. pneumoniae by screening the p&dgCRISPRi libraries in rich and minimal media, and in rich media supplemented with an antibiotic from one of the four major classes. Networks are analyzed in detail and are combined and fused with additional (omics)data to provide context, and mined for new biological insights, while 30-50 interactions are validated to confirm high- confidence interactions. Most importantly, these GINs will proof central to developing an integrated understanding of all processes in an organism and may for instance aid in the design of new antimicrobial strategies.
A genetic interaction network represents a global view of the relationships between the genes in a genome that work together to establish a specific phenotype. While genetic interaction networks are key to develop an integrated understanding of all processes in an organism, there is no widely applicable approach that can do this in high-throughput and genome-wide for bacterial pathogens. Here we solve this challenge by developing and employing pooled and dual-guided CRISPRi, Streptococcus pneumoniae. and build the first genome-wide genetic interaction network for