Biochemistry Metal and sulfur composition of iron - molybdenum cofactor of nitrogenase ( molybdenum - iron protein / dinitrogen reduction / iron - sulfur clusters )

The sulfur content of N-methylformamide solutions of cofactor from Clostridium pasteurianum nitrogenase has been determined to be 11.9 (+0.9) mol per mol of molybdenum. This number was determined radiochemically, using iron-molybdenum cofactor isolated from molybdenum-iron protein from bacteria grown on 35SO4. A total of 3.2 (±0.2) mol of sulfur per mol of molybdenum was found to be present in cysteine and methionine, probably arising from contaminating proteins not intrinsic to the cofactor. Combined with accumulated evidence that is discussed, these results lead to an updated stoichiometry of MoFe6S8orq, not MoFe6S4 as previously thought, for this cluster. The molybdenum-iron protein of nitrogenase (MoFe protein) is thought to contain the catalytic site for the reduction of dinitrogen to ammonia (1). A cofactor containing both iron and molybdenum (FeMo cofactor) has been isolated from acid-denatured protein (2). This cofactor may be used to activate inactive, molybdenum-free, "MoFe" protein from a mutant strain of Azotobacter vinelandii, UW 45, suggesting an intrinsic role for the FeMo cofactor in the catalytic process. Physical studies of the FeMo cofactor and the MoFe protein have shown the following: (i) the characteristic EPR spectrum ofthe MoFe protein is elicited in a broadened form from the FeMo cofactor (3); (ii) the Mossbauer spectra of iron atoms associated with this S = 3/2 center in the protein similarly are reproduced in studies of the cofactor (3); and (iii) the Mo x-ray absorption fine structure (EXAFS) analyses suggest that the structure ofthe molybdenum site in the cofactor is closely similar to that in the protein (4). This indicates that the structure of the iron and molybdenum cluster in the MoFe protein is relatively well represented in the isolated FeMo cofactor and that structural studies ofthe cofactor may help to solve the more difficult problem of the function of that site in the protein. The structure ofthis species almost certainly is unique among biological metal-sulfur clusters. Common iron-sulfur clusters [Fe2S2(SR)4 and Fe4S4(SR)4] have not been detected as subspecies of the FeMo cofactor (3). Mossbauer and EPR spectroscopies of the MoFe protein (5, 6) show that each of the six irons coupled to the S = 3/2 spin system is in a distinctive magnetic environment. Several clusters containing iron, molybdenum, and sulfur have been synthesized (7-9), but as yet none has duplicated the unusual physical and chemical properties of the FeMo cofactor. Despite the absence of crystallographic information, some appreciation of the structure of this cluster might be gleaned from the body ofspectroscopic data that is rapidly accumulating. However, a prerequisite of the design of models based on such information is an accurate knowledge ofthe stoichiometry ofthe FeMo cofactor. To date, the best information on that score arises from the rationalization of spectroscopic experiments, rather than from chemical analysis. Electron-nuclear double resonance (ENDOR) and EPR data suggest one molybdenum per spin system and two noninteracting spin systems per MoFe protein, implying the presence of one molybdenum in each cluster (10, 11). M6ssbauer and ENDOR studies show at least six iron atoms coupled to the spin system in the protein (5, 6), and chemical analyses of solutions of isolated FeMo cofactor reveal six to eight iron atoms per molybdenum (2-4, 12, 13). Isolation of the FeMo cofactor in the presence of dithionite makes quantitation of the amount of sulfur especially difficult. Initially, the FeMo cofactor was reported to contain six acid-labile sulfides per molybdenum (2), whereas more recent assays have suggested the number is closer to four (13). So small a number of sulfur atoms (less than the number of iron atoms) is incompatible with the composition of synthetic FeS clusters, as well as being contrary to recent Fe