Evidence for a tetranuclear iron-sulfur center in glutamine phosphoribosylpyrophosphate amidotransferase from Bacillus subtilis.

Glutamine phosphoribosylpyrophosphate amidotransferase from Bacillus subtilis is an unusual enzyme because it contains an essential iron-sulfur center but does not catalyze an obvious oxidation-reduction reaction. In this communication, results of revised sulfide analyses, iron-sulfur cluster displacement studies, and Mössbauer spectroscopy are presented that lead to the conclusion that the native enzyme contains a tetranuclear [4Fe-4S] center in a diamagnetic state.

* This research was supported by National Science Foundation United States Public Health Service Grant GM 16406 (P. D.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section I734 solely to indicate this fact.
11 To whom correspondence should be addressed. HCI, 10 mg of CaC12, 5 mg of MnCL, and 0.2 mg of "Fe (prepared by dissolving "Fe metal, >90% "Fe from Oak Ridge National Laboratories, in concentrated HCI). The pH of the medium was maintained at 6.5 to 6.8 by addition of KOH during growth.
Iron content was determined by complexation with 2,4,6-tripyridyls-triazine using a modification (6) of the procedure of Fischer and Price (7) and by measurements with a Perkin-Elmer 303 atomic absorption spectrophotometer employing Fe(NO:,).j.9H20 in 0.2 N HNOJ as a standard. Inorganic sulfide was measured using the method of King and Morris (8) with the following modifications. Assays were performed in 1 . 5 -4 self-capping polypropylene centrifuge tubes with a final fluid volume of 1.0 ml. Incubation of protein samples in the alkaline zinc reagent was increased to 2 h, during which time the stoppered tubes were blended on a Vortex mixer for 5 to 10 s every 30 min. Sulfide content was corrected for S" released from cysteinyl residues as determined from identical treatment of amidotransferase apoprotein (1). Na2S was used as a standard as previously described (1). The absorbance yield at 670 nm was 32.3 cm" pmol" (1.0 ml final volume), which is about 10% higher than most previously reported values (8). Amidotransferase concentrations were determined from an extinction coefficient at 278 nm based on dry weight (1); a molecular weight for the subunit of 50,000 was used.
The methods used for diplacement (extrusion) of Fe-S clusters have been described in detail (4, 9, IO). Amidotransferase was examined using both (A) the simple displacement method (9) with thiophenol, and (B) the double displacement reaction (4, lo), in which the Fe-S chromophores are initially removed with o-xyl(SH)?,' followed by ligand substitution with thiophenol. In both cases, quantitation was by spectrophotometric methods (4, S), using a Cary 17 spectrophotometer. Samples of amidotransferase (7. The Mossbauer samples were immersed in liquid nitrogen for freezing and storage. They were transferred for measurement to a variable temperature cryostat (Janis Research Corp.) or to a 4.2 K cryostat equipped with a superconducting magnet. The Mossbauer spectrometer was of the constant acceleration type, All Doppler shifts, in particular the chemical shifts, &,, are referred to metallic iron at room temperature.

RESULTS
Iron and Sulfide Analysis-The unconventional values of 3 Fe and 2 S" atoms/amidotransferase obtained in previous studies (1) prompted us to examine these values in several different preparations of the enzyme by a variety of procedures. Table I presents some typical results. The previous iron analyses by chemical and neutron activation analysis have been confirmed with atomic absorption spectroscopy. A range from 2.6 to 3.7 atoms of iron/subunit has been observed in various preparations. The procedure of King and Morris (8) for S" analysis was found to be extremely sensitive to the amount of mixing and volume of air above the sample during the incubations at each step. When these were minimized, as described under "Experimental Procedures," the values in I The abbreviations used are: ~-x y l ( S H )~, o-xylyl-a,a'-dithiol; HMPA. hexamethylphosphoramide; HiPIP, high potential iron protein. 3.68 f 0.06 3.12 f 0.09 Table I were obtained. The moles of S"/amidotransferase subunit was as high as 3.3 and was generally, but not always, within 5% of the iron content. Although no preparation of the enzyme was found to contain 4 atoms of Fe and S2per subunit, the finding of equimolar Fe and S" a t values greater than 3 suggests that the native enzyme contains a tetranuclear Fe-S center.

