On the Mechanism of Inactivation of Xanthine Oxidase by Cyanide*

oxidase When the cyanide is partial reduction of the equivalent to approximately a Z-electron uptake per eq of thiocyanate Cyanide-inactivated enzyme largely reactivation reincorporation

The inactivation of xanthine oxidase by cyanide is accompanied by the extraction of sulfur from the protein which is eliminated as thiocyanate. When the cyanide inactivation is carried out under anaerobic conditions, partial reduction of the enzyme occurs, equivalent to approximately a Z-electron uptake per eq of thiocyanate released. Cyanide-inactivated enzyme can be largely reactivated by incubation with NaS. Experiments with 3%labeled NaS reveal that the reactivation is accompanied by the reincorporation of sulfur into the protein. Treatment of such 35S-labeled enzyme with cyanide results in inactivation again and the elimination of YS-labeled thiocyanate.
The apparently irreversible inactivation of xanthine oxidase by incubation with cyanide was first discovered by Dixon and Keilin (1) and has been the focus of considerable attention since. In 1958, Fridovich and Handler (2) reported that incubation with l*C-labeled NaCN resulted in the tight binding of i*CN to the inactivated enzyme with a stoichiometry of 1 molecular eq of l*CN bound per eq of enzyme flavin. They also reported that dithionite prevent'ed both inactivation and fixation of i*CN. More recently, Rajagopalan and Handler (3) and Coughlan, Rajagopalan, and Handler (4) have reported similar results with xanthine oxidase and two other molybdenum-containing ironsulfur flavoproteins, aldehyde oxidase and xanthine dehydrogenase. These workers concluded that cyanide inactivation is due to cyanide forming a complex with the enzyme-bound molybdenum, thus preventing substrate binding and electron transfer through the sequence of molybdenum, flavin and iron-sulfur chromophores.
The existence of a molybdenum-cyanide complex is inconsistent with the failure to observe an isotope effect on the molybdenum electron paramagnetic resonance signals of I'%-and 13C-cyanide-inactivated xanthine oxidase (5). This paper describes further properties of the cyanide inactivation of xanthine oxidase and a method for reactivation of the cyanideinactivated enzyme.
* Supported by Research Grant GAI-11106 from the United States Public Health Service.

AND MATERIALS
Milk xanthine oxidase was prepared as described previously (6). Xanthine-oxygen reductase activity was measured spectrophotometrically at 295 rnp and at a temperature of 25" (6). The catalytic activity is expressed as AFRz5" values (7). This value is obtained by dividing the change in absorbance per min at 295 rnp by the absorbance at 450 rnp of the xanthine oxidase used in the assay.
Kl*CN and NaZ3% were obtained from Amersham/Searle (Arlington Heights, Illinois) and had respective specific activities of 1.01 and 4.5 mCi per mmole.

