Many intercellular signals are synthesised as transmembrane precursors that are released by proteolytic cleavage (‘shedding’) from the cell surface. ADAM17, a membrane-tethered metalloprotease, is the primary shedding enzyme responsible for the release of the inflammatory cytokine TNFα and several EGF receptor ligands. ADAM17 exists in complex with the rhomboid-like iRhom proteins, which act as cofactors that regulate ADAM17 substrate shedding. Here we report that the poorly characterised FERM domain-containing protein FRMD8 is a new component of the iRhom2/ADAM17 sheddase complex. FRMD8 binds to the cytoplasmic N-terminus of iRhoms and is necessary to stabilise iRhoms and ADAM17 at the cell surface. In the absence of FRMD8, iRhom2 and ADAM17 are degraded via the endolysosomal pathway, resulting in the reduction of ADAM17-mediated shedding. We have confirmed the pathophysiological significance of FRMD8 in iPSC-derived human macrophages and mouse tissues, thus demonstrating its role in the regulated release of multiple cytokine and growth factor signals.

The canonical members of the Jagged/Serrate and Delta families of transmembrane ligands have an extracellular, amino-terminal C2 domain that binds to phospholipids and is required for optimal activation of the Notch receptor. Somatic mutations that cause amino substitutions in the C2 domain in human JAGGED1 (JAG1) have been identified in tumors. We found in reporter cell assays that mutations affecting an N-glycosylation site reduced the ligand’s ability to activate Notch. This N-glycosylation site located in the C2 domain is conserved in the Jagged/Serrate family but is lacking in the Delta family. Site-specific glycan analysis of the JAG1 amino terminus demonstrated that occupancy of this site by either a complex-type or high-mannose N-glycan was required for full Notch activation in reporter cell assays. Similarly to JAG1 variants with defects in Notch binding, N-glycan removal, either by mutagenesis of the glycosylation site or by endoglycosidase treatment, reduced receptor activation. The N-glycan variants also reduced receptor activation in a Notch signaling–dependent vascular smooth muscle cell differentiation assay. Loss of the C2 N-glycan reduced JAG1 binding to liposomes to a similar extent as the loss of the entire C2 domain. Molecular dynamics simulations suggested that the presence of the N-glycan limits the orientation of JAG1 relative to the membrane, thus facilitating Notch binding. These data are consistent with a critical role for the N-glycan in promoting a lipid-binding conformation that is required to orient Jagged at the cell membrane for full Notch activation.

An accurate and reliable susceptibility testing method for polymyxins is urgently needed not only for the clinical laboratory but also for new polymyxin-like lipopeptide development. Reference broth microdilution (rBMD), which was the recommended method by CLSI-EUCAST in clinics, has been proven not to be ideal, while the agar dilution (AD) method that was widely used in new antibiotics discovery has been neglected. In the present study, the AD method was compared with rBMD and broth macrodilution (BMAD) in susceptibility testing of polymyxin B and colistin against >200 Gram-negative isolates. AD showed strong agreement with BMAD for colistin (except for Klebsiella aerogenes and Pseudomonas aeruginosa); however, its performance was poor for polymyxin B or compared to rBMD. MICs of AD method were not affected when different types of Petri dishes were used, while glass-bottom microtiter plates could lower the MIC of polymyxins 2–8 times compared to tissue-culture-treated polystyrene plates when using rBMD, which demonstrated that tissue-culture-treated plates were not suitable. It was then validated with non-tissue-culture-treated plates. The culture volume was another influencing factor of accuracy for rBMD, and 200 μL seemed to be the most suitable volume for MIC detection of polymyxins. Additionally, no lack of growth phenomenon (skipped well) was observed for AD when it frequently occurred for both BMAD and rBMD. As for strains carrying mcr-1 gene, 100% of AD results were in essential agreement (EA) and categorical agreement (CA) with both rBMD and BMAD. Overall, rBMD is convenient and widely accepted for susceptibility testing of polymyxins. Although it may be too early to say that AD is superior compared to rBMD and BMAD, it did show some advantages in repeatability and anti-interference ability.