Plant Physiol. (1992) 99, 693-699 Received for publication September 20, 1991 0032-0889/92/99/0693/07/$01 .00/0 Accepted December 11, 1991

Expression in Escherichia coli of Cytochrome c Activity from a Maize NADH: Complementary DNA'

Wilbur H. Campbell Phytotechnology Research Center and Department of Biological Sciences, Michigan Technological University, Houghton, Michigan 49931-1295

ABSTRACT the deduced amino acid sequences of NR to related mam- A cDNA clone was isolated from a maize (Zea mays L. cv malian sequences because of its similarity to the Mo- W64AxW183E) scutellum Xgtl 1 library using maize leaf pterin domain of sulfite oxidase, and the C-terminal region NADH:nitrate reductase Zmnrl cDNA clone as a hybridization of NR has sequence homology to NADH:Cyt b5 reductase, probe; it was designated Zmnr1S. ZmnrlS was shown to be an which is a FAD-containing (3, 22). In between the NADH:nitrate reductase clone by nucleotide sequencing and com- Mo-pterin and FAD domains, NR has similarity to mam- parison of its deduced amino acid sequence to Zmnrl. Zmnr1S, malian Cyt b5. This central Cyt b domain of NR is bridged to which is 1.8 kilobases in length and contains the code for both the other two domains on either side of it by highly variable the cytochrome b and flavin adenine dinucleotide domains of sequence regions that appear to be hinges (3). These hinge nitrate reductase, was cloned into the EcoRI site of the Esche- regions are probably exposed in native NR and susceptible to richia coli expression vector pET5b and expressed. The cell attack by proteolytic (15, 22). Hence, native NR can lysate contained NADH:cytochrome c reductase activity, which is cleaved by into stable fragments that a characteristic partial activity of NADH:nitrate reductase depend- readily be proteinases ent on the cytochrome b and flavin adenine dinucleotide domains. retain partial catalytic activities of holo-NR, such as methyl Recombinant cytochrome c reductase was purified by immunoaf- viologen NR, NADH:Cyt c reductase, or NADH:ferricyanide finity chromatography on monoclonal antibody Zm2(69) Sepha- reductase (15, 22). The stable fragments appear to be the rose. The purified cytochrome c reductase, which had a major domains of the native NR or combinations of two domains; size of 43 kilodaltons, was inhibited by polyclonal antibodies for for instance, the fragment with methyl viologen NR activity maize leaf NADH:nitrate reductase and bound these antibodies is composed of the Mo-pterin and Cyt b domains, the Cyt c when blotted to nitrocellulose. Ultraviolet and visible spectra of reductase fragment contains the Cyt b and FAD domains, oxidized and NADH-reduced recombinant cytochrome c reduc- and the ferricyanide reductase fragment contains only the tase were nearly identical with those of maize leaf NADH:nitrate FAD domain (3, 15, 22). These catalytically active and stable reductase. These two enzyme forms also had very similar kinetic NR in out some of the properties with respect to NADH-dependent cytochrome c and fragments of have been useful sorting ferricyanide reduction. complexities of the biochemistry of NR (2, 3, 22). However, it is difficult to produce these fragments in sufficient amounts for many studies, especially those involving physical and chemical measurements. An alternative approach is to produce the NR domains and their catalytically active combinations as recombinant pro- Higher plant NADH:NR2 (EC 1.6.6.1) catalyzes the first teins in coli cDNA clones for NR. The a of Escherichia using step in nitrate assimilation and is key site for regulation FAD domain of maize leaf NADH:NR was overexpressed in this pathway (22). The enzyme is a homodimer composed of E. coli using part of the Zmnrl cDNA (9) and purified to two approximately 100-kD polypeptide chains with each con- homogeneity by NADH elution from blue Sepharose (11). taining one equivalent of FAD, -Fe, and Mo-pterin (2, The recombinant FAD domain of NR has an Mr of 30,000, 3, 17). Each of these electron-carrying cofactors appears to be NADH:ferricyanide reductase activity, and a UV-visible spec- bound into a specific region of the enzyme's polypeptide, trum essentially identical with mammalian NADH:Cyt b5 which folds into distinct domains. The N-terminal region of reductase, which demonstrated that FAD is bound to the NR can be recognized as the Mo-pterin domain by comparing recombinant protein in a functional manner (1 1). The recom- binant FAD domain has now been crystallized, and analysis This research was supported by National Science Foundation of its three-dimensional structure is underway (G. Lu, W.H. grant DCB-9020773 and U.S. Department ofAgriculture Competitive Campbell, Y. Lindqvist, and G. Schneider, unpublished re- Research grant 90-37280-5474. 