6007
Fe-S Cluster Displacement (Extrusion) Studies-In view of the unexpected finding of -3Fe and 3S2-/enzyme subunit, the nature of the Fe-S prosthetic groups was investigated using the cluster displacement (core extrusion) (4,9,10) technique. Treatment of the protein in denaturing organic solvents with an excess of organic thiol, RSH, results in the release of the Fe-S clusters liganded by organic thiolates, [Fe2S2(SR)4]2-or [Fe4S4(SR)4]2-, or both. The displaced clusters can be identified by comparison of their physical properties to those of the synthetic complexes prepared separately. In the absence of interfering protein chromophores, the displaced Fe-S clusters may be identified and quantitated simply using electronic absorption spectra (4,9).
The optical spectrum of amidotransferase after sequential treatment with o-xyl(SH), and thiophenol in 80% v/v HMPA is compared in Fig. 1 to the spectra of samples of C. pasteurianum and spinach ferredoxins under identical conditions, all at the same total iron concentration. Since C. pasteurianum ferredoxin is known to contain 2 [4Fe-4S] cores and spinach ferredoxin to contain a single [2Fe-2S] core, the results clearly establish that the Fe-S prosthetic group of amidotransferase is extruded as a tetranuclear cluster. Similar spectra were obtained upon treatment of amidotransferase with thiophenol alone in HMPA, but the displacement reaction was much slower, requiring 260 min for completion. Final concentrations of thiophenol above -20 mM, while increasing the rate of displacement, also caused some turbidity, apparently due to precipitation of apoprotein. The optical spectrum of amidotransferase treated with o-xyl(SH)? in 80% v/v HMPA (not shown) was very similar in both peak position and overall shape to that obtained with C. pasteurianum ferredoxin, and bore no similarity to the characteristic spectrum (4)  Mossbauer Spectroscopy-Mossbauer measurements were made on frozen amidotransferase (20.6 mg/ml) that was prepared from bacteria grown in "Fe-enriched medium. The Mossbauer spectrum of native amidotransferase consisted of two quadrupole doublets (Fig. 2 A ) . The major component with 94.5% of the total intensity had a chemical shift These values are very close to the Mossbauer parameters 6Fe = 0.42 mm/s and A E = 1.12 mm/s observed for reduced HiPIP of Chromatium uinosum, but differ from those of any other type of known Fe-S center (11)(12)(13)(14). The minor quadrupole doublet in Fig. 2A with 5.5% of the total intensity is probably an impurity due to partial oxidation of the sample. Deliberate exposure of the enzyme to air produced an increase in intensity of a signal with similar Mossbauer parameters (6). Dithionite treatment did not alter the spectrum of the native enzyme. The chemical shift of the "impurity" was = 0.18 mm/s, the quadrupole splitting A E = 0.86 mm/s. Application of a strong magnetic field (Fig. 2B) yielded no evidence for either internal magnetic fields or any noticeable heterogeneity in the quadrupole interaction. In fact, a satisfactory simulation was obtained assuming a diamagnetic iron complex of net spin zero, quadrupole splitting AE = 1.17 mm/s, and asymmetry parameter 7 = 0.7. EPR experiments between 3 and 77 K c o n f m t h a t less than 3% of all native amidotransferase molecules have a detectable electronic spin (l). 2 In view of the very recent discovery of three-iron centers (13)(14)(15), the evidence for and against such a cluster in amidotransferase must be closely examined. Accepting the facts that (i) the iron sites in native amidotransferase are not distinguishable by Mossbauer spectroscopy, (ii) that they display isomer shifts and quadrupole splittings halfway between the values typical for ferrous and ferric high spin iron in tetrahedral sulfur coordination, and (iii) that they couple to a total spin of zero, 29, = 0, the only plausible spin assignments are S1 = S2 = 5/2, Sa = S 4 = 2 for a tetranuclear, and S1 = S2 = 5/2, S3 = 2 for a trinuclear cluster. While the former is the accepted assignment for reduced HiPIP, the latter corresponds to the reduced state of the three-iron centers that have been studied to date (13,14), which are  paramagnetic, however, and reveal two distinct iron environments in the Mossbauer spectra. Together with the core extrusion experiments, which yield a two-iron complex in the case of the three-iron centers rather than the four-iron complex extruded from amidotransferase, all the evidence is compatible with the hypothesis that native amidotransferase contains a tetranuclear Fe-S cluster of the reduced HiPIP type.

