Denitrification, a novel type of respiratory metabolism in fungal mitochondrion.

Subcellular localization and coupling to ATP synthesis were investigated with respect to the denitrifying systems of two fungi, Fusarium oxysporum and Cylindrocarpon tonkinense. Dissimilatory nitrate reductase of F. oxysporum or nitrite reductase of C. tonkinense could be detected in the mitochondrial fraction prepared from denitrifying cells of each fungus. Fluorescence immunolocalization, cofractionation with mitochondrial marker enzymes, and cytochromes provided evidence that the denitrifying enzymes are co-purified with mitochondria. Respiratory substrates such as malate plus pyruvate, succinate, and formate were effective donors of electrons to these activities in the mitochondrial fractions. Moreover, nitrite and nitrate reduction were shown to be coupled to the synthesis of ATP with energy yields (P:NO3− or P:2e ratios) of 0.88 to 1.4, depending upon whether malate/pyruvate or succinate were provided as substrates. Nitrate or nitrite reductase activity was inhibited by inhibitors such as rotenone, antimycin A, and thenoyltrifluoroacetone. Thus, fungal denitrification activities are localized to mitochondria and are coupled to the synthesis of ATP. The existence of these novel respiration systems are discussed with regard to the origin and evolution of mitochondria.

Denitrification is a process whereby nitrate or nitrite is reduced to a gaseous form of nitrogen (N 2 or N 2 O). Biological denitrifying activities, which play important roles in the global nitrogen cycle, were believed to be uniquely characteristic of prokaryotes (1-3) until we discovered their presence in several fungi (4 -6). Whether the fungal denitrification systems resemble bacterial systems for anaerobic respiration and, if so, whether they are localized to mitochondria has not been established. Bacterial denitrification usually involves four sequential reactions, which are catalyzed by nitrate reductase (Nar), 1 nitrite reductase (Nir), nitric oxide (NO) reductase (Nor), and nitrous oxide (N 2 O) reductase, respectively (1-3). Among fungal denitrifying systems, those of Fusarium oxysporum and Cylindrocarpon tonkinense have been best characterized (6 -11). Fungal systems usually lack N 2 O reductase, and consequently evolve N 2 O as the final denitrification product (4 -6). Many systems such as that of C. tonkinense also lack Nar. However, many bacterial denitrifying (or nitrate-respiring) systems may also lack part of the reductase chain (1)(2)(3). Involvement of cytochrome P450 (P450nor) for Nor is typical in the fungal systems that have been analyzed to date (4 -12), whereas bacterial Nor is of the cytochrome bc type (1)(2)(3). In contrast, Nir isolated from F. oxysporum is very similar to its copper-containing counterparts of bacteria (11). Further, denitrification by C. tonkinense has been shown to support anaerobic cell growth (6). These findings are highly indicative that fungal denitrification is an energy-yielding process like the bacterial counterpart.
We proposed previously that the reactions catalyzed by Nar in F. oxysporum and by Nir in C. tonkinense might be coupled to the synthesis of ATP (4,6,11). We now demonstrate the intracellular localization of the reductases, characterize their electron donors, and show they are coupled to the synthesis of ATP.

EXPERIMENTAL PROCEDURES
Fungal Strains-F. oxysporum MT 811, first identified as a fungal denitrifier (4), and C. tonkinense IFO 30561, that was shown to grow under denitrifying conditions (6), were used throughout this work.
Subcellular Fractionation of F. oxysporum-Denitrifying cells (5 g, wet weight), cultivated in a glycerol-containing medium under microaerobic condition (4,11), were disrupted by grinding in a mortar on ice with quartz sand (12.5 g) and 10 ml of sucrose buffer (0.8 M sucrose, 2 mM EDTA, 0.1% bovine serum albumin, and 10 mM Tris-HCl, pH 7.2) that contained 0.3 mM phenylmethylsulfonyl fluoride (PMSF). After addition of another 10 ml of the same buffer, the slurry was centrifuged 3 times at 1,500 ϫ g for 15 min to remove the quartz sand and undisrupted cells. The subsequent pellet obtained by two sequential centrifugations at 10,000 ϫ g for 40 min was suspended in the sucrose buffer and then subjected to ultracentrifugation in a discontinuous sucrose density (43.1, 50.4, and 68.9%) gradient, at 105,000 ϫ g for 180 min. The 10,000 ϫ g supernatant was further fractionated into the soluble and microsomal fractions by centrifugation at 150,000 ϫ g for 60 min. Each fraction obtained was incubated under aerobic conditions for 60 min at room temperature in the sucrose buffer containing 4 mM ADP, 4 mM MgCl 2 , and 10 mM sodium phosphate to deplete endogenous electron donors. Finally, each fraction was pelleted by centrifugation at 150,000 ϫ g and subsequently resuspended in the sucrose buffer.
