Genetic Expeditions with Haploid Human Cells

Random mutagenesis followed by phenotypic selection (forward genetics) is among the most powerful tools to elucidate the molecular basis of intricate biological processes and has been used in a suite of model organisms throughout the last century. However, its application to cultured mammalian cells is complicated by their diploid nature, as for most genes this creates a genetic buffer preventing the occurrence of a loss-of-function phenotype. Here, we highlight the potential of haploid human somatic cells for random mutagenesis-based genetic screens. First, using the hemorrhagic fever agent Lassa virus and specific antibodies, we profile genes needed for the functional glycosylation of the Lassa virus receptor, dystroglycan, in a genome-wide screen. This reveals all previously known genes needed for dystroglycan modification and implicates five new factors in the process. Surprisingly, these include the putative kinase SGK196/POMK – unusual for the generation of glycan structures. Using nuclease-generated knockout cell lines and genetic complementation, we confirm the function of these genes in dystroglycan glycosylation and Lassa virus entry and describe loss-of-function mutations for SKG196 and TMEM5 in patients suffering from hereditary dystroglycanopathy of unknown cause, underscoring the relevance of our observations for human disease. In-depth analysis of Lassa virus entry with iterative haploid genetic screens reveals a role for the intracellular factor lysosome-associated membrane protein 1 (LAMP1) in Lassa virus entry that is independent of dystroglycan. We expose an interaction of the viral envelope glycoprotein with human LAMP1 that is regulated by pH and not shared by avian LAMP1 – rationalizing the 30-year-old observation that bird cells resist infection. Additional haploid genetic screens point out a specific requirement of Lassa virus for α2,3-linked sialic acid, which we map to LAMP1. Serial ablation of LAMP1 glycosylation sites reveals a critical function of glycosylated Asn76 for glycoprotein binding and infection. Remarkably, this site is absent in avian LAMP1, explaining its inability to accommodate virus entry. Finally, we confirm the importance of LAMP1 for Lassa virus infection in vivo using knockout mice. Our findings suggest an unusual entry mechanism for Lassa virus that involves a pH-induced receptor switch from dystroglycan to LAMP1 as the virus traverses the endocytic compartment. Last, we describe how haploid genetics can be used to identify human genes needed for fitness and expose genetic interactions. By mapping tens of millions of independent gene-trap integrations, we assign a fitness score to virtually every expressed gene. This reveals approximately 2,000 genes needed for cellular fitness in KBM7 and HAP1 cells. With this information we investigate the purpose of essential ‘orphan genes‘. This reveals a role for C11orf10 in the functionally conserved oligosaccharyltransferase complex needed for N-glycosylation. Using depletion screens in isogenic knockout cells, we assemble a genetic interaction map of the secretory pathway that exposes a role for PTAR1 and C10orf76 in this cellular process. Our data demonstrate the existence of numerous synthetic lethal interactions between human genes and their high degree of compartmentalization. This parallels observations in yeast and encourages a systematic search for synthetic lethal interactions in genes frequently mutated in human malignancies.