DISCUSSION
The combined weight of evidence presented above clearly indicates that native amidotransferase contains a [4Fe-4S] cluster as its prosthetic group. The iron and sulfide analyses of 3 and 2 atoms/subunit, respectively, reported previously (1) did not agree with this conclusion. The present studies have shown, however, that seemingly minor variations in the sulfide analysis procedure can lead to substantial underestimation of the sulfide content. We believe that present analyses of up to 3.3 sulfides/subunit are more nearly correct, because the procedure used gives both the highest color yield with standard sulfide solution and the highest sulfide content with amidotransferase.
Although the consistently high iron and sulfide content suggested the presence of [4Fe-4SJ clusters in less than stoichiometric amounts, the possibility that amidotransferase might contain two [2Fe-2S] clusters was examined using the cluster extrusion (displacement) method. Initial results with thiophenol in HMPA gave clear evidence for the presence of tetranuclear Fe-S clusters in amidotransferase; the quantity of [4Fe-4SJ centers agreed very well with the measured iron and sulfide contents. Because the reaction with thiophenol was sluggish, we also examined the reaction of amidotransferase with o-xyl(SH)z, a reagent known to stabilize [2Fe-2S] centers markedly (9). Since this also caused rapid r e l e a d of the Fe-S chromophores, this method should minimize the possibility that [2Fe-2S] centers are present in native amidotransferase, but are converted to [4Fe-4S] clusters during removal from the protein. Results using both o-xyl(SH)z and the thiophenolate derivatives resulting from sequential ligand exchange reaction, which are spectroscopically more characteristic, were in excellent agreement with the presence of a [4Fe-4S] center in most of the isolated enzyme molecules.
When the Azotobacter uinelandii ferredoxin, which has been recently reported to contain a novel [3Fe-3S] center (13,15), is treated with o-xyl(SH)p, the center is extruded as a [2Fe-2S] cluster. Furthermore, the clusters extruded from the A. vinelandii ferredoxin with thiophenol show distinctive spectral features (cf. Fig. 12 of Ref. 9) which were not detected in the present case. Thus, the results of our cluster displacement studies, like the Mossbauer data, argue against the presence of a [3Fe-3S] cluster in amidotransferase. Whether the [4Fe-4S] centers are lost from a variable fraction of amidotransferase molecules during isolation or the cells normally produce a small amount of apoprotein molecules is not known. Isolation of amidotransferase from cells grown on limiting amounts of iron, however, also yielded enzyme containing about 3 atoms of iron/subunit.
The Mijssbauer data indicate that native amidotransferase is diamagnetic, in good agreement with preliminary EPR studies. 2 The values of the Miissbauer parameters agree very well with those expected for a [4Fe-4S] cluster in the t 2 (C") oxidation state, with an average iron valence of +2.5 and the electronic spins of the iron atoms coupled and delocalized over four essentially equivalent ir& atoms (12). Ferredoxins and high potential iron proteins in this oxidation state can be readily reduced or oxidized by 1 electron/4 irons to yield centers with characteristic EPR and Mossbauer spectra. Such behavior has been difficult to observe in the cast? of amidotransferase, but preliminary experiments now indicate that it may be possible to carry out limited, reversible oxidation and reduction of the Fe-S cluster of the e n~y m e .~ Characterization of such reactions may allow a more detailed comparison to be made between the [4Fe-4S] cluster of amidotransferase and the more thoroughly studied [4Fe-4S] clusters of other proteins. The iron-sulfur center in amidotransferase is essential for enzymatic activity, because reaction with 0 2 (16) or ophenanthroline (1) causes complete inactivation. Activity can be partially restored by incubation with sulfide, ferrous ion, and thiols. Whether the center participates directly or indirectly in catalysis remains unknown; it seems unlikely, however, that the [4Fe-4S] center is oxidized or reduced during catalysis. The only other enzyme that catalyzes a non-oxidation-reduction reaction and contains iron and acid-labile sulfide, aconitase, has been claimed to contain a [2Fe-2S] center (4), but more recent evidence suggests the presence of a [3Fe-3S] cluster (13). These results suggest that Fe-S centers may have general importance in other than oxidation-reduction reactions.