2 Abbreviations: NR, nitrate reductase; anti-FD, rabbit polyclonal sults). The Cyt b domain of Chlorella NADH:NR has also antibody for SDS-denatured recombinant FAD domain of NR; anti- been expressed in E. coli, and although this domain has no NR, rabbit polyclonal antibody for nitrate reductase; FAD, flavin catalytic activity by itself, the recombinant protein has a adenine dinucleotide; IgG, immunoglobulin G; IPTG, isopropyl 3-D- visible spectrum very similar to native NR (5). thiogalactopyranoside. Reported here is the expression in E. coli of a fragment of 693 694 CAMPBELL Plant Physiol. Vol. 99, 1992

NADH:NR with NADH:Cyt c reductase activity. The recom- the chromosome of the DE3 cell strain and is under the binant protein was purified using monoclonal antibody-based control of the inducible lac UV5 promoter (24). Cells were immunoaffinity chromatography, which was developed for grown to 0.6 A600nm before induction with IPTG. After 2.5 to purification ofholo-NR from maize leafextracts (12). Recom- 3 h of induction with IPTG, the cells were harvested by binant Cyt c reductase has immunochemical, kinetic, and centrifugation. spectral properties virtually identical with NADH:NR purified from maize leaves by the same method. Purification of Recombinant Cyt c Reductase MATERIALS AND METHODS E. coli cells harvested from two 1-L cultures were suspended in 1 L of 0.1 M KPO4 (pH 7.5) and 1 mM EDTA and frozen Isolation of Maize Scutellum cDNA Clones for NR at -70°C for at least 12 h before thawing. No additives were Total RNA was isolated from nitrate-induced scutella of required to lyse the cells because they were expressing T7 maize (Zea mays L. cv W64AXW183E) seedlings and shown lysozyme from the pLysS plasmid (23). After thawing, cell to contain a transcript that hybridized to the Zmnrl cDNA debris was removed by centrifugation for 30 min at 20,000g, clone, as previously described (10). cDNA was synthesized and the supernatant fluid was mixed with 7 to 10 mL of from polyadenylated RNA isolated on mAP paper (Amer- Zm2(69) Sepharose (monoclonal antibody Zm2(69) coupled sham, Chicago, IL) and cloned into EcoRI-cut Xgtl 1 arms to cyanogen bromide-activated Sepharose 4B [Pharmacia] at (Stratagene, San Diego, CA) after EcoRI-NotI linker-adaptor 1 mg of IgG/mL gel) at 4°C for 45 min (12). The gel was (GAATTCGCGGCCGC) ligation (Librarian cDNA cloning recovered by vacuum filtration and washed with a total of 2 kit; Invitrogen, San Diego, CA). The maize scutellum cDNA L of suspension buffer using three washes with the gel col- library was screened by hybridization with radioactive Zmnrl lected by filtration in between each wash. The washed gel was insert DNA, and four positive clones were isolated and plaque mixed with a minimum volume of suspension buffer and purified (9, 18). Insert DNA was excised with EcoRI and packed into a column (0.9 x 15 cm). The column was eluted subcloned into the EcoRI site of Bluescript II vector (Strata- with 10 mm glycine (pH 11), and 1. 1-mL fractions were gene). Nucleotide sequencing was carried out by the dide- collected into tubes containing 15 ML of 2 M NaPO4, pH 7 oxynucleotide method on the double-stranded DNA of these (12). Crude extract, the first filtrate obtained immediately subclones using T7, T3, SK, and KS primers (Stratagene), as after the binding step, and purified fractions eluted from the well as six custom primers derived from the nucleotide se- immunoaffinity column were assayed for Cyt c reductase quence of Zmnrl (9). One of these clones, designated activity (21), protein (16), and A413nm (purified fractions only). ZmnrlS, is very similar to Zmnr1 in nucleotide sequence and contains 1804 base pairs, which makes it 300 base pairs Purification of Other Enzyme Forms and Miscellaneous shorter than Zmnrl (9). The other three clones were either Methods smaller or contained some nucleotide sequence regions not related to Zmnrl. Previously described methods