Info

Nitrogen Fixation

4. Acknowledgements

The authors wish to thank Drs Gary Roberts and Dennis Dean for stimulating discussions. Work from the authors' laboratory was supported by NIGMS grant 35332.

5. References

Allen RM et al. (1993) J. Biol. Chem. 268, 23670-23674

Bulen WA, LeComte JR (1966) Proc. Natl. Acad. Sci. USA 56, 979-86

Filler WA et al. (1986) Eur. J. Biochem. 160, 371-377

Georgiadis MM et al. (1992) Science 257, 1653-1659

Homer MJ et al. (1995) J. Biol. Chem. 270, 24745-24752

Jeffery C (1999) Trends in Biochem. Sci. 24, 8-11

Jeffery C et al. (2000) Biochem. 39, 955-964

Kim J, Rees DC (1992) Science 257,1677-1682

Paustian TD et al. (1990) Biochem. 29(14), 3515-3522

Pope MR et al. (1985) Proc. Natl. Acad. Sci. USA 82, 3173-3177

Rangaraj P et al. (2001) J. Biol. Chem. 276, 15968-15974

Rangaraj P et al. (1997) Proc. Natl. Acad. Sci. USA 94, 11250-11255

Robinson AC et al. (1987) J. Biol. Chem. 262, 14327-14332

Ruttimann-Johnson C et al. (2001) J. Biol. Chem. 276, 4522-4526

Ruttimann-Johnson C et al. (2000) In Pedrosa FO, Hungria M, Yates MG, Newton WE (ed) Nitrogen Fixation: From Molecules To Crop Productivity, pp. 35-36, Kluwer Academic Publishers, Dordrecht

NITROGENASE MECHANISM: CAN OLD LAGS TEACH US NEW TRICKS?

R.N.F. Thorneley1, H. Angove1, G.A. Ashby1, M.C. Durrant1, S.A. Fairhurst1,

S.J. George1, P.C. Hallenbeck2, A. Sinclair1, J.D. Tolland1

'Department of Biol. Chem., John Innes Centre, Norwich, NR4 7UH, UK

2Départ. de Micro, et Immunol., Université de Montréal, Montréal, H3C 3J7 Canada

1. Introduction

Our understanding of the mechanism of nitrogenase is still informed by the concepts, experimentally determined rate constants and simulations of Lowe-Thorneley (1984). The Fe-protein cycle describes the minimum number of partial reactions necessary to effect the transfer of an electron from the Fe protein to the MoFe protein coupled to the hydrolysis of two molecules of MgATP. A major challenge is to understand at atomic resolution the energy transducing reactions (electron transfer, hydrolysis of MgATP and associated conformation changes) that only occur within the transient complex formed between the Fe and MoFe proteins. X-ray crystallography (Chiu et al. 2001 and references therein) and low angle X-ray scattering (Grossman et al. 1997) using AIF4", Be l'Y and the A127-Fe-protein to trap out different conformations of this complex have contributed greatly to our understanding of these events and form the basis for the interpretation of our spectroscopic and kinetic data. We have probed the structural determinants of the kinetic and spectroscopic profiles of reactions occurring in the Fe-protein cycle in three ways: (i) by use of the ATP analog 2'-deoxyATP; (ii) by site-directed mutagenesis of residues that interact with MgATP; and (iii) by replacement of the 4Fe-4S cluster by a 4Fe-4Se cluster in the Fe protein.

In addition, stopped-flow Fourier transform infrared spectroscopy has been used to monitor azide reduction and carbon monoxide inhibition in both the pre-steady state and steady state. These data are beginning to contribute to our understanding of the chemistry that occurs at the FeMo-cofactor consequent on the eight sequential electron and proton transfers that comprise the MoFe-protein cycle of the Lowe-Thorneley model.

2. The ATP Analog, 2'-deoxyATP

The 2'-hydroxy group of the ribose element of ADP is hydrogen bonded to Glu221 in the Fe protein (Figure 1). This hydrogen bond can be easily removed by using 2'-deoxyATP (or ADP) in which the hydroxy-

group is replaced by a hydrogen atom. We have studied a number of the partial reactions occurring in the Fe-protein cycle of Klebsiella pneumoniae (Kp) nitrogenase using 2'-deoxyATP/ADP and compared the kinetics and EPR data with those obtained with

ATP and ADP. The increase in the rate of Fe-chelation from the Fe protein (Kp2) by bathophenanthroline induced by 2'-deoxyATP binding is very similar to that observed with ATP. 2'-DeoxyATP, like ATP, induces the rhombic to axial change in the EPR signal of Kp2. The reduction by S02" of Kp2ox(MgADP)2 and Kp2ox(Mg2'-deoxyADP)2 occurs at essentially the same rate (k = 4 and 5 x 106 M'V1, respectively).

f Thr 45

Figure 1. MgADP binding site of Av2 (cartoon by D. Lawson using coordinates of Jang et al. 2000).

