Human Amidophosphoribosyltransferase AN OXYGEN-SENSITIVE IRON-SULFUR PROTEIN*

Glutamine 5-phosphoribosyl-1-pyrophosphate amidotransferase (EC 2.4.2.14), amidophosphoribosyltransferase, was partially purified from human placenta. Upon exposure to oxygen, both the glutamine and ammonia activities were lost in parallel. Inactivation by oxygen increased as the temperature of incubation rose and the partial pressure of oxygen increased. Molecular oxygen rather than a radical derivative was responsible for inactivation since scavengers of oxygen radicals did not protect against inactivation. AMP,GMP,PP-ribose-P, and inorganic phosphate partially protected both the glutamine and ammonia activities from inactivation by oxygen. Incubation with 1,10-orthophenanthroline, but not 1,7-metaphenanthroline or tiron, led to inactivation of amidophosphoribosyltransferase. Both the 1,10-orthophenanthroline- and oxygen-inactivated enzymes could be reconstituted by incubation with ferrous iron and inorganic sulfide in the presence of dithiothreitol under anaerobic conditions. The iron requirement could not be replaced by zinc, copper, cobalt, nickel, magnesium, or calcium. The sulfide requirement could not be replaced by higher concentrations of dithiothreitol. It is concluded from these studies that human amidophosphoribosyltransferase is an iron-sulfur protein and oxidation of this structure may be responsible for the marked lability of this enzyme in vitro.


Glutamine
Sphosphoribosyl-1-pyrophosphate amidotransferase (EC 2.4.2.14), amidophosphoribosyltransferase, was partially purified from human placenta. Upon exposure to oxygen, both the glutamine and ammonia activities were lost in parallel. Inactivation by oxygen increased as the temperature of incubation rose and the partial pressure of oxygen increased. Molecular oxygen rather than a radical derivative was responsible for inactivation since scavengers of oxygen radicals did not protect against inactivation. AMP, GMP, PP-ribose-P, and inorganic phosphate partially protected both the glutamine and ammonia activities from inactivation by oxygen. Incubation with l,lOorthophenanthroline, but not 1,7-metaphenanthroline or tiron, led to inactivation of amidophosphoribosyltransferase.
Both the l,lO-orthophenanthroline-and oxygen-inactivated enzymes could be reconstituted by incubation with ferrous iron and inorganic sulfide in the presence of dithiothreitol under anaerobic conditions. The iron requirement could not be replaced by zinc, copper, cobalt, nickel, magnesium, or calcium. The sulfide requirement could not be replaced by higher concentrations of dithiothreitol. It is concluded from these studies that human amidophosphoribosyltransferase is an iron-sulfur protein and oxidation of this structure may be responsible for the marked lability of this enzyme in vitro.
Glutamine 5-phosphoribosyl-1-pyrophosphate amidotransferase (EC 2.4.2.14), amidophosphoribosyltransferase, catalyzes the Fist and probable rate-limiting step in the pathway of purine biosynthesis de nouo. Although the kinetic and physical properties of amidophosphoribosyltransferase have been studied in extracts or partially purified preparations from many cells (for reviews, see Refs. 1 to 4), only four laboratories have reported studies with highly purified or homogeneous preparations of the protein (5-8). Marked instability of catalytic activity in vitro may account for the paucity of studies with the purified enzyme.
In 1975, Turnbough and Switzer (9, 10) provided one explanation for the instability of this enzyme. They reported that amidophosphoribosyltransferase from Bacillus subtilis was rapidly inactivated after exposure to oxygen, t1,2 of 13 min, 1 from B. subtilis was an iron-sulfur protein and presented evidence indicating that oxidation of the iron-sulfur center of the enzyme was responsible for inactivation following exposure to oxygen. However, this explanation may not be applicable to all forms of amidophosphoribosyltransferase since the pigeon liver enzyme, which is also unstable in vitro (ll), is reported not to contain sulfide (6). The amidophosphoribosyltransferase isolated from Escherichia coli is probably different from both the pigeon liver and B. subtilis enzymes in that it does not contain iron (8), a metal found in the avian (5, 6) and B. subtilis (7) proteins.
Human amidophosphoribosyltransferase, like that from bacteria and avian liver, is also unstable in vitro and this property of the enzyme has retarded our attempts to purify the protein. Since amidophosphoribosyltransferase from different sources appears to be heterogenous in structure, we have initiated studies to explain the instability of the human enzyme in uitro. In this communication we present findings which indicate that human amidophosphoribosyltransferase is inactivated by molecular oxygen and report data which suggest it, too, is an iron-sulfur protein. with 50 mM Tris-HCl buffer, pH 7.4 at 25"C!, and the peak protein fractions were pooled. This and all subsequent steps were performed at 4°C. Two hundred microliters of enzyme plus ligand were placed in glass tubes (1.3 x 9.5 cm) and the tubes were then capped with an air-tight rubber stopper. Two 21gauge needles were inserted into the rubber stopper and the contents of the tube were gassed with 100% oxygen or 100% nitrogen.
