Fig 2: Expression and stability of NIT2 C146 mutants.(A) Western blot analysis (using an anti-His-tag antibody) for the His-tagged mutants of NIT2 C146 expressed in HEK 293 cells. (B) Aggregation points of NIT2 and NIT2 C146A as determined by thermal shift assay. n = 3; **P < 0.01, Welchs’ correction. (C) Melting curve of purified NIT2 and NIT2 C146A protein. Source data are available online for this figure.

Fig 3: Generation of Nit2 knockout mice.(A) LoxP sites flanking exon 2 of the NIT2 gene were inserted by CRISPR/cas9. (B) Nit2flox/flox mice were crossed with Tg(Cdh5-Cre-ERT2)1Rha to generate endothelial-specific, tamoxifen-inducible knockout mice of Nit2. (C) Ex vivo endothelial cell outsprouting from aortic segments isolated from CTL and ecNit2−/− mice with quantification, normalized to CTL –VEGF (D). *P < 0.05, Mann–Whitney test (-VEGF versus +VEGF). (E) Validation of NIT2 deletion by Western blot in aortic endothelial cells as well as other tissue of a global, constitutive knockout mouse of Nit2 (Nit2wt/wt and Nit2ko/ko, generated by CRISPR/cas9 deletion of ~500 bp. Nit2wt/wt and Nit2ko/ko are littermates in a heterozygous breeding of Nit2wt/ko with Nit2wt/ko mice.). (F) Retina angiogenesis in neonatal mice with quantification (G). *P < 0.05 Nit2wt/wt as compared to Nit2ko/ko, Mann–Whitney test. SMC aortic smooth muscle cells. Source data are available online for this figure.

Fig 4: Endothelial knockout of NIT2 impairs angiogenesis in mice.Generation of a tamoxifen-inducible, endothelial cell-specific knockout mouse of Nit2 (ecNit2−/−). (A) Validation of the knockout efficiency by Western blotting of aortic endothelial cells. (B) LC-MS/MS for αKGM and αKG in plasma of CTL and ecNit2−/− mice. *P < 0.05 as compared to CTL, Mann–Whitney test. (C, D) Untargeted metabolomics from CTL and ecNit2−/− mice from lung tissue (C) and plasma (D). (E) Retina angiogenesis in neonatal mice (P6), isolectin and EdU (5-ethynyl-2’-deoxyuridine) staining from CTL and ecNit2−/− mice. (F) Representative images of retinas from CTL and ecNit2−/− mice stained with Isolectin. (G) In all, ×40 magnification images of vessels between retinal arteries and veins. (H–K) quantification of vascular parameters as indicated. *P < 0.05, Mann–Whitney test. (L, M) Choroidal neovascularization upon a laser-induced injury in adult mice. Quantification of isolectin staining. *P < 0.05, Mann–Whitney test. Source data are available online for this figure.

Fig 5: A human single-nucleotide variant (SNV) decreases NIT2 expression and elevates plasma levels of αKGM.(A) Co-localization analyses and regional association plots of αKGM mQTL (CLSA metabolite, n = 8203) with cis-eQTL for NIT2 in artery tibial (n = 475) and aortic tissue (n = 329) (GTEx v8 study). The sentinel variants rs3830303 and rs277627 are indicated. (B) Strategy for CRISPR/cas9 deletion of the regulatory element containing rs277627 in HEK 293 cells. (C) Genomic PCR showing a 437 bp deletion of the rs277627 containing locus in one single clone (after clonal expansion of HEK 293). NIT2 expression in NTC and rs27767−/− HEK 293 cells by RT-qPCR. P = 0.0286, Mann–Whitney test. (D) and Western blot (E). (F) RT-qPCR of NIT2 after CRISPR/Cas9-mediated homology-directed repair to generate the NIT2 intron containing SNP rs277627 (G → A). HUVEC CTL were electroporated with Cas9-GFP only. *P = 0.003 paired t test. Expression is normalized to beta-actin. TSS transcription start site. RT-qPCR real-time polymerase chain reaction, FP forward primer, RP reverse primer, NTC non-targeting control. Source data are available online for this figure.