Back

TENT5A anchors STING at the endoplasmic reticulum to block SEC24B-dependent trafficking and suppress cGAS-STING signaling

Zheng, S.; Wang, P.-h.

2026-07-25 immunology
10.64898/2026.07.24.739811 bioRxiv
Show abstract

The cGAS-STING pathway is a central sensor of cytosolic DNA that drives type I interferon responses during infection and tumor surveillance. Activation of STING requires coordinated trafficking from the endoplasmic reticulum (ER) to the Golgi, yet how this spatial transition is regulated by ER-resident factors remains incompletely understood. Here we identify TENT5A as an ER-localized negative regulator of STING signaling. Genetic depletion of TENT5A markedly enhances cGAS-STING-dependent type I interferon production in response to cytosolic DNA and DNA virus infection. Conversely, TENT5A overexpression suppresses downstream signaling and antiviral gene expression. Mechanistically, TENT5A directly interacts with the C-terminal tail of STING and competes with the COPII adaptor SEC24B, thereby preventing STING incorporation into COPII vesicles and blocking its ER-to-Golgi trafficking. This ER retention inhibits STING oligomerization, TBK1 recruitment and IRF3 activation, ultimately attenuating type I interferon responses. In vivo, myeloid-specific deletion of TENT5A enhances systemic interferon responses, restricts herpesvirus replication, and improves host survival. Moreover, TENT5A deficiency promotes CD8 T cell effector activation within the tumor microenvironment and suppresses tumor growth in a melanoma model. Together, our findings define TENT5A as an ER checkpoint that spatially controls STING trafficking through competition with COPII machinery, revealing vesicular sorting as a regulatory layer of innate immune activation.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.