|Home | About | Journals | Submit | Contact Us | Français|
A series of site-directed mutants of the ferredoxin-dependent spinach nitrite reductase has been characterized and several amino acids have been identified that appear to be involved in the interaction of the enzyme with ferredoxin. In a complementary study, binding constants to nitrite reductase and steady-state kinetic parameters of site-directed mutants of ferredoxin were determined in an attempt to identify ferredoxin amino acids involved in the interaction with nitrite reductase. The results have been interpreted in terms of an in-silico docking model for the 1:1 complex of ferredoxin with nitrite reductase.
Ferredoxin:nitrite oxidoreductase (EC 184.108.40.206, hereafter referred to as nitrite reductase) catalyzes the six-electron reduction of nitrite to ammonia during an early step in the nitrogen assimilation pathway of oxygenic photosynthetic organisms, using reduced ferredoxin as the physiological electron donor (Hase et al., 2006). These ferredoxin-dependent assimilatory nitrite reductases are soluble enzymes with molecular masses of approximately 65kDa and are located in the stromal space of chloroplasts in photosynthetic eukaryotes (Hase et al., 2006). Ferredoxin-dependent nitrite reductases are characterized by a unique active site that contains a siroheme coupled to a [4Fe–4S] cluster via a bridging sulfur atom from a cysteine residue (Swamy et al., 2005; Hase et al., 2006). Electron paramagnetic resonance (EPR) and resonance Raman studies (Kuznetsova et al., 2004a, 2004b) have elucidated the electronic state of the coupled prosthetic group arrangement found in oxidized nitrite reductase (and of the similar active site found in assimilatory sulfite reductases, Krueger and Siegel, 1982a, 1982b). This ‘as-isolated’ form of nitrite reductase contains an oxidized, EPR-silent (S=0) cluster in the [4Fe–4S]2+ state and a siroheme that contains a high-spin, six-coordinate Fe3+ with water or some other weak ligand in the sixth axial position (Kuznetsova et al., 2004a, 2004b), with the fifth axial position being occupied by the cysteinyl sulfur that provides the bridge to the iron–sulfur cluster (Swamy et al., 2005). Although the [4Fe–4S] cluster and the siroheme are magnetically coupled, they behave as independent one-electron couples during electrochemical redox titrations, with Em values of –290 and –370mV for the siroheme and [4Fe–4S] 2+,1+ cluster, respectively (Hirasawa et al., 1994). Flash photolysis measurements, using the deazaflavin radical as a non-physiological reductant to the purified spinach enzyme, indicate that the first electron delivered to nitrite reductase passes transiently through the iron–sulfur cluster and then produces a reduced (i.e. Fe2+) siroheme (Hirasawa et al., 1994). The subsequent delivery of a second electron produces the doubly reduced enzyme with the iron–sulfur cluster now reduced to the EPR detectable (S=½) [4Fe–4S]1+ cluster. The x-ray crystal structure of the enzyme, which indicates that the enzyme's [4Fe–4S] cluster lies between the siroheme and the likely binding site for ferredoxin, is consistent with the hypothesis that electron flow proceeds initially from reduced ferredoxin to the enzyme's [4Fe–4S] cluster and subsequently from the reduced cluster to the siroheme (Swamy et al., 2005).
The fact that ferredoxin is a one-electron donor, coupled with the fact that ferredoxin-dependent nitrite reductases contain only a single high-affinity binding site for ferredoxin (see below), means that the enzyme must accumulate six electrons in one-electron steps (there is no evidence for any significant release of partially reduced intermediate products during the catalytic cycle). A substantial amount of evidence has accumulated indicating that nitrite binds too slowly to the fully oxidized enzyme for the nitrite adduct of the oxidized enzyme to be a kinetically competent intermediate (Kuznetsova et al., 2004a, 2004b). Instead, it is likely that the enzyme is first reduced by one electron and the Fe2+ siroheme then binds nitrite rapidly (Kuznetsova et al., 2004a, 2004b). Delivery of a second electron from reduced ferredoxin produces a well characterized Fe2+ siroheme/NO adduct (with the iron–sulfur cluster in the [4Fe–4S]2+ state), which has recently been shown to behave as expected for a true reaction intermediate (Kuznetsova et al., 2004a, 2004b). Very much less is known about the subsequent electron transfer steps but there is some evidence that hydroxylamine (which is only two electrons more oxidized than the final product ammonia) could perhaps function as a late intermediate (Hase et al., 2006). In contrast, nothing is known about the pathway for the eight protons that are also involved in the reaction.