We conclude that with free Fe protein, the hydrogen bond from the 2'-hydroxy group of the ribose to the carboxylate of Glu221 does not significantly determine the conformation of the protein in the vicinity of the 4Fe-4S cluster. The overlay of the x-ray structures of the Azotobacter vinelandii Fe protein (Av2) and Av2(MgADP)2 (Jang et al. 2000) shows little or no movement of this hydrogen bond when free Av2 undergoes the ADP-induced conformation change. Reactions occurring within the Fe-MoFe protein complex are, however, significantly affected by the presence of this hydrogen bond as is shown by the kinetic data summarized in Table 1.

Table 1. Comparison of kinetic data for ATP and 2'-deoxyATP with Kp nitrogenase (23°C)

ATP 2'-DeoxyATP

Kd nucleotide (mM) 0.43 1.3

Complex dissociation (k-3, s"1) 6.4 0.8

Reductant independent ATPase (s"1) 0.6 1.2

Replacement of ATP by 2'-deoxyATP decreases the first-order rate constants for electron transfer and complex dissociation by a factor of eight. Complex dissociation is still rate-limiting because the specific activity for proton reduction is also decreased by a factor of eight. The affinity for Mg2'-deoxyATP is about four-fold less than for MgATP calculated from the hyperbolic dependence of the electron-transfer rate (k2) on nucleotide concentration. The rate of MgATP hydrolysis must also be decreased by a factor close to eight since the ATP/2e" ratio of 5.4 for 2'-deoxyATP is only 25% higher than that for MgATP. Indeed, this difference can be entirely-accounted for by an increased rate of reductant-independent ATP hydrolysis catalyzed by the complex of oxidized Kp2 with Kpl, which is about two-fold higher for Mg2'-deoxyATP than for MgATP. These data show that electron transfer, ATP hydrolysis and complex dissociation are co-modulated by changes in Fe protein structure consequent on loss of the hydrogen bond between Glu221 and the 2'-hydroxyl group of ribose element of ATP. Inspection of the overlay of the x-ray structures of Av2(MgADP)2 and Av2(MgADP.AlF4)2 presented by Jang et al. (2000) shows that this hydrogen bond is broken in the latter structure. We, therefore, suggest that, since ATP hydrolysis only occurs when the Fe protein is bound to the MoFe protein, this hydrogen bond is sensitive to and contributes to the free energy change on docking the Fe protein onto the MoFe protein prior to electron transfer.

3. Asp43Glu, Asp43Asn and Thr45Ser: Switch 1 Region Mutants of Av2

Figure 1 shows that Asp43 and Ser44 are hydrogen bonded to water molecules coordinated to the Mg2+ ion that, in turn, is coordinated to the terminal phosphate group of ADP. Thr45 is hydrogen bonded to Gln54. All three of these residues are part of the signal transduction network in Av2 termed "Switch 1" by analogy to P21ras, which has high structural homology in this region (Jang et al. 2000). We have constructed and characterized three Av2 variants, Asp43Glu, Asp43Asn and Thr45Ser.

EPR spectroscopy showed that the Av2-Asp43Glu and Av2-Asp43Asn mutants do not undergo the MgATP-induced rhombic to axial change in form of the g = 1.95 signal, whereas the Av2-Thre45Ser clearly does. An increase in the initial rate of iron chelation by bathophenanthroline is another criterion by which to estimate the extent of MgATP-induced conformation changes in the Fe protein. The data in Table 2 confirm that the Av2-Asp43Glu and Av2-Asp43Asn mutants do not undergo the MgATP-induced conformation change whereas the Av2-Thr45Ser does. The more than three-fold increase in rate for Av2-Thr45Ser (760) relative to wild-type Av2 (220) is quite remarkable and indicates significant structural changes in the protein environment of the 4Fe-4S cluster. However, the effect of these mutations appears to be much less with respect to the reactivity

Table 2. Increase in initial rates of Fe chelation Table 3. Rate of reduction of oxidized Av2

by bathophenanthroline induced by MgATP with MgADP bound by S02 " (L0

Av2-native 220 Av2-native 4xl06M"1s"1

Av2-Asp43Glu 8 Av2-Asp43Glu 6x 106 M"'s''

Av2-Asp43Asn 7 Av2-Asp43Asn 8 x 106 M'V1

Av2-Thr45Ser 760 Av2-Thr45Ser 4xl06M"1s"1

of the oxidized proteins. The rates of reduction of native Av2ox(MgADP)2 and the mutated forms by dithionite (SO2") are all within a factor of two of each other (Table 3). These clearly show the differential effect of these mutations depending on the oxidation levels of Av2 and on whether MgADP or MgATP is bound. This is important since electron transfer and ATP cleavage are early events in the series of reactions occurring within the Fe-MoFe protein complex, and subsequent energy-transducing reactions and rate-limiting complex dissociation, must involve oxidized Fe protein with 2MgADP bound. Native Av2, Av2-Asp43Glu, Av2-Asp43Asn and Av2-Thr45Ser have specific activities of 2360, 18.5 and 1100 nmole H2 (min mg)"1 and primary electron-transfer rates to Avl of 180, 5.4, not detectable, and 250 s"1, respectively. The significantly higher rate of electron transfer (extrapolated value of k0bS at infinite MgATP concentration) for Av2-Thr45Ser (250 s"1) compared to native Av2 (180 s"1) is intriguing and consistent with its increased rate of Fe-chelation induced by MgATP. It appears as though this mutant undergoes a somewhat larger MgATP-induced conformation change than native Av2. However, the rate of complex dissociation involving the oxidized MgADP-bound form is slower (k.3 = 3.7 s"1) compared to that of native Av20X(MgADP)2 (k.3 = 6.5 s"1). This results in the two-fold decrease in specific activity for this mutant.

