Functional dissection and site-directed mutagenesis of the structural gene for NAD(P)H-nitrite reductase in Neurospora crassa.

Neurospora crassa NAD(P)H-nitrite reductase, encoded by the nit-6 gene, is a soluble, alpha2-type homodimeric protein composed of 127-kDa polypeptide subunits. This multicenter oxidation-reduction enzyme utilizes either NADH or NADPH as electron donor and possesses as prosthetic groups two iron-sulfur (Fe4S4) clusters, two siroheme groups, and two FAD molecules. The native activity of the enzyme is the NAD(P)H-dependent reduction of nitrite to ammonia. In addition, N. crassa nitrite reductase displays several partial activities in vitro, including a siroheme-independent NAD(P)H-cytochrome c reductase activity and an FAD-independent dithionite-nitrite reductase activity. These partial activities are presumed to be manifestations of discrete functional domains within the protein. A full-length nit-6 cDNA was constructed and used in developing an expression system within E. coli capable of yielding high levels of NADPH-nitrite reductase activity. Maximal expression was obtained in nirB- E. coli cells grown anaerobically at 22 +/- 1 degrees C, in conjunction with co-expression of a plasmid-borne cysG gene (encoding the rate-limiting enzyme in siroheme synthesis) and co-transformation with plasmid pGroESL (encoding bacterial chaperonins GroES and GroEL). Dissection of gene segments encoding putative functional domains within the nit-6 gene was performed. Expression of a partial cDNA construct encoding the FAD-/NAD-binding domain yielded extracts with NADPH-cytochrome c reductase activity but no NADPH-nitrite reductase activity or dithionite-nitrite reductase activity. Expression of a cDNA construct encoding the (Fe4S4)-siroheme-binding domain resulted in extracts possessing dithionite-nitrite reductase activity but no NADPH-nitrite reductase or NADPH-cytochrome c reductase activity. Analysis of site-directed mutations altering amino acid residues Cys-331 within the FAD-/NAD-binding domain and Ser-755 within the (Fe4S4)-siroheme-binding domain of the nitrite reductase demonstrated that these residues were not essential for native or partial enzyme activity. Cys-757 within the (Fe4S4)-siroheme-binding domain was essential for native enzyme activity.

Nitrate assimilation is a two-step enzymatic process: the reduction of nitrate to nitrite, a two-electron transfer reaction catalyzed by the enzyme nitrate reductase, followed by the reduction of nitrite to ammonium, a six-electron transfer reac-tion catalyzed by the enzyme nitrite reductase (1)(2)(3)(4). In the filamentous fungus Neurospora crassa, NAD(P)H-nitrite reductase (EC 1.6.6.4) is an ␣ 2 -type homodimer of M r 290,000, which utilizes either NADPH or NADH as electron donor and possesses two iron-sulfur (Fe 4 S 4 ) clusters, two siroheme cofactor groups, and two FAD molecules (5,6). The enzyme requires three equivalents of NADH or NADPH to pass six electrons to the substrate nitrite, converting it to ammonia. Besides this native enzyme activity, N. crassa nitrite reductase catalyzes in vitro partial activities, which include an FAD-dependent NAD(P)H-cytochrome c reductase activity, and dithionite (S 2 O 4 2Ϫ )-nitrite reductase activity (5)(6)(7)(8). The NAD(P)H-cytochrome c reductase activity appears to short cut the (Fe 4 S 4 )siroheme-containing portion of the enzyme, while the dithionite-nitrite reductase activity bypasses the requirement for NAD(P)H and FAD. The electron flow within the enzyme complex has been deduced primarily through inhibition studies on the various activities of the enzyme (5)(6)(7) and is illustrated in the following scheme where brackets define the enzyme complex and (-SH) represents a sulfhydryl moiety presumed to be involved in early electron transfer reactions (6). Deduced amino acid sequence data suggest that functional domains of fungal nitrite reductases as well as Escherichia coli NADH-nitrite reductase are laid out in a linear fashion along the polypeptide chain, beginning with the FAD-and NAD(P)Hdependent functions at the N-terminal end and ending with the nitrite-reducing functions at the C terminus (9). Similarly, the functional domains of nitrate reductases from fungi and plants are laid down in a linear fashion along the protein, but in reverse order (9). In the case of nitrate reductases, the domains are well defined. Hinge regions bridge the domains within the protein, as supported by the fact that proteolytic degradation of nitrate reductases yields polypeptide fragments with partial enzyme activities (2,9). In contrast, the domain structures of nitrite reductase are not well defined, and no hinge regions bridging putative domains have been identified through proteolysis (2,9).
Exley et al. (10) previously reported the nucleotide sequence of the nit-6 gene and the deduced amino acid sequence of the nit-6 product (10). In that study, putative functional domains and amino acid residues within the N. crassa NAD(P)H-nitrite reductase polypeptide were identified based on sequence conservation between the N. crassa nitrite reductase and the nitrite reductases from Aspergillus nidulans (a related ascomycete) and E. coli (9,11,12). Two putative functional domains were identified within the N. crassa nitrite reductase: an FAD-/NAD-binding domain in the N-terminal portion and an (Fe 4 S 4 )-siroheme-binding domain in the C-terminal part of the polypeptide (10).
A conserved feature found within the FAD-/NAD-binding domain is a Cys residue at position 331 of N. crassa nitrite reductase (9,10). This Cys residue is found within an 8-residue sequence, YAIGECAS, perfectly conserved in all three nitrite reductases mentioned above (9). Based on this conservation, this Cys residue has been postulated to be essential for enzyme function and NADH binding, in analogy with an essential Cys in nitrate reductase (9,10).