AND DISCUSSION
Our interest in the mechanism of cyanide inactivation arose from the observation that when the inactivation is carried out under anaerobic conditions, substantial reduction of the flavin and iron-sulfur chromophores is obtained.
This phenomenon is illustrated in Fig. 1. The time course of the reduction is found to be very similar to that of the aerobic reaction as monitored either by difference spectral changes or by enzyme activity.
On admitting air to the anaerobic cuvette there is an immediate reoxidation to the spectrum characteristic of cyanide-inactivated enzyme.
The inset of Fig. 1 shows the difference spectrum between cyanide-inactivated enzyme before and after reoxidation (Curve A) and that between untreated enzyme and cyanide-inactivated enzyme after reoxidation (Curve U). It can be seen that the latter difference spectrum is very similar to that reported previously by Coughlan et al. (4). The magnitude of the extinction changes on anaerobic treatment with cyanide indicate considerable reduction.
By reference to the spectra obtained previously by anaerobic titration with dithionite and substrates (6), the extent of reduction obtained can be estimated to be of the order of 1.5 electron eq per eq of enzyme-bound flavin. As the AFRZ5" value of the enzyme used in these experiments was 153, the enzyme contained 730/, of its active sites in a functional form (8,9). If the reaction with cyanide was only with functional enzyme, the observed 1.5 electron reduction obtained on cyanide inactivation would thus correspond to a 2-electron reduction per functional active site.
It is shown in Fig. 2 that the reaction of cyanide giving rise to the characteristic difference spectrum between active and inactive enzyme (cf. Fig. 1 be seen that there is a close correlation between the AFR2b" value a,nd the decrease in extinction at 320 rnp characteristic of cyanide inactivation. From these results it can be concluded that the xanthine oxidase employed by Coughlan et al. (4) contained approximately 60% of its active sites in a functional form. When xanthine oxidase was incubated at 25O with 0.009 M KlaCN the inactivation was found to proceed in a pseudo first order fashion, with a half-time of inactivation of 5.7 min. When the reaction was carried out for 60 min and the products chro-reaction with cyanide under anaerobic conditions; Curve S, immediately after admitting air. The inset shox-s difference spectra.
Curve A, the difference between Curves 3 and 2. The different symbols represent results from two separate experiments. Curve B, the difference between Cttrves 1 and 2.
matographed on a column of Sephadex G-25, it was found to our surprise that the amount of %N fixed to the completely inactivated protein was only 0.28 eq per eq of enzyme flavin, even though the enzyme used initially contained 7356 of its active sites in a functional form. Analysis of the column fractions showed that unreact'ed KXX emerged as a fairly sharp peak centered around 34 ml, while a third peak of radioactivity emerged at a position centered around 52 ml. This latter peak was found to correspond in elution volume to that of ammonium thiocyanate run separately through the column and assayed calorimetrically by its absorbance at 460 mp as ferric thiocyanate Fhen mixed with Siirbo's reagent (10). The same technique was used to identify the third peak of radioactivity as thiocyanate. By comparison of the spectrophotometric and radio assay it was found that the specific activity was 1.03 mole of I"CN per mole of CNS-, and that the amount of thiocynnate libemted corresponded to 0.70 eq per eq of enzyme flavin applied to the column. The reason for the discrepancy between these results and those reported earlier (2,3,4), which claimed the binding of 1 eq of 1%X per eq of flavin, is due apparently to reaction of CN-at sites other than that responsible for inactivation.
For example, incubation with '4Ci? for periods long in excess of that required for inactivation results in the fixation of more 14CN to the protein ( Table I). The fixation reaction may well be due to the well known cyanolysis of disulf?de bridges. (2) that when the cyanide reaction is carried out anaerobically in the presence of dithionite, no inactivation occurs. However, our results do not support their claim that these conditions prevent fixation of WN to the protein.
As shown in Table I, in our hands fisation of more than 1 eq of XX to the reduced enzyme can be obtained without any loss of catalytic activity, presumably through Reaction 1 above. As it was evident from the above results that cyanide inactiration was due to elimination of sulfur from the protein as thiocyanate, it was of obvious interest to determine if the sulfur could be replaced with return of activity. It was found that incubation with NasS would accomplish this under suitable conditions.
The degree and speed of reactivation depends significantly on pH and temperature.
Maximal reactivation was obtained with good reproducibility by incubation with 0.01 nr Na2S at 45", pH 9.0. Fig. 3 shows the results of two such reactivation INACTIVATED ENZYME SCIIIZ,U 1. Proposed mechanism of inactivation of xanthine oxidase by cyanide.
The bar is taken to represent the protein molecule, the symbols (Fe/S),, and (Fe/S),,d, the oxidized and reduced forms of iron-sulfur chromophores, Fl and FlH*, the oxidized and semiquinoid forms of flavin adenine dinucleotide.
latter experiment was found at the end of 4 hours of incubation to have incorporated 3 moles of 35S per eq of enzyme flavin. When this labeled enzyme was subjected to ammonium sulfate precipitation and dialysis, it was found to have retained 2.76 moles of 3% per eq of enzyme flavin.
Further treatment with unlabeled KCN resulted in the usual cyanide inactivation and the loss of 3% as CN3%- (Table II).
It is evident from these results that in addition to the reincorporation of sulfur to the active site, extraneous sulfur had been introduced into the protein, presumably by formation of persulfide by reaction with disulfides.  (11) and more recently with protein by Cavallini, Frederici, and Barboni (12). The results presented suggest very strongly that the active site of xanthine oxidase contains a persulfide grouping that reacts readily with cyanide to be liberated as thiocyanate.
Protein-S-S-+ CN--+ protein-S-+ CNS- That the thiocyanate sulfur does not originate from the acidlabile sulfur associated with the nonheme iron chromophore is substantiated by little difference in spectral properties in the visible absorption spectrum upon cyanide inactivation and by identical sulfide analysis for both native and cyanide-inactivated enzymes. If another sulfhydryl group were in the vicinity of the persulfide, an attractive possibility to account for the 2electron reduction accompanying the cyanide inactivation would be available through the formation of a new disulfide bond. Scheme 1 presents this hypothesis in diagrammatic form. In addition to explaining the elimination of thiocyanate accompanying cyanide inactivation, this scheme would account for the reduction of the enzyme found when the cyanide inactivation is carried out anaerobically, as well as for the reformation of the active center persulfide on incubation with Na&. In addition it could account for other facets of this enzyme.
Arsenite has been shown to be a powerful reversible inhibitor of the oxidized native enzyme and to interfere with cyanide inactivation (4). This could be due to arsenite forming a complex with the persulfide and the neighboring sulfhydryl group. Finally, the possibility exists that the presence of nonfunctional active sites in all known preparations of xanthine oxidase (8,9) is due to destruction (during preparation or storage) of the persulfide. The nature of the cyanide reaction seems more complex than a simple nucleophilic displacement as other nucleophiles such as sulfite and hypotaurine (a sulfinate which readily attacks persulfide (12)) do not mimic the cyanide reaction. This is indicative that a group(s) on the enzyme (possibly molybdenum) stabilizes the postulated persulfide as well as mediating its attack by cyanide.
Work is now in progress to elucidate these points.