Search

In mammals, cyclic dinucleotides (CDNs) bind and activate STING to initiate an anti-viral type I interferon response. CDNs and STING originated in bacteria and are present in most animals. By contrast, interferons are believed to have emerged in vertebrates; thus, the function of CDN signaling in invertebrates is unclear. Here, we use a CDN, 2’3’-cGAMP, to activate immune responses in a model cnidarian invertebrate, the starlet sea anemone Nematostella vectensis. Using RNA-Seq, we found that 2’3’-cGAMP induces robust transcription of both anti-viral and anti-bacterial genes in N. vectensis. Many of the anti-viral genes induced by 2’3’-cGAMP are homologs of vertebrate interferon stimulated genes, implying that the interferon response predates the evolution of interferons. Knockdown experiments identified a role for NF-κB in specifically inducing anti-bacterial genes downstream of 2’3’-cGAMP. Some of these putative anti-bacterial genes were also found to be induced during Pseudomonas aeruginosa infection. We characterized the protein product of one of the putative anti-bacterial genes, the N. vectensis homolog of Dae4, and found that it has conserved anti-bacterial activity. This work suggests that a broad anti-bacterial and anti-viral transcriptional response is an evolutionarily ancestral output of 2’3’-cGAMP signaling in animals.Significance statementCyclic dinucleotides are signaling molecules that originated in bacteria and were subsequently acquired and co-opted by animals for immune signaling. The major cyclic dinucleotide signaling pathway in mammals results in the production of anti-viral molecules called interferons. Invertebrates such as sea anemones lack interferons, and thus it was unclear whether cyclic dinucleotide signaling would play a role in immunity in these animals. Here we report that in the anemone Nematostella vectensis, cyclic dinucleotides activate both anti-viral and anti-bacterial immune responses, and do so through a conserved pathway. These results provide insights into the evolutionary origins of innate immunity, and suggest a broader ancestral role for cyclic dinucleotide signaling that evolved toward more specialized anti-viral functions in mammals.

The Stimulator of Interferon Genes (STING) pathway initiates potent immune responses upon recognition of DNA derived from bacteria, viruses and tumors. To signal, the C-terminal tail (CTT) of STING recruits TBK1, a kinase that phosphorylates serine 365 (S365) in the CTT. Phospho-S365 acts as a docking site for IRF3, a transcription factor that is phosphorylated and activated by TBK1, leading to transcriptional induction of type I interferons (IFNs). IFNs are essential for antiviral immunity and are widely viewed as the primary output of STING signaling in mammals. However, other more evolutionarily ancestral responses, such as induction of NF-κB or autophagy, also occur downstream of STING. The relative importance of the various outputs of STING signaling during in vivo infections is unclear. Here we report that mice harboring a serine 365-to-alanine (S365A) point mutation in STING exhibit normal susceptibility to Mycobacterium tuberculosis infection but, unexpectedly, are resistant to Herpes Simplex Virus (HSV)-1, despite lacking STING-induced type I IFN responses. Likewise, we find Irf3-/- mice exhibit resistance to HSV-1. By contrast, resistance to HSV-1 is abolished in mice lacking the STING CTT or TBK1, suggesting that STING protects against HSV-1 upon TBK1 recruitment by the STING CTT, independent of IRF3 or type I IFNs. Interestingly, we find that STING-induced autophagy is a TBK1-dependent IRF3-independent process that is conserved in the STING S365A mice, and autophagy has previously been shown to be required for resistance to HSV-1. We thus propose that autophagy and perhaps other ancestral interferon-independent functions of STING are required for STING-dependent antiviral responses in vivo.