The putative (Fe 4 S 4 )-siroheme-binding domain is located in the C-terminal half of the N. crassa nitrite reductase polypeptide (residues 715-763 (10)). This domain was identified based on the presence of two conserved motifs, each of which includes a pair of cysteine residues (13)(14)(15). The sequences of nitrite reductases from E. coli, N. crassa, and A. nidulans have a conserved serine 2 residues away from the C-terminal cysteine motif (10).
Other enzymes utilizing an (Fe 4 S 4 )-siroheme cluster, including nitrite reductase from spinach (16), maize (17), and birch (18) and sulfite reductase from Salmonella typhimurium (19) all possess a Ser or Thr at the analogous position. The conservation of the amino acid side chain at this position led Exley et al. to speculate that this residue (Ser-755 in the N. crassa nitrite reductase) might be important to the functional integrity of the (Fe 4 S 4 )-siroheme cluster (10).
In this study, an expression system in E. coli capable of yielding high levels of functional NAD(P)H-nitrite reductase is described. The results of a dissection of nit-6 gene segments encoding functional domains within N. crassa nitrite reductase are reported. Effects on native and partial enzyme activities due to the alteration of amino acid residues Cys-331, Ser-755, and/or Cys-757 through site-directed mutagenesis are also presented.

MATERIALS AND METHODS
Neurospora Strains and Culture Conditions-N. crassa wild-type strain 74-OR23-1 A (stock number 987, Fungal Genetics Stock Center, Department of Microbiology, University of Kansas, Kansas City, KS) was utilized in preparing mycelial cell-free extracts for controls in nitrite reductase enzymatic assays and immunoblotting studies. Cultures of mycelia were grown aerobically in 1 ϫ Vogel's medium (20) plus 3% sucrose at 30°C, with the modification that NH 4 NO 3 was omitted.
Instead, different nitrogen sources were added to the medium depending on the type of nitrogen nutrition desired. Mycelia routinely were grown for 18 -24 h from conidial inocula in medium containing 20 mM glutamine and then harvested by filtration, washed with distilled water, and transferred into media containing either glutamine or sodium nitrate as the sole nitrogen source. Mycelia were cultured for 6 h in media containing 20 mM NaNO 3 to induce nitrate assimilation or 20 mM glutamine to repress it. Harvested mycelia were washed in distilled water, frozen in liquid nitrogen and stored at Ϫ80°C.
Preparing Cell-free Extracts of N. crassa-Frozen pads of mycelia were homogenized on ice in 7-ml Ten Broeck homogenizers in approximately 4 volumes of ice-cold extraction buffer (20 mM KPO 4 , pH 7.5, 10% glycerol, 5 mM EDTA, and 10 M FAD; just before use, fresh dry cysteine was added to a final concentration of 5 mM, and dry phenylmethylsulfonyl fluoride was added to a final concentration of 2 mg/ml). The homogenates were immediately centrifuged at 20,000 ϫ g for 30 min at 0 -4°C. Supernatants were decanted into fresh tubes and assayed for NADPH-nitrite reductase, NADPH-cytochrome c reductase, and dithionite-nitrite reductase activity.
E. coli Strains and Culture Conditions Used for Protein Expression-E. coli strains and protein expression vectors used in this study are shown in Table I. E. coli colonies grown with appropriate antibiotic selection were picked from Luria broth (LB) 1 plates and used to inoculate overnight cultures grown at 37°C in LB containing the appropriate antibiotic. These cultures were grown to stationary phase and used to inoculate the liquid medium sustaining growth during heterologous protein expression (by inoculating fresh medium with 1 ⁄250 volume of stationary culture). Minimal medium was prepared by making a 20 ϫ salt solution containing 210 g of dibasic potassium phosphate, 90 g of monobasic potassium phosphate, and 20 g of ammonium sulfate per liter. This solution was diluted to give a final volume of 1 ϫ after adding 0.25 ml of trace metal solution (20) per liter and then sterilized. After cooling, the solution received 25 ml of a sterile 20% solution of glucose per liter (final volume) and 2 ml of 1 M magnesium sulfate per liter (final volume). The appropriate antibiotics were then added, followed by inoculation with the stationary culture.
Growth of E. coli-Aerobic E. coli cultures were grown in 500 ml to 1 liter of medium in 2-liter shaker flasks at 250 rpm in a New Brunswick series 25 shaker incubator. Room temperature during aerobic growth was consistently 27 Ϯ 1°C. Aerobic cultures were induced for protein expression when the cultures reached an A 550 of 0.3-0.5 by adding isopropyl ␤-D-thiogalactopyranoside (IPTG) to a final concentration of 0.8 mM. For anaerobic growth, E. coli cells were propagated in large culture volumes of 8 -10 liters in a New Brunswick Microferm fermentor. The air intake valve of the fermentor was attached to a tank of pure nitrogen. Medium was flushed with N 2 for 1 min while mixing at 300 rpm, both prior to and after inoculating the media with starter culture and after adding IPTG. Room temperature during anaerobic growth was consistently at 22 Ϯ 1°C. Induction of anaerobic cultures was carried out when the cultures reached an A 550 of 0.1-0.2 (typically after 24 h of growth on minimal medium at room temperature) by adding IPTG to a final concentration of 0.8 mM. Maximal levels of NADPHnitrite reductase were found in cells grown anaerobically in minimal medium, 15 h following IPTG induction. Therefore, these conditions were used for routine protein expression.
