Contamination of adenosine deaminase by superoxide dismutase. Stabilization of endothelium-derived relaxing factor.

We have measured cyclic GMP accumulation in co-cultures of bovine aortic endothelial cells and rat smooth muscle cells as an index of endothelium-derived relaxing factor (EDRF) production. Adenosine deaminase (EC 3.5.4.4, Sigma type VI) produced a 5- to 10-fold increase in the basal and bradykinin-stimulated cyclic GMP content of co-cultures but had no effect on smooth muscle cells alone. Cyclic GMP accumulation in response to adenosine deaminase was not blocked by adenosine deaminase inhibitors or affected by adenosine, the products of adenosine deamination (inosine and ammonia), or adenosine receptor antagonists. Since superoxide anion is known to destroy EDRF and nitric oxide (NO) (which is similar or identical to EDRF in composition), we tested for superoxide dismutase (SOD, EC 1.15.1.1) in single lots of eight commercial sources of adenosine deaminase by measuring inhibition of the superoxide-mediated reduction of cytochrome c. SOD activity was found in all sources of adenosine deaminase, but varied widely. One lot of Sigma type VI enzyme contained 0.08 units SOD/unit adenosine deaminase. The EC50 values of purified SOD (0.23 units/mL) and Sigma type VI adenosine deaminase (2.1 units/mL) needed to increase the cyclic GMP content of co-cultures differed by a similar factor, 0.11. Thus, the SOD activity in adenosine deaminase is sufficient to account for its effect on cyclic GMP accumulation. One lot of Boehringer Mannheim adenosine deaminase contained much less SOD contamination (0.006 units SOD/unit adenosine deaminase) and produced much less accumulation of cyclic GMP in co-cultures. Cyclic GMP accumulations in response to adenosine deaminase and SOD were both abolished by the NO synthetase inhibitor NG-monomethyl-L-arginine (0.1 mM), consistent with the idea that these enzymes act by stabilizing EDRF. Adenosine deaminase and the SOD activity contaminating it were found to have similar molecular masses of 33-34 kD as assessed by gel permeation chromatography. When run under reducing conditions to dissociate homodimeric SOD into monomers, a 16.6 kD peptide which co-migrates with purified cupro-zinc SOD was visible in silver-stained sodium dodecyl sulfate-polyacrylamide gels of the Sigma type VI but not the Boehringer Mannheim adenosine deaminase. We conclude that commercial sources of adenosine deaminase are variably contaminated by SOD. Since EDRF is synthesized by many tissues, the use of adenosine deaminase contaminated with SOD may produce numerous effects not attributable to the deamination of adenosine.