Subcellular Fractionation of C. tonkinense-Denitrifying cells were cultured under initially aerobic condition (6). Subsequently, glucose in the reported culture medium was replaced with an equivalent level of glycerol. Cells were disrupted as above except that the buffer was 10 mM Tris-HCl, pH 7.4, containing 0.6 M mannitol, 0.1 mM EDTA, 0.1% bovine serum albumin, 0.3 mM PMSF, 0.3 mM leupeptin, and 0.5 mM tosylphenylalanyl chloromethyl ketone. Subcellular fractions were prepared as in the case of F. oxysporum above except that the 10,000 ϫ g pellet was fractionated in a sucrose density gradient of 25, 35, 42, and 53%.
Fluorescence Microscopy-All manipulations were performed at room temperature. A drop of the putative mitochondrial fraction from F. oxysporum was smeared on a glass slide, air-dried, and fixed for 10 min with 4% paraformaldehyde in phosphate-buffered saline (10 mM sodium * This work was supported by a grant for Research Projects from the University of Tsukuba and Grant 05556013 from the Ministry of Education, Science, Culture, and Sports of Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. § To whom correspondence should be addressed. Tel.: 81 298 53 4603; Fax: 81 298 53 4605. 1 The abbreviations use are: Nar, nitrate reductase; Nir, nitrite reductase; Nor, nitric oxide reductase; P450, cytochrome P450; P450nor and P450nor2, cytochrome P450 nitric oxide reductase and its isozyme; PMSF, phenylmethylsulfonyl fluoride; PMS, phenazine methosulfate; TMPD, N,N,NЈ,NЈ-tetramethylphenylenediamine; MVH, reduced methylviologen; TTFA, thenoyltrifluoroacetone; Mes, 4-morpholineethanesulfonic acid. phosphate, pH 7.2, 150 mM NaCl). The sample was washed for 10 min in phosphate-buffered saline and blocked with BAT (10% BlockAce (Dainihonseiyaku Inc., Osaka, Japan), 0.1% Triton X-100) for 30 min. Rabbit antiserum against Nir was diluted 1:200 with BAT and incubated with the sample for 2 h. The sample was then washed for 15 min in TBS-TX (10 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.1% Triton X-100) and incubated for 1 h with a drop of Texas Red-linked antibodies against rabbit IgG raised in a donkey (Amersham Corp., Bucks, UK) and diluted 1:100 with BAT containing 1 g/ml Hoechst 33342 (Hoechst, Germany). The sample was washed for 15 min in TBS-TX, rinsed in TBS (TBS-TX without Triton X-100), and mounted in antifade solution (0.1% paraphenylenediamine, 10 mM Tris-HCl, pH 7.4, 150 mM NaCl, 90% glycerol). Confocal fluorescence images were obtained with a laser scanning confocal system (MRC-500; Bio-Rad) attached to a microscope with epifluorescence optics (BHS-RFC; Olympus, Tokyo, Japan). Fluorescence due to Hoechst 33342 was recorded with a CD camera (model XC-73CE; Sony, Tokyo, Japan) and the MRC-500 system. The antiserum against Nir was produced by immunizing a rabbit with purified Nir from F. oxysporum (11) (Biologica, Nagoya, Japan). The specificity of the antiserum was confirmed by Western blot analysis (see Fig. 4C). A negative control was also prepared in which antiserum against Nir was replaced by preimmune rabbit serum. No staining with Texas Red-conjugated secondary antibody was observed (data not shown).
Determination of ATP-ATP was quantitated by the luciferin-luciferase bioluminescence technique with a Lucifer-LU kit (Kikkoman, Chiba, Japan) and an ATP photometer (NU-620; Niti-ion, Tokyo, Japan). An aliquot (0.1 ml) was taken for each time point and added to 0.4 ml of 0.125 N perchloric acid on ice. The solution was then neutralized with 0.5 ml of 0.1 N NaOH. After centrifugation (10,000 ϫ g, 60 min, 4°C) the sample was diluted to a concentration below 1 M ATP with 20 mM HEPES-NaOH buffer (pH 7.55) that contained 10 mM MgSO 4 and 2 mM EDTA. The amount of ATP synthesized is indicated as the difference between the total amount at each time point and that at 0 time.