EXAFS spectroscopic studies of the FeMo cofactor (14). Furthermore, the only published study that suggests the presence of sulfur in any form but sulfide (15) has been disputed recently (13). Radiochemical analyses for sulfur (using `5S as a tracer) have the advantage of being insensitive to the presence of dithionite. They also are inherently more accurate and more precise than calorimetric techniques. A major problem is that radiochemical assays require some method to determine the specific activity of the 35s in the proteins as isolated. The source of the difficulty is the potential for the presence of unevaluated amounts of contaminating sulfate in the growth medium, sulfate being rather difficult to assay in the presence of the other components of the inedium. In order to facilitate understanding of the structure of the FeMo cofactor, we have reexamined the Mo:Fe:S stoichiometry of this cluster, using atomic absorption spectroscopy to assay for Fe and Mo, and radiochemical analyses for S. To obtain the specific activity ofthe sulfur in the isolated proteins and cofactor we have used Clostridium pasteurianum as the source of the MoFe protein, and we have simultaneously purified to homogeneity the well-characterized eight-iron ferredoxin from that organism (16). This ferredoxin (16 mol of S/8 mol of Fe/mol of protein) serves as an excellent internal standard for the specific activity of35S. MATERIALS AND METHODS Carrier-free H235SO4 was obtained from New England Nuclear. [14C]Toluene was the product of Amersham-Searle. Reagent grade nitric acid was distilled twice prior to use. Oxygen was Abbreviations: MoFe protein, molybdenum-iron protein of nitrogenase; FeMo cofactor, iron-molybdenum cofactor of the MoFe protein; ENDOR, electron-nuclear double resonance; EXAFS, x-ray absorption fine structure. * To whom reprint requests should be addressed. 147 The publication costs ofthis article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U. S. C. §1734 solely to indicate this fact. 148 Biochemistry: Nelson et al removed from carbon monoxide by equilibration ofthe gas with a solution of 5% pyrogallic acid in 20% NaOH, and from other gases by passage through Oxysorb (Messer-Griesheim) and Deoxo (Engelhardt) cartridges. All chemicals used were of reagent grade. C. pasteurianum strain W6 was the gift of Leonard Mortenson. Glassware used during analysis for Mo and Fe was triply acid-washed (17); glassware thus treated and the plastieware used introduced essentially no Fe or Mo contamination according to atomic absorption analysis. Activity assays with the MoFe and Fe proteins were performed as published (12). Protein concentrations were determined by biuret assay (18). Growth of 'S-Enriched C. pasteurianum. The minimal sulfate requirement for these bacteria was estimated by growth of 100-ml cultures on various amounts of sulfate. The medium used for the growth of labeled C. pasteurianum was as previously described (16), except that it contained 0.2 mM sulfate (19). To 14 liters of medium containing 10 mCi (1 Ci = 3.7 X 1010 becquerels) of H2'SO4 was added a 1-liter inoculum of bacteria grown on natural-abundance materials. After growth of the bacteria (OD6m = 1.7) they were harvested in a Sharples centrifuge, yielding 43.6 g of cell paste. This was stored overnight at 186 K. Isolation of 3S-Enriched Proteins. The isolation of the nitrogenase proteins was performed anaerobically. All steps were carried out under 1.1 atmosphere of 5% H2 in Ar. All solutions were degassed and contained 2 mM Na2S204 unless it is stated otherwise. Columns were made anaerobic by rinsing with buffer containing dithionite until the effluent was reducing to methyl viologen, and they were loaded by cannulation. Other manipulations were essentially as described (12). Cells were lysed anaerobically under an atmosphere of 20% CO in H2/Ar by addition of 40 mg of lysozyme and 5 mg of DNase I to a suspension in 450 ml 50 mM Tris HCI, pH 8.0. After 1.5 hr at 37°C the debris was pelleted by centrifugation, and the supernatant was