Isolation and Characterization of an Acetylene-resistant Nitrogenase*

A genetic strategy was developed for the isolation of a mutant strain of Azotobacter vinelandii that exhibits in vivo nitrogenase activity resistant to inhibition by acetylene. Examination of the kinetic features of the altered nitrogenase MoFe protein produced by this strain, which has serine substituted for the a -subunit Gly 69 residue, is consistent with other studies that indicate the MoFe protein normally contains at least two acetylene binding/reduction sites. The first of these is a high affinity site and is the one primarily accessed during typical acetylene reduction assays. Results of the present work indicate that this acetylene binding/reduction site is not directly relevant to the mechanism of nitrogen reduction because it can be eliminated or severely altered without significantly affecting nitrogen reduction. Elimination of this site also results in the manifes-tation of a low affinity acetylene-binding site to which both acetylene and nitrogen are able to bind with approximately the same affinity. In contrast to the normal enzyme, nitrogen and acetylene binding to the altered MoFe protein are mutually competitive. The location of the a -Ser 69 substitution is interpreted to indicate that the 4Fe-4S face of the FeMo cofactor capped by the a -subunit Val 70 residue is the most likely region within FeMo cofactor to which acetylene binds with high affinity.

A genetic strategy was developed for the isolation of a mutant strain of Azotobacter vinelandii that exhibits in vivo nitrogenase activity resistant to inhibition by acetylene. Examination of the kinetic features of the altered nitrogenase MoFe protein produced by this strain, which has serine substituted for the ␣-subunit Gly 69 residue, is consistent with other studies that indicate the MoFe protein normally contains at least two acetylene binding/reduction sites. The first of these is a high affinity site and is the one primarily accessed during typical acetylene reduction assays. Results of the present work indicate that this acetylene binding/reduction site is not directly relevant to the mechanism of nitrogen reduction because it can be eliminated or severely altered without significantly affecting nitrogen reduction. Elimination of this site also results in the manifestation of a low affinity acetylene-binding site to which both acetylene and nitrogen are able to bind with approximately the same affinity. In contrast to the normal enzyme, nitrogen and acetylene binding to the altered MoFe protein are mutually competitive. The location of the ␣-Ser 69 substitution is interpreted to indicate that the 4Fe-4S face of the FeMo cofactor capped by the ␣-subunit Val 70 residue is the most likely region within FeMo cofactor to which acetylene binds with high affinity.
Nitrogenase is composed of two component proteins designated the Fe protein and the MoFe protein. It catalyzes the MgATP-dependent reduction of nitrogen (N 2 ), protons, and a variety of triply bonded substrates, including acetylene (1,2). During catalysis, the Fe protein delivers electrons one at a time to the MoFe protein in a process that requires hydrolysis of two MgATP for each electron transfer event (3). The available evidence indicates that electrons are initially delivered from a [4Fe-4S] cluster located within the Fe protein to an [8Fe-7S] "P cluster" (4) contained within the MoFe protein. Intramolecular electron transfer then occurs from the P cluster to a [7Fe-9S-Mo-homocitrate] cluster called FeMo cofactor (5). Because FeMo cofactor has been identified as providing the substrate reduction site (6), its structure and reactivity has received considerable attention (7)(8)(9)(10). Of particular interest to our laboratories is the contribution of the FeMo cofactor polypeptide environment toward its ability to bind and effect the reduction of various substrates. In previous work we gained insight about the functioning of nitrogenase by characterizing an altered Azotobacter vinelandii MoFe protein for which the ␣-subunit His 195 residue was substituted by glutamine (11). This altered MoFe protein does not significantly reduce nitrogen (11) but is still able to reduce acetylene and does so with essentially unaltered kinetic parameters. Even though nitrogen cannot serve as an effective substrate for the ␣-Gln 195 MoFe protein, it remains a potent inhibitor of both acetylene and proton reduction. The major conclusion from this previous work was that nitrogen and acetylene compete for occupancy of a common or shared site. This interpretation is complicated, however, by other observations. First, in the case of the normal MoFe protein, acetylene is a noncompetitive inhibitor of nitrogen reduction, whereas nitrogen is a competitive inhibitor of acetylene reduction (12)(13)(14). Second, kinetic and spectroscopic evidence obtained using the unaltered MoFe protein (15,16) as well as a MoFe protein altered by amino acid substitution (17) indicates that there are two acetylene-binding sites located within the MoFe protein. In light of these observations we became interested in determining if it is possible to obtain a MoFe protein that is altered in its ability to reduce acetylene but remains capable of normal nitrogen reduction.