A prediction of the Lowe-Thorneley model is that the Av2-Thr45Ser mutant should be more efficient than native Av2 with respect to the percentage of electron flux going to hydrogen evolution under one atmosphere of nitrogen but be less efficient with respect to the ATP/2e" ratio. Both of these predictions are based on an increased steady-state concentration of the Av2ox(MgADP)2Avl complex due to the slower rate of complex dissociation. This suppresses hydrogen evolution but increases the reductant-independent ATPase contribution. The percentage of electron flux into hydrogen evolution under nitrogen for Av2-Thr45Ser and native Av2 (27 ± 2.0 and 32 ± 1.9%, respectively), the ATP/2e ratio under argon (6.1:1 and 4.5:1, respectively), and the kcat for reductant-independent ATPase activity (1.2 and 0.65, respectively) provide some evidence in support of this prediction. Evolution has presumably selected a Thr at position 45 in order to maximize the total rate of ammonia production and minimize the amount of ATP hydrolyzed at the expense of a slightly less efficient enzyme with respect to hydrogen evolution.

4. Replacement of the 4Fe-4S Cluster by a 4Fe-4Se Cluster in the Fe Protein (Kp2)

The 4Fe-4S cluster of Kp2 can be removed, after the addition of MgATP, by chelation of the iron with bathophenanthroline, followed by gel filtration on a Sephadex G-50 under strictly anaerobic conditions in a glove box. The apo-Kp2 can then be reconstituted with a 4Fe-4Se cluster using seleno-D,L-cystine, ferrous ion, dithiothreitol and cysteine-desulphurase (Nif S, supplied by Dr D.R. Dean). The Se-Kp2 protein was purified on DEAE cellulose in Tris-buffer (50 mM, pH 7.4) and eluted with 0.4 M NaCl, 2mM dithionite. The protein contains equivalent amounts of Fe and Se in the range 3 to 5 g atoms per mole of Kp2.

Figure 2 shows the EPR spectra of Se-Kp2 with features at g = 5.2 (S = 7/2) and g = 1.95 (S = 1/2) that are characteristic of 4Fe-4Se clusters (Gaillard et al. 1986; Yu et al. 1991). Modeling of the 4Fe-4Se cluster into the native Av2 x-ray crystal structure shows that replacing S by Se should barely change the Fe-Fe distance [2.76 A (S), 2.78 A (Se)]. The difference between the Fe-S (2.29 A) and Fe-Se (2.42 A) bond lengths causes the 4 Se atoms to move outwards with the Fe atoms fixed by their protein cysteinyl sulfur ligands. EXAFS analysis of Se-Kp2 is in progress (with G. George and R. Prince).

Se-Kp2 has a specific activity about 15% of wild type but with an ATP/2e = 17 (cf. 4.4 for native-Av2). The rate of primary electron transfer is decreased to 38 s"1 vs. 180 s"1 for native Av2, which is associated with a three-fold tighter binding of MgATP (Kd = 0.15 mM vs. 0.44 mM for native Av2).

5. Time-Resolved Fourier Transform Infrared Spectroscopy of CO Inhibition of Azide Reduction

The Lowe-Thorneley model predicts a lag phase of ca. 1 s for the eight-electron reduction of azide to ammonia. The reaction can be studied by stopped-flow FTIR using the apparatus previously described (George et al. 1997, 2000) by monitoring the loss of azide at 2050 cm"1. A time course is shown (Figure 3), which clearly exhibits the predicted lag phase and an extrapolated, rapid initial loss of 8 |xM azide that occurs within the first 200 ms. This amount of azide is stoichiometric with the Mo concentration and indicates that each FeMo-cofactor binds and reduces one equivalent of azide. The reduction of bound azide by eight electrons has to be completed before the second azide binds and a steady state is established at times longer than ca. 1 s. We have also studied the inhibition of this reaction by carbon monoxide under conditions of sub-stoichiometric CO w.r.t Mo that gave ca. 50% inhibition of the rate of azide reduction. No lag phase was observed (<200 ms) for the onset of inhibition. This is intriguing since we have previously monitored the time course for the appearance of an IR band at 1904 cm"1 under low CO conditions (George et al. 1997). This band, which is the first to appear in the time course, takes ca. seven seconds to reach its maximum intensity and is less than 10% developed at 1 s. These data indicate that the 1904 cm"1 species cannot be responsible for the rapid onset of inhibition of azide reduction. We are now seeking, in the technically more difficult region of the IR spectrum where strong protein amide bands absorb g = 5.2

500 1500 2500 3500 4500 Magnetic Field (Gauss)

Figure 2. EPR spectra of Se-Kp2 (bottom) with MgADP (middle) and MgATP (top) at 18 K, 10 mW with buffer blank subtracted.