Ferredoxin-dependent nitrite reductases have been shown, using a variety of techniques, to form a high-affinity, 1:1 complex with ferredoxin at low ionic strength (e.g. the purified enzyme isolated from spinach leaf forms a complex with a Kd of 600nM in 10mM phosphate buffer at pH7.7—Hirasawa et al., 1986). The observation that the nitrite reductase/ferredoxin complex dissociates at higher ionic strengths, as is also the case for the ferredoxin complexes of other ferredoxin-dependent enzymes (Hase et al., 2006), led to the conclusion that electrostatic forces play an important role in stabilizing the nitrite reductase/ferredoxin complex (Hase et al., 2006). The fact that ‘plant-type’ ferredoxins are in general very acidic proteins and thus carry a substantial negative surface charge at neutral pH values led to the suggestion that ferredoxin supplied the negative charges involved in stabilizing the complex with nitrite reductase and that the enzyme supplied a spatially matching set of positively charged residues that would form a network of salt bridges binding the two proteins together (Hase et al., 2006). Mutational studies of the ferredoxin-dependent nitrite reductase from the cyanobacterium Anabaena sp. PCC 7120 (Curdt et al., 2000) and chemical modification studies, designed to eliminate the negative charges from surface-exposed ferredoxin aspartate and glutamate residues (see, e.g. Hirasawa et al., 1986) and the positive charges on surface-exposed nitrite reductase lysine and arginine residues of spinach nitrite reductase (Hirasawa et al., 1993; Dose et al., 1997) provided support for this hypothesis. The recent availability of a 2.8-Å-resolution structure for spinach nitrite reductase (Swamy et al., 2005) made in-silico modeling of the complex formed between these two proteins possible (to date, attempts to crystallize a complex of nitrite reductase and ferredoxin have proven unsuccessful). This model (Figure 1) suggested specific details of the interaction between ferredoxin and nitrite reductase and we report below a site-directed mutagenesis study designed to test this model and to identify specific amino acids on the enzyme involved in binding to ferredoxin.
The in-silico docking model shown in Figure 1 (Swamy et al., 2005) identified four basic amino acids (shown in blue in Figure 1) on the surface of spinach nitrite reductase, Lys80, Lys83, Lys100, and Arg502, as likely to be involved in binding ferredoxin. (The numbering system used throughout is based on designating the Cys at the N-terminus of the mature form of the native chloroplast protein as amino acid number 1. The ferredoxin structure is based on that for the E92K variant of spinach ferredoxin I (Binda et al., 1998) with the N-terminal amino acid of the mature chloroplast form of the protein designated as amino acid number 1.) Three additional conserved basic amino acids, Lys436, Arg375, and Arg556 (shown in red in Figure 1), that had been identified as possibly being involved in ferredoxin binding in earlier chemical modification studies (Dose et al., 1997), but which appear to be incompatible with ferredoxin binding site derived from the more recent structural information, were also targeted for mutagenesis. Two other highly conserved basic amino acids, Lys49 and Lys268, and one highly conserved, polar but uncharged residue, Asn304, were also targeted in order to make this mutagenesis survey more complete.
The results of ferredoxin-binding experiments at low ionic strength (Table 1) indicate that none of the mutated nitrite reductase variants shows large decreases in the strength of ferredoxin binding when compared to that of the wild-type enzyme. Kd values for the variants ranged from a low of approximately 1μM (for K83E, K268E, and R502E) to a high of 50μM (K49E), compared to the 12-μM value found for the wild-type enzyme. As can be seen in Table 1, double or even triple amino acid replacements of basic amino acids with oppositely charged amino acids (R375/R556E, K436/R556E, and R375/K436/R556E) had little effect on the Kd value for ferredoxin binding. It should be noted that some of the variants, in their as-isolated, ferri-siroheme states, exhibited significantly altered binding affinities for nitrite (e.g. a five-fold increase in affinity was observed for the R502E variant and a decrease in affinity by a factor of at least 10 was observed for the K80E variant). The reduced affinity of the K80E variant for nitrite is of particular interest, as Lys80 is located close to the ‘hole’ in the protein that is the likely entry pathway for nitrite to the active site and amino acid replacements at this position may hinder access for this substrate (Arg502, although located at some distance from Lys80, may also lie on the access pathway for nitrite). A series of additional site-specific variants are currently being investigated in an attempt to further understand the role of specific amino acids in nitrite binding.
