Erdem Karatekin

Type IV secretion systems (T4SSs) are versatile molecular machines used by bacteria to secrete protein effectors into host cells, promoting pathogenesis, and to transfer DNA between bacteria through conjugation, driving horizontal gene transfer. Most, like Dot/Icm of the pathogen Legionella pneumophila ( L. pneumophila) or Escherichia coli ( E. coli) RK2, are primed for substrate delivery only upon contact with a target membrane, but mechanisms are unknown. A pilus could bind a receptor to initiate priming, but many T4SSs, especially those that deliver effectors, lack a pilus. Here, we present evidence that T4SSs are primed by direct contact with target membrane lipids. Combining fluorescence assays with genetics and biochemistry, we found that Dot/Icm drives lipid exchange between bacterial cells and between bacteria and synthetic membranes containing only lipids. Lipid exchange requires membrane contact but does not require ATP hydrolysis or even full complex assembly. Minimally, the outer membrane core complex protein DotG needs to be present in at least one of the apposed membranes. We similarly observed lipid mixing with the simpler E. coli RK2 T4SS, where we could follow lipid mixing and plasmid transfer simultaneously. We found that lipid mixing always preceded or accompanied plasmid transfer, suggesting it may be part of the contact-dependent priming mechanism. Lipid mixing was inhibited or promoted by lipids that inhibit or promote membrane fusion, respectively. Lipids inhibiting lipid mixing also inhibited substrate transfer. Together, our results suggest that initial contact between DotG outer segments and target membrane lipids promotes lipid mixing as part of the mechanism that primes T4SS for substrate translocation. Highlights: The Legionella pneumophila Dot/Icm secretion system requires contact with a target membrane for effector translocation, but how membrane contact primes the machinery for this is unknown. We found that Dot/Icm drives contact-dependent lipid mixing between bacteria or between bacteria and liposomes, showing that Dot/Icm priming does not require a protein receptor. Lipid mixing also occurs with the simpler Type IVA system ( Escherichia coli RK2) during conjugation, suggesting lipid mixing is a general feature of Type IV Secretion System function. The core complex protein DotG is sufficient to drive lipid mixing, suggesting DotG-target membrane interactions may destabilize the target membrane.

Membrane fission is an energy-consuming process, critical for all domains of life. Prototypical fission machineries use local energy input such as nucleoside triphosphate hydrolysis to constrict and cut membranes. However, some membrane fission reactions paradoxically rely on protein scaffolds that by themselves stabilize rather than cut membranes. It turns out these proteins do not work alone; they use nonlocal energy input that generates a membrane tension gradient. Such a gradient mobilizes membrane flow that in turn tends to relax the membrane tension gradient. By interfering with membrane flow, the protein scaffold causes the membrane tension to increase unchecked to the point of mechanical failure, membrane fission. This friction-driven scission (FDS) mechanism is generic, conserved from bacteria to humans, and only requires two ingredients: a membrane tension generating process and a protein scaffold that hinders the associated membrane flow. Because both are often present in cells, it is likely that FDS contributes to membrane fission more frequently than previously appreciated.

Hair cells of the zebrafish lateral line have proven to be a good model for studying hair cell function in a system that is easily genetically manipulated, rapidly develops and is experimentally accessible. However, characterization of potential developmental changes, and possible differences along lateral line position are lacking. Here, we used in vivo patch clamp to investigate the electrophysiological and exocytic properties of neuromast hair cells over early development across body location. Long depolarizations led to steady increases in membrane capacitance, presumably due to exocytosis of vesicles localized to ribbon synapses. The magnitude and kinetics of capacitance changes did not vary significantly across the L1 to L6 position of neuromasts along lateral line, but the magnitudes were found to be significantly smaller in hair cells found in the tail region across all developmental time points. For each region, we found no significant changes in capacitance responses between 3 and 7 days after fertilization. Hair cell capacitance responses were greatly reduced in animals injected with CRISPR/Cas9 with gRNAs targeted to otoferlin b. These results confirm the essential role of otoferlin b in neuromast hair cell function, and they establish the fidelity of CRISPR/Cas9 to rapidly mediate genetic removal of critical genes to study their impact on synaptic release.