(1500-1700 cm"1), evidence for a new CO species that forms at very short times and is responsible for the inhibition of azide reduction.

20.02

20.00

19.96

20.00

Figure 3. Stopped-flow FTIR time course for azide reduction showing a one-second lag phase (at 23°C, pH 7.4).

6. Conclusions

In this paper, we have shown how three methods of perturbing the structure of the Fe protein, i.e. by ATP analogs, site-directed mutagenesis of ATP-binding site residues, and selenium substitution of sulfide in the 4Fe-4S cluster, can give new insights into how protein structure modulates the partial reactions that comprise the Fe-protein cycle. In addition, substrate-reduction reactions occurring in the MoFe-protein cycle are becoming accessible to study by real time monitoring by stopped-flow FTIR spectroscopy.

7. References

Gaillard J et al. (1986) Biochem. 25, 464-468

George SJ et al. (1997) J. Amer. Chem. Soc. 119, 6450-6451

George SJ et al. (2000) J. Biol. Chem. 275, 33231-33237

Grossman JG et al. (1997) J. Mol. Biol. 266, 642 -648

Jang SB et al. (2000) Biochem. 39, 14745-14752

Lowe DJ, Thorneley RNF (1984) Biochem. J. 224, 877-886

8. Acknowledgements

RNFT is supported by the UK Biotechnology and Biological Sciences Research Council. HA and JT are grateful to the BBSRC, and AS to the Gatsby Foundation for financial support. PCH was supported on sabbatical leave at the JIC by NSERC and the Royal Society.

AZIDE, CYANIDE AND NITRITE ARE NEW SUBSTRATES OF THE OXYGEN-DEPENDENT NITROGENASE OF THE THERMOPHILIC BACTERIUM STREPTOMYCES THERMOAUTOTROPHICUS

Chair of Microbiology, University of Bayreuth, D-95440 Bayreuth, Germany

1. Introduction

Streptomyces thermoautotrophicus is a thermophilic, aerobic and obligately chemolithoautotrophic bacterium (Gadkari et al. 1990). It is able to fix dinitrogen with CO or H2 plus CO2 as growth substrates (Gadkari et al. 1992, Ribbe et al. 1997). Nitrogenase of S. thermoautotrophicus comprises three enzymes, a heterotrimeric Mo-dinitrogenase (Stl), a dimeric Mn-superoxide oxidoreductase (St2) and a heterotrimeric Mo-carbon monoxide dehydrogenase (CODH or St3). These three enzymes are dioxygen insensitive; in fact, dioxygen is an essential intermediate (Ribbe et al. 1997). CODH oxidizes CO and transfers the electrons released to oxygen (O2), thereby producing superoxide (O2 ")• The Mn-superoxide oxidoreductase reoxidizes O2" to O2 and transfers the electrons to a Mo-dinitrogenase which, in turn, reduces dinitrogen to ammonium. In contrast to the electronic coupling in known nitrogenases via ferredoxin/flavodoxin, the S. thermoautotrophicus enzyme establishes a molecular coupling via O2 . The amino acid sequence of superoxide oxidoreductase, designated SdnO, was highly similar to superoxide dismutases (SOD) of bacilli (Hofmann-Findeklee et al. 2000). The amino acid sequences of the three SdnM, SdnS and SdnL polypeptides (Hofmann-Findeklee et al. 2000) were very similar to the corresponding subunits of structurally characterized CODH from Oligotropha carboxidovorans (Gremer et al. 2000).

"Conventional" Mo-containing nitrogenases are versatile and reduce low-molecular-weight compounds containing N-N, N-O, C-N, and C-C double or triple bonds (Burgess 1993) and carbon monoxide (CO) is a non-competitive inhibitor of all substrates except the H+ (Hardy et al. 1965; Rivera-Ortiz, Burris 1975; Pham, Burgess 1993). In contrast, the nitrogenase of S. thermoautotrophicus cannot reduce acetylene to either ethylene or ethane (Ribbe et al. 1997). Schöllhorn and Burris (1967) showed the reduction of azide and Kelly et al. (1967) the reduction of cyanide, which supplies both HCN as a substrate and CN" as a potent inhibitor. The ratio of these two species is pH dependent (Li et al. 1982). Reduction of nitrite to NH^ by nitrogenase has been reported (Vaughn, Burgess 1989), however, very little is known about NO2" as a substrate.

Carbon monoxide dehydrogenase (CODH) is the key enzyme of carboxidotrophic metabolism. It is a copper-containing molybdenum-iron-sulfur flavoprotein (Gremer et al. 2001), which oxidizes CO to CO2 using water as oxidant to gain the energy for the carboxidotrophic growth (Meyer, Schlegel 1980). The crystal structure of CODH has been solved (Dobbek et al. 1999; Meyer et al. 2000). Until now, N3" reduction to NH/ by CODH, using dithionite as the electron donor, has not been demonstrated. Here, we report that the purified nitrogenase of S. thermoautotrophicus can reduce azide, cyanide, and N02~ to NH4!, in the presence of MgATP and dithionite, and show the effect of CO on substrate reduction and H2 formation. In addition, we show that the CODH (St3 enzyme) of S. thermoautotrophicus can reduce azide to NH4+, which further adds to the analogy between the two enzyme systems.