Before measuring the effects, if any, of amino acid replacements on the kinetic properties of the mutated nitrite reductase variants, it was important to measure the siroheme and iron–sulfur content of the variants. Although all of the residues chosen for replacement lie on the surface of the protein and none appears to be directly involved in prosthetic group binding (Swamy et al., 2005), it was essential to ascertain whether any loss in activity displayed by any of the variants might not simply arise from incomplete occupancy of the prosthetic group binding sites. Table 1 shows that, while a few of the nitrite reductase variants have siroheme and iron–sulfur cluster contents similar to those of the wild-type recombinant enzyme, several variants have very low prosthetic group contents. The A280/A390 ratios of the nitrite reductase variants, when compared to that for the wild-type enzyme, also indicate sub-stoichiometric prosthetic group content for many of the variants, as do estimates of the siroheme and [4Fe–4S] cluster contents measured from integration of EPR signals for several of the variants and the magnitudes of the circular dichroism features in the visible region that arise from these prosthetic groups (data not shown). The reasons for this surprising loss of prosthetic group in some of these variants is not clear but, as indicated by circular dichroism measurements in the ultraviolet (data not shown), none of the variants differed appreciably in alpha-helix or beta-sheet content from the wild-type enzyme and so incomplete occupancy of the prosthetic group binding site does not appear to be accompanied by large changes in secondary structure. Another indication that the amino acid replacements that result in a loss of prosthetic group content do not produce global changes in the conformation of nitrite reductase was provided by a series of tryptophan fluorescence measurements (nitrite reductase has eight highly conserved tryptophan residues—Tripathy et al., 2007). It is not only the fact that the nitrite reductase variants with low prosthetic group content exhibit tryptophan fluorescence excitation and emission spectra that are very similar to those of the wild-type enzyme, but these variants also exhibit patterns of tryptophan fluorescence quenching by either nitrite or the non-physiological quencher acrylamide (Tripathy et al., 2007) that are quite similar to that of wild-type, recombinant nitrite reductase (data not shown). The data of Table 1 also show that the low occupancy of the prosthetic group sites in these variants does not seem to affect the affinity of the enzyme for ferredoxin in any dramatic fashion.
The largest loss of the cofactors is for the substitution of Arg 375 by Glu. While this amino acid residue is on the surface of the protein, it lies relatively close to the siroheme (Figure 1) and is within hydrogen-bonding distance of Gln 445, which is part of a loop formed by amino acid residues 440–454. This loop includes two cysteines, Cys441 and Cys447, that coordinate the iron–sulfur cluster. Thus, while not directly involved in coordinating the cofactors, Arg 375 does interact with a key loop and mutation of this residue to Glu appears to destabilize the neighboring loop resulting in disruption of the binding site for the siroheme and iron–sulfur cofactors.
Although some of the amino acid replacements made result in a substantial loss of siroheme and/or iron–sulfur cluster content, the environment of the prosthetic groups in those nitrite reductase molecules that have incorporated the prosthetic groups appears to be similar to that found in the wild-type enzyme. For example, EPR studies of the as-isolated and cyanide-bound forms of five nitrite reductase variants that contain sub-stoichiometric amounts of siroheme (i.e. K49E, K268E, N304E, R502E, and R556E) reveal that all exhibit the high-spin ferri-siroheme resonance characteristic of the wild-type active enzyme (g=6.89, 5.15, 1.95) that is converted to a low-spin siroheme signal (g=2.68, 2.33, ~1.5) on cyanide binding (Hirasawa et al., 1987); see Figure 2. Some of the variants, N304E and K49E in particular, exhibit an additional resonance, with g=2.47, 2.40, 1.75, that is not responsive to cyanide addition and corresponds to a non-functional siroheme conformation. As shown in Figure 3, the dithionite-reduced, cyanide-bound samples of each of these variants exhibit EPR spectra identical to that of the wild-type enzyme. This signal, with g values of 2.04, 1.94, and 1.92, arises from the magnetically isolated, S=½ reduced [4Fe–4S]+ cluster (Aparicio et al., 1975).
