RNA interference screening identifies host factors involved in brucella infection

Intracellular pathogens rely to varying extents on cellular functions of the host cell for their own propagation. A number of bacteria have evolved strategies to invade human cells and to establish an intracellular niche, which often consists of a cellular compartment that is modified by the pathogen to its own benefit. To understand the infection strategies of such organisms and eventually design new medical interventions, knowledge on the host factors exploited by the pathogens is critical. To this end, the InfectX consortium attempts to decipher the human infectome for a number of bacterial and viral pathogens.
In this framework, we study the zoonotic pathogen Brucella, which is able to invade phagocytic as well as non-phagocytic cells. The molecular mechanisms by which Brucella enters cells, evades lysosomal degradation, and finally replicates in an endoplasmic reticulum-like compartment, remain largely unknown.
To identify host factors involved in these processes, genome-wide microscopy-based RNA interference (RNAi) screens of Brucella entry and replication in HeLa cells were performed. To assign the function of the hits from the genome-wide screen to either early or late stages of Brucella infection, a follow-up assay suitable for high-throughput screening of Brucella entry was established. Both screening protocols are described in detail in research article I.
In-depth analysis of the genome-wide siRNA data generated within InfectX found that all screens including the one for Brucella infection show signs of miRNA-like off-target effects. Research article II focuses on the discovery and validation of this phenomenon in siRNAs screens and illustrates the potential of such an analysis to discover natural miRNAs and synthetic miRNA-like molecules that regulate the process of study. These findings motivated the screening of a library of human miRNA mimics for their involvement in Brucella infection. We identified miR-103 and miR-107 (miR-103/107), which strongly promote Brucella entry in non-phagocytic cells as presented in research article III. Interestingly, also the infection of other pathogens tested within InfectX was promoted by these miRNAs. Proteome and transcriptome analyses of cells with high levels of miR-103/107 indicated that this alters endocytic properties which manifested in reduced clathrin dependent uptake of transferrin in these cells. Furthermore, the abundance of several surface receptors required by different pathogens is increased. TGF-β receptor 2 showed elevated expression upon miR-103/107 transfection and independent experiments could confirm that high levels of this transmembrane kinase promote Brucella infection.
Having analyzed the full scale of off-target effects, we set out to determine a strategy to validate candidate genes of the genome-wide screens. We thus assayed a set of human kinases with a total of eleven individual siRNAs and one siRNA pool. Research article IV shows that the true discovery rate is directly proportional to the number of siRNAs tested and that siRNA pools tend to give more reliable results than individual siRNAs. As a consequence of these findings, we used six independent siRNAs and one siRNA pool for the validation of genes discovered in the primary screen with one pooled and five single siRNAs. This allowed the identification of several host cell pathways relevant for Brucella infection. Besides previously known functions, which include actin cytoskeleton remodeling or maturation of endocytic vesicles, also novel ones such as FGF and TGF-β signaling were found. While most of these networks were connected to Brucella entry into HeLa cells, we were also able to identify retrograde trafficking between endosomes and the Golgi apparatus to regulate a post-entry process as presented in research article IV.
Altogether, the results of our studies presented here point out limitations as well as the potential of siRNA technology. If off-target effects are accounted for and experimental confirmation is applied carefully on identified factors, RNAi allows to successfully reveal genes and pathways hitherto unrelated to the mechanism of interest. Additionally, and if analyzed accordingly, off-target effects also constitute a rich source of information for the discovery of miRNAs and miRNA-like molecules that regulate a certain process. Applied to the presented screens for human factors taking part in Brucella infection, this led to the description of miRNAs and several host pathways, which support pathogenicity. By that our results contribute to the expansion of the currently described infectome for this intracellular pathogen.