2. Experimental Procedures

S. thermoautotrophicus UBT1 (DSM 41605, ATCC 49746) was grown chemolitho-autotrophically with CO as a sole source of carbon and energy in mineral medium (Meyer, Schlegel 1980) containing NH4C1 (28 mM) under a gas mixture of (v/v) 78% air, 13% CO and 9% C02 (Ribbe et al. 1997).

Crade extracts were prepared by passing bacterial suspensions (50-60 g of cell wet weight suspended in 50-60 ml of 50 mM potassium phosphate buffer, pH 7.7, containing 0.5 mM phenylmethylsulfonyl fluoride and a few crystals of DNase) 5 to 6 times through a French pressure cell at maximum pressure under oxic conditions and then subjected to low spin centrifugation. Cytoplasmic fractions were obtained by ultra centrifugation and then (70-80 mL) loaded on a 15 cm x 2.6 cm anion-exchange Source 30Q (Amersham Pharmacia Biotech) column, equilibrated with 25 mM potassium phosphate buffer (pH 7.7). Elution with 225 mL 25 mM potassium phosphate buffer (pH 7.7) was followed by a linear gradient of 0 to 0.6 M NaCl in phosphate buffer. The St2 protein does not bind to Source 30Q and is eluted in the first 80-160 mL. The proteins, Stl and St3, are eluted close together at about 0.3 M NaCl. Fractions with ammonium-forming activity (Stl protein) were pooled, supplemented with 1.2 M K2HPO4/HCI buffer (pH 7.7), and stirred for 15 min. Precipitated protein was removed by low-spin centrifugation. The supernatant (80-100 mL) was loaded onto a hydrophobic interaction chromatography column (13 cm by 2.6 cm; Source 15ISO, Amersham Pharmacia Biotech), and equilibrated with 1 M K2HPO4/HCI buffer (pH 7.7). Unbound proteins were eluted with 225 mL of equilibration buffer, followed by 450 mL of a decreasing linear gradient (1.2 to 0 M). The St2 protein was purified as described (Ribbe et al. 1997).

Nitrogenase assays were performed as described previously by following NFLi+ formation from N2 by the indophenol method (Ribbe et al. 1997). CODH activity was measured spectroscopically with iodonitro-tetrazolium chloride (INT) as electron acceptor (Kraut et al. 1989). Hydrogen evolution was analyzed by conversion of HgO to mercury in a trace analytical RGD2 reduction gas detector as described (Gadkari et al. 1990). Dinitrogen was analyzed by gas chromatograph employing standard methods. Solutions of KCN, NaN3 and KNO2 were prepared in 50 mM KH2P04/Na0H buffer (pH 7.5).

3. Azide as a Substrate

Under appropriate assay conditions with dithionite as electron donor, the purified nitrogenase components, Stl and St2, are capable of catalyzing the reduction of azide to ammonium (Figure 1). Maximum NH4+ formation activities of 1.29 (imole NH4+ h"1 x mg-1 was at 25 mM azide. Concentrations above 25 mM neither increased nor inhibited activity.

Figure 1 shows the time course of ammonium formation by purified S. thermoautotrophicus nitrogenase in the presence or absence of N2 or azide. After 1 h of incubation at 65 °C, the ammonium formed in the control was 0.10 mM, whereas ammonium formation with azide under either nitrogen or helium atmosphere was 0.46 mM and 0.51 mM, respectively. After 6 h of incubation, the activity with azide under dinitrogen or helium increased 3.9-fold and 4.6-fold, respectively, compared to the control. Apparently, azide is a substrate and it produces only N2 and NH4+ in a molar ratio 1.4. Azide reduction to NFLt+ had a Kr of 2.38 mM and a Fmax of 26.6 nmole of NFLt+ formed (min mg)"1.

Figure 1. Ammonium formation from azide. Reaction mixtures contained 2.3 mL KH2P04/Na0H buffer (50 mM, pH 7.5), 2.5 mM ATP, 5.0 mM MgCl2, 10.0 mM dithionite, 0.82 mg Stl, 0.13 mg St2 and 25 mM NaN3 where indicated. Assays were performed at 65° C. ■, without azide under 100% N2; □, with azide under 100% N2; and O, with azide under 100% helium.

Figure 1. Ammonium formation from azide. Reaction mixtures contained 2.3 mL KH2P04/Na0H buffer (50 mM, pH 7.5), 2.5 mM ATP, 5.0 mM MgCl2, 10.0 mM dithionite, 0.82 mg Stl, 0.13 mg St2 and 25 mM NaN3 where indicated. Assays were performed at 65° C. ■, without azide under 100% N2; □, with azide under 100% N2; and O, with azide under 100% helium.

