Major histocompatibility complex (MHC) class I and II are crucial for the adaptive immune system because they are involved in peptide presentation to T cells. Until recently, it was believed that MHC genes and their associated immune components had been conserved since their evolutionary emergence in jawed fish. However, sequencing of the Atlantic cod (Gadus morhua) genome revealed a loss of MHC class II genes, and an extreme expansion of MHC class I genes. These findings lead to the hypothesis that a loss of the MHC class II pathway coincided with a more versatile use of MHC class I, but so far there is no direct experimental evidence in support of this. To gain a deeper understanding of the function of the expanded MHC class I, we selected five MHC class I gene variants representing five of the six clades identified in previous studies and investigated their intracellular localization in human and Atlantic cod larval cells. Intriguingly, we uncovered that all selected MHC class I variants localize to endolysosomal compartments in Atlantic cod cells. Additionally, by introducing point mutations or deletions in the cytosolic tail, we found that hypothetical sorting signals in the MHC class I cytosolic tail do not influence MHC class I trafficking. Moreover, we demonstrated that in Atlantic cod, tapasin and MHC class I colocalize on endolysosomes suggesting that peptide-loading assistance and stabilization of MHC class I occurs outside the endoplasmic reticulum. Altogether, our results demonstrate that MHC class I from Atlantic cod is sorted to the endolysosomal system, which may indicate that it interacts with exogenous peptides for potential cross presentation.

Stimulator of interferon genes (STING) is essential for the type I interferon response induced by microbial DNA from viruses or self-DNA from mitochondria/nuclei. Recently, gain-of-function mutations in STING have been identified in patients with STING-associated vasculopathy with onset in infancy (SAVI). The SAVI patients exhibit complex systemic vascular inflammation and interstitial lung disease, resulting in pulmonary fibrosis and respiratory failure. SAVI mouse models have recently developed, harbouring common SAVI mutations, such as N153S and V154M, which correspond to the human N154S and V155M, respectively. Interestingly, crosses of heterozygous SAVI mice did not yield homozygous SAVI mice as of embryonic day 14, indicating that homozygous SAVI embryos were not viable and that wild-type (WT) allele would function dominantly over SAVI alleles in terms of viability. However, the molecular mechanism underlying the dominance has not been understood. In the present study, we show that STING (WT) and STING (SAVI) can form heterocomplex. The heterocomplex localized primarily in the endoplasmic reticulum (ER) and failed to reach the trans-Golgi network (TGN), where STING activates the downstream kinase TBK1. SURF4 is the essential protein functioning in the retrieval of STING from the Golgi to the ER. The amount of SURF4 bound to STING (SAVI) significantly increased in the presence of STING (WT). These results suggest that STING (WT) can suppress the activity of STING (SAVI) by tethering STING (SAVI) to the ER through heterocomplex formation. The dormant heterocomplex formation may underlie, at least in part, the dominance of STING WT allele over SAVI alleles in the STING-triggered inflammatory response.