Preparing Cell-free Extracts of E. coli-E. coli cultures grown under either aerobic or anaerobic conditions were harvested by centrifugation at 4500 ϫ g for 15 min. The cell pellets were resuspended and washed in 0.5 times the original culture volume with 0.9% NaCl. The cells were again collected by centrifugation, and drained cell pellets were stored at Ϫ70°C. Frozen cell paste was thawed on ice, mixed with approximately 4 -10 volumes of ice-cold extraction buffer, and sonicated at 70 mA for 50 s (Branson sonifier, Branson Sonic Power Co.) in ice-jacketed 50-ml stainless steel beakers. Cell debris was removed by centrifugation at 30,000 ϫ g for 30 min. The supernatants were decanted into fresh tubes on ice and immediately assayed for NADPH-nitrite reductase, NADPHcytochrome c reductase, and dithionite-nitrite reductase activities. Enzyme Assays-NADPH-nitrite reductase activity was measured spectrophotometrically by following the nitrite-dependent oxidation of NADPH (6). NADPH-cytochrome c reductase activity was measured spectrophotometrically by following the reduction of cytochrome c, as described previously by Vega (7). Dithionite-nitrite reductase activity was measured by following nitrite disappearance, essentially as described by Lafferty and Garrett (5).
Protein Determination-The protein concentration of each cell-free extract was determined using the Protein Dye Concentrate system from Bio-Rad.
DNA Isolation, Purification, and Subcloning-Plasmid DNA was propagated in E. coli strain DH5␣ or JM109 with the appropriate antibiotic selection overnight. Plasmid DNA was isolated and purified as described by Sambrook,Fritsch,and Maniatis (21) or by using DNA purification columns (Quiagen). DNA restriction fragments were purified from low melt gels using the GeneClean kit from Bio 101. Standard techniques were used in DNA subcloning (21).
PCR-Rapid Amplification of cDNA Ends and Full-length nit-6 cDNA Construction-PCR-rapid amplification of cDNA ends (22) was used in constructing a full-length nit-6 cDNA (Fig. 1). Poly(A) ϩ RNA was isolated from nitrate-induced N. crassa wild-type mycelia as described previously (10). Three separate reverse transcription reactions were performed using three different oligomer primers designed to hybridize to nit-6 mRNA in the poly(A) ϩ RNA fraction. One g of poly(A) ϩ RNA was combined with 10 pmol of primer (16 nucleotides in length) designed to hybridize downstream of introns 1 and 2, intron 3, and at the 3Ј end of the nit-6 transcript. Reverse transcription reactions were performed as recommended by the manufacturer (Life Technologies, Inc.). Upon completion of the reaction, 2 units of RNase H and 9 units of RNase T 1 were added, and the mixture was incubated at 37°C for 15 min. This reaction mixture was extracted with phenol/chloroform, and the DNA was ethanol-precipitated. The DNA was then dissolved in 500 l of TE (10 mM Tris⅐Cl, pH 8, 1 mM EDTA). A 10-l aliquot from each first-strand cDNA reaction was used as template for PCR to produce and amplify double-stranded DNA.
Oligonucleotide primers were designed to hybridize adjacent to existing restriction sites in order to facilitate ligations and cloning ( Fig. 1). Primers hybridizing to the 5Ј and 3Ј ends of the nit-6 cDNA were designed to introduce restriction sites (NdeI and BamHI) for subcloning into the expression vector pET11a. Each PCR reaction contained a 10-l aliquot of first-strand cDNA diluted in TE, 60 pmol of primer, 100 M dNTPs, in a total volume of 100 l of 1ϫ Pfu polymerase buffer (Stratagene). 1 unit of Pfu polymerase (Stratagene) was added, and then the reaction was covered with mineral oil and placed in a Perkin-Elmer Thermocycler. Conditions were programmed to heat the solution to 95°C for 5 min and then to cycle it to 60°C for 1 min, 74°C for 1.5 min, and then 95°C for 1 min (for a total of 35 cycles), followed by a final incubation at 74°C for 10 min. The reactions were chloroform-extracted, and the amplified DNA was ethanol-precipitated. After washing in 70% ethanol and drying, the DNA was cleaved with the appropriate restriction enzymes using the manufacturer's recommended conditions. The DNA was then fractionated by gel electrophoresis on 2% low melt agarose gels, and the appropriate size band was excised and purified. The three restricted PCR products isolated consisted of a 0.83-kb SalI/NspV fragment, a 0.28-kb SunI/AatII fragment, and a 0.14-kb NheI/BamHI fragment. These fragments were ligated together with equimolar amounts of a 1.05-kb AatII/NheI restriction fragment from the nit-6 cDNA (pNiR-3), a 1.3-kb NspV/SunI restriction fragment of the nit-6 genomic clone (pnit-61), and the 3-kb vector pBluescript SK ϩ (Stratagene) cleaved with SalI/BamHI. Restriction fragment analysis was used to identify a full-length nit-6 cDNA. Dideoxy sequencing was carried out over the PCR-derived portions and ligation boundaries to ensure no mutations had occurred. The resulting clone was designated pnit6.5, and the insert was named nit6.5. The full-length nit-6 cDNA (nit6.5) was subcloned as a single fragment from pnit6.5 into expression vector pET11a via NdeI/BamHI cleavage, yielding plasmid pETnit6.5. In this construct, the ATG translational start codon of nitrite reductase falls within the NdeI site, and the stop codon lies upstream from the BamHI site.
