As key executers of biological functions, the activity and abundance of proteins are subjected to extensive regulation. Deciphering the genetic architecture underlying this regulation is critical for understanding cellular signalling events and responses to environmental cues. Using random mutagenesis in haploid human cells, we apply a sensitive approach to directly couple genomic mutations to protein measurements in individual cells. Here we use this to examine a suite of cellular processes, such as transcriptional induction, regulation of protein abundance and splicing, signalling cascades (mitogen-activated protein kinase (MAPK), G-protein-coupled receptor (GPCR), protein kinase B (AKT), interferon, and Wingless and Int-related protein (WNT) pathways) and epigenetic modifications (histone crotonylation and methylation). This scalable, sequencing-based procedure elucidates the genetic landscapes that control protein states, identifying genes that cause very narrow phenotypic effects and genes that lead to broad phenotypic consequences. The resulting genetic wiring map identifies the E3-ligase substrate adaptor KCTD5 (ref. 1) as a negative regulator of the AKT pathway, a key signalling cascade frequently deregulated in cancer. KCTD5-deficient cells show elevated levels of phospho-AKT at S473 that could not be attributed to effects on canonical pathway components. To reveal the genetic requirements for this phenotype, we iteratively analysed the regulatory network linked to AKT activity in the knockout background. This genetic modifier screen exposes suppressors of the KCTD5 phenotype and mechanistically demonstrates that KCTD5 acts as an off-switch for GPCR signalling by triggering proteolysis of Gβγ heterodimers dissociated from the Gα subunit. Although biological networks have previously been constructed on the basis of gene expression, protein-protein associations, or genetic interaction profiles, we foresee that the approach described here will enable the generation of a comprehensive genetic wiring map for human cells on the basis of quantitative protein states.

The phospholipase PLA2G16 is required for the entry of picornaviruses, and in its absence, virus infection is prevented by a galectin-8-mediated process. Thijn Brummelkamp and colleagues use a genome-wide haploid genetic screen to identify the host factors required for the replication of picornaviruses. They identify the small phospholipase PLA2G16 as being required for the cytoplasmic delivery of the viral genome. In a second screen to find mutants that restored virus susceptibility to PLA2G16-deficient cells, the authors identify galectin-8, a sensor previously implicated in the autophagic clearance of intracellular bacteria. The precise function of PLA2G16 remains unclear, but the authors suggest that it facilitates the displacement of the viral genome from galactin-8-positive vesicles. Picornaviruses are a leading cause of human and veterinary infections that result in various diseases, including polio and the common cold. As archetypical non-enveloped viruses, their biology has been extensively studied1. Although a range of different cell-surface receptors are bound by different picornaviruses2,3,4,5,6,7, it is unclear whether common host factors are needed for them to reach the cytoplasm. Using genome-wide haploid genetic screens, here we identify the lipid-modifying enzyme PLA2G16 (refs 8, 9, 10, 11) as a picornavirus host factor that is required for a previously unknown event in the viral life cycle. We find that PLA2G16 functions early during infection, enabling virion-mediated genome delivery into the cytoplasm, but not in any virion-assigned step, such as cell binding, endosomal trafficking or pore formation. To resolve this paradox, we screened for suppressors of the ΔPLA2G16 phenotype and identified a mechanism previously implicated in the clearance of intracellular bacteria12. The sensor of this mechanism, galectin-8 (encoded by LGALS8), detects permeated endosomes and marks them for autophagic degradation, whereas PLA2G16 facilitates viral genome translocation and prevents clearance. This study uncovers two competing processes triggered by virus entry: activation of a pore-activated clearance pathway and recruitment of a phospholipase to enable genome release.

Amplification of MYCN is a driver mutation in a subset of human neuroendocrine tumors, including neuroblastoma. No small molecules that target N-Myc, the protein encoded by MYCN, are clinically available. N-Myc forms a complex with the Aurora-A kinase, which protects N-Myc from proteasomal degradation. Although stabilization of N-Myc does not require the catalytic activity of Aurora-A, we show here that two Aurora-A inhibitors, MLN8054 and MLN8237, disrupt the Aurora-A/N-Myc complex and promote degradation of N-Myc mediated by the Fbxw7 ubiquitin ligase. Disruption of the Aurora-A/N-Myc complex inhibits N-Myc-dependent transcription, correlating with tumor regression and prolonged survival in a mouse model of MYCN-driven neuroblastoma. We conclude that Aurora-A is an accessible target that makes destabilization of N-Myc a viable therapeutic strategy.

The Myc transcription factor family affects the expression of a large group of genes that play an important role during tumourigenesis. Therefore, deregulated expression of the MYC gene family is a hallmark of multiple human tumours. An indirect reduction of Myc-protein levels due to interference with protein-stabilizing mechanisms could be a promising therapeutic approach targeting Myc-dependent tumours.

Multiple proof-of-principle experiments suggest that targeting members of the Myc family of oncoproteins will have significant therapeutic benefit in human tumors. Deregulated expression of MYCN, one of the three family members implicated in human tumors, is prevalent in both pediatric and adult neuroendocrine tumors. We have shown previously that MYCN-amplified neuroblastoma cells depend on high levels of Aurora-A expression for maintaining N-Myc function. Similarly, Aurora-A and N-Myc bind each other and together drive an oncogenic gene expression program in neuroendocrine prostate tumors. Although stabilization of N-Myc does not require the catalytic activity of Aurora-A, we found that two inhibitors of Aurora-A disrupt the Aurora-A/N-Myc complex and promote degradation of N-Myc via the Fbxw7 ubiquitin ligase. Disruption of the Aurora-A/N-Myc complex inhibits N-Myc-dependent transcription, correlating with tumor regression and prolonged survival in a mouse model of MYCN-driven neuroblastoma, suggesting that Aurora-A is an accessible target that makes destabilization of N-Myc a viable therapeutic strategy. Citation Format: Markus Brockmann, Evon Poon, Louis Chesler, Martin Eilers. Targeting the dependence of N-Myc on interaction with Aurora-A with small molecules. [abstract]. In: Proceedings of the AACR Precision Medicine Series: Synthetic Lethal Approaches to Cancer Vulnerabilities; May 17-20, 2013; Bellevue, WA. Philadelphia (PA): AACR; Mol Cancer Ther 2013;12(5 Suppl):Abstract nr IA13.

In human neuroblastoma, amplification of the MYCN gene predicts poor prognosis and resistance to therapy. In a shRNA screen of genes that are highly expressed in MYCN-amplified tumors, we have identified AURKA as a gene that is required for the growth of MYCN-amplified neuroblastoma cells but largely dispensable for cells lacking amplified MYCN. Aurora A has a critical function in regulating turnover of the N-Myc protein. Degradation of N-Myc requires sequential phosphorylation by cyclin B/Cdk1 and Gsk3. N-Myc is therefore degraded during mitosis in response to low levels of PI3-kinase activity. Aurora A interacts with both N-Myc and the SCF(Fbxw7) ubiquitin ligase that ubiquitinates N-Myc and counteracts degradation of N-Myc, thereby uncoupling N-Myc stability from growth factor-dependent signals.