Assays of Nar and Nir-Nar activity was detected by quantitation of its product, nitrite, by the method of Nicholas and Nason (13). The reaction mixture for Nar (final volume, 2 ml) contained 10 mM sodium nitrate, 50 mM methylviologen, 2 mM dithionite, and a fraction that contained Nar in 100 mM potassium phosphate buffer (pH 6.7). Reactions were incubated anaerobically at 30°C. For assays of mitochondrial Nar, each fraction was incubated anaerobically at 25°C in the sucrose buffer plus 10 mM sodium nitrate, 4 mM MgCl 2 , 10 mM sodium phosphate, 4 mM ADP, and an electron donor (5 mM) (total volume, 2 ml). The anaerobic technique was described previously (11). The gas phase was replaced with carbon monoxide to suppress the activities of Nir and cytochrome oxidase.
Nir activity derived by physiological electron donors was assayed by quantitating its reaction product (NO) either by the P450nor-trapping method (11) or by gas chromatography. The gas chromatography method employed quantification of N 2 O that was converted from NO by an NAD(P)H-dependent reaction catalyzed by P450nor of F. oxysporum or P450nor2 of C. tonkinense (6,11). The reaction mixture (2 ml) for assays by the P450nor-trapping method contained the mitochondrial fraction (0.09 mg protein), 2 mM sodium nitrite, 10 mM potassium phosphate, 2 M P450nor from F. oxysporum (7), 5 mM electron donor, and a buffer containing 0.6 M mannitol, 1 mM EDTA, 0.1% bovine serum albumin, 0.3 mM PMSF, and 10 mM Tris-HCl (pH 7.4). Reactions were incubated at 25°C in a Thunberg-type cuvette and the increase in absorbance at 432 nm was monitored. The gas phase in the cuvette was replaced with helium and the reaction was initiated by mixing a condensed solution of the electron donor in a side arm containing to other reaction mixture components. The reaction mixture (1.75 ml) for assays by the gas chromatography method contained 5 mM sodium nitrite, 0.8 M P450nor2, 10 mM potassium phosphate, 1.5 mM sodium azide (to inhibit aerobic respiration), 5 mM ADP, 2.5 mM NADPH, 5 mM electron donor, and the mitochondrial fraction (0.33 mg protein). Reactions were incubated at 25°C in a test tube sealed with a butyl rubber stopper (7 ml). The gas phase in the tube was replaced with helium.
Nir activity driven by artificial electron donors was assayed either by the gas chromatography method or P450nor-trapping method above except NADH-phenazinemethosulfate (PMS) or ascorbate-N,N,NЈ,NЈtetramethylphenylenediamine (TMPD) were provided as electron donors.
Other Determinations-The difference absorption spectrum was measured with an RSP-701 multichannel spectrophotometer (Unisoku, Osaka, Japan) equipped with a xenon lamp. Cytochrome oxidase activity was assayed aerobically in 100 mM Mes-KOH buffer (pH 6.7) with 10 M reduced cytochrome c from Saccharomyces cerevisiae (Sigma), and the change in absorbance at 550 nm at 25°C was monitored.

Subcellular Localization of Nitrate Reductase and Nitrite
Reductase--Denitrifying cells of F. oxysporum and C. tonkinense were disrupted and fractionated, and the distribution in the various subcellular fractions of Nar (F. oxysporum) and Nir (C. tonkinense) was investigated (Fig. 1). Fig. 1A shows distribution of Nar activity in fractions from F. oxysporum. In addition to an artificial electron donor (reduced methylviologen (MVH)), respiratory substrates, malate plus pyruvate, were effective donors of electrons for the activity. Although MVHdependent Nar and the marker enzyme (cytochrome oxidase) activities were recovered in various fractions, Nar activity supported by malate plus pyruvate was enriched in fraction II (50.4% sucrose). The result is highly indicative that the fraction contained intact mitochondria. Nir activity was also detected in fraction II of F. oxysporum (data not shown). This fraction additionally contained cytochromes a and b (Fig. 2), as well as cytochrome oxidase activity (Fig. 1A). Fig. 1B shows distribution of cytochrome oxidase and Nir among various subcellular fractions of C. tonkinense. Most of the NADH-PMS-dependent Nir activity was recovered in the 35% sucrose fraction (fraction 2). Most of cytochrome oxidase activity was also recovered in this fraction, whereas catalase (a marker enzyme of peroxisome) activity was primarily in the 42% sucrose fraction (fraction 3) (data not shown). Apparently, mitochondria of C. tonkinense grown under micro-aerobic con- ditions are less dense and/or have a smaller sedimentation coefficient than those of F. oxysporum.