Synaptotagmin-1 (Syt1) is a major calcium sensor for rapid neurotransmitter release in neurons and hormone release in many neuroendocrine cells. It possesses two tandem cytosolic C2 domains that bind calcium, negatively charged phospholipids, and the neuronal SNARE complex. Calcium binding to Syt1 triggers exocytosis, but how this occurs is not well understood. Syt1 has additional roles in docking dense core vesicles (DCV) and synaptic vesicles (SV) to the plasma membrane (PM) and in regulating fusion pore dynamics. Thus, Syt1 perturbations could affect release through vesicle docking, fusion triggering, fusion pore regulation, or a combination of these. Here, using a human neuroendocrine cell line, we show that neutralization of highly conserved polybasic patches in either C2 domain of Syt1 impairs both DCV docking and efficient release of serotonin from DCVs. Interestingly, the same mutations resulted in larger fusion pores and faster release of serotonin during individual fusion events. Thus, Syt1’s roles in vesicle docking, fusion triggering, and fusion pore control may be functionally related.

Cell membrane tension affects and is affected by many fundamental cellular processes, yet it is poorly understood. Recent experiments show that membrane tension can propagate at vastly different speeds in different cell types, reflecting physiological adaptations. Here we briefly review the current knowledge about membrane tension gradients, membrane flows, and their physiological context.

Over the past 15 years, hundreds of previously undiscovered bacterial small open reading frame (sORF)-encoded polypeptides (SEPs) of fewer than fifty amino acids have been identified, and biological functions have been ascribed to an increasing number of SEPs from intergenic regions and small RNAs. However, despite numbering in the dozens in Escherichia coli, and hundreds to thousands in humans, same-strand nested sORFs that overlap protein coding genes in alternative reading frames remain understudied. In order to provide insight into this enigmatic class of unannotated genes, we characterized GndA, a 36-amino acid, heat shock-regulated SEP encoded within the +2 reading frame of the gnd gene in E. coli K-12 MG1655. We show that GndA pulls down components of respiratory complex I (RCI) and is required for proper localization of a RCI subunit during heat shock. At high temperature GndA deletion (ΔGndA) cells exhibit perturbations in cell growth, NADH⁺/NAD ratio, and expression of a number of genes including several associated with oxidative stress. These findings suggest that GndA may function in maintenance of homeostasis during heat shock. Characterization of GndA therefore supports the nascent but growing consensus that functional, overlapping genes occur in genomes from viruses to humans. Significance Statement Same-strand overlapping, or nested, protein coding sequences optimize the information content of size-constrained viral genomes, but were previously omitted from prokaryotic and eukaryotic genome annotations. It was therefore surprising when dozens of nested sORFs were recently discovered in bacteria. Our case study of E. coli GndA supports the hypothesis that overlapping genes may exist because they encode proteins with related functions. More broadly, characterization of nested sORFs may revise our understanding of the architecture of bacterial and eukaryotic genes.

Understanding protective immunity to COVID-19 facilitates preparedness for future pandemics and combats new SARS-CoV-2 variants emerging in the human population. Neutralizing antibodies have been widely studied; however, on the basis of large-scale exome sequencing of protected versus severely ill patients with COVID-19, local cell-autonomous defence is also crucial^1–4. Here we identify phospholipid scramblase 1 (PLSCR1) as a potent cell-autonomous restriction factor against live SARS-CoV-2 infection in parallel genome-wide CRISPR–Cas9 screens of human lung epithelia and hepatocytes before and after stimulation with interferon-γ (IFNγ). IFNγ-induced PLSCR1 not only restricted SARS-CoV-2 USA-WA1/2020, but was also effective against the Delta B.1.617.2 and Omicron BA.1 lineages. Its robust activity extended to other highly pathogenic coronaviruses, was functionally conserved in bats and mice, and interfered with the uptake of SARS-CoV-2 in both the endocytic and the TMPRSS2-dependent fusion routes. Whole-cell 4Pi single-molecule switching nanoscopy together with bipartite nano-reporter assays found that PLSCR1 directly targeted SARS-CoV-2-containing vesicles to prevent spike-mediated fusion and viral escape. A PLSCR1 C-terminal β-barrel domain—but not lipid scramblase activity—was essential for this fusogenic blockade. Our mechanistic studies, together with reports that COVID-associated PLSCR1 mutations are found in some susceptible people^3,4, identify an anti-coronavirus protein that interferes at a late entry step before viral RNA is released into the host-cell cytosol.