Molecular Identification of Diabetogenic Viral Gene

The best evidence that viruses have a causative role in the pathogenesis of insulin-dependent diabetes mellitus comes from experiments in mice infected with encephalomyocarditis (EMC) virus. When SJL/J male mice were inoculated with a highly diabetogenic EMC-D virus, diabetes developed in 95% of the animals. In contrast, none of the mice inoculated with a nondiabetogenic EMC-B virus became diabetic. Tissue culture experiments showed that EMC-B induces considerable amounts of interferon, whereas EMC-D does not. Despite these differences, EMC-D and EMC-B could not be distinguished antigenically by a sensitive plaque-neutralization assay. Furthermore, the buoyant density in CsCI density gradients and the capsid proteins of these two variants on polyacrylamide gels could not be distinguished. Molecular-hybridization studies with radiolabeled DNA complementary to EMC-D and EMC-B RNAs failed to distinguish them. Determination of complete nucleotide sequences of EMC-D and EMC-B revealed that EMC-D (7829 bases) differs from EMC-B (7825 bases) by only 14 nucleotides. The differences consist of two deletions of five nucleotides, one base insertion, and eight point mutations. The first deletion of three nucleotides and the second deletion of two nucleotides are located in the 5'-poly(C) tract and the 3'-end polyadenylation site, respectively. One base insertion in EMC-B occurs in the 5'-noncoding region. The eight point mutations are located in the polyprotein-coding region. Two of them are silent, whereas the other six mutations, one located on the L gene and five on the VP1 gene, introduce amino acid changes. Our findings indicate that a maximum of 14 of 7829 genomic nucleotides are critical in determining the diabetogenicity of EMC virus.