Electron Donors-As shown above, malate plus pyruvate were effective donors of electrons to the Nar activity of the mitochondrial fraction of F. oxysporum. This suggests that electrons were supplied to Nar from these respiratory substrates via the mitochondrial electron transport system. As shown in Fig. 3, formate and succinate were also effective electron donors for Nar. In contrast, NADH is ineffective in donating electrons to the activity, consistent with the fact that NADH does not permeate into the mitochondrial matrix. Formate is a physiological electron donor for the reduction of nitrate in the nitrate respiration system of Escherichia coli (14,15) and the fungal Nar system resembles the bacterial system in this regard. Malate plus pyruvate and succinate were also effective donors of electrons to the mitochondrial Nir activity of C. tonkinense (see Figs. 6 and 8).
Mitochondrial Fluorescence Microscopy-Further evidence that fraction II from F. oxysporum really contained mitochondria was obtained by fluorescence microscopy (Fig. 4). Staining by the indirect fluorescent antibody technique showed that particles in fraction II contained Nir (Fig. 4A). These also reacted with the DNA-specific dye (Fig. 4B) showing that they were mitochondria. The results of immunostaining (Fig. 4A) further supported our previous assertion that the Nir of this fungus is located in the intermembrane space of the mitochondrion (11).
Inhibition of the Mitochondrial Nar and Nir Activities by Detergent and Respiratory Inhibitors-The effectiveness of various respiratory substrates as electron donors for the mitochondrial activities of Nar and Nir are strong evidence that the fungal denitrifying system was associated with the mitochondrial respiratory chain. This was confirmed by the effects of treatments with a detergent or respiratory inhibitors on the mitochondrial Nar and Nir activities. The malate plus pyruvatedependent mitochondrial Nar activity of F. oxysporum was lost upon treatment with deoxycholate, while the MVH-dependent activity was unaffected (Fig. 5). Inhibitors of respiration, namely rotenone, antimycin A, and thenoyltrifluoroacetone (TTFA), strongly inhibited the mitochondrial Nir activity of C. tonkinense (Fig. 6, A and B), while they did not affect the ascorbate-TMPD-dependent activity (Fig. 6C). Deoxycholate (Fig. 5) apparently did not inhibit Nar activity itself but destroyed the aligned electron transport system for oxidation of pyruvate and malate. Thus, pyruvate-and malate-dependent Nir activity was inhibited by rotenone and antimycin but was not inhibited by TTFA, while succinate-dependent Nir activity was inhibited by antimycin and TTFA but not by rotenone (Figs. 6, A and B). Therefore, the fungal Nir shares the mito-chondrial electron transport system for aerobic respiration and requires electrons from complex III as do the bacterial nitrite reductases (1) .   FIG. 2. Presence of cytochromes a and b in the mitochondrial fraction of F. oxysporum. Difference absorption spectrum of dithionite-reduced minus nonreduced fraction II in Fig. 1 was recorded.   FIG. 3. Reconstitution of the mitochondrial Nar activity of F. oxysporum with respiratory substrates. Fraction II (0.25 mg of protein/assay) in Fig. 1 was assayed for Nar activity employing various respiratory substrates as electron donors. Electron donors: (E), malate plus pyruvate; (Ç), formate; (q), succinate; (f), glutamate; (Ⅺ), NADH; (å), minus electron donor (no activity; hidden behind the former two symbols).

FIG. 4. Fluorescence images of a mitochondrion prepared from denitrifying cells of F. oxysporum.
A, confocal fluorescence image after staining with Texas Red shows the distribution of Nir. Scale bar, 2 m. B, the image due to fluorescence from Hoechst 33342 reveals DNA. C, immunoblot analysis shows the high specificity of the antibodies to Nir. Lanes 1-3 are for: 1, cell-free extracts obtained from nondenitrifying cells (grown aerobically in the absence of nitrate/nitrite) (55 g protein); 2, cell-free extracts from denitrifying cells (140 g of protein); 3, purified Nir (0.092 g protein). In lane 2 three bands could be observed. The top one must be the native Nir. Other two bands with lower M r might depend on Nir degraded by intracellular proteases.