EXPERIMENTAL PROCEDURES
Isolation of Acetylene-resistant Strains and DNA Biochemistry-Strains of A. vinelandii used in this study were grown under diazotrophic (nitrogen-fixing) conditions using a modified Burk medium previously described (18). Spontaneous acetylene-resistant mutants were isolated from strain DJ939. This strain has the MoFe protein ␤-subunit tyrosine residue 98 substituted by histidine. The construction of this strain and the biochemical characterization of the altered MoFe protein it produces were previously described (19). Acetylene-resistant strains were isolated by plating approximately 10 9 cells of DJ939 on Burk medium Petri plates and then incubating these plates in gas-tight jars that included 0.025 atm of acetylene in air. Acetylene-resistant colonies that appeared after about 10 days were picked and streaked on fresh Burk's medium Petri plates and again incubated under 0.025 atm of acetylene in air. In this way, five different independently isolated acetylene-resistant strains were obtained. To determine whether or not the acetylene-resistant phenotype for a particular strain was the result of a mutation within the nifD gene (the nifD gene encodes the MoFe protein ␣-subunit), the genomic segment corresponding to the nifD gene of each strain was inserted into the pUC119 cloning vector. Isolated DNA from each strain was then used in attempts to transform DJ939 to the acetylene-resistant phenotype. Of the five strains, two of them (DJ1250 and DJ1252) were found to carry a mutation within the nifD gene. Sub-segments of the nifD gene from each of these two clones were then used to determine that a specific region within nifD endows the acetylene-resistant phenotype. In both cases this segment was found to include the region encoding residues 37 to 214. DNA sequence analysis revealed that both strains carried a single nucleotide change so that the ␣-subunit residue-69 codon (GGC) was substituted by a serine codon (AGC).
To determine whether or not the ␣-Ser 69 substitution can suppress the acetylene-sensitive phenotype in other mutant strains, plasmid pDB1096 was used as the donor DNA for transformation experiments using either strain DJ266 or DJ1036 as the recipient. In these transformation experiments, selection was for the acetylene resistance phe-notype. Plasmid pDB1096 contains an approximately 500-base pair segment of nifD that includes the coding region including ␣-subunit residues 37-214. This plasmid also carries the ␣-subunit residue-69 serine codon. Strain DJ266 (20) is mutated within the nifH gene so that the Fe protein Arg 100 residue is substituted by Leu 100 , and strain DJ1036 is mutated within the nifD gene so that the MoFe protein ␣-subunit Phe 381 residue is substituted by Leu 381 (21). Both DJ266 and DJ1036 exhibit the acetylene-sensitive phenotype, and the acetyleneresistant strains derived from each of them were, respectively, designated DJ1257 and DJ1269. Plasmid pDB1096 was then used in the same way to reconstruct an acetylene-resistant derivative of DJ939 that was designated DJ1270. The phenotypes and genotypes of all the strains used in this work are summarized in Table I.
A strain that produces an altered MoFe protein for which ␣-Ser 69 is the only substitution was isolated in two steps. In the first step a strain derived from strain DJ995 (22) was constructed using plasmid pDB243 as the donor DNA, and it is deleted for the nifD sequence encoding residues 69 through 94. This strain was designated DJ1259. Strain DJ1259 was then used as the recipient in a transformation experiment where plasmid pDB1096 described above was the source of donor DNA. The resultant strain was designated DJ1262, and it produced an altered MoFe protein having the ␣-Gly 69 residue substituted by serine. Methods and strategies used for strain constructions in the present work have been previously described (23)(24)(25). None of the recombinant plasmids used in this work are capable of autonomous replication in A. vinelandii.
