Through forward genetic screening of ENU-mutagenized mice combined with automated mapping, we identified 5 Ampd3 mutant alleles associated with reduced naive T-cell populations in peripheral blood (Figure 1A). The nature of Ampd3 point mutations in all mutants was revealed by exome sequencing (Figure 1B), and all 5 mutations affected amino acid residues in the AMPD catalytic domain. All mutations are predicted to be “probably damaging” by Polymorphism Phenotyping v228 scores (Figure 1B). To evaluate how these mutations affect protein stability, we cloned the wild-type Ampd3 coding sequence, as well as these Ampd3 mutant isoforms, into a FLAG-tagged vector to test the protein expression in HEK293 cells. Expressions of the resulting FLAG-tagged proteins show that the gdg and pao mutations greatly affect AMPD3 protein stability, but the csn, tos, and cme mutations do not (Figure 1C). A 3-dimensional structure of mouse AMPD3 generated based on homology predicts that W449 and I470 are located in a hydrophobic area inside the protein that is close to the active site. They interact with each other through hydrophobic interactions. Mutating these residues to charged residues could affect the stability of the hydrophobic center and, thus, the protein stability. T335, S555, and K644 are on or near the surface of the protein, and mutations in these residues could affect oligomerization of the protein or other protein-protein interactions (supplemental Figure 1).
To validate causation, clustered regularly interspaced short palindromic repeats/Cas9-mediated gene targeting was used to generate Ampd3-knockout mice. The resulting Ampd3−/− allele bears a frame-shifting 7-bp deletion in exon 5 of Ampd3, which is expected to generate a truncated form of AMPD3 completely lacking the AMPD domain. The ablation of full-length protein in Ampd3-knockout mice was confirmed by western blot (supplemental Figure 2). Ampd3−/− mice are viable and fertile, with no obvious changes in the size or shape of the lymphoid organs. Examination of the peripheral blood cells with flow cytometry showed that Ampd3−/− mice exhibited a marked reduction in CD62LhiCD44lo naive CD4+ T cells and naive CD8+ T cells and subtle, but statistically significant, changes in CD3+ T cells, CD4+ T cells, CD8+ T cells, and CD4+ and CD8+ memory T cells when large groups of animals were analyzed (Figure 2A,C). These data confirm that the Ampd3 loss-of-function mutations caused the reduction of naive CD4+ and CD8+ T-cell populations in the peripheral blood, as identified in the forward genetic screen (Ampd3gdg, Ampd3csn, Ampd3pao, Ampd3tos, and Ampd3cme). Ampd3−/− CD4+ and CD8+ T cells express lower levels of the surface marker CD62L, as detected by lower CD62L mean fluorescence intensity (MFI). However, expression of the T-cell activation marker CD44 is not affected, as evidenced by comparable CD44 MFI of wild-type and Ampd3−/− CD4+ or CD8+ cells (Figure 2B), suggesting that the reduction in naive T-cell populations in the peripheral blood in Ampd3-deficient mice is likely not due to T-cell activation.