Denitrification by the Fungus Fusarium oxysporum and Involvement of Cytochrome P-450 in the Respiratory Nitrite Reduction*

From conditions for production in Fusarium oxysporum of the unique nitratelnitrite-inducible cyto- chrome P-450, tentatively called P - 4 5 0 d N I R , it was expected that the fungus is capable of metabolizing ni- trate dissimilatively. Here we report that F. oxysporum exhibits a distinct denitrifying ability which results in the anaerobic evolution of nitrous oxide (NzO) from nitrate or nitrite. Comparison of the cell growth during denitrification indicated that the dissimilatory reduction of nitrate to nitrite is an energetically favorable process in F. oxysporum; however, further reduction of nitrite to N2O might be energy-exhausting and may function as a detoxification mech-anism. A potent nitrite reductase activity to form N20 could be reconstituted by combination of the cell-free extract prepared from the denitrifying cells and an NADH-phenadinemethosulfate-dependent reducing system. The activity was strongly inhibited by carbon monoxide, cyanide, oxygen ( 0 2 ) , and the antibody against P-45OdNIR. The results, along with those con-cerning inducing conditions of P’&OdNIR, were highly indicative that the cytochrome is involved in the deni- trifying nitrite reduction. This work has thus pre-sented not only the first demonstration that a eukaryote exhibits a marked denitrifying ability, but also the first instance of a cytochrome P-450 that is involved

by high aeration (6) but not by simultaneous addition of ammonium ions (7), suggesting a critical role for the hemoprotein in a dissimilatory metabolism of nitrate/nitrite by the fungal cells. Among eukaryotic organisms only the fungus Neurospora crassa has so far been suggested to catalyze the dissimilatory reduction of nitrate to nitrite, but it cannot reduce nitrite further to gaseous forms of nitrogen (9) and thus it does not result in denitrification. In the case of F. oxysporum, however, P -4 5 0 d N I R is induced by nitrite as well as by nitrate, suggesting involvement of the P-450 in further reduction of nitrite. These facts prompted us to investigate whether the fungus exhibits denitrifying ability.
Here we report that F. oxysporum exhibits a potent denitrifying ability to reduce nitrate or nitrite to N20 and that P-4 5 0 d~1~ might be involved in the anaerobic, respiratory nitrite reduction.

MATERIALS AND METHODS
Fungal Strain-F. oxysporum MT-811, previously isolated from soil and identified (IO), was used throughout this work. Prior to following experiments the fungal strain was purified by a single spore isolation followed by a submerged culture in the presence of antibiotics (tetracycline, streptomycin, and ampicillin; 10 pg/ml each).
Cultiuation-F. oxysporum was cultivated (incubated) under conditions with limited aeration to examine denitrifying ability. The fungus was incubated in 500-or 300-ml volume Erlenmeyer flasks with two side arms containing 150 or 90 ml (final) of a medium that consisted of 3% glycerol, 0.2% peptone, 10 mM sodium nitrate (or nitrite), and inorganic salts (6,8), on a rotary shaker (150 rpm) a t 26.5 "C. The seed was prepared by cultivation for 5 days in 5-liter Erlenmeyer flasks containing 2 liters of the same medium without nitratejnitrite and inoculated by one-fifth of the frnal volume (30 ml of the 150-ml culture). The top and side arms of the flask were sealed with a butyl rubber stopper and double butyl rubber stoppers, respectively, after inoculation. When necessary, the air in flasks was substituted just after inoculation by helium or carbon monoxide (CO) gas that was passed through a sterilized cotton filter.
Sampling-At each indicated time during the time course of cultivation, three flasks were harvested. Mycelia were collected by filtration (filter paper) and washed with distilled water. A portion of mycelia was dried by heating a t 90 "C for 1 day and weighed. Other portions were disrupted and fractionated (see below), and then assayed for P-450aNiR (soluble fraction) or nitrate and nitrite reductases. The first filtrate (broth) was further filtrated with a microfilter (Millipore, UFC 30GVO0) to remove proteins, and residual nitrate or nitrite was determined.