Cell Growth and Protein Purification-A. vinelandii cells were grown, derepressed for nif gene expression, and harvested as described previously (22). Crude extracts were prepared by osmotic shock, and all MoFe proteins used in this study were purified using the immobilized metal-affinity chromatography method described previously (22). MoFe protein produced by strain DJ995 (wild type) and DJ1262 each has a polyhistidine tail located near the carboxyl terminus of their respective ␣-subunits. Protein was quantitated by a modified biuret method using bovine serum albumin as the standard (26), and protein purity was monitored by SDS-polyacrylamide gel electrophoresis (27). The Fe protein used in all experiments did not possess a polyhistidine tag and was purified from wild type A. vinelandii cell extracts as described previously (28). Purified products were pelleted and stored in liquid nitrogen until used. All protein manipulations were carried out anaerobically using a Schlenk apparatus equipped with a BASF catalyst tower (29).
Assays-The composition of the assay mixture and overall techniques used are described elsewhere (11,28). For each assay, 0.05 mg of MoFe protein was used, whereas the Fe protein quantity was adjusted to give the component protein molar ratio shown in Table II. Acetylene was freshly prepared for each assay run by the reaction of calcium carbide with water. Upon initiation of the assay by the addition of Fe protein, samples were incubated for 8 min with shaking in a 30°C water bath, and the reaction terminated by the addition of 250 l of a 0.4 M EDTA solution. H 2 production was monitored by injection of 200 l of the gas phase into a Shimadzu GC-14 gas chromatograph equipped with a Supelco 80/100 molecular sieve 5A column and a thermal conductivity detector. Ethylene production was monitored using a Hewlett-Packard 5890A gas chromatograph equipped with an Al 2 O 3 capillary column and a flame ionization detector. Ammonia production was monitored using the colorimetric indophenol assay previously described (11,30). To reduce background contamination, glass vials used for NH 3 determination were soaked overnight in 1 M NaOH, rinsed thoroughly with distilled water and then ethanol, and oven-dried. During the assays, samples were degassed using He gas in the vial headspace. All sets of kinetic data were collected simultaneously with parallel wild type controls when necessary.
Inhibition patterns were evaluated by examination of Lineweaver-Burk double-reciprocal plots of the kinetic assay data. Michaelis constants (K m ) were derived by fitting the assay data to a hyperbolic equation of the form, where v is the specific activity of the substrate being examined, [S] is the concentration of substrate, and V max is the theoretical maximum specific activity, which was also treated as a variable parameter during fitting.
Electron Paramagnetic Resonance (EPR) 1 Spectroscopy-EPR spectra were recorded on a Bruker ESP300E spectrometer equipped with a dual mode cavity and an Oxford ESR 900 liquid helium cryostat. Perpendicular mode spectra were recorded at ϳ12 K with a microwave power of 2.01 mW, microwave frequency of 9.65 GHz, modulation amplitude of 12.63 G, modulation frequency of 100 kHz, and a time constant and conversion time of 20.48 ms.

Rationale and Design of Experiments for Isolation of Acetylene-resistant
Strains-Our initial goal was to determine the possibility of isolating a mutant strain of A. vinelandii that is altered in its ability to reduce acetylene without a concomitant impairment in its ability to reduce nitrogen. Because acetylene is known to be a specific inhibitor of nitrogen fixation, a straightforward genetic approach to address this issue would be to isolate mutant strains exhibiting in vivo acetylene resistance when cultured under diazotrophic growth conditions. Although acetylene has been known for many years to specifically inhibit A. vinelandii diazotrophic growth (1), the explosive levels of acetylene required for a clean genetic selection has denied the use of this direct approach. We therefore developed an indirect approach that relied on the flux-dependence of nitrogenase substrate discrimination as described below.