Synthesis of ATP Coupled to the Mitochondrial Nar and Nir
Activities-The mitochondrial reactions catalyzed by Nar and Nir were coupled to the synthesis of ATP (Figs. 7 and 8). The energy yield (P:NO 3 Ϫ or P:2e ratio) (1, 16) was 1.4 for the pyruvate/malate-dependent reaction catalyzed by Nar (at 3 min), and it was 1.0 and 0.88 for the pyruvate-and malatedependent and the succinate-dependent reaction catalyzed by Nir (between 2 and 6 min), respectively. The synthesis of ATP was also inhibited by antimycin A (data not shown). No synthesis of ATP was observed in replicate experiments in which the mixture of pyruvate and malate together or succinate was replaced by Nar or Nir artificial electron donors, even though formation of the reaction products, namely nitrite or NO, were detected (data not shown). DISCUSSION Our results clearly demonstrate the presence of novel forms of respiratory metabolism in mitochondria of denitrifying fungi. While the mitochondrion has been regarded as the aerobic organelle of eukaryotes, the occurrence of anaerobic respiration in fungal mitochondria evokes interesting questions. First, what is the evolutionary relationship between the fungal and bacterial denitrifying systems? We previously determined that Nir of F. oxysporum bears close resemblance to the copper- was incubated with ADP in the presence (E), or absence (q) of nitrite, employing malate plus pyruvate (A and C) or succinate (B and D) as the electron donor. ATP synthesis in C and D are coupled, respectively, to the NO formation in A and B. NO was quantitated by the gas chromatography method. As in the case of NADH in Fig. 3, NADPH added to support the reaction catalyzed by P450nor2 had no effect on the mitochondrial Nir activity (data not shown).
containing counterpart in bacteria (11). Intracellular localization in the intermembrane space in mitochondria (see Fig. 4) would also parallel localization in the periplasm of most of the bacterial enzymes (1-3). Since formate can donate electrons to the mitochondrial Nar activity (Fig. 3), we suppose there is a formate dehydrogenase system that is directly associated with the respiratory chain that can support the anaerobic reduction of nitrate (14,15). Thus the fungal Nar system generally resembles the nitrate-reducing system in E. coli and other nitrate-respiring bacteria. Further, Nar of F. oxysporum resembles the bacterial counterpart in that it is membrane-bound and composed of 3 (or 4) subunits which are similar in molecular size to those of bacteria. 2 While the Nar and Nir systems in F. oxysporum seem to resemble those of the bacterial denitrifying systems, the P450-dependent Nor system of fungi contrasts with the cytochrome bc type Nor of bacteria (1)(2)(3). Nor of the P450 type has not been detected bacteria, although we have suggested horizontal transfer of the P450 gene from bacteria to the denitrifying fungi during evolution (17).
According to John and Whatley (18), the denitrifying bacterium Paracoccus denitrificans has many features of mitochondria that render it as a direct descendant of the endosymbiotic protomitochondria that gave rise to the mitochondrion. The mitochondria of eukaryotic cells are believed, therefore, to have lost adaptive systems, such as denitrification, during their evolution. This hypothesis is supported by systematic comparisons of sequences of 16S rRNA that show that the ␣ subclass of proteobacteria, to which P. denitrificans belongs, are most closely related to mitochondria (2, 19 -21). In view of this hypothesis, the present results may indicate that fungal mitochondria have retained the adaptive denitrifying system of the protomitochondrion. However, this hypothesis cannot account for the unique involvement of P450 in the fungal Nor system. While P450nor is recovered from fungal cells in a soluble form, there are two isoforms of P450nor, one of which contains in its gene a presequence resembling the targeting and sorting signals for import into mitochondria. 3,4 While mitochondria of denitrifying fungi seem to contain a set of reductases (Nar, Nir, and Nor in F. oxysporum, and Nir and Nor in C. tonkinense) it is unclear whether the fungal denitrifying system is truly a remnant of the protomitochondrion or it is system that was acquired at a later evolutionary stage.
The morphology of mitochondria during the anaerobic growth also deserves attention. The mitochondria of yeasts exist in a "promitochondria" form when cells are growing under anaerobic, fermenting conditions (21). In contrast, fungal mitochondria are only moderately modified during exposure to denitrifying conditions (Fig. 4). The fungal cells used in this study were grown under a micro-aerobic or an initially aerobic (4, 6) condition and cytochrome oxidase was also detected ( Figs.  1 and 2). However, the denitrifying systems of F. oxysporum and C. tonkinense differ in response to oxygen tensions. The system of F. oxysporum seems to be induced to supplement the aerobic respiration (22), while that of C. tonkinense acts more positively to support cell growth even under much more restricted aeration (6). The morphological and electron transport capabilities of mitochondria in C. tonkinense that grown under denitrification, restricted aeration conditions deserves further study.