Gas Analyses-The upper space gas of cultivation flasks or of the reaction vessel for reductase assays was analyzed by gas chromatography (GC) and gas chromatography-mass spectrometry (GC-MS) with a Shimadzu gas chromatograph GC 12A and with a Shimadzu gas chromatograph-mass spectrometer GCMS-9000C, both of which were equipped with a Porapack Q column (3 mm X 2 m (inner diameter) for analysis of NeO) and a molecular sieve 5A column (same size, for N, and O?), a thermal conductivity detector, and helium carrier, respectively. Each sample gas was taken off through the double butyl rubber stopper with a syringe and applied to analyses. Each peak was identified and quantified by comparison with standard gases (GL Sciences, Tokyo).
Fractionation of the Cell-free Extract-The fungal cells were disrupted as reported (8,11) by grinding with alumina, and the extract was fractionated by differential centrifugation (first at 10,000 X g, and then at 105,000 X g). Each precipitate was washed once, filled up to the same volume as the supernatant with the buffer used for disruption, and utilized for each assay.
Nitrate and Nitrite Reductase Assays-Cell-free nitrate and nitrite reductase activities were determined as follows employing an NADHphenadinemethosulfate (PMS) or reduced methylviologen (MVH) system as an electron donor. The reaction mixture (final, 10 ml) containing 1 ml of a subcellular fraction, 2.0 mM sodium nitrite, 5% glycerol, 0.02 mM PMS, or 0.1 mM methylviologen, in 60 mM potassium phosphate buffer (pH 7.5), was put in a vial (27-ml volume) and degassed. The vial was sealed with a butyl rubber stopper under flush of helium gas. The reaction was initiated by adding with a syringe 1 ml of a solution containing 50 mM NADH or 1% sodium hydrosulfite, 1% sodium bicarbonate, at 30 "C. The increase in the gas phase of evolved gas (such as N,O) and/or the decrease of nitrate or nitrite in the reaction mixture was determined.
Determination of Nitrate, Nitrite, and Ammonium Ions-The concentration of nitrate or nitrite in culture media or the reaction mixture was determined with a high performance liquid chromatograph (HPLC; Toso CCP and 8010 Series) equipped with a TSK gel IC-Anion PW column. Eluting solution was 50 mM boric acid containing 0.7 mM sodium I-octanesulfonate. The anions were monitored with a UV detector at 210 nm. Ammonium ions were determined colorimetrically (12).
Other Determinations-P-450 was determined according to Omura and Sat.0 (13); that is, the difference spectrum of (dithionite-reduced plus CO) minus (dithionite-reduced). Protein was determined by the method of Lowry et al. (14). The antibody against P -4 5 0 d~1~ was prepared from rabbit as described previously (11). Nonimmune rabbit I& was obtained from Sigma.

N,O Evolution by Intact Cells from
Nitrate-Since dissimilatory nitrate reduction is usually an anaerobic process, F. oxysporum was cultivated (incubated) in a nitrate-containing medium under the condition where oxygen supply was discontinued, and its denitrifying ability was examined. After oxygen that was derived from air and initially present in the flask was consumed, a rapid N 2 0 evolution was observed, as shown in Fig. 1. Added nitrate was stoichiometrically (i.e. with 100% yield) converted to N20. Slight (as compared with that due to oxygen) but distinct cell growth was observed during the incubation period in which the N20 evolution was continuing, suggesting that the dissimilatory nitrate reduction depends on an energy-yielding system (respiration). The extent of growth during the nitrate respiration seems reasonable on considering the lower amount of nitrate in the medium (total, 1.5 mmol) as compared with that of oxygen (initial, -4 mmol). N 2 0 evolution was preceded by the onset of nitrate uptake and induction of both P-450dNIR and the cell-free N20-forming nitrite reductase (dNIR) activity (cf. Fig. 3) in the cells. The concomitant induction of P -4 5 0 d~~~ and dNIR activity suggested that the P-450 is connected with the activity. The N 2 0 evolution was not observed at all when nitrate was omitted from the medium (not shown). The results clearly showed that the fungus exhibits a potent denitrifying ability.