The rate at which electrons are delivered to the nitrogenase substrate reduction site is usually referred to as flux and can be controlled in vitro by adjusting the relative ratio of Fe protein to the MoFe protein in the assay mixture. High flux conditions are obtained by using a high Fe protein/MoFe protein ratio. In their analysis of the in vitro catalytic properties of nitrogenase, Davis et al. (31) show that under low flux conditions, acetylene is a more effective inhibitor of nitrogen reduction than under high flux conditions. The explanation offered for this phenomenon is that a more reduced state of the enzyme might be required for nitrogen binding than for acetylene binding. In the case of acetylene reduction measured in typical assays, this suggestion has been strongly supported by a series of detailed kinetic analyses described by Burgess and Lowe (9) and Thorneley and Lowe (32). We therefore reasoned that strains that exhibit low flux under in vivo conditions as a result of an amino acid substitution in one of the nitrogenase component proteins might exhibit diazotrophic growth that is hypersensitive to acetylene. This prediction proved to be correct because many of the mutant strains from our collection having amino acid substitutions within either the Fe protein or the MoFe protein and that otherwise exhibit nearly normal diazotrophic growth were found to be unable to grow in the presence of 0.025 atm of acetylene. For example, certain mutants that produce nitrogenases that are only modestly impaired in their capacity for intercomponent electron transfer (20) or intramolecular electron transfer between the P cluster and FeMo cofactor (19) or are slightly altered in their FeMo cofactor environments (21,33) all exhibit diazotrophic growth that is highly sensitive to acetylene ( Fig. 1 and Table I). In contrast, there is little or no effect on diazotrophic growth when the wild type strain is cultured in the presence of 0.025 atm of acetylene ( Fig. 1 and Table I).
A mutant strain that produces a MoFe protein having histidine substituted for the MoFe protein ␤-subunit Tyr 98 residue (19) and that exhibits diazotrophic growth that is hypersensitive to acetylene was used as the parental strain for the selection of spontaneous mutants resistant to acetylene ( Fig. 1 and Table II). Two of five independently isolated acetylene-resistant strains retain the original ␤-His 98 substitution and also carry a second substitution within the MoFe protein ␣-subunit as determined by genetic mapping experiments. DNA sequence analysis of the nifD gene isolated from these two mutants revealed that both of them encode a MoFe protein ␣-subunit having the Gly 69 residue substituted by serine. The other three acetylene-resistant mutants were true revertants.
Nature of the Acetylene Resistance Phenotype Elicited by the ␣-Ser 69 Substitution-Acetylene resistance resulting from the ␣-Ser 69 substitution could be explained by either a compensation in the electron-flux defect caused by the original ␤-His 98 substitution or by alteration in the accessibility of an acetylenebinding site such that acetylene is no longer an effective physiological inhibitor of nitrogen reduction. These possibilities were distinguished by showing that the ␣-Ser 69 substitution can effectively suppress the acetylene-sensitive phenotype when placed in combination with a variety of other individual substitutions, each of which lowers flux in a different way ( Table I). The most striking example of this phenomenon is illustrated by the observation that the ␣-Ser 69 substitution within the MoFe protein effectively suppresses the acetylenesensitive phenotype of a mutant strain that has leucine substituted for the Fe protein Arg 100 residue. The Fe protein Arg 100 residue is located at the component protein-docking interface (34,35), and substitutions at this position have been shown to affect intercomponent electron transfer (20,36). The general ability of the ␣-Ser 69 substitution to suppress the acetylenesensitive phenotype is best explained by a structural barrier that prevents acetylene from effectively binding to the active site of the altered protein.