PCR Subcloning-Standard reaction conditions were used for PCR amplification of cloned DNA (22). These techniques were routinely used to alter flanking sequences and introduce novel restriction sites to facilitate subcloning or otherwise manipulate DNA. DNA containing FIG. 1. Strategy for construction of full-length nit-6 cDNA. Poly(A) ϩ RNA was isolated and reverse transcribed in three separate reactions using primers 1, 2, and 3. First-strand cDNA was then used in conjunction with primers 4 -7 in PCR to obtain dsDNA flanking intron locations in the nit-6 gene. All PCRs were designed to cover existing restriction sites in order to simplify subcloning in conjunction with portions of the nit-6 cDNA and genomic clones. Portions of the genomic clone and partial cDNA clone were used in the full-length cDNA construction to avoid large PCR products. Primers 4 and 9 contained the linkers shown to simplify subcloning of the entire cDNA into expression vectors.
the E. coli cysG gene was amplified by PCR from plasmid pRSM10 kindly provided by Nicholas Kredich of Duke University Medical Center (23), using two oligomeric primers flanking the gene. One primer was designed to hybridize upstream of the Shine-Dalgarno sequence of the cysG gene. This primer introduced a BamHI site. The other primer hybridized just downstream of the 3Ј end of the gene. This primer introduced a SalI site and BamHI site. The resulting 1.8-kb product was cleaved with BamHI and isolated from a 1% low melt agarose gel following gel electrophoresis. This fragment was ligated into the BamHI site downstream of the nit6.5 insert in pETnit6.5 (Fig. 2). Restriction analysis was used to identify clones containing the cysG gene in the same orientation as the nit-6 cDNA. The resulting plasmid was designated pETnit6.5cysG, and the insert was named nit6.5cysG.
The nit6.5cysG insert was also subcloned into vector pTRC99A for expression in E. coli strain JCB3878, a nirB Ϫ (deletion) strain lacking the gene encoding the E. coli NADH-nitrite reductase. Strain JCB3878 was kindly provided by Jeffrey Cole of the University of Birmingham (United Kingdom). Cloning into vector pTRC99A required a new restriction site at the 5Ј end of nit6.5cysG in order to ligate the insert in the proper orientation and reading frame for protein expression. Using a primer designed to introduce an NcoI site that hybridized to the 5Ј end of the nit-6 cDNA and a primer designed to hybridize downstream of the NspV site in the nit-6 cDNA, PCR was performed to amplify the intervening DNA between these sites within nit6.5cysG, creating an 0.84-kb fragment with an NcoI site at the 5Ј end. After chloroform extraction of the reaction mixture, the DNA was ethanol-precipitated, dried, and dissolved in TE. The DNA was then digested with NcoI, followed by digestion with NspV. The digested DNA was fractionated by gel electrophoresis on a 1.5% low melt agarose gel, followed by excision and purification. The NcoI/NspV fragment was ligated together with an NspV/SalI fragment of pETnit6.5cysG (containing a 2.76-kb portion of the nit-6 cDNA and the entire cysG gene) into expression vector pTRC99A cleaved with NcoI and SalI. The resulting plasmid was designated pTRCnit6.5cysG.
Partial constructs of the nit-6 cDNA were made in conjunction with restriction digest products of a single PCR reaction. Two oligonucleotide primers were synthesized. One primer hybridized 0.38 kb downstream of the EclXI site within nit6.5. This primer was designed to introduce an early stop codon and BamHI site within the cDNA. Another primer was designed to introduce an NdeI site and translational start site about 80 nucleotides upstream of the EclXI site. The intervening sequence between the primers was amplified using PCR. The reaction mixture was chloroform-extracted, and the DNA was ethanol-precipitated and dissolved in TE. Half of the product was digested with EclXI and BamHI.
The other half was digested with EclXI and NdeI. The EclXI/BamHIdigested DNA was ligated to a 4.12-kb fragment of pETnit6.5 cleaved with EclXI and BamHI (this effectively removed more than half of the nit-6 cDNA at its 3Ј end). The resulting plasmid was designated pET5N, for 5Ј (N-terminal) construct. The NdeI/EclXI-digested PCR product was ligated to the 5.48-kb fragment of pETnit6.5 digested with NdeI and EclXI (thereby removing about 1.12 kb of DNA at the 5Ј end of the nit-6 cDNA), producing the plasmid designated pET3C, for 3Ј (C-terminal) construct. Fig. 2 displays a restriction map of the cDNA constructs. The sites shown are restriction sites exploited in DNA manipulations and subcloning.