One-half liter of gas was flushed into the tube over a l-min period through one needle and evacuated through a second needle. At the end of the gas exchange, both needles were removed and the sealed tubes were incubated for 0 to 60 min at temperatures ranging from 4 to 45'C. Following this incubation, the tubes were opened and enzyme catalytic activity was determined in room air by the assays described above.
I,lO-Orthophenanthroline Inactivation of Amidophosphoribosyltransferase-The enzyme preparation was incubated with 20 mM l,lO-orthophenanthroline in Buffer B for 15 min at 37°C in room air. Because of the insolubility of l,lO-orthophenanthroline in aqueous buffer, it was necessary to dissolve this material in ethanol; the final concentration of ethanol was 2% during the incubation. Following this incubation, 1 ml of the orthophenanthroline-treated enzyme was passed through a Sephadex G-25 column (I X 21 cm) equilibrated with 50 mM Tris-HCl, pH 7.4 at 25°C. The peak protein fractions were pooled and transferred to the anaerobic chamber. Reconstitution of Amidophosphoribosyltransferase-The enzyme used for these studies was inactivated either by exposure to oxygen for 120 min at 37"C, 1.0 atm by the technique described above (-16% residual activity) or by incubation with l,lO-orthophenanthroline as described above (-17% residual activity). The inactivated enzyme was transferred to the anaerobic chamber for the reconstitution stud&.
The Oxygen Inactivation of Amidophosphoribosyltransferuse--In Tris-HCl buffer, timidophosphoribosyltransferase was rapidly inactivated upon exposure to oxygen; glutamine and ammonia activity were lost in parallel (Fig. 1). There was no loss of glutamine or ammonia activity during incubation for 1 h at 37°C under 100% nitrogen. -The enzyme was passed through a Sephadex G-25 column equilibrated with 50 mM Tris-HCl buffer, pH 7.4, just prior to use and 160 ~1 of enzyme were transferred to a sealed glass tube and gassed with 100% oxygen as described under "Methods." Incubation under oxygen was performed at 37°C for the tim'e indicated on the abscissa. Glutamine (M) and ammonia (M) activities were assayed as described under "Methods." Each assay was performed in duplicate and the mean is reported here.
Inactivation by oxygen was temperature-dependent as shown in Table I. Residual activity for the glutamine and ammonia functions decreased as the incubation temperature increased.
In addition to temperature, the rate of inactivation was dependent upon the partial pressure of oxygen to which the enzyme was exposed (Fig. 2). Conditions were selected for this experiment so that approximately 50% of the catalytic activity was lost during the period of exposure to the highest oxygen concentration (1.0 atm for 25 min at 37°C). Had the incubation been continued for a longer time or performed at a higher temperature, all of the catalytic activity would have been lost. As depicted in Fig. 2, the plot of percentage inactivation versus atmospheres of oxygen was hyperbolic with saturation occurring at approximately 1 atm of oxygen. No apparatus was readily available for performing experiments under hyperbaric conditions so it was not possible to determine experimentally the partial pressure of oxygen which was saturating in this experiment. Concentrations of superoxide dismutase (5 pg/ml), catalase (500 pg/ml), L-histidine (5 mM), glucose (5 mM), and inosine (5 ITIM), which are effective in scavenging 02-, O('Ag), OH-, and HzOz (15), did not significantly prolong the tl,z of amidophosphoribosyltransferase during the exposure to oxygen (Table II). These results suggest that molecular oxygen, rather than an oxygen radical, is responsible for inactivation.

Protection
Against Oxygen Inactivation-A number of Data from these experiments are presented in Tables III and IV. Both the glutamine and ammonia activities were more resistant to oxygen inactivation in the presence of PP-ribose-P, inorganic phosphate, and purine ribonucleotides, allosteric effecters of human amidophosphoribosyltransferase (12,16,17). The protective effect of these ligands was concentration-dependent and the values presented in Table III were obtained with the concentrations that provided maximal protection.
Glutamine and ammonia, alone or in combination, did not protect against oxygen inactivation in the absence of PPribose-P (Table IV). When these substrates were combined with PP-ribose-P, the t1,2 was prolonged, but the protection provided by the combined substrates was no greater than that afforded by PP-ribose-P alone. However, azaserine, a glutamine analogue, did provide protection. Unlike glutamine, azaserine binds irreversibly to amidophosphoribosyltransferase and this interaction between the enzyme and azaserine is enhanced by simultaneous incubation with PP-ribose-P (18, 19). As shown in Table IV, azaserine alone at high concentrations prolonged the half-life of amidophosphoribosyltransferase catalytic activity and this effect was enhanced in the presence of PP-ribose-P.