5 10

% CO in assay vial

Figure 2. Effect of CO on NH4+ and H2 formation from azide by Stl and St2. Reaction mixtures were as in Figure 1, but with 25 mM NaN3 and various CO concentrations. NH4+ was analyzed after 4 h at 65°C under N2. ■, NH4+ formed without azide; □, H2 formed without azide; NH4+ formed with azide; O, H2 formed with azide. For NH4~, 1.6 |imol (h.mg)"1 and 0.28 jtmol (h.mg)"1 were set as 100% with and without azide, respectively. For H2, 0.13 |amol (h.mg)"1 and 0.43 umol (h.mg)"1 were set as 100% with and without azide, respectively.

3.1. Influence of carbon monoxide on azide reduction. CO is a powerful noncompetitive inhibitor of all substrates except protons of "conventional" nitrogenases (Hardy et al. 1965; Bulen et al. 1965; Hwang et al. 1973). It specifically influences the substrate reduction sites without interfering in the enzyme-catalyzed hydrolysis of MgATP.

Figure 2 shows that, with increasing CO, the reduction of azide (25 mM) correspondingly decreased. With azide at 65°C, NH4+ formation without CO was 1.61 (¿mole h"1 x mg"1. With 15% CO, however, it was only 0.19 (j.mole h"1 x mg"1, i.e. only 11.8% of that without CO. In an assay with N2 (control) in the absence of CO, the activity was 0.28 fimole h"1 x mg"1 and, in the presence of 15% CO, it was only 0.061 (¿mole h"1 x mg"1, which was 21.8% of that without CO. In both assays (with and without azide) with CO present, there was an increase in H2 evolved, which corresponded to the CO concentration. After 4 h incubation, the ratio of NH44" to H2 in the control assay without CO was 1:0.18, which changed with increasing CO. At 15% CO, the ratio was 1:1.6 (Figure 2), indicating that most of the electrons have been transferred to protons. A similar change in this ratio was observed with added azide from 1:0.015 to 1:0.45 without and with CO (15%), respectively. Although these results show a similar pattern of CO inhibition as for the "conventional" nitrogenase, the amount of CO required for inhibition was distinctly higher.

4. Cyanide as a Substrate

Cyanide can be also reduced by the Stl and St2 components (Figure 3). The maximum concentration of cyanide, which is tolerated by the Stl and St2

components, is 20 mM. Above this concentration, activity declined and, at 25 mM, ceased entirely.

5 10

% CO in assay vial

Figure 2. Effect of CO on NH4+ and H2 formation from azide by Stl and St2. Reaction mixtures were as in Figure 1, but with 25 mM NaN3 and various CO concentrations. NH4+ was analyzed after 4 h at 65°C under N2. ■, NH4+ formed without azide; □, H2 formed without azide; NH4+ formed with azide; O, H2 formed with azide. For NH4~, 1.6 |imol (h.mg)"1 and 0.28 jtmol (h.mg)"1 were set as 100% with and without azide, respectively. For H2, 0.13 |amol (h.mg)"1 and 0.43 umol (h.mg)"1 were set as 100% with and without azide, respectively.

Figure 3. Ammonium formation from cyanide catalyzed by purified Stl and St2 components. Reaction mixtures were as Figure 1 plus 10 mM KCN where indicated. Assays at 65°C under 100% N2 or He: ■, without cyanide under N2; with cyanide under N2; O, with cyanide under 100% He.

Figure 3. Ammonium formation from cyanide catalyzed by purified Stl and St2 components. Reaction mixtures were as Figure 1 plus 10 mM KCN where indicated. Assays at 65°C under 100% N2 or He: ■, without cyanide under N2; with cyanide under N2; O, with cyanide under 100% He.

At 10 mM KCN, the maximum activity was 0.42 pmole NH4

ammonium formation in the presence of either KCN or N2. The reaction occurred with similar efficiency in the presence and in the absence of N2

(Figure 3). After 6 h of incubation at

65°C, the total NH4+ formation with

KCN was 0.88 mM and with N2

(control) it was 0.45 mM, suggesting that cyanide may be preferred. KCN

was reduced to NH4+ with a Km of 5.21

mM and a Fraax of 16.7 nmole of NH4+

formed (min"1 x mg"1). These values fall into the range (Km of 1.1 mM to

4.5 and Vmax 113 to 158 nmole min"1 x -1

h"1 x mg"1. Figure 3 shows

% CO in assay vial

Figure 4. Influence of CO on NH4+ and H2 formation with cyanide present by the Stl and St2 components under N2. Reaction mixtures were as in Figure 1 with 10 mM KCN and CO as indicated. NH4+ was analyzed after 4 h at 65° C. ■, NH4+ formed without KCN; □, H2 formed without KCN; NH/ formed with KCN; O, H2 formed with KCN. For NH4+, 0.67 nmole (h.mg)"1 and 0.28 ¡xmol (h.mg)"1 were set as 100% with and without KCN, respectively. For H2, 0.11

mg ) reported for "conventional" nitrogenases (Li et al. 1982; Shen et al. 1997).

4.1. Influence of carbon monoxide on cyanide reduction. Increasing amounts of CO (up to 15%) hardly affected KCN reduction (Figure 4). In the absence of CO, NH4+ formation was 0.67 j-imol (h.mg)"1, whereas with 15% CO, it was 0.57 ^imol (h.mg)"1. Moreover, even 47% CO did not affect NH4+ formation (results not shown), however, FL formation increased; 4-fold with 15% CO.

