The protein STING (stimulator of interferon genes) has an important role in the mammalian innate immune system. STING was initially thought to function as an adaptor by linking the cytosolic detection of microbial DNA to downstream signaling components and promoting the production of type I interferons. Mice deficient in STING show higher susceptibility to infection by different DNA viruses. More recently, STING was shown to be a sensor for the cyclic dinucleotide c-di-GMP, used by bacteria as secondary messengers. Now, Holm et al. find that STING is also involved in the interferon response of cells after viral membrane fusion, in a manner independent of nucleic acids or virus capsid proteins. The authors purified virus-like particles (VLPs) derived from herpes simplex virus I, which lacked capsid or viral genomic DNA but were still able to fuse with cellular membranes. The VLPs could induce an interferon response in both mouse and human cells without causing an inflammatory response. Using VLPs that lacked specific envelope glycoproteins, the authors established that viral membrane fusion is required to induce the interferon response. Other membrane-fusion events, such as cell-cell fusion or cationic liposome fusion, could cause a similar response. Using liposomes with different lipid composition, the authors could see a direct correlation between fusogenic potential and the level of cellular responses. They next investigated the signaling events involved in the cellular response to membrane fusion. Cells from STING-deficient mice did not show interferon responses upon treatment with VLPs or liposomes, implying that protein was needed for the response to membrane fusion. In addition, the roles of phospholipase C-γ and phosphatidylinositol-3-OH kinase, both involved in membrane-proximal signaling, were also demonstrated by using specific inhibitors. The authors note that many membrane-fusion events occur as part of normal physiological processes (for example, ER-membrane fusion or natural syncytium formation) and do not induce immune responses. Thus, cells must be able to differentiate those events from 'unscheduled' membrane fusions that are sensed as danger signals. How exactly membrane fusion is sensed remains to be established, but this work uncovers a novel innate detection mechanism and extends the role of STING in immune responses. (Nat. Immunol. doi:10.1038/ni.2350, published online 17 June 2012)