N20 Evolution by Intact Cells from Nitrite-A similar result
was obtained upon incubation of the fungal cells with a nitritecontaining medium (Fig. Z ) , although N20-evolving activity was somewhat lower. Nitrogen atoms in N,O was shown derived from nitrite by the GC-MS analysis (cf. Fig. 4). More potent activity (ie. 100% conversion) was observed when air in the flask was replaced by helium before incubation (not shown). In contrast to the growth on the nitrate medium above, however, dry cell matter began to decrease when the N,O evolution was initiated. It seems therefore that the denitrifying process from nitrite to N 2 0 is not an energyyielding but an energy-exhausting process. The N20 evolution was strongly inhibited or repressed by CO or by intermittent supply of oxygen. The inhibition by CO is a unique feature of the fungal denitrification, which is consistent with the hypothesis that P-450dNIR is involved in the process.
These results (Figs. 1 and 2) unequivocally demonstrated that the fungus F. oxysporum can metabolize both nitrate and nitrite in a dissimilative manner, resulting in denitrification.

Fungal Denitrification
This is the first demonstration that eukaryotic cells exhibit a distinct denitrifying activity. It is also clear that the activity is inducible with nitrate or nitrite only when oxygen applicable to aerobic respiration is exhausted or not present.
Cell-free Nitrite Reductase Activity-Since a potent denitrifying ability of F. oxysporum was demonstrated above, we then sought to detect a cell-free dNiR activity. As shown in Fig. 3, the cell-free extract (10,000 X g supernatant; cf. Table  I) obtained from the denitrifying cells exhibited a potent, NADH-PMS-dependent nitrite reductase activity. The substrate nitrite was stoichiometrically converted to NzO within 20-30 min. The initial rate was approximately 1,000 min" (reduction of nitrite) or 500 min" (N,O evolution) with respect to involved P -4 5 0 d~1~. MVH could also serve as an electron donor (not shown). Fig. 4 shows results of identification of the product. Nitrogen atoms in the product N 2 0 were proved to be derived from nitrite by GC-MS (m/z = 46), employing a heavy isotope (["Nlnitrite) as substrate. Other possible intermediates or products, such as nitric oxide (NO) or Nz, could not be detected. This was also the case with the denitrification by intact cells above (Figs. 1 and 2; data not shown). The activity did not appear in the absence of the artificial electron donor. The cell-free extract, obtained from  Table I) was prepared from the denitrifying cells (Fig. l), which contained 1.0 P M P-45OdN1~. The assay was done at one-half of the scale of that in A , GC analyses after the reaction of the cell-free extracts from denitrifying (left; from the culture in Fig. 1) and non-denitrifying (right; from the 0-min-incubated cells in Fig. 1 non-denitrifying cells (the seed culture) and thus containing no P-450dNIR, did not possess NzO-forming activity at all (Fig.  44). As was expected, the cell-free activity did not produce detectable amounts of ammonium ions, the product of assimilatory reduction of nitrite (not shown).
The activity was strongly inhibited by CO, cyanide, or O2 (Fig. 3). These substances usually interact with heme-containing terminal oxidases or reductases. When purified P-4 5 o d~1~ was incubated with the NADH-PMS-reducing system in the same manner, a slight but distinct amount of N 2 0 was formed (Fig. 3). In addition, the rate of NzO evolution was somewhat increased by combination of the active extract and purified P-450dNIR (not shown). It is therefore highly possible t h a t P -4 5 0 d~1~ is the nitrite reductase itself. The low recovery of product by the purified P -4 5 0 d~l~ system is probably due to lack of supporting component(s). Cytochromes in the Extract-Since the soluble fraction above (Fig. 3; cf. Table 1) was fully active in forming N20, cytochrome components in the fraction seemed worth determining (Fig. 5). It is rather surprising that the soluble fraction from denitrifying cells contained a large amount of cytochromes. The spectra for non-denitrifying cells were typical of cytochrome b6 (15), suggesting that the cytochrome is not involved in the denitrification. The spectra for denitrifying cells indicated that the cells contained higher amounts of bor C-type cytochromes and CO-binding heme protein (P-420) in addition to P-450. This would suggest the possible involvement of cytochrome b or c in the nitrite reductase system in addition to P-450.