5. Nitrite as a Substrate

Under appropriate assay conditions with dithionite as the electron donor, the nitrogenase components Stl and St2 are able to catalyze the reduction of NO2" to NH4+ linearly up to 9 mM N02". At 65°C, the specific activity of a control without NO2" was 0.82 |_miole NH4+ h"1 x mg"1, whereas with 9mM NO2", it was 1.60 ^imole NH4+ h"1 x mg"1. Higher concentrations were inhibitory and, with 35 mM N02", ammonium formation was 27% less than the control without NO2". Vaughn and Burgess (1989) observed that, at 60 mM, nitrite is an inhibitor of the Fe protein. In our experiments in the presence of MgATP and dithionite, the St2 protein alone was able to reduce NO2" slowly (0.3 jimole NH4+ h"1 x mg"1), even at a concentration of 35 mM NO2"

% CO in assay vial

Figure 4. Influence of CO on NH4+ and H2 formation with cyanide present by the Stl and St2 components under N2. Reaction mixtures were as in Figure 1 with 10 mM KCN and CO as indicated. NH4+ was analyzed after 4 h at 65° C. ■, NH4+ formed without KCN; □, H2 formed without KCN; NH/ formed with KCN; O, H2 formed with KCN. For NH4+, 0.67 nmole (h.mg)"1 and 0.28 ¡xmol (h.mg)"1 were set as 100% with and without KCN, respectively. For H2, 0.11

(imol (h.mg)"1 and 0.43 (imol (h.mg)"1 and without KCN, respectively.

were set as 100% with

Figure 5. Ammonium formed from azide (10 mM) catalyzed by CODH. Reaction mixtures contained 2 mL of KH2P04/Na0H buffer (50 mM, pH 7.5), 10 mM dithionite and 2.8 mg CODH. Assays were under 100% N2 65°C. CODH + dithionite + N3; □, CODH + N3"; CODH + dithionite; A, dithionite + N3".

Figure 5. Ammonium formed from azide (10 mM) catalyzed by CODH. Reaction mixtures contained 2 mL of KH2P04/Na0H buffer (50 mM, pH 7.5), 10 mM dithionite and 2.8 mg CODH. Assays were under 100% N2 65°C. CODH + dithionite + N3; □, CODH + N3"; CODH + dithionite; A, dithionite + N3".

, suggesting that 35 mM NO2" inhibited only the activity of the Stl protein. NO2" was reduced to NH4+ with a Km of 2.79 mM and a Vmax of 27.9 nmole of NH4+ formed (min.g)"1.

6. Reduction of Azide by Carbon Monoxide Dehydrogenase (St3 protein)

Using appropriate conditions with dithionite as electron donor, the purified CODH of S. thermoautotrophicus is capable of catalyzing the reduction of azide to NH4+ and N2 (Figure 5). After 25 h at 65°C, total NH4+ measured was 10.18 mM. In control vials, where only CODH or CODH plus dithionite or CODH plus N3" were present, neither NH4+ nor N2 was formed.

During azide (5 mM) reduction, 5.16 mM NH4+ and 5.20 mM N2 were formed. According to Dilworth and Thorneley (1981), azide can be reduced by one of the following three equations:

Because NH4+ and N2 were formed in a stoichiometric ratio, this enzyme appears to operate in accord with reaction (1).

7. Discussion

Many alternative substrates are known for the "conventional" nitrogenases. The structural genes and cofactors of all these nitrogenases are highly similar. The oxygen-insensitive nitrogenase of S. thermoautotrophicus (St-nitrogenase) is genetically and structurally different from "conventional" nitrogenases (Hofmann-Findeklee et al. 2000). The primary sequence of Stl shows high homology to molybdenum-containing hydroxylases, especially with CODH from Oligotropha carboxidovorans, Hydrogenophaga pseudoflava and Pseudomonas thermocarboxydovorans (Hofmann-Findeklee et al. 2000). St2 is highly homologous to Mn-containing SODs (Hofmann-Findeklee et al. 2000). A functional difference between the Stl/St2 system and the "conventional" nitrogenase is the inability of the former to reduce acetylene. Therefore, it was not obvious that the St-nitrogenase would reduce substrates of the "conventional" nitrogenases, but azide, cyanide and nitrite will indeed serve as substrates for St-nitrogenase. However, there are some differences.

First, the CO concentration (15%) required for inhibition of azide reduction was much higher than for the "conventional" nitrogenases (0.5-2.0%). Second, even with 47% CO, the reduction of cyanide was hardly inhibited. Third, nitrite did not inhibit St2 activity; on the contrary, St2 is able to reduce nitrite to NH4+. Fourth, the results obtained with azide were more or less consistent with the results known from other nitrogenases (Figure 1, Figure 2). Fifth, CO inhibits the reduction of azide to NH4+ by St-nitrogenase and also simultaneously stimulates the production of H2. However, the CO concentrations required for inhibition were much higher than for "conventional" nitrogenases, indicating that, for St-nitrogenase, CO appears to act not as a non-competitive inhibitor, but as a competitive inhibitor. Inhibition by CO is dependent on the ratio of azide and CO concentrations. Sixth, CO is a non-competitive inhibitor of cyanide reduction by "conventional" nitrogenase, however, 47% CO hardly affected cyanide reduction by St-nitrogenase. It appears that CO probably does not function as an inhibitor of St-nitrogenase but does stimulate H2 evolution (Figure 4).

