We have recently identified a new member of the Ras/GTPase superfamily termed Rad which has unique sequence features and is overexpressed in the skeletal muscle of humans with type II diabetes (Reynet, C., and Kahn, C. R. (1993) Science, 262, 1441-1444). When expressed in bacteria as a glutathione S-transferase fusion protein, Rad bound [alpha-32P]GTP quickly and saturably. Binding was specific for guanine nucleotides and displayed unique magnesium dependence such that both GTP and GDP binding were optimal at relatively high Mg2+ concentrations (1-10 mM). Rad had low intrinsic GTPase activity which was greatly enhanced by a GTPase-activating protein (GAP) activity present in various tissues and cell lines. Several known GAPs had no stimulatory effect toward Rad. Conversion of Ser to Asn at position 66 in Rad (equivalent to position 12 in Ras) resulted in a total loss of GTP binding. Mutation of Pro61 (equivalent to Gly12 in Ras) or Gln109 (equivalent to Gln61 in Ras) had no effect on Rad GTPase activity, whereas creation of a double mutation at these positions resulted in exceptionally high intrinsic GTPase activity. In vitro, Rad was phosphorylated by the catalytic subunit of cAMP-dependent protein kinase (PK). Phosphopeptide mapping indicated two PKA phosphorylation sites near the COOH terminus. Rad also co-precipitated a serine/threonine kinase activity from extracts of various tissues and cell lines which catalyzed phosphorylation on Rad but was not inhibited by PKA inhibitor. Thus, Rad is a GTP-binding protein and a GTPase which has some structure/function similarities to Ras, but displays unique features. Rad may also be phosphorylated on serine/threonine residues by PKA and other kinases, as well as regulated by its own GAP which is present in many tissues and cell types.

Rad is a low molecular weight GTPase that is overexpressed in skeletal muscle of some patients with type 2 diabetes mellitus and͞or obesity. Overexpression of Rad in adipocytes and muscle cells in culture results in diminished insulin-stimulated glucose uptake. To further elucidate the potential role of Rad in vivo, we have generated transgenic (tg) mice that overexpress Rad in muscle using the muscle creatine kinase (MCK) promoterenhancer. Rad tg mice have a 6-to 12-fold increase in Rad expression in muscle as compared to wild-type littermates. Rad tg mice grow normally and have normal glucose tolerance and insulin sensitivity, but have reduced plasma triglyceride levels. On a high-fat diet, Rad tg mice develop more severe glucose intolerance than the wild-type mice; this is due to increased insulin resistance in muscle, as exemplified by a rightward shift in the dose-response curve for insulin stimulated 2-deoxyglucose uptake. There is also a unexpected further reduction of the plasma triglyceride levels that is associated with increased levels of lipoprotein lipase in the Rad tg mice. These results demonstrate a potential synergistic interaction between increased expression of Rad and high-fat diet in creation of insulin resistance and altered lipid metabolism present in type 2 diabetes. diabetes mellitus ͉ glucose transport ͉ RGK GTPase ͉ transgenic mouse

A 10-allele polymorphism was identified in rad (ras associated with diabetes), a gene that is overexpressed in non-insulin-dependent diabetes mellitus (NIDDM) muscle. The polymorphism, designated RAD1, consists of a variable number of trinucleotide repeats (GTT and ATT) located in the poly(A) region of an intronic Alu sequence. Based on the number of GTT and ATT repetitions, the alleles can be grouped into four classes (I-IV). RAD1 allele frequencies were determined in 210 NIDDM patients and 133 nondiabetic control subjects, all Caucasians. One allele (number 8, class III) accounted for > 80% of the chromosomes in both groups. However, an excess of minor alleles, all belonging to class I, II, or IV, was observed among NIDDM chromosomes (P < 0.025), suggesting a possible association between RAD1 and NIDDM predisposition. To promote further studies to test the hypothesis that genetic variability at the rad locus contributes to NIDDM, we mapped rad on the human genome. Using the fluorescence in situ chromosomal hybridization technique, rad was unequivocally assigned to chromosomal band 16q22. In families that were informative for RAD1, the rad locus was mapped within a 3-cM region defined by the markers D16S265, D16S186, and D16S397 (logarithm of odds scores = 10.08, 10.9, and 10.84 at recombination fractions of 0.024, 0.001, and 0.03, respectively). The high degree of heterozygosity of these markers will allow large-scale family studies to be performed to test the presence of linkage between rad and NIDDM.