Characterization of the ␣-Ser 69 MoFe Protein-To assess the catalytic consequences of the ␣-Ser 69 substitution without interference by the effects of any other amino acid substitution, a strain was constructed that produces an altered MoFe protein for which ␣-Ser 69 is the only substitution. A comparison of the catalytic properties of the isolated ␣-Ser 69 MoFe protein and the wild type MoFe protein are summarized in Fig. 2 and Table  II. The acetylene reduction behavior is graphically illustrated in the Lineweaver-Burk plot shown in Fig. 2, where the wild type MoFe protein and the ␣-Ser 69 MoFe protein data exhibit very different slopes yet approach the same y intercept, indic-ative of a large change in K m accompanied by almost no change in the V max between the two proteins. The inset for Fig. 2 shows the actual enzyme saturation data with the accompanying fit to the hyperbolic form of the Michaelis-Menten equation (as described under "Experimental Procedures"). The resulting values for K m and V max derived from these fits are given in Table  II. These results show that the ␣-Ser 69 MoFe protein exhibits a K m of ϳ0.14 Ϯ 0.01 atm, an approximately 20-fold increase from the apparent K m observed for acetylene binding to the wild type MoFe protein. Also shown in Table II are the interactions among some of the substrates and inhibitors of nitrogenase catalytic activity. Acetylene inhibition of nitrogen reduction catalyzed by the ␣-Ser 69 MoFe protein is changed from a noncompetitive inhibition pattern to a competitive pattern. The ␣-Ser 69 MoFe exhibits proton reduction specific activity and nitrogen reduction specific activity very similar to that of the unaltered MoFe protein. Also, like the wild type MoFe protein, the nitrogen reduction activity of the ␣-Ser 69 MoFe protein is inhibited by hydrogen. Finally, the ␣-Ser 69 MoFe protein exhibits an S ϭ 3/2 EPR spectrum that is identical to the wild type in both lineshape and intensity. DISCUSSION An explanation of the effect of the ␣-Ser 69 substitution on the ability of the altered MoFe protein to reduce acetylene can be considered in the context of previous work. First, Davis et al. (15) report evidence for the presence of high affinity and low affinity acetylene-binding sites on the Clostridium pasteurianum nitrogenase enzyme. The K m they report for the low affinity acetylene-binding site (0.23 Ϯ 0.1 atm) is similar to the one reported here for the ␣-Ser 69 MoFe protein (0.14 Ϯ 0.01 atm). Second, EPR studies by Lowe et al. (16) were interpreted to indicate the presence of two binding sites for acetylene on the MoFe protein from Klebsiella pneumoniae (16). Finally, we previously reported that substitution of the ␣-Arg 277 residue by histidine elicits CO-induced cooperativity of acetylene binding to the altered MoFe protein (17). Thus, there is clear evidence that at least two acetylene binding sites are located within the MoFe protein. In the case of the ␣-Ser 69 MoFe protein, the high affinity acetylene-binding site appears to have been eliminated or severely altered without significantly altering the capacity of the altered enzyme to reduce nitrogen. Also, the remaining low affinity acetylene binding site in the ␣-Ser 69 MoFe protein does not appear to be significantly altered, maintaining a K m value similar to the value previously reported by Davis et al. (15) and a V max that is only slightly less than the value derived for wild-type MoFe protein (Table II).