All of the nitrite reductase samples used in this study displayed Michaelis-Menten kinetics, with respect to the dependence of initial rate on ferredoxin concentration, and Table 2 summarizes the steady-state kinetic parameters, obtained in rate measurements at varying ferredoxin concentrations but a constant, saturating nitrite concentration, for wild-type enzyme and 13 site-specific variants. Table 2 also contains, for comparison, the relative rates obtained when reduced ferredoxin, the physiological electron donor, was replaced by the non-physiological electron donor, reduced methyl viologen (all rate data were calculated on an equal siroheme content basis, to compensate for the sub-stoichiometric prosthetic group content of many of the variants). With the exception of the K268E variant, which retains 98% of the methyl viologen-linked activity of the wild-type enzyme, all of the variants show significant losses in the ability to reduce nitrite to ammonia, with reduced methyl viologen serving as the electron donor. While some of these losses are relatively modest (37% in the case of the N304E variant and 42% in the case of the R502E variant), most show rates that are 50% or less than that observed with wild-type nitrite reductase. In the case of the K80E variant (a 92% decrease in activity) and the R375E variant (a 95% decrease in activity), less than 10% of the wild-type activity is retained. Surprisingly, the methyl viologen-linked activity of the R375E/K436E/R556E triple mutant is significantly higher than that of the R375E single mutant (as these rates are normalized on the basis of equal siroheme content, this result cannot be attributed to a simple difference in number of nitrite binding sites present).
Despite the fact that most of the single-site amino acid replacements examined in this study display more than a single characteristic difference, when compared to the wild-type enzyme, the data of Table 2 do identify some striking effects on the ferredoxin-dependent activity of nitrite reductase when compared to the effects on activity with the non-physiological electron donor, reduced methyl viologen. While all of the variants examined in this study show lower Vmax values with reduced ferredoxin as the electron donor, when compared to the value observed with wild-type enzyme, the two variants with the highest methyl viologen-linked activity (i.e. K268E and N304E) show substantially higher loss of activity when ferredoxin serves as the donor (i.e. the limiting rate is 31% that observed for the wild-type enzyme for K268E and 17% for N304E). Furthermore, while the N304E variant exhibits only a modest increase in the KM for ferredoxin, the 20-fold increase in the KM for ferredoxin of the K268E variant results in a catalytic efficiency with ferredoxin as the electron donor for this variant that is only 1.5% of that observed with the wild-type enzyme. Charge-reversal variants at two of the nitrite reductase amino acids identified in the in-silico docking model of Figure 1, namely K80E and K83E, exhibit large increases in the KM for ferredoxin (20-fold in the case of K80E and an even larger 1000-fold in the case of K83E), which, when combined with their lowered Vmax values, produce catalytic efficiencies that are, respectively, only 0.5 and 0.02% of those measured for wild-type nitrite reductase. In contrast, the K49E, K100E, and R375E variants, although showing low catalytic efficiencies with ferredoxin as the electron donor, show comparable losses of methyl viologen-linked activity. Thus, the kinetic properties of these three variants do not provide any direct evidence for a role for these three conserved basic amino acids in the interaction of the enzyme with ferredoxin. The same might be said of R502, given the rather modest effects of the charge-reversing R502E replacement on the ferredoxin-dependent kinetic parameters of the enzyme.
