Oxidation of NADH catalysed by human xanthine oxidase: generation of superoxide anion.

Xanthine Oxidase is of importance because of its potential to generate superoxide anion and hydrogen peroxide; reactive oxygen species that have been implicated in many instances of ischaemia-reperfusion injury [I]. The proposed pathogenic mechanism (21 is largely based on the well known properties of the bovine milk and rat liver enzymes [3] and depends on the ischaemia-induced proteolytic conversion of the dehydrogenase form of the enzyme (predominant in vivo and preferentially reducing NAD) to the oxidase form which reduces only molecular oxygen. The oxidase form is believed to oxidise hypoxanthine, accumulated during ischaemia, generating reactive oxygen species. We have recently described the characterisation of human xanthine oxidase purified fiom breast milk [4]. This human enzyme dfiers from other known forms in having very low activity to xanthine or hypoxanthine, a fact that poses problems in the context of the above hypothesis. Non-human xanthhe oxidase is known to use NADH as a reducing substrate and, in view of the accumulation of NADH during ischaemia, we have investigated the capacity of both dehydrogenase (HXDH) and oxidase (HXO) forms of human milk xanthine oxidase to oxidise NADH and to generate superoxide. Cream was separated from human milk by centfigation at 4000g for 20 minutes. Xanthine oxidase/dehydrogenase was extracted by resuspension of the cream in 0.2 M KzmOi followed by further centfigation as above. A crude enzyme extract was produced from the supernatant by treatment with butanol, and ammonium sulphate precipitation, which was then resuspended in 25 mM Na'-phosphate buffer pH 7.4 and applied to a column packed with heparin covalently linked to 4% crosslinked beaded agarose (type 1, Sigma). The enzyme was eluted from the column with 0.4 M NaCl in the above buffer and was routinely obtained with a ratio of Azso to of 6.5 to 7.5. This enzyme was used without fiuther purification. In order to obtain the enzyme as pure HXDH, the entire preparation was camed out in the presence of 2.5 mM DTT and 0.1 mM PMSF. Pure HXO was obtained by limited proteolysis of the purified enzyme with trypsin [5]. AU activity assays were camed out in 25 mM Na'phosphate buffer pH 7.4 at 37 f 0.2OC in air-saturated buffer. The rate of NADH oxidation was calculated by using an E~~~ of 340 N ' c m ' . The rate of superoxide production was monitored at 550 nm in the presence of 25 pM cytochrome c and calculated by using an &5SO of21 mM-lcm-'. In the absence of NAD, the rate of oxidation of NADH by HXDH followed Michaelis Menten kinetics with a V,w of 0.33 f 0.006 pmol min'lmg-l and a Kn, of 1.26 k 0.11 pM. Inhibition of this activity by NAD was of a mixed type. In the presence of 5 pM and 10 pM NAD, Michaelis Menten kinetics were followed, with V, = 0.22 f 0.008 and 0.17 f 0.006 pmol min-lmg-' respectively, and Km = 2.73 f 0.44 and 4.41 f 0.74 respectively. Inhibition constants K(comp) = 1.63 pM and K,(uncomp) = I 1.43 p M were obtained from these data. 02, as the oxidising substrate, can accept either one electron resulting in the superoxide anion, or two electrons, resulting in H202. In the absence of NAD, the rate of HXDH catalysed superoxide production fiom NADH oxidation reached 0.23 pmol min-'mg-' at 6 p M NADH. At kgher [NADH], the rate of superoxide production decreased, with a rate of 0.17 pmol min-lmg-' obtained at 100 pM NADH. A plot of superoxide production rates against [NADW log [NADH] gave a symmetrical peak, which is usually indicative of substrate inhibition. In the presence of NAD, the rate of superoxide production showed only small deviations from Michaelis Menten kinetics. Apparent K,,, values of 0.69 pM, 0.75 pM, and 1.0 pM were estimated at 5 pM, 10 pM, and 20 pM NAD, with apparent V, values of 0.14, 0.1, and 0.08 pmol min-'mg'l respectively. From these figures, inhibition constants of K,(comp) = 4.4 pM and K,(uncomp) = 16.6 pM were estimated. The kinetics of NADH oxidation catalysed by HXO deviated significantly from Michaelis Menten kinetics, showing substrate activation at high [NADH]. HXO was found to be less active than HXDH, with approximately 30 % of the latters activity at low [NADH] (< 10 pM) and 48 % at 100 pM NADH. The activity of HXO was also inhibited by NAD, although inhibition constants could not be determined from these data because of similar deviations from Michaelis Menten kinetics. Qualitatively, the effectiveness of NAD as an inhibitor of HXOcatalysed NADH oxidation with O2 is of the same order as for HXDH. The rate of superoxide production catalysed by HXO also deviated from Michaelis Menten kinetics with apparent substrate activation at high [NADH], both in the presence and absence of NAD. Superoxide production reached a rate of 0.10 pmol mi6'mg-I in the presence of 100 pM NADK and 0.038 pmol min'lmg-l at 100 pM NADH and 50 pM NAD. This trend contrasts with the apparent substrate inhibition observed for superoxide production by HXDH. While the kinetics of NADH oxidation catalysed by HXDH and HXO are clearly complex, the human enzyme is capable of generating superoxide anion at significant rates. Moreover, HXDH appears to be more effective than HXO in this respect. These findings suggest an alternative mechanism for ischaemia-induced reperfusion injury, whereby ischaemia-induced elevation of NADH levels leads to generation of superoxide anions on reoxygenation of the tissues. Such a mechanism does not depend on dehydrogenase to oxidase conversion (a controversial issue), and is potentially applicable to aU vertebrate species including human.