Site-directed Mutagenesis-Site-directed mutagenesis was achieved using the Altered Sites kit from Promega. The full-length nit-6 cDNA (nit6.5) was ligated into the vector pALTER-1 (supplied with the kit) via the BamHI and SalI site introduced at the ends of the cDNA during its construction (Fig. 1). Mutagenic DNA oligomeric primers were synthesized and purified, as described previously (21). The design of these oligonucleotides is shown in Fig. 5. A 0.1-pmol aliquot of oligonucleotide was phosphorylated using T4 polynucleotide kinase, as recommended by the manufacturer (Boehringer Mannheim). Mutagenesis was carried out as described in the Altered Sites manual. Mutagenized plasmid DNA was analyzed by direct nucleotide sequence determination or, where applicable, by restriction endonuclease analysis (see Fig. 5), followed by direct nucleotide sequence determination of the DNA in the region of the mutation. The DNA mutated within the region encoding the FAD-/NAD-binding domain of nitrite reductase was transferred to plasmid pTRCnit6.5cysG as a 0.3-kb fragment. This fragment was isolated by cleaving the DNA with restriction endonuclease NspV and EclXI, followed by electrophoresis and excision from a 3% low melt agarose gel. This fragment was ligated into the 9.57-kb NspV/EclXIdigested fragment of DNA from plasmid pTRCnit6.5cysG (which lacked the comparable segment of DNA). The DNA mutated within the region encoding the (Fe 4 S 4 )-siroheme-binding domain of nitrite reductase was subcloned into plasmid pTRCnit6.5cysG as a 0.3-kb fragment obtained by cleaving with restriction endonucleases AatII and SplI, followed by electrophoresis and excision from a 3% low melt agarose gel. The fragment was ligated together with a 6.3-kb SalI/SplI fragment and a 4-kb SalI/AatII fragment from pTRCnit6.5cysG (a single fragment derived from an AatII/SplI digest was not applicable because there are several AatII sites within the nit-6 cDNA).
SDS-Polyacrylamide Gel Electrophoresis and Immunoblot Analysis-7% SDS-polyacrylamide gels were prepared and gel electrophoresis was performed as described previously (24). Prestained protein molecular weight markers were purchased from Boehringer Mannheim. to bottom are the full-length nit-6 cDNA (nit6.5), the 5Ј (N-terminal) construct, the 3Ј (C-terminal) construct, and the full-length cDNA-cysG construct (nit6.5cysG). The restriction sites at the 5Ј and 3Ј ends of each construct were designed to facilitate cloning into expression vector pET11a. Asterisks mark the approximate locations of site-directed mutagenesis. SD represents the approximate location of the Shine-Dalgarno sequence for the cysG gene. Shine-Dalgarno sequences promoting ribosomal binding for translation of the nit-6 cDNA and partial constructs are located in the plasmid vectors used for expression (not shown). Insertion into each expression vector placed the 5Ј end of each DNA fragment just downstream of the Shine-Dalgarno sequence in these plasmids. The 5Ј end of the nit-6 cDNA-cysG was also altered using PCR subcloning techniques to introduce an NcoI site to facilitate cloning into expression vector pTRC99A (Pharmacia).
Equivalent amounts of protein from cell-free extracts were fractionated on each gel (20 g/lane). Fractionated protein was transferred to a nitrocellulose filter using an LKB Electro-Blot apparatus as described previously (24). Filters were air-dried and used immediately or stored at 4°C. Immunoblot analysis was performed using an immunoblot assay kit (goat anti-rabbit IgG (H ϩ L) alkaline phosphatase) from Bio-Rad. Kit components were used as recommended by the manufacturer. The filters were probed with primary antibody specific for N. crassa nitrite reductase. This antibody was generated and purified as described previously (10,19,24). Antibody concentrations of 1% were incubated with the filters for 1 h at room temperature.
Densitometry-Densitometry was performed by scanning the images of immunoblots into an Adobe Photoshop workstation. Individual bands corresponding to protein recognized by anti-nitrite reductase antibody were selected, and the data for each band were saved as an EPS file. Using a computer program written by David Tolbert (Biology Department, University of Virginia), the numerical values corresponding to the optical density of each pixel in a given band were summed (the sum of pixel intensities due to background was also determined and subtracted from this figure). Relative intensities were calculated in comparison with a specified band, as described under "Results."

Dissection of NAD(P)H-Nitrite Reductase Functional Domains-
A full-length nit-6 cDNA, designated nit6.5, was constructed and subcloned into expression vector pET11a, producing pETnit6.5. The full-length cDNA encodes a polypeptide of 1176 amino acid residues, the same as that deduced from the genomic clone (10). The amino acid sequence of this polypeptide highlighting the putative functional motifs relevant to this study is shown in Fig. 3. As described under "Materials and Methods," two partial cDNA constructs inserted in pETlla were also generated. pET5N (the N-terminal expression construct) encodes amino acid residues 1-492 with respect to the full-length nit-6 product (Fig. 3). These residues compose the putative FAD-/NAD-binding domain of the protein but lack the putative (Fe 4 S 4 )-siroheme-binding domain. pET3C (the C-terminal expression construct) encodes amino acid residues 348-1176 with respect to the full-length nit-6 product (Fig. 3). These residues compose the region of the protein including the putative (Fe 4 S 4 )-siroheme-binding domain but lack the putative FAD-/NAD-binding domain.
pETnit6.5, pET5N, and pET3C were each transformed into E. coli strain CB926 growing at 37°C, and the cells were induced for protein expression under anaerobic or aerobic conditions. Cell-free extracts from these cultures did not possess detectable native or partial enzyme activity, although successful expression of protein was deduced by immunoblot analysis.
However, enzyme activity was detected in cell-free extracts from transformed cells grown at room temperature (Table II). Higher levels of nitrite reductase activity were obtained in extracts of cells grown anaerobically in minimal media; therefore, these conditions were used routinely when expressing protein in E. coli. Activity within extracts was further increased by introduction of the cysG gene into the nit-6 expression plasmid and introduction of plasmid pGroESL (25) into E. coli host JCB3878 (Table III). The cysG gene of E. coli and S. typhimurium encodes uroporphyrinogen III methyl transferase, which converts uroporphyrinogen III into siroheme. This enzyme is likely to be the rate-limiting step in siroheme production (23). Cell-free extracts from cultures of E. coli strain CB926 transformed with pETnit6.5cysG and grown under anaerobic conditions demonstrated NADPH-nitrite reductase activity about 20% higher than extracts from cells expressing the construct without cysG (Table III).