As mentioned above, differential chemical modification experiments, carried out before a structure for nitrite reductase was available, identified three conserved basic amino acids, R375, K436, and R556, as possibly being involved in binding ferredoxin (Dose et al., 1997). While the kinetic parameters observed for the R375E charge-reversal variant show little effect on the KM for ferredoxin, compared to the wild-type enzyme value, and do not point to any markedly greater decrease of ferredoxin-dependent activity than of methyl viologen-dependent activity, both the K436E and the R556E variants show both a greater decrease in the rate of the ferredoxin-dependent activity than of the methyl viologen-dependent activity and extremely large increases in the KM values for ferredoxin (1000 and 2000-fold, respectively). These combine to produce the lowest ferredoxin-dependent catalytic efficiencies of any of the variants tested (i.e. 1.9×10−5 and 1.1×10−5, respectively, of the value measured for the wild-type enzyme).
The docking model shown in Figure 1 also allows one to make predictions about the possible involvement of specific ferredoxin amino acids in the interaction of this electron-donating protein with nitrite reductase. As an extensive set of mutated, well characterized spinach ferredoxin variants was not available, we used instead the wild-type ferredoxin from the cyanobacterium Anabaena sp. PCC 7120 and a set of its variants. The amino acid sequence and the three-dimensional structure of this wild-type Anabaena ferredoxin are quite similar to those of spinach ferredoxin I (Rypniewski et al., 1991; Hurley et al., 1997; Binda et al., 1998) and the two cyanobacterial and plant ferredoxins have similar oxidation-reduction midpoint potential (Em) values (Aliverti et al., 1995; Piubelli et al., 1996; Hurley et al., 1997). It has also been shown that this Anabaena ferredoxin serves as an electron donor to the ferredoxin-dependent glutamate synthase of spinach chloroplasts with kinetic parameters similar to those measured for spinach ferredoxin (Hirasawa et al., 1998). For these reasons and because the Anabaena sp. PCC 7120 site-specific ferredoxin variants have been particularly well characterized with respect to their three-dimensional structures, their Em values and their interactions with another ferredoxin-dependent enzyme, Anabaena ferredoxin:NADP+ oxidoreductase, hereafter abbreviated FNR (Hurley et al., 1997), these variants were used in the current study. It should also be mentioned that the vegetative cell forms of ferredoxin from Anabaena sp. PCC 7120 and from Anabaena sp. PCC 7119 have identical amino acid sequences and structures (Morales et al., 1999) and so information found in the literature about one of these proteins is also true of the other. Table 3 shows the binding constants for the Anabaena ferredoxins available for this study to spinach leaf nitrite reductase at low ionic strength. Charge-reversal substitutions at the two highly conserved C-terminal glutamate residues, Glu94 and Glu95 (these correspond to Glu92 and Glu93 in the spinach ferredoxin I sequence, respectively, and are shown in green in Figure 1), had no significant effect on the affinity of the Anabaena ferredoxin for spinach nitrite reductase. In the case of the more conservative E94D replacement, effects on the Kd for binding to nitrite reductase, the KM for ferredoxin and the Vmax similar to those seen with the E94K variant were observed. The same absence of any major effect on the Kd for binding to nitrite reductase was observed when the highly conserved aromatic residue, Phe65, was replaced by the non-aromatic amino acids alanine or isoleucine (Phe65 in the Anabaena ferredoxin corresponds to Phe63 in spinach ferredoxin I). All of the ferredoxins tested displayed Michaelis-Menten kinetics when tested with spinach leaf nitrite reductase and the steady-state kinetic parameters for these ferredoxins (measured at a saturating nitrite concentration) are shown in Table 3. Perhaps the most surprising observation was the fact that most of the variants exhibited significantly lower KM values with nitrite reductase than did the wild-type Anabaena ferredoxin (values for five variants ranged from 10μM for the E94K variant to 23μM for the F65I variant, compared to the 100-μM value observed for the wild-type ferredoxin). Only the replacement of Phe65 by the non-aromatic and much smaller alanine resulted in an increase in KM, namely to 200μM. In contrast, replacing Phe65 with the non-aromatic, but significantly larger isoleucine produced a variant with a KM value 4.3-fold lower than that of the wild-type ferredoxin and the conservative replacement of Phe65 with another aromatic residue in the F65W variant resulted in a 6.7-fold decrease in KM. What is also noteworthy is that C-terminal charge-reversing substitutions produced similar decreases at both positions 94 and 95, with the E94K variant showing a Vmax that is only 12% of the Vmax value obtained for