Elimination of the high affinity acetylene-binding site by the ␣-Ser 69 substitution can be considered in light of the FeMo cofactor structure and its polypeptide environment (Fig. 3a). This structure has a central prismatic waist of six Fe atoms, each of which is ligated to three sulfides in a nearly trigonal planar geometry. These six Fe atoms are arranged such that three pairs of Fe atoms are shared at the intersection of three geometrically identical 4Fe-4S faces within the FeMo cofactor (Fig. 3b). The ␣-Gly 69 residue is located immediately adjacent to the ␣-Val 70 residue that caps one of the 4Fe-4S faces of FeMo cofactor. These two residues are located near the end of a short helix that is positioned between the P cluster and FeMo cofactor. We propose that the high affinity acetylene-binding site is provided by the 4Fe-4S face capped by ␣-Val 70 and that the ␣-Ser 69 substitution results in movement of ␣-Val 70 so that the high affinity acetylene-binding site is no longer accessible. An  alternative interpretation that invokes either a kinetic or electronic explanation for the effect of the ␣-Ser 69 substitution on acetylene reduction is probably not reasonable. The basis for this conclusion is that reduction of nitrogen to ammonia requires six electrons, whereas reduction of acetylene requires only two electrons. Thus, any kinetic or electronic perturbation that impairs acetylene reduction is also expected to impair nitrogen reduction, yet the ␣-Ser 69 MoFe protein is not significantly impaired in its ability to reduce nitrogen. For this reason and because there is no apparent perturbation in the S ϭ 3/2 EPR signal of the ␣-Ser 69 MoFe protein when compared with the wild type, it is also unlikely that a global rearrangement in the polypeptide environment of FeMo cofactor has been elicited by the ␣-Ser 69 substitution. Our interpretation of the reciprocity in the competitive inhibition of nitrogen and acetylene reduction exhibited by the ␣-Ser 69 MoFe protein is that the low affinity acetylene-binding site and the nitrogen-binding site are the same. For simplicity, this site is hereafter referred to only as the "nitrogen binding" site. If the 4Fe-4S face of FeMo cofactor that is capped by ␣-Val 70 provides the high affinity acetylene-binding site, where is the nitrogen binding site? Our results do not provide a definitive answer to this question. Nevertheless, based on the results described here and results previously obtained with an altered MoFe protein having the ␣-His 195 MoFe protein substituted by glutamine, we argue that this site is probably provided by the same 4Fe-4S face that is capped by ␣-Val 70 . Inspection of the FeMo cofactor polypeptide environment shows that the ⑀ imidazole nitrogen of the ␣-His 195 residue is hydrogen-bonded to a central sulfide that is a member of the 4Fe-4S face capped by ␣-Val 70 ( Ref. 37 and Fig. 3). Substitution of ␣-His 195 by glutamine results in an altered MoFe protein that can bind nitrogen with normal affinity but does not significantly reduce it (11). Nevertheless, the ␣-Gln 195 MoFe protein is unaltered in its ability to either bind or reduce acetylene, so the high affinity acetylene-binding/reduction site must remain intact. An important observation is that the ␣-Gln 195 MoFe protein cannot significantly reduce nitrogen, yet nitrogen remains an effective competitive inhibitor of acetylene reduction. Thus, acetylene cannot be bound and reduced at the high affinity site when nitrogen is bound to the active site. A simple working model that explains this situation is that the nitrogen binding site overlaps with the high affinity acetylene-binding site, both of which are located within the same 4Fe-4S face of FeMo cofac-  2. Lineweaver-Burk plot comparing the kinetic behavior of the ␣-Ser 69 MoFe protein (filled circles) and the wild type MoFe protein (filled squares) for the reduction of acetylene. The plot shows the reciprocal of the specific activity (expressed in nmol of C 2 H 2 reduced/min/mg of MoFe protein) versus the reciprocal of the partial pressure of acetylene (in argon) used for the assay. The inset shows the raw saturation curve data for the reduction of acetylene that was fit to a hyperbolic function as described under "Experimental Procedures." The resulting Michaelis constants (K m ) are shown in Table II. tor. It should be noted, however, that we have no data to show that the binding of acetylene at the high affinity site necessarily prevents binding of nitrogen. Thus, although the nitrogenbinding site appears to overlap with the high affinity acetylenebinding site, it is not established that the converse is true. For reasons described below, the presence of a nitrogen binding site that overlaps with both the high affinity and low affinity acetylene binding sites within the same 4Fe-4S face is a model we favor. Nevertheless, other possibilities that explain the available data have not been excluded. For example, because certain Fe and S molecules are shared between adjacent 4Fe-4S faces provided by FeMo cofactor (Fig. 3), binding of a substrate at one face could, in principle, prevent binding of a second molecule at an adjacent face.