Under anaerobic conditions, the host E. coli nirB gene product, NADH-nitrite reductase, becomes expressed. This enzyme also requires siroheme, and therefore its expression may reduce cofactor availability to the N. crassa nitrite reductase. To circumvent this, the nit6.5cysG insert was subcloned into expression vector pTRC99A, resulting in plasmid pTRC nit6.5cysG for expression in the nirB Ϫ strain JCB3878. Strain JCB3878 lacks the lac I q gene necessary to repress expression from pET11a, but pTRC99A carries the lac I q gene on the plasmid itself. pTRCnit6.5cysG was transformed into E. coli strain JCB3878, and the cells were induced for protein expression. Extracts of such cells possessed levels of NADPH-nitrite reductase activity almost 3 times as high as those of pETnit6.5cysG expressed in strain CB926 (Table III).
N. crassa NAD(P)H-nitrite reductase must associate as a dimer to execute its native NAD(P)H-nitrite reductase activity (5,24). To address the possibility that some fraction of the nit-6 polypeptide expressed in E. coli was incorrectly folding or not properly associating into its correct oligomeric form, plasmid pGroESL (25) was transformed along with plasmid pTRCnit6.5cysG into strain JCB3878. pGroESL contains the genes encoding the chaperonin proteins GroES and GroEL, which assist the folding and assembly of some proteins into their biologically active oligomeric structures (25). Extracts from cells transformed with both plasmid pGroESL and pTRCnit6.5cysG and induced under anaerobic conditions in minimal media exhibited levels of NADPH-nitrite reductase activity 3 times higher than transformed cells otherwise identical but lacking pGroESL (Table III).

Domain Analysis of NAD(P)H-Nitrite Reductase-
The FAD-/NAD-binding domain and the (Fe 4 S 4 )-siroheme-binding domain of the nit-6 polypeptide were tested independently of one another for functional activity by examining nitrite reductase activities in extracts from cells expressing the partial constructs pET5N and pET3C. Extracts from induced cell cultures The 1176 amino acid residues are numbered from N terminus to C terminus. The amino acid residues forming the dinucleotide-binding folds as well as residue Cys-331 within the FAD-/NADbinding domain are underlined. The stretch of amino acid residues encompassing the pair of Cys residue motifs presumed to be involved with iron coordination of the (Fe 4 S 4 )-siroheme couple is illustrated in boldface type. In relation to this sequence, the N-terminal partial construct encodes amino acid residues 1-492, and the C-terminal partial construct encodes amino acid residues 349-1176. transformed with pET5N, encoding the FAD-/NAD-binding domain, possessed NADPH-cytochrome c reductase activity but no NADPH-nitrite reductase or dithionite-nitrite reductase activity above background (Table II). Extracts from induced cell cultures transformed with pET3C, encoding the (Fe 4 S 4 )-siroheme-binding domain, possessed dithionite-nitrite reductase activity but no NADPH-cytochrome c reductase or NADPHnitrite reductase activity above background (Table II). Proteins in extracts from cultures expressing the partial nit-6 products and full-length nit-6 product were fractionated by SDS-polyacrylamide gel electrophoresis, transferred to a nitrocellulose filter, and probed with anti-nitrite reductase antiserum. Bands were detected corresponding to proteins of M r 57,000, M r 98,000, and M r 136,000 in extracts from cells expressing the 5Ј (N-terminal) construct, the 3Ј (C-terminal) construct, and the full-length nit-6 cDNA, respectively (Fig. 4). These apparent molecular weights correspond to the approximate size of each protein product as deduced from their amino acid sequences (53.3, 90.2, and 127.4 kDa, respectively). Apparent degradation products are also evident in extracts from cultures expressing the full-length and C-terminal construct (Fig. 4).
Amino Acid Replacements within Nitrite Reductase-As described under "Materials and Methods," mutagenic primers were designed and used to introduce mutations into the nit-6 cDNA that caused specific amino acid replacements within the N. crassa nitrite reductase. A restriction fragment containing the altered DNA was used to replace the comparable wild-type fragment in plasmid pTRCnit6.5cysG. This procedure altered specific codons within the nit-6 cDNA but otherwise left this gene as well as the downstream cysG gene intact. The four oligonucleotide primers synthesized for this procedure and the resulting alterations with respect to the wild-type DNA and nitrite reductase polypeptide are shown in Fig. 5. One mutation was located in the region of DNA encoding the FAD-/NADbinding domain. This mutation resulted in the replacement of residue Cys-331 with an Ala (Fig. 5). As discussed earlier, this cysteine has been suggested to be essential for the functioning of this domain (9). The expression construct encoding this amino acid change within the nit-6 cDNA was termed pTRC.C331A.
The other three mutations were located within the region of DNA encoding the (Fe 4 S 4 )-siroheme-binding domain (Fig. 5). In one of these mutations, Ser-755 was replaced with alanine, generating plasmid pTRC.S755A. A second mutation in this region replaced Cys-757 with an alanine, generating plasmid pTRC.C757A. Each of these mutations was constructed to determine whether one or both of these residues was essential to the function of this domain. The third mutation involved two amino acid changes; Ser-755 was replaced with Cys and Cys-757 was replaced with Ser, generating plasmid pTRC.S755C/ C757S, in order to determine if the position of these two residues relative to one another was essential for function.