wild-type ferredoxin and the E95K variant showing a Vmax that is only 18% of the Vmax value obtained for wild-type ferredoxin. This is somewhat surprising, given the fact that the in-silico docking model of Figure 1 predicts that Glu95 of Anabaena ferredoxin (which corresponds to Glu93 of spinach ferredoxin I) would interact with nitrite reductase, while Glu94 of Anabaena ferredoxin (which corresponds to Glu92 of spinach ferredoxin I) would not. The results of Table 2 are also quite different from those obtained in studies of the interaction between spinach ferredoxin I (Piubelli et al., 1996; Aliverti et al., 1997) and Anabaena sp. PCC 7120 ferredoxin (Hurley et al., 1993b, 1997) with the FNRs from the same organisms. In this case, while a charge-reversing substitution at one position in ferredoxin had a very large effect on electron transfer to FNR (Glu92 in spinach ferredoxin and Glu94 in Anabaena ferredoxin), a similar replacement at the adjacent position (Glu93 in spinach ferredoxin and Glu95 in Anabaena ferredoxin) had virtually no effect. It should also be pointed out that replacing Glu91 in the native ferredoxin from the green alga Chlamydomonas reinhardtii (this residue is equivalent to Glu94 in Anabaena ferredoxin) with either lysine or glutamine significantly decreases the rates of the reactions catalyzed by the C. reinhardtii nitrite reductase, glutamate synthase, and ferredoxin:thioredoxin reductase, while similar replacements at the adjacent Glu92 position produce smaller losses in activity (Garcia-Sánchez et al., 1997, 2000; Jacquot et al., 1997).
Studies on the interaction of Anabaena ferredoxin with Anabaena FNR have also demonstrated that the conserved aromatic residue, Phe65, is essential for rapid electron transfer from ferredoxin to this target enzyme (Hurley et al., 1993a, 1993b, 1997). Furthermore, our previous studies showed that this amino acid was essential for optimal electron transfer from Anabaena ferredoxin to spinach glutamate synthase (Hirasawa et al., 1998). The results of Table 3 suggest that this conserved aromatic amino acid (shown in green in Figure 1) may also be required for optimal electron transfer from ferredoxin to nitrite reductase.
Although the original experimental design of this study was based on the expectation that variants could be found that were impaired in their interaction with ferredoxin, while displaying full occupancy of the prosthetic group binding site and a methyl viologen-linked activity very similar to that of the wild-type, most of the 13 site-specific variants characterized showed either sub-stoichiometric prosthetic group content or diminished rates of methyl viologen-linked activity, or both. While these multi-parameter effects make an unambiguous mapping of the ferredoxin-binding site by a simple mutagenic approach difficult, the results obtained above have identified some nitrite reductase residues that cannot be altered without very large effects on the activity of the enzyme. It should also be pointed out that the unexpected effects of mutational replacements on prosthetic group site occupancy open a new avenue for exploring the role of different regions in the protein in cofactor assembly. Finally, it should be mentioned that the results obtained with site-specific ferredoxin mutants do appear, when additional variants become available for study, to provide an opportunity for mapping the nitrite reductase binding domain on ferredoxin. These variants are currently being expressed, as are additional site-specific variants of ferredoxins in an attempt to map the binding site on ferredoxin for nitrite reductase in greater detail.
It is clear that much additional work remains to be done before a definitive identification of the ferredoxin-binding site on nitrite reductase is unambiguously identified. One obvious approach would be to obtain crystals of the complex between the two proteins but, despite many attempts, we have not been able to obtain co-crystals of spinach nitrite reductase and spinach ferredoxin (Swamy et al., 2005). In an attempt to circumvent the lack of success in co-crystallizing the spinach proteins, we have recently shifted the focus of these efforts to attempts at co-crystallizing the most abundant isoform of ferredoxin from the green alga Chlamydomonas reinhardtii with a recombinant form of the ferredoxin-dependent nitrite reductase from the same organism (the C. reinhardtii nitrite reductase expresses at considerably higher levels and is considerably more stable than the spinach enzyme). It is hoped that these efforts will lead to an x-ray structure for the complex. The greater stability of the C. reinhardtii nitrite reductase may also offer an opportunity to locate variants in which the effects of single amino acid replacements are confined to a single characteristic of the enzyme. To that purpose, a mutational study of C. reinhardtii nitrite reductase and the most abundant C. reinhardtii ferredoxin isoform has been started in our laboratory.