We consider the 4Fe-4S face capped by ␣-Val 70 an attractive candidate for providing the physiologically relevant substratebinding site for three other reasons. First, ␣-Val 70 is located at the end of a short helix that is directly connected to the P cluster through the cluster coordinating ␣-Cys 62 residue (Fig.  3). This helix provides a potential through-protein route for electron transfer directly from the P cluster to the FeMo cofac-tor. Second, the P cluster is known to undergo a redox-dependent structural rearrangement that could be coupled to electron or proton transfer to the FeMo cofactor (4,38). The short helix that spans from the P cluster to the FeMo cofactor is an ideal candidate for providing communication between the two cluster types. For example, slight movement of ␣-Val 70 residue as a consequence of events occurring at the P cluster could play an important role in the correct positioning of substrate for reduction (Fig. 3a). From this perspective it is interesting that the amino acid sequence between and including ␣-Cys 62 to ␣-Val 70 is among the most highly conserved primary sequences among all known MoFe proteins. The conservation in primary sequence of this segment also extends to the corresponding region from the alternative nitrogenases (39). However, in this context it is interesting to note that the residue within the VFe protein corresponding to ␣-Gly 69 from the MoFe protein is substituted by a leucine. The VFe protein also exhibits a K m for acetylene binding that is approximately 10 times higher than for the MoFe protein (40). Arguments made for a possible functional role for the ␣-Val 70 residue in nitrogenase catalysis can also be made for the ␣-Arg 96 residue. The ␣-Arg 96 residue is hydrogenbonded to the same FeMo cofactor 4Fe-4S face capped by the ␣-Val 70 (Fig. 3c). This residue is also connected by a short helix to the ␣-Cys 88 residue, which provides a bridging ligand to the two subcluster fragments of the P cluster. Whether or not either or both of these residues directly participate in catalysis, the 4Fe-4S face they approach is distinguished from the two other 4Fe-4S faces in its potential for communication with the P cluster. Third, of the three 4Fe-4S faces provided by FeMo cofactor, the richest source of potential proton donors necessary for substrate reduction is found in proximity to the 4Fe-4S face capped by ␣-Val 70 (Fig. 3, a and c). Such potential proton donors include ␣-His 195 , ␣-Arg 96 , and a pool of water molecules organized around the ␣-carboxylate group of homocitrate.
In summary, substitution of the MoFe protein ␣-Gly 69 residue by serine results in elimination of a high affinity acetylenebinding site located within the MoFe protein. This substitution also eliminates the nonreciprocity in the mutual inhibition patterns of acetylene and nitrogen reduction exhibited by the wild type protein. The location of the ␣-Ser 69 substitution leading to the acetylene-resistant phenotype indicates that the 4Fe-4S face capped by ␣-Val 70 is a likely candidate to provide the high affinity acetylene-binding site. We believe that the genetic approach described here should also prove useful for the evaluation of the binding of other nitrogenase substrates and inhibitors. FIG. 3. a, illustration of the relative positions of the nitrogenase metal cluster types and the location of the short helix that spans from the P cluster ligand ␣-Cys 62 to ␣-Val 70 , which approaches a 4Fe-4S face of the FeMo cofactor. The P cluster is shown in the dithionite-reduced, all-ferrous form. Also shown is the ␣-His 195 residue, which is hydrogenbonded to the FeMo cofactor. b, schematic view of the FeMo cofactor, depicting the atoms included in an individual 4Fe-4S face. It can be seen that each Fe-S-Fe group that bridges the 4Fe3S and 1Mo3Fe3S subclusters that compose the cofactor will be shared by two 4Fe-4S faces. c, "end-on" view of the FeMo cofactor, showing the approaching residue ␣-Val 70 along with hydrogen-bonded residues ␣-His 195 and ␣-Arg 96 . The figures were generated using the programs MOLSCRIPT (41) and Raster3D (42,43).