These experiments were performed with an enzyme preparation that had been passed through a Sephadex G-25 column before incubation under oxygen. Consequently, the effect of azaserine plus PP-ribose-P was the result of azaserine irreversibly bound to the enzyme and not simply the protective effect of PP-ribose-P.
Further proof that azaserine was bound to the enzyme was demonstrated by the failure to detect glutamine activity in the effluent from the G-25 column. Since ammonia activity of human amidophosphoribosyltransferase is not inhibited by azaserine (19), it was possible to demonstrate that this activity was protected from oxygen inactivation by azaserine. In addition to the effect of ligands, substrates, and substrate analogues, dithiothreitol and ,&mercaptoethanol were found to protect amidophosphoribosyltransferase from oxygen inactivation. Dithiothreitol (5 InM) prolonged the half-life approximately 3-fold and 60 mM P-mercaptoethanol prolonged the half-life approximately 1.5-fold. Both the glutamine and ammonia activities were comparably protected.
Those forms of the enzyme demonstrated to contain iron are inhibited by l,lO-orthophenanthroline but not by 1,7-metaphenanthroline (5-7). As shown in Fig. 3, human amidophosphoribosyltransferase catalytic activity was inhibited by l,lOorthophenanthroline but not by 1,7-metaphenanthroline. Al- though not depicted here, l,lO-orthophenanthroline inhibited ammonia as well as glutamine utilization. The inhibitory effect of this metal chelator was not reversed by increasing the Mg2+ concentration from 5 to 10 mM during the assay, nor was it reversed by passing the orthophenanthroline-treated enzyme through a Sephadex G-25 column before assay. Ten millimolar tiron, a chelator of Fe"+, did not inhibit amidophosphoribosyltransferase activity.

Reconstitution of Amidophosphoribosyltransferase
Catalytic Actiuity-Both the oxygen-inactivated and orthophenanthroline-inactivated enzyme could be reconstituted by incubation with iron plus sulfide in the presence of dithiothreitol (Tables V and VI). The glutamine and ammonia activities were both restored, but for simplicity of data presentation only the glutamine results are shown in the tables. Reconstitution studies were performed in an anaerobic chamber since preliminary experiments indicated that the best results were obtained in an oxygen-free environment.
Other studies demonstrated that the highest percentage of reconstitution was obtained with 250 PM ferrous ammonium sulfate, 1.0 mM sodium sulfide, and 10.0 mM dithiothreitol when the incubation was carried out for 20 to 24 h anaerobically at room temperature.
Consequently, all of the data presented here  were obtained with this standard set of conditions. Reconstitution experiments were performed with the orthophenanthroline-inactivated enzyme on six occasions and the results of each experiment are presented in Table V. Incubation with 10 mM dithiothreitol alone resulted in a mean restoration of 12.5 + 8.0% (1 SD.) of the amidophosphoribosyltransferase activity. Increasing the dithiothreitol concentration above 10 mM did not produce a greater percentage of reconstitution.
Addition of sulfide to dithiothreitol increased the percentage of reconstitution slightly, mean of 18.9 f 8.0%. Addition of iron to dithiothreitol increased the percentage of reconstitution to 28.7 f 8.8% (p < 0.001 when compared to dithiothreitol alone). The combination of iron plus sulfide and dithiothreitol produced the greatest percentage of reconstitution, mean of 40.6 + 7.5% (p < 0.001 when compared to dithiothreitol plus iron alone). The requirement for iron could not be replaced by nickel, cobalt, copper, zinc, magnesium, or calcium.
In the absence of dithiothreitol, neither ferrous nor ferric iron produced significant reconstitution. The effect of sulfide could not be attributed to sulfhydryl reduction because all samples contained 10 mM dithiothreitol and higher concentrations of dithiothreitol did not replace the requirement for sulfide. Three experiments were performed with the oxygen-inactivated enzyme and the results of these experiments are presented in Table VI. Dithiothreitol alone produced a mean percentage reconstitution of 7.6 + 2.0%; dithiothreitol plus sulfide, 11.2 f 3.0%; and dithiothreitol plus iron, 8.0 k 1.8%. The combination of iron plus sulfide and dithiothreitol produced the greatest percentage reconstitution, mean of 18.7 f 2.3% (p < 0.01 when compared to dithiothreitol alone). For the oxygen-inactivated enzyme, the percentage of reconstitution was less than that obtained for the orthophenanthrolineinactivated enzyme but the characteristics of reactivation were similar in that maximal reconstitution was obtained with iron plus sulfide in the presence of dithiothreitol.
However, in the case of the orthophenanthroline-inactivated enzyme, there was a significant effect of iron alone on reconstitution of catalytic activity, whereas the oxygen-inactivated enzyme demonstrated only a minimal effect of iron alone on reconstitution.