Dissimilatory Nitrate Reductase (dNaR) Activity-Along with dNiR activity above, a distinct NADH-PMSor MVHdependent dNaR activity could be detected only in the large particle fraction (10,000 X g precipitate; cf. Table I), as shown in Fig. 6. Although the rate of N 2 0 evolution was lower than that due to the nitrite reductase system (Fig. 3), a distinct NzO evolution could be observed when the NADH-PMSreducing system was employed. In contrast, only a minor amount of N 2 0 was evolved as compared with decreased nitrate when MVH was used. The discrepancy in stoichiometry between substrate and product might depend on trapping of intermediates or products by some nonenzymatic reactions caused by the potent reducing force of dithionite in the system. Nitrite, the expected product due to dNaR reaction, could also be detected to form in the MVH-dependent system.  Table I and diluted with 60 mM potassium phosphate buffer (with 20% glycerol and 2 mM mercaptoethanol) by 5-fold (reduced minus oxidized) or 10-fold (CO, inset). The reference in the CO spectra was not reduced to avoid interference by cytochrome oxidase. FIG. 6. Cell-free nitrate reductase activity. The reaction was done in the same manner as in Fig. 3, except that nitrite was replaced by the same amount (10 pmol) of sodium nitrate and that the extract was replaced by the large particle fraction (10,000 X g, precipitate). Nitrate or nitrite in the reaction mixture was determined by HPLC, and the decreased amount of nitrate or the amount of formed nitrite was indicated (left anis). Symbols: 0, MVH-dependent nitrate reduction; (3, nitrite formation due to the nitrate reduction by MVH; 0 , N,O evolution due to the MVH system; A, N20 evolution due to the NADH-PMS system. The nitrite level was lower than that of utilized nitrate. The discrepancy seems reasonable, however, because a portion of the formed nitrite might have been utilized for the subsequent dNiR reaction. The N20-evolving activity of the particulate fraction indicated that all components essential for reduction of nitrate to N 2 0 are connected with the large particles. Localization of Nitrate and Nitrite Reductase Actiuities-The highly active cell-free extract was fractionated by centrifugation, and both reductase activities were determined with each subcellular fraction ( Table I). As noted above, nitrate reductase activity was recovered only in the large particle fraction. In contrast, nitrite reductase activity was recovered in both the soluble (105,000 x g supernatant) and the large particle fractions. This suggests that components corresponding to the activity are connected to or have interactions with particles and that they are easily released into the soluble fraction. The cell-free activity shown above (Figs. 1-4) thus depended on these soluble (or solubilized) components; however, the reconstituted activity was highly active (Fig. 3).
Inhibition of Nitrite Redu.ctase by Antibody-The involvement of P-450dNIH in the dissimilatory nitrite reductase activity was conclusively shown by the inhibition experiment with the antibody that was raised against purified P-450dNIK (Fig.   7).
Other Inhibitors-It has been shown from the above results that the cell-free nitrite reductase activity was inhibited by CO, 02, and cyanide, and that the N20 evolution by intact cells was inhibited or repressed by CO and 0,.
The effects of other inhibitors were also examined ( Table  11). Antimycin A, azide, salicylhydroxamic acid (an inhibitor of cyanide-insensitive respiration; Ref. 16), metyrapone (a P -FIG. 7. Inhibition of the cell-free nitrite reductase activity by antibody. The dNiR activity in the soluble fraction was assayed as in Fig. 3 in the presence of indicated amounts of antibodies raised against P -4 5 0 d~r~ (0) or of nonimmune IgG fraction of rabbit (O), and the initial N20-forming rate was compared. Prior to the assay, the extract and IgG were preincubated a t 0 "C for 24 h.

TABLE I1
Inhibitors against dNiR and denitrification The indicated amount of each inhibitor was added to the assay syst,ems, and t,he extent of inhibition was expressed as a percentage (i.e. 0, no inhibition and 100, complete inhibition). The denitrifying activity was determined with respect to intact cells as in Fig. 2 using nitrite as substrate, except that the air in flasks was replaced by helium (with the exception of the CO inhibition). dNiR was assayed as in Fig. 3 and Table I. 450 inhibitor), and menadione (an inhibitor against NADPHcytochrome P-450 reductase system) showed little inhibition of the cell-free activity. The lack of inhibition by metyrapone or menadione is not surprising, considering the uniqueness of the electron transport system of P-450dNIR. Although antimycin did not inhibit the cell-free activity at all, it inhibited to a considerable extent the N 2 0 evolution from nitrite by intact cells. The results suggest that the nitrite reduction is connected to some respiratory chain, although the process does not seem energetically favorable. This is consistent with the above observation that nitrite reductase was connected to large particles.