A recombinant form of spinach nitrite reductase, with an N-terminal His-tag, was expressed in Escherichia coli and purified as described previously (Tripathy et al., 2007). Site-specific variants of nitrite reductase were prepared and purified as described previously (Tripathy et al., 2007). A list of the mutagenic primers used to replace specific nitrite reductase amino acids is given in Table 4. Yields of nitrite reductase ranged from 0.3 to 1.0mg of pure nitrite reductase from 1L of E. coli culture for the different variants. All of the nitrite reductase variants used in this study showed a single Coomassie Blue-staining band after SDS–PAGE, with an apparent molecular mass of approximately 68kDa. Ferredoxin and nitrite reductase were isolated from spinach leaves and purified as described previously (Hirasawa et al., 1993). Wild-type and site-specific variants of the vegetative cell form of ferredoxin from the cyanobacterium Anabaena sp. PCC 7120 were a generous gift from Prof. Gordon Tollin (Department of Biochemistry, University of Arizona). The Anabaena wild-type ferredoxin and the site-specific Anabaena ferredoxin variants used in this study have all been extensively characterized previously (Hurley et al., 1997). The spinach leaf ferredoxin and the Anabaena ferredoxins used in this study all displayed a single Coomassie Blue-staining band after SDS–PAGE. Concentrations of wild-type proteins were measured as described previously (Tripathy et al., 2007). Because many of the site-specific nitrite reductase variants had altered prosthetic group content, the protein concentrations of these variants were measured using the method of Bradford (1976), which was also used for estimating the protein concentration of the wild-type nitrite reductase samples used for prosthetic group quantitation.
Steady-state kinetic parameters for the nitrite-reductase-catalyzed reaction with reduced ferredoxin as the electron donor and nitrite reductase activity with reduced methyl viologen as the electron donor were measured as described previously (Hirasawa and Knaff, 1985). Kd values for the complexes formed between ferredoxin and nitrite reductase and between nitrite and nitrite reductase were measured using the previously described spectral perturbation method (Hirasawa and Knaff, 1985). Total iron (Miller and Massey, 1965), siroheme (Siegel et al., 1973), and acid-labile sulfide (Siegel et al., 1973) contents were measured using standard methods. Absorbance spectra in the visible and ultraviolet regions were measured at 0.5nm spectral resolution using a Shimadzu Model 2401PC spectrophotometer. Tryptophan fluorescence was measured as described previously (Tripathy et al., 2007). Circular dichroism (CD) spectra were obtained at 1.0nm spectral resolution using an Olis DCM-10 UV/Vis Spectropolarimeter. EPR spectra were recorded on frozen samples with a Bruker ESP300E spectrometer equipped with an ER-4116 dual-mode cavity and an Oxford instruments ESR-9 flow cryostat. Spectra were typically recorded at 15K with a microwave power of 2mW, a modulation frequency of 100KHz, and a modulation amplitude of 6.4Gauss. A 10-fold molar excess of KCN was added to oxidized samples prior to dithionite reduction in order observed the EPR spectrum of the magnetically isolated [4Fe–4S]+ cluster in reduced nitrite reductase samples.
This work was supported by a contract from the Chemical Sciences, Geosciences and Biosciences Division, Office of the Basic Energy Sciences, Office of Sciences, U.S. Department of Energy (Contract No. DE-FG03-99ER20346 to DBK) and from a grant the U.S. National Institutes of Health (GM62524 to MKJ), which provided funds for the EPR portions of the study.
The authors would like to thank Prof. Gordon Tollin and Dr John Hurley (Department of Biochemistry, University of Arizona) for their gift of wild-type and site-specific variants of the vegetative ferredoxin from Anabaena sp. PCC 7120 and Prof. Sabeeha Merchant (Department of Chemistry and Biochemistry, The University of California at Los Angeles) for her generous gifts of genes encoding C. reinhardtii nitrite reductase and ferredoxin 1. No conflict of interest declared.