Three functionally distinct classes of cGAS proteins in nature revealed by self-DNA-induced interferon responses

Mosallanejad, K. et al. (BioRxiv) doi: 10.1101/2022.03.09.483681


  • cGAS

  • Cytosolic self-DNA

  • Interferon response

  • Innate sensing


Main Findings

In this preprint, Mosallenejad et al. showed that mammalian cGAS could be grouped into three classes, each functionally distinct in their induction of class I interferon (IFN) and the interferon stimulated gene (ISG) RSAD2 in reaction to self-DNA. In class 1 cGAS, the N-terminus from human and orangutan cGAS suppressed self-DNA reactivity in human and mouse cells, as it has been described in humans (Barnett et al. 2019). This phenotype for class 1 was cell-type specific (in mouse but not human cells) for gibbon and marmoset cGAS, thus it would be interesting to see whether this same phenotype occurs in gibbon and marmoset host cells. In contrast, the N terminus of class 2 cGAS from mice promoted the production of IFN from self-DNA stimuli while class 3 cGAS from chimpanzee, crab-eating macaques, rhesus macaques and white-handed gibbons was not self-DNA reactive at all. 

The authors went on to show that human class 1 cGAS lacking the N-terminus (cGASΔN) is self-DNA reactive after mislocalisation of the protein from the cytosol to the mitochondria, as has recently been shown in human cells (Li et al. 2021). However, mislocalisation of cGAS to the mitochondria was not observed in self-DNA reactive full length class 2 mouse cGAS, and chimpanzee class 3 cGAS was mislocated to the mitochondria, despite not being self-DNA reactive. Therefore, it is evident that further studies are needed to explain the cellular mechanisms behind self-DNA reactivity in each class of cGAS.

The study finally showed that type I IFN responses from class 1 cGAS could be released by anchoring human cGASΔN to the outer mitochondrial membrane and subsequently releasing it using a virus protease-mediated mechanism. Using this mechanism, cGAS could be repurposed as a guard-like protein that induced IFN responses in response to viral virulence activity. 

Overall this study identified three distinct classes of cGAS proteins from different mammals and showed that innate immune sensing mechanisms are not conserved throughout nature. The study provided important evidence for the differences between cGAS reactivity to self-DNA and showed how cGAS induction of IFN could be repurposed using molecular techniques to act as a sensing mechanism of intracellular viral protease activity. 


  • Limited analysis of different cell types, or analysis of primary cells 

  • Lack of functional analysis of different classes of cGAS proteins from different mammalian species in homologous cell types (for example, rhesus macaque cGAS in rhesus macaque cells)



The preprint shows for the first time that mammalian cGAS may not respond to self-DNA and induce interferon responses similarly across mammalian species. The preprint also provides evidence to clarify certain discrepancies between previous studies as to the activity of human cGASΔN constructs – showing that fusing epitope tags the N-terminus of cGASΔN rendered it inactive compared to preserving cGASΔN self-DNA reactivity when it was tagged at the C-terminus. This preprint highlights the need to consider the use of mouse and NHP models for studying cGAS sensing in human disease. This authors provide significant evidence for immunological research that innate sensing mechanisms of endogenous DNA may not be conserved across mammals, which will be interesting to explore further in depth for example in the sensing of DNA deriving from infection or cellular injury. 


Reviewed by Beth Charlton as part of the cross-institutional journal club of the Immunology Institute of the Icahn School of Medicine, Mount Sinai, the Kennedy Institute of Rheumatology and the Oxford Centre for Immuno-Oncology  (OXCIO) (University of Oxford, GB) and Karolinska Institute’s Center for Infectious Medicine (CIM) & Center for Molecular Medicine (CMM). You can follow her on Twitter.