loaded onto a DEAE-cellulose DE-32 column (Whatman; 3 X 20 cm) equilibrated with 50 mM Tris HCl, pH 7.1 (at 20°C) and cooled to 4°C. Elution by a gradient of 300 ml each of 0.15 M NaCI/25 mM Tris HCl, pH 7.1, and 0.60 M NaCl/25 mM Tris HCl, pH 7.1 (both equilibrated under 20% CO in 5% H2 in Ar), yielded successively MoFe protein, Fe protein, and the ferredoxin. MoFe protein was purified further, essentially as published (19). After concentration by ultrafiltration (Amicon XM-100A membrane) the protein was purified by chromatography on Sepharose 6B (Pharmacia; 2.5 X 70 cm) with 50 mM Tris HCI, pH 8.0, as elution buffer. The product was loaded onto a DEAEcellulose DE-32 column (3 x 20 cm) equilibrated with 50 mM Tris HCI, pH 8.0, and eluted with a gradient of 500 ml each of 50 mM Tris HCl, pH 8.0, and 0.3 M NaCl/50 mM Tris HCl, pH 8.0. After concentration to approximately 35 mg/ml by ultrafiltration, this protein was frozen and stored under liquid nitrogen. Fe protein also was purified further, essentially as published (19). After dilution with an equal volume of 50 mM Tris HCl, pH 8.0, it was loaded onto a DEAE-cellulose DE-32 column (2 x 8cm) equilibrated with that buffer. After elution in a small volume of 0.5 M NaCl/50 mM Tris HCl, pH 8.0, the protein was purified by chromatography on a Sephacryl S-200 column (Pharmacia; 2.5 x 70 cm) eluted with 0.5 M NaCl/50 mM Tris HCl, pH 8.0. After concentration by ultrafiltration (Amicon PM-30 membrane), the protein was diluted with an equal volume of 50 mM Tris HCl, pH 8.0, and loaded onto a DEAE-cellulose DE-32 column (3 x 20 cm) equilibrated with the same buffer. Elution was accomplished by a gradient of 350 ml each of 0.15 M NaCl/50 mM TrisHCl, pH 8.0, and 0.4 M NaCl/ 50 mM TrisHCl, pH 8.0. The product was concentrated by ultrafiltration, frozen, and stored under liquid nitrogen. The ferredoxin was concentrated by ultrafiltration (Amicon UM-2 membrane) and desalted by chromatography on Bio-Gel P-2 (Bio-Rad) with 50 mM Tris HCI, pH 8.0 (no dithionite), in an inert atmosphere box. The dithionite-free protein was oxidized by exposure to air. Further manipulations of this protein were performed aerobically in the absence of dithionite, as described by Rabinowitz (16). The protein was loaded onto a DEAE-cellulose DE-32 column (3 X 20 cm) equilibrated with 50mM Tris HCl, pH 8.0. After a rinse of 500 ml of 0. 15M NaCl/ 50 mM Tris HCl, pH 8.0, elution was accomplished by a gradient of 1 liter each of 0.15 M NaCl/50 mM Tris HCl, pH 8.0, and 0.35 M NaCl/50mMTris HC1, pH 8.0. Fractions withA390/ Am 2 0.75 were combined, concentrated by ultrafiltration (Amicon UM-2 membrane), and further purified by chromatography on a Sephadex G-75 column (Pharmacia; 2.5 x 90 cm) eluted with 50 mM Tris HC1, pH 8.0. Samples with A390/Am 0.77 were combined, concentrated by ultrafiltration, and purified by ammonium sulfate fractionation. The initial cut between 65% and 85% saturated ammonium sulfate was purified further by repeated crystallization in 80% saturated ammonium sulfate. Isolation of FeMo Cofactor from MoFe Protein. FeMo cofactor was isolated by a procedure similar to that of Shah and Brill (2). All solutions contained 5 mM sodium dithionite added as a 1 M solution in 1 M Tris HCl, pH 8.0. A 0.5-ml aliquot of 3S-labeled MoFe protein (29 mg/ml) was diluted first with 0.5 ml of 0.25 M NaCl/0.025 M Tris HCl, pH 7.4, and then with 2 ml of water. Addition of 100 ,ul of 1 M citric acid was followed after 150s by 400 ,ul of0.5 M Na2HPO4. At this point the protein precipitated. The suspension was allowed to stand at ice temperature for 1 hr. after which the protein was separated by centrifugation at 120 X g for 10 min. The supernatant was removed and replaced with 1 ml of dimethylformamide (Aldrich, vacuum distilled in the absence of grease), which was layered onto the protein. The sample was resuspended by brief (5-s) agitation on a Vortex mixer, the protein was separated by centrifugation at 520 X g for 5 min, and the supernatant was removed. This dimethylformamide wash was