were used to purify maize leafNADH:NR (12) and the recombinant FAD domain (1 1). Construction of the Expression Vector Spectra ofpurified Cyt c reductase and maize leafNADH:NR were analyzed with a Shimadzu UV- 1201 UV/visible spectro- ZmnrlS insert was excised with EcoRI and cloned into the photometer at 25°C. These were reduced with solid EcoRI site of the E. coli expression vector pET5b (Novagen, NADH to generate their reduced spectra. Kinetic analyses Madison, WI). After the clone was transformed into E. coli were done at 25°C with a Gilford 2400-2 UV/visible spectro- strain DH5a- competent cells, three clones were selected photometer measuring the change in absorbance with time containing the ZmnrlS insert, and the orientation was deter- for Cyt c reduction at 550 nm and for ferricyanide reduction mined by nucleotide sequencing using the T7 primer, which at 340 nm (NADH), which were converted to micromoles per binds to the T7 promoter in the pET vector series (24). Two minute using the following millimolar extinction coefficients: ofthe clones were inframe with the AUG translation start site NADH, 6.2 (340 nm); Cyt c, 20 (550 nm). SDS-PAGE and of the pET5b vector, designated pZCRl and pZCR2, and the protein blotting to nitrocellulose were done as previously other was in the opposite orientation and designated pZCR3. described (4, 17). Polyclonal antibodies for SDS-denatured recombinant FAD domain of maize leaf NADH:NR were Expression of ZmnrlS in E. coli elicited in a rabbit using two subcutaneous injection sets, 3 The constructs pZCRl, pZCR2, and pZCR3 were trans- weeks apart, with 1 mg of protein for each injection set, as formed into E. coli strain BL21(DE3)pLysS-competent cells, previously described for preparation of anti-NR (4). The and colonies harboring both the expression plasmid and the rabbit was exsanguinated 10 d after the second injection set. pLysS plasmid were selected by restriction analysis (18, 24). Anti-NR, anti-FD, and monoclonal antibody Zm2(69) were The cells were grown in LB medium or LBM9 medium the purified IgG fractions obtained from a protein-A Sepha- modified by the addition of 0.14 mm Fe-citrate, with both rose column (4). The specificity ofthe anti-NR was previously containing 20 ,M ampicillin (24). Cells were grown at 37°C, described (4). Protein blots were developed with anti-rabbit except where noted otherwise. To express the Cyt c reductase IgG conjugated to alkaline phosphatase (Cappel, Organon fragment encoded by the insert DNA, cells were induced with Teknika Corp, West Chester, PA) using a standard alkaline 0.4 jM IPTG to express the T7 polymerase, which resides in phosphatase blot assay (9). EXPRESSION OF CYT c REDUCTASE OF MAIZE NITRATE REDUCTASE 695

RESULTS AND DISCUSSION A. Target sequence 1 10 20 520 Construction of the Expression Vector MASMTGGQQMGRDPNSRPLVCHLD ... SPVVF < pET5b > < L > < ZmnrlS > When ZmnrlS3 insert is excised with EcoRI and cloned into the EcoRI site of the pET5b expression vector, Zmnrl S B. Cyt b domain would be inframe with the translation initiation codon of the 151 160 226 FTMSEVRKHASQ ... GELITTG vector (24). This construct, designated pZCRl, was made, and the predicted alignment was confirmed by nucleotide C. FAD domain sequencing. When expressed in E. coli, pZCRl is predicted 259 270 520 to yield a polypeptide with 520 amino acid residues and an PRDKIHCPLVGK ... DMANSFWF Mr of 58,067. The N-terminal 18 residues would be derived D. Potential internal cleavage site from the vector and the EcoRI-NotI linker-adaptor of the insert, and the remaining 502 residues would be from 131 140 150 520 ZmnrlS, except for the first nucleotide of residue 19 being PGLKR STSTPFMNTTDVGKQF ... SFVVF contributed by the last nucleotide of the linker-adaptor (Fig. Trypsin site 1A). The amino acid sequence deduced from Zmnr 1S differs Figure 1. Partial amino acid sequence of the Cyt c reductase of NR from the sequence of Zmnrl (9) in 18 residues (data not expressed in E. coli. The target polypeptide of pZCR1 is predicted to shown). In most cases, these differences result from single- have 520 amino acid residues. A, The first 14 residues are derived base changes in the nucleotide sequence of ZmnrlS relative from the pET5b vector (24), and the next four residues are from the to the Zmnrl sequence (9), except near the 5' end, where EcoRI-NotI linker-adaptor ("L"). Residues encoded by Zmnrl S begin three nucleotides are inserted, and in the FAD domain, where with residue 19, which has as its first nucleotide the final one of the a nine-nucleotide insertion is found. linker-adaptor. Only the first and last five residues encoded by Interestingly, most of these substitutions in the deduced Zmnrl S are shown. B, The Cyt b domain of the recombinant protein amino acid sequence of ZmnrlS relative to Zmnrl (9) result is predicted to have 76 residues. C, The FAD domain of the recom- in residues found in either the rice or barley NR-deduced binant protein is predicted to have 262 residues. D, Potential cleavage site for or an E. coli with similar is shown. amino acid sequences (6, 7, 19). Because these maize NR trypsin protease properties Single-letter codes are used for the amino acids. cDNA clones have virtually the same 3'-untranslated nucleo- tide sequence, Zmnr and ZmnrIS appear to be alleles rather than separate genes for NADH:NR; however, this is not completely established. Based on an restriction fragment binant ZmnrlS protein would bind compatible cofactors and length polymorphism map of segregation data pooled from fold in stable conformations, which might protect the hinge four F2 populations of maize inbred lines, Zmnrl maps to between them against degradation. two loci: one on chromosome 2 near the centromere and a However, the N-terminal 150 residues of the recombinant second undetermined location (D. Grant, personal commu- protein would not be expected to bind the Mo-pterin nication). Consequently, further analysis is required to estab- synthesized in E. coli, which differs from that found in maize lish the relationship between Zmnrl and ZmnrlS. Because NR (8, 13). The N-terminal region might fold into a random these two cDNA clones were isolated from different cultivars coil, which may be susceptible to proteolysis, especially in the of maize, one of which is an inbred line (9) and the other a hinge region between the incomplete Mo-pterin domain and hybrid of two inbred lines apparently unrelated to the first, the stably formed Cyt b domain (Fig. ID). This region of NR they may represent the degree of natural variation occurring has been shown to be cleaved by trypsin when spinach in the maize NADH:NR gene. NADH:NR is subjected to mild proteolysis (15), and it is The 58-kD protein predicted to be expressed from pZCRl cleaved at the same site shown in Figure 1D as a potential would contain the complete Cyt b and FAD domains of proteolytic cleavage site in the recombinant ZmnrIS protein NADH:NR, as well as a portion of the Mo-pterin domain (3, (8). The larger polypeptide fragment resulting from such a 9, 17, 22). Based on the amino acid sequence alignments of cleavage of the 58-kD ZmnrlS recombinant protein would all available NR sequences and related proteins (3, 22), the be predicted to have an Mr of 42,797 or molecular mass of Cyt b domain of the predicted recombinant protein encoded 43 kD. by pZCR would begin at residue 151 and end at 226 (Fig. 1B, 76 residues total), and the FAD domain would begin at Cyt c Reductase Activity Expression in E. coli residue 259 and end at the natural C terminus shared by all NADH:NR forms (Fig. IC, 262 residues total). The hinge Lysates of E. coli cells containing pZCRl, grown at 37°C region between these two domains appears to be 33 residues in LB medium and induced with IPTG, were prepared using long, but the assignments of the C terminus of the Cyt b a single freeze/thaw cycle, and after centrifugation, the super- domain and the N terminus of the FAD domain are only natant fluid was found to have Cyt c reductase activity. A approximations. Because both the Cyt b and FAD domains single freeze/thaw cycle is sufficient to break the cell walls of of NR have been expressed separately in E. coli (5, 1 1), it can E. coli when the cells express the pLysS plasmid, which be expected that the Cyt b and FAD domains of the recom- encodes T7 lysozyme (23). Thus, it appeared that pZCR1 was being expressed in E. coli to yield an active Cyt c reductase 'The nucleotide sequence of ZmnrlS has been deposited in fragment of NR; however, the level of expression was not as GenBank and assigned accession No. M77792. high as had been achieved with the recombinant FAD domain 696 CAMPBELL Plant Physiol. Vol. 99, 1992

(1 1). Studier et al. