Additional file 1 of Aicardi-Goutières syndrome-associated mutation at ADAR1 gene locus activates innate immune response in mouse brain

Additional file 1: Figure S1. Illustration of the mutation sites in ADAR1 protein.The mutation sites and the relative positions in ADAR1 protein of the 8 originally reported missense mutations are shown. All these mutations were designed to be introduced to mouse genome. The successful sites are indicated by arrows above the protein (blue), other 6 below the protein (black). Figure S2. Sequence alignment of mouse and human ADAR1 protein sequences. Mouse and human ADAR1 protein sequences were compared using Promals3D software ( http://prodata.swmed.edu/promals3d/promals3d.php ). The K948 in mouse is equivalent to K999 in human, highlighted in red. Sequences of mouse ADAR1, Q99MU3 and human, P55265 in UniPro protein database were used for alignment. Figure S3. Human and mouse ADAR1 structure comparison. 3A, the highly conserved protein structure between mouse and human in the catalytic domain flanking the K999N mutation site is shown by the protein modeling, showing a RMSD of 1.06 Å in the protein main chains, supportive of similar structure. 3B, Protein structure analysis predicted a salt bridge (green lines) between K948 (K999 in human) and E964 (E1004 in human), which stabilize the loop in which K948 and E964 localize in the catalytic domain. Figure S4. The RNA editing activities of the K999N mutant protein in mouse brain. 4A, C and E, RNAs from whole brain tissues were amplified by RT-PCR for GRIA2, GRIA3, GRIK1 mRNAs. and the PCR products were subject to Sanger sequencing analysis. The representative chromatograms for the sequences flanking the edited adenosines at +60 site and R/G site in GirA2 mRNA, R/G site in GirA2 mRNA, and Q/R site in Girk1 mRNA are shown. 4B, D and F, The RNA editing activities as measured by the relative ratio of the G peak (edited) over the total A (unedited) and G peak areas at each editing site of the ADAR1 K999N mutant (n=5) vs wild type (n=4) are compared, and no statistical difference was observed as analyzed using Wilcoxon rank-sum test. Figure S5. RNA editing at the ADAR2-targeted GRIA-2 Q/R site. The efficiency of RNA editing on the mRNA coding GRIA-2 subunit at the Q/R site was assayed. Total RNAs were isolated from whole brain tissues of wt and ADAR1K999N mice, then RT-PCR was performed, followed by Sanger sequencing. The chromophotographs of Sanger sequencing flanking the Q/R site, known edited by ADAR2, are shown. The genomic coded adenosine was almost completely converted to inosine, which is read as guanosine. The editing site is pointed by the arrow. Figure S6. Luminex assays of the mouse brain protein extracts. Protein extracts from wild type and ADAR1 K999N mice were subject to Luminex assays for cytokine and chemokine levels. The panel profiles 45 cytokines/chemokines. Increases in measured concentration of each cytokine in the ADAR1K999N mice were calculated in terms of fold changes relative to wild type mice. n=3 (wild type), n=6 (K999N), * P<0.05 with statistical analyzed using Wilcoxon rank-sum test. Figure S7. ISH for ISG-15 on mouse brain sections. In situ hybridization for ISG-15 transcripts in paraffin sections from wildtype (A-C) and mutant (D-F) mouse brains. A&D, Low power coronal section of whole mouse brain demonstrates no staining in wildtype (A) and multiple foci of hybridization (red staining) in mutant brain (D). B&E, Medium power of periventricular region demonstrates no staining in wildtype (B) and abundant staining in ependyma and parenchymal regions in mutant brain (E). C&F, Medium power of neocortex demonstrates no staining in wildtype (C) and abundant staining in ependyma and parenchymal regions in mutant brain (F). Bar = 1 mm in A&D; =100 microns in B, C, E &F. Figure S8. ISH for CXCL10 on mouse brain sections. In situ hybridization for CXCL10 transcripts in paraffin sections from wildtype (A-C) and mutant (D-F) mouse brains. A&D, Low power coronal section of whole mouse brain demonstrates no staining in wildtype (A) and multiple foci of hybridization (red staining) in mutant brain (D). B&E, Medium power of periventricular region demonstrates no staining in wildtype (B) and abundant staining in ependyma and parenchymal regions in mutant brain (E). C&F, Medium power of neocortex demonstrates no staining in wildtype (C) and abundant staining in ependyma and parenchymal regions in mutant brain (F). Bar = 1 mm in A&D; =100 microns in B, C, E &F. Figure S9. Immunohistochemically stained for GFAP and IBA1. FFPE sections of hippocampus of wildtype (A&C) and ADAR1K999N (B&D) mice immunohistochemically stained for GFAP (A&B) and IBA1 (C&D) show no difference in astrocytosis or microgliosis. Counterstained with hematoxylin. Bar = 200 microns. Figure S10. H-E, Luxol Fast Blue (LFB) and Von Kossa staining on brain sections. FFPE sections of mutant (A&B) mice and wildtype (C&D) mice following Von Kossa (A&C) and LFB (B&D) staining. There is no evidence of calcification on Von Kossa staining, nor evidence of demyelination on LFB staining. Bar = 1mm. Figure S11. Original image of Western Blot for Fig 1 panel C. Western blot was stained with ADAR1 and beta-Actin antibodies on parallel blots with same quantity of protein loading.