Each of the mutant plasmid constructs was transformed along with pGroESL and expressed in host strain JCB3878. Cell-free extracts were prepared and assayed for NADPH-nitrite reductase, NADPH-cytochrome c reductase, and dithionite-nitrite reductase activities (Table IV). Based on these enzyme assays, none of the altered amino acid residues appeared to be essential for NADPH-cytochrome c reductase activity (Table IV). Only residue Cys-757 was essential for NADPHnitrite reductase and dithionite-nitrite reductase activity. The position of residue Cys-757 and Ser-755 relative to each other was also essential for NADPH-nitrite reductase and dithionitenitrite reductase activity. Cys-331 and Ser-755 were not essential for NADPH-nitrite reductase, NADPH-cytochrome c reductase, or dithionite-nitrite reductase activity (Table IV). The S755A mutation did abolish NADPH-nitrite reductase and dithionite-nitrite reductase activity in extracts from cultures grown under aerobic conditions (data not shown). A protein band cross-reactive with the anti-nitrite reductase antibody with an M r of approximately 136,000 was detected in each of the mutant extracts as shown by immunoblot analysis (Fig. 4).

Functional Analysis of Nitrite Reductase from N. crassa
Using densitometry, relative amounts of cross-reacting material recognized by the nitrite reductase antibody were determined for the wild-type and mutant full-length nit-6 products on the same immunoblot (Fig. 4). The amount of cross-reacting material in each of these extracts was assigned a value relative to that determined for the extract from cells expressing the wild-type nit-6 cDNA under the optimal anaerobic conditions. This extract produced the highest level of cross-reacting material, which was arbitrarily assigned a value of 100.

DISCUSSION
The association of NADPH-cytochrome c reductase activity with the FAD-/NAD-binding domain independently of the (Fe 4 S 4 )-siroheme-binding domain and the association of the dithionite-nitrite reductase activity with the region of the protein possessing the (Fe 4 S 4 )-siroheme-binding domain independently of the FAD-/NAD-binding domain clearly defines these domains as the functional entities responsible for the partial activities of N. crassa nitrite reductase.
The ratio of NADPH-cytochrome c reductase activity to NADPH-nitrite reductase activity (Cyt/Nit, Table IV) indicates the relative effect that an amino acid replacement has on the functioning of the FAD-/NAD-binding domain in comparison with the (Fe 4 S 4 )-siroheme-binding domain. The Cyt/Nit ratio of native nitrite reductase isolated from N. crassa increases during final purification of the protein. This increase is attributed to the instability of the nitrite-reducing function of the enzyme (associated with the (Fe 4 S 4 )-siroheme cluster) during the purification procedure (7). Conditions resulting in the highest heterologous expression of NADPH-nitrite reductase enzyme activity within cell-free extracts of E. coli (anaerobic growth, cysG downstream of the nit-6 cDNA in pTRC99A, and pGroESL co-transformed into the host strain JCB3878) yields a Cyt/Nit ratio virtually identical to that found in cell-free extracts derived from nitrate-induced N. crassa mycelia (8.5 versus 8.6, respectively (Table III). In this aspect, the E. coli expression system under optimal conditions yields N. crassa nitrite reductase with the functional properties of N. crassa nitrite reductase obtained from its native source.
Several amino acid residues within nitrite reductases from fungi and E. coli have been suggested to play vital roles in nitrite reductase activity, based on inhibitor studies and/or sequence conservation and location within putative functional domains (9,10). These residues were chosen for amino acid replacements within N. crassa nitrite reductase in an effort to establish their functional importance by direct biochemical evidence. Replacement of the highly conserved residue Cys-331 did not abolish NADPH-nitrite reductase activity, NADPHcytochrome c reductase activity, or dithionite-nitrite reductase activity. However, extracts of cultures expressing pTRC.C331A showed a reduction in the Cyt/Nit ratio compared with that of the wild-type full-length construct expressed under the same conditions (5.7 versus 8.5 (Table IV)). The dithionite-nitrite reductase activity within extracts expressing pTRC.C331A was diminished to a lesser extent than either NADPH-dependent activity. Apparently, this mutation influences electron flow within the FAD-/NAD-binding domain more than within the (Fe 4 S 4 )-siroheme domain. Whatever its function, the highly conserved Cys-331 is not essential for the electron transfer during catalysis.
Expression of the S755A mutant protein yielded extracts possessing NADPH-nitrite reductase, NADPH-cytochrome c reductase, and dithionite-nitrite reductase activity, demonstrating that Ser-755 is not essential for the electron transfer reactions of nitrite reductase. The Cyt/Nit ratio for S755A increased from 5.5 to 26, indicating a greater impairment of function in the nitrite-reducing properties of the enzyme compared with its cytochrome c-reducing properties. Perhaps Ser-755 plays a peripheral role in stability of the nearby (Fe 4 S 4 )siroheme cluster.
In the C757A mutant, Cys-757, one of the four Cys residues that coordinates the Fe 4 S 4 cluster, is replaced with Ala. Extracts from cultures expressing pTRC.C757A were essentially devoid of NADPH-nitrite reductase or dithionite-nitrite reductase activity (Table IV). Cys-757 is thus essential for the nitrite-reducing function of nitrite reductase but is not essential for NADPH-cytochrome c reductase activity (Table IV). Recent crystallographic data show that the bridging ligand in the (Fe 4 S 4 )-siroheme cluster of E. coli sulfite reductase is Cys-483 (13), which corresponds to Cys-761 in the N. crassa nitrite reductase, not Cys-757.