DISCUSSION
Results presented here demonstrate that human amidophosphoribosyltransferase is inactivated upon exposure to oxygen. This may explain the lability of catalytic activity in vitro and account for some of the difficulty experienced in attempting to purify human amidophosphoribosyltransferase.
Switzer and colleagues have shown that amidophosphoribosyltransferase purified from B. subtilis contains 3 mol of iron and 2 mol of sulfide/m01 of protein, and they have reported that oxidation of the unique iron-sulfur center of this protein is responsible for oxygen inactivation (7). Studies with other iron-sulfur proteins also suggest that oxygen inactivation is related to oxidation of sulfide or Fe", or both, with subsequent disruption of the iron-sulfur center (20, 21). If human amidophosphoribosyltransferase were demonstrated to contain iron and sulfide, this might explain the sensitivity of this enzyme to oxygen inactivation.
Results of the orthophenanthroline experiments indicate that human amidophosphoribosyltransferase contains a heavy metal, and the reconstitution studies suggest that this metal is iron, possibly Fe'+.
Reconstitution of the orthophenanthroline-inactivated and oxygen inactivated enzyme suggests that human amidophosphoribosyltransferase contains sulfide as well as iron. For the oxygen-inactivated enzyme incubation with iron alone or sulfide alone had little effect on catalytic activity, but incubation with the combination of iron plus sulfide restored a significant amount of catalytic activity. These results demonstrate a requirement for both iron and sulfide for reconstitution of catalytic activity. For the orthophenanthroline-inactivated enzyme, incubation with iron alone was partially effective in restoring activity, incubation with sulfide alone had little effect, and incubation with the combination of iron plus sulfide was more effective than iron alone. It is not surprising that iron alone was partially effective in restoring activity to the orthophenanthroline-inactivated enzyme since removal of iron from other proteins with chelating agents does not always lead to a loss of the sulfide. In some cases, sulfide remains in the protein but in a higher state of oxidation (21,22). Under these conditions, dithiothreitol can partially reduce the oxidized sulfur and relieve the absolute requirement for an exogenous source of sulfide (21). Thus, the reconstitution studies presented here demonstrate that iron, probably the ferrous salt, and inorganic sulfide are required for restoration of catalytic activity of human amidophosphoribosyltransferase following oxygen and orthophenanthroline inactivation. These findings suggest this enzyme, like that from B. subtilis, is also an iron-sulfur protein.
A still unresolved question is the role of the iron-sulfur center in the physiological function of amidophosphoribosyltransferase. This enzyme catalyzes two separate but related reactions; glutamine hydrolysis and synthesis of phosphoribosylamine from NH3 produced in the former reaction or from an exogenous source of NH, (19). Other reports have demonstrated that oxygen inactivation or orthophenanthroline in-activation, or both, lead to a loss of the first activity, i.e. glutamine hydrolysis (5-7, 9, 10). In this report, we demonstrate that oxygen inactivation and orthophenanthroline inactivation also lead to a loss of the second activity, i.e. synthesis of phosphoribosylamine from NH:s. Moreover, reconstitution studies demonstrate that incubation with iron and sulfide leads to a restoration of the latter, as well as former, activity of amidophosphoribosyltransferase.
These findings suggest that the iron-sulfur center plays an important role in the utilization of NH3 for phosphoribosylamine synthesis. This conclusion is supported by other data which suggest the iron-sulfur center is near the catalytic site of amidophosphoribosyltransferase.
PP-ribose-P, Pi, azaserine, and purine ribonucleotides, ligands which either bind to the catalytic site or produce a conformational change in the catalytic site, protect the enzyme against oxygen and orthophenanthroline" inactivation.
The potential location of the iron-sulfur center at the catalytic site and its demonstrated importance for NH3 utilization suggest that one function of the iron-sulfur center may be to provide a site for NH3 binding, a mechanism proposed to explain the function of heavy metals in enzymes which utilize NH3 (23).
Another potential function of the iron-sulfur center of amidophosphoribosyltransferase is to act as a sensor of the oxidation-reduction potential in the cell. Ruzicka and Beinert have presented data which suggest that aconitase, a high potential iron-sulfur protein, may function as a sensor of the oxidation-reduction state of the cell and postulate that reversible reduction and oxidation of the iron-sulfur cluster of other enzymes may be a general principle used in nature to regulate the activity of some enzyme systems (24). Reversible oxidation and reduction of the iron-sulfur center of amidophosphoribosyltransferase could potentially lead to changes in the activity of this enzyme, and this might provide yet another mechanism for the control of amidophosphoribosyltransferase activity. In addition, this type of reaction might provide a link between amidophosphoribosyltransferase and other iron-sulfur proteins since all studies reported to date indicate that iron-sulfur proteins function as electron carriers (25-29) or participate in oxidation-reduction reactions (29).