Cyanide, CO, and oxygen strongly inhibited the cell-free activity, in contrast to the inhibition by antimycin. This can be explained by that P-450dNIR is the nitrite reductase itself, as noted above. Inhibition by cyanide is one of unique features of P-450dNIK. This is consistent with our results that cyanide induced a spectral change in the P-450 which closely resembled that caused by cyanide in cytochrome oxidase.2 Oxygen might exhibit both repression (Figs. 1 and 2) and inhibitory effects (Fig. 3)  the intact cell activity than the cell-free activity. This may depend on rather high cell population in the culture.

DISCUSSION
The results reported above provide evidence that the fungus F. oxysporum is capable of reducing nitrate and nitrite to a gaseous form of N,O when oxygen supply is limited or discontinued and that P-450dNIR might be involved in this process. It is also likely that the denitrification from nitrate is for a dissimilatory purpose and thus is, as a whole, an energyyielding process. On the other hand, the denitrification did not seem energetically favorable when nitrite was employed as substrate. It may function as a detoxification mechanism. These features somewhat resemble those of the denitrifying bacteria, Propionibacterium acnes, an obligate anaerobe (17). It has been reported that several yeasts and fungi can form N 2 0 from nitrate (18). The rate (requiring more than 10 days) and recovery (below 0.2%) of nitrate-form nitrogen are, however, far from physiologically significant in these cases. Therefore, F. oxysporum is the first eukaryote that has been shown to catalyze denitrification, an anaerobic process.
We did not find significant evolution of N, in this study due to the fungal dissimilatory metabolism of nitrate or nitrite. It will be important to conclude by additional studies whether the fungus can further reduce N,O to N2 or not, although the fungus does not seem to possess the ability in so far as the present results are employed for the judging.
The present results also suggested a novel function of P-450, which has been known as a group of hemoproteins catalyzing a variety of monooxygenation reactions (19)(20)(21).
Although the reconstitution experiment employing purified components is required to show conclusively the real function of P-450dNIR, the present results are highly indicative that the P-450 is a dissimilatory nitrite reductase and thus must not be a monooxygenase. it must be connected with the fungal denitrification, an anaerobic process. Involvement of P-450 in reductive reactions has been known (22)(23)(24); however, physiological significance of these reactions has not been fully elucidated. In contrast, the reducing system by P-450dNIR unequivocally has physiological significance and is quite unique with many respects. For example, it appears to be connected to the respiratory chain and thus composed of quite different electron transport system from that of the traditional P-450 system; the substrate nitrite is a hydrophilic and inorganic substance that is quite different from so far known, hydrophobic, organic substrates.
We have recently succeeded in obtaining a cDNA clone for P-450,]N1H."J The deduced amino acid sequence revealed many interesting properties of the P-450, i.e. that P-450dNIR exhibited higher sequence homology against bacterial P-450s (in particular those of Streptomyces (Ref. 25) rather than against eukaryotic P-450s, and that the N terminus region contained neither the signal peptide-like, hydrophobic domain that is characteristic of microsomal P-450s nor the tagging sequence that is essential for the localization of mitochondrial P-450s. These results strongly support our previous claim that P-450aNIR is the first soluble P-450 of eukaryotes (6). The result along with the present data have promised us unique features of the P-450 to be clarified in many respects with high potential interests. It remains to be elucidated how the soluble P-450 interacts with organella such as mitochondria, although some interaction has been demonstrated above (Table I, Fig.  6).
We have recently found that several strains of F. oxysporum and related genera produce similar, nitrate/nitrite-inducible P-450 and that these P-450 species show properties immunologically and spectrally similar to those of P-450dNrR (7). It seems highly possible that at least these fungal strains also exhibit a similar denitrifying ability. It would be also ecologically important to study the distribution among eukaryotic microorganisms of such an ability to evolve the greenhouse effect gas (N20), since the recent tendency toward acidification in soil will potentiate the activity of fungi.