The relative position of Ser-755 and Cys-757 is essential for nitrite reductase activity, since extracts from cultures expressing pTRC.S755C/C757S lack detectable NADPH-nitrite reductase and dithionite-nitrite reductase activity (Table IV). NADPH-cytochrome c reductase activity in the S755C/C757S double mutant is essentially unaffected.
Several factors affect the activity of N. crassa nitrite reductase during expression in E. coli. Coomassie Blue staining of proteins in extracts fractionated by SDS-polyacrylamide gel electrophoresis demonstrated that the different levels of enzyme activity under various growth conditions are not simply due to the level of protein being expressed (data not shown). The lowered temperatures necessary to obtain NADPH-nitrite reductase, NADPH-cytochrome c reductase, and dithionite-nitrite reductase expression under all conditions examined suggest that the enzymatic activities of the nit-6 encoded protein are unstable at 37°C. Perhaps this thermal instability of N. FIG. 5. Sequence of mutagenic primers used to create site-specific mutations within the nit-6 cDNA, as described under "Materials and Methods." The reverse complement of each primer (shown) corresponds to the sequence of the nontemplate strand of the nit-6 gene. The respective amino acid residues encoded by this sequence are shown below in boldface type, with numbering at each end corresponding to the full-length protein. The analogous coding sequence of wild-type DNA and respective amino acid residues are also shown below each mutant sequence. The altered nucleotides and amino acid residues within the mutant sequence are underlined. Dots represent positions of novel restriction sites (AccI and NaeI) created by the mutations. These novel restriction sites facilitated rapid screening for mutations. Primer mutS755A was designed to create a silent GGG to GGC mutant codon (as well as the TCT to GCC mutant codon) in order to generate the NaeI restriction site. Primer mutS755C/C757S resulted in the novel AccI site without having to mutate any other codons. The other primers did not create a novel restriction site, and the mutations generated were screened by direct nucleotide sequencing. crassa nitrite reductase is related to protein folding or assembly. Since all three activities are affected by temperature, temperature sensitivity is not localized within a single domain. It should be noted that N. crassa grows well at 25°C but poorly above this temperature. The increase in nitrite reductase activities resulting from nit-6 expression in E. coli under anaerobic conditions versus aerobic conditions may result from a sensitivity of siroheme to aerobic conditions. Aerobic conditions are more detrimental to the nitrite-reducing activity of the protein than the NADPHcytochrome c reductase activity (data not shown). Under aerobic conditions in vitro, N. crassa nitrite reductase is capable of generating peroxide, which in turn damages its nitrite-reducing capability (4).
Introduction of the cysG gene into pETnit6.5 resulting in plasmid pETnit6.5cysG and expression of this plasmid in E. coli strain CB926 under anaerobic conditions yielded extracts with a relatively small increase in NADPH-nitrite reductase activity (Tables II and III). Under aerobic conditions the effect of introducing cysG into plasmid pETnit6.5 was much greater (a 3-fold increase in nitrite reductase activity was seen in extracts from aerobically grown cells upon introduction of cysG (data not shown)). However, expression of pTRCnit6.5cysG in strain JCB3878 under anaerobic conditions yielded extracts with NADPH-nitrite reductase activity almost 3-fold higher than that achieved by expressing pETnit6.5cysG in strain CB926 under the same conditions. This was predominantly due to a specific increase in the nitrite-reducing function of the protein expressed, since the specific activity for NADPH-cytochrome c reductase was only increased by 15% (Table III). A possible reason is that competition for siroheme manifested by the presence of E. coli NADH-nitrite reductase under anaerobic conditions limited siroheme acquisition by the heterologously expressed N. crassa enzyme.
The transformation of pGroESL into cells in which pTRCnit6.5cysG was expressed resulted in extracts with increased NADPH-nitrite reductase, dithionite-nitrite reductase, and NADPH-cytochrome c reductase activities (Table III). This elevation in activity and the similarity of the Cyt/Nit ratio (Table IV) to that of the enzyme isolated from N. crassa indicates that GroES and/or GroEL may assist in the proper folding and/or assembly of N. crassa nitrite reductase within E. coli.
This study provides direct evidence that N. crassa NAD(P)Hnitrite reductase possesses discrete functional domains and that these domains are laid out in a linear fashion along the polypeptide. The N-terminal portion of the polypeptide is responsible for the NAD(P)H-and FAD-dependent functions of the enzyme, while the C-terminal portion contributes the nitrite-reducing capabilities. Thus, the sequence of electron transfer in N. crassa nitrite reductase from NAD(P)H to nitrite is reflected in the domain organization of the polypeptide sub-unit from N-terminal to C-terminal and in the nucleotide sequence of the nit-6 gene from its 5Ј-end to its 3Ј-end. Nevertheless, it would be misleading to assume that the nitrite reductase holoenzyme in its catalytically active homodimeric state mediates electron transfer down two parallel (or antiparallel) tracks defined by pairs of subunits. It is equally plausible that electron transfer takes place from the FAD-/NAD-binding domain of one subunit to the juxtaposed (Fe 4 S 4 )-sirohemebinding domain of the other subunit.
In conclusion, the highly conserved Cys-331 and the conserved Ser-755 are not essential for native nitrite reductase activity or partial activities, whereas Cys-757 is essential for the nitrite-reducing catalytic functions of the enzyme.