CO can produce many different effects. An altered A. vinelandii MoFe protein, where aGly69 was replaced by serine, suffers competitive inhibition of substrate reduction rather than noncompetitive by CO (Christiansen et al. 2000). Also, a mutant of A. vinelandii, where the MoFe protein a-glutamine-191 residue was replaced by lysine, exhibited 50% inhibition of proton reduction by CO (Kim et al. 1995). Furthermore, at low concentrations, CO acts as a stimulant for V-nitrogenase of A. vinelandii (Cameron, Hales 1996). In contrast, S. thermoautotrophicus is a carboxidotrophic bacterium that grows under 45-50% CO and, as expected, CO does not exert much, if any, inhibitory effect on St-nitrogenase as exemplified by the high CO concentration required for inhibition of azide reduction. Thus, azide and cyanide may not bind with similar affinity to the Stl component. Nitrite reduction by Stl is consistent with a previous report (Vaughn, Burgess 1989), however, although nitrite inhibits the Fe-protein activity of "conventional" nitrogenase by affecting its 4Fe-4S center, the St2 component of St-nitrogenase, which does not contain Fe-S centers, is not inhibited by nitrite. On the contrary, the St2 protein alone was able to reduce nitrite to NH/.

The CODH (St3 protein) can also reduce azide to NFLt+. The difference between St3 and Stl is that, for substrate reduction, Stl requires St2 plus MgATP and dithionite, whereas St3 requires only dithionite for azide reduction. Because the structural genes of Stl and St3 are highly homologous, it is not surprising that St3 can also reduce azide to NH4+. Possibly many other compounds, which serve as substrates for nitrogenases, could be reduced by St-CODH. However, in general, CODHs of aerobic bacteria have oxidizing characters whereas nitrogenases are reductive. The Stl component is a heterotrimeric monomer and resembles at least one of the CODH species occurring in S. thermoautotrophicus. Similar to the "conventional" nitrogenases, Stl can reduce N2, azide, cyanide, and nitrite. The St-nitrogenase is entirely different from the "conventional" nitrogenases (Ribbe et al. 1997) and it remains of interest to know why these two enzymes, in spite of these differences, react more or less similarly with respect to reduction of these substrates.

8. References

Bulen WA et al. (1965) In San Pietro (ed) Non-heme Iron Protein: Role In Energy Conversion, pp. 107-112, Antioch Press, Yellow Springs, OH Burgess B, Lowe DJ (1996) Chem. Rev. 96, 2983-3011 Cameron LM, Hales BJ (1996) J. Am. Chem. Soc. 118, 279-280 Christiansen J et al. (2000) J. Biol. Chem. 275, 36104-36107 Dilworth MJ, Thorneley RNF (1981) Biochem. J. 193, 971-983 Dobbek et al. (1999) Proc. Natl. Acad. Sci. USA 96, 8884-8889 Gadkari D et al. (1990) Appl. Environ. Microbiol. 56, 3227-3234 Gadkari D et al. (1992) J. Bacterid. 174, 26627-26633 Gremer L et al. (2000) J. Biol. Chem. 275,1864-1872 Gremer L et al. (2001) Biospectrum

Hardy RWF et al. (1965) Biochem. Biophys. Res. Commun. 20, 539-544 Hofmann-Findeklee C et al. (2000) In Pedrosa FO, Hungaria M, Yates G, Newton WE (ed) Nitrogen Fixation: From Molecules to Crop Productivity, pp. 23-30, Kluwer, Dordrecht Hwang JC et al. (1973) Biochem. Biophys. Acta 292, 256-270 Kelly M et al. (1967) Biochem. J. 102, lc-3c Kim CH et al. (1995) Biochem. 34, 2798-2808 Kraut M et al. (1989) Arch. Microbiol. 152, 335-341 Li JG et al. (1982) Biochem. 21, 4393-4402 Meyer OM, Schlegel GH (1980) J. Bacteriol. 141, 74-80

Meyer OM et al. (1993) In Murrell JC, Kelly DP (ed) Microbial Growth on Ci Compounds, pp. 83-90, American Society for Microbiology, Washington, DC Meyer OM et al. (2000) Biol. Chem. 381, 865-876 Pham DN, Burgess BK (1993) Biochem. 32, 13725-13731 Ribbe M et al. (1997) J. Biol. Chem. 272, 26627-26633 Rivera-Ortiz JM, Burns RH (1975) J. Bacteriol. 123, 537-545 Schollhorn R, Burris RH (1967) Proc. Natl. Acad. Sci. USA 57, 1317-1323 Shen J et al. (1997) Biochem. 36, 4884-4894 Vaughn SA, Burgess BK (1989) Biochem. 28, 419-424

Section 2: Bacterial Genomics

Was this article helpful?

0 0

Post a comment