repeated. No color was detected in these washes. After the second dimethylformamide wash, 0.5 ml of N-methylformamide (Aldrich, vacuum distilled in the absence of grease) containing 5 mM Na2HPO4 (added as a 0.5 M solution in water) was added. The sample was agitated for 5 min on a Vortex mixer and allowed to stand on ice approximately 10 min. The protein was separated by centrifugation at 1,450 X g for 10 min, and the dark-green clear cofactor solution was removed and stored in a double-septum vial. A second wash with N-methylformamide/phosphate yielded a light-green solution, and a third was colorless. The first two washes were combined and concentrated in vacuo to approximately 0.2 ml. After filtration, samples were taken for analysis. Assay of FeMo Cofactor Activity. These were performed in a manner similar to that previously described (12). Crude extract of A. vinelandii UW45 was provided by Paul Lindahl and stored frozen in liquid nitrogen. The concentrated FeMo cofactor solution was diluted 1:10 into 0.025 M sodium phosphate, pH 7.5, containing 20 mM sodium dithionite. Aliquots of this solution (2-20 ,ul) were added to 300-,u1 samples of the crude extract, incubated at 30°C for 90 min, and then placed on ice. Of these samples, 100 ul was taken with a 20-fold excess of the iron protein from A. vinelandii for assay of nitrogenase activity. Specific activity was taken from the slope of the line fit to a plot of activity vs. molybdenum concentration. Assay of 'S. The amount of 'S in samples was determined by liquid scintillation counting. A solution of 0.05% 1,4-bis[2(5-phenyloxazolyl)]benzene, 0.65% 2,5-diphenyloxazole, 33% Triton X-100 in toluene was used as scintillation fluid. A small amount of the sample was added to 70 A.l of 0.3 M KOH in a Proc. Natl. Acad. Sci. USA 80 (1983) Proc. Natl. Acad. Sci. USA 80 (1983) 149 scintillation vial and was followed by the scintillation fluid. Methanol (200 pl), dioxane, or both were used to clear turbid solutions. Efficiency was assessed by adding a known amount of ['4C]toluene and recounting. Disintegrations per minute were corrected for decay (t112 = 87.2 days) and all were standardized to the same date. Assay ofFe and Mo. These elements were assayed by atomic absorption spectroscopy with a Perkin-Elmer model 2380 spectrometer, equipped with a graphite furnace, at 248.3 nm (Fe) and 313 nm (Mo). Samples were digested by incubation overnight at 1000C in 0.5 M HNO3 and were evaporated to dryness under a flow of N2. Samples were dissolved in 1.00 ml of either 0.5 or 0.1 M HNO3. For assay ofMo, enough NH4Cl was added to yield a2% solution in order to suppress interference from Fe. The amount of Fe and Mo in these samples was determined by the method of standard additions. Lines were fit to the data by least-squares analysis, and standard deviations were assigned to the slope and intercept. Standard deviations were propagated throughout the analysis by standard techniques (20). Analysis of Amino-Acid-Bound 35S. Aliquots of both the labeled FeMo cofactor and the ferredoxin were treated with performic acid (21), Iyophilized, and hydrolyzed by heating for 24 hr with 6 M HC1 at 1200C in sealed containers. The solutions were Iyophilized and the residues were redissolved in 100 Al of 0.01 M HCI. Aliquots (10 Al each) of each sample were subjected to thin-layer chromatography on silica gel (1BWF, Baker) using 1-propanol/concentrated ammonium hydroxide, 70:30 (vol/vol), as eluting solvent. Authentic cysteic acid and methionine sulfone were used as standards. The radioactivity was quantified by two methods. One procedure involved scraping off the ninhydrin spots corresponding to the standards and directly measuring the amount of 'S in each spot by scintillation counting. For the second method, the radioactivity was first visualized by fluorography (22). The intensities of the spots were quantitated by comparison of the optical density of the darkened portions of the film with spots produced by known quantities of [3S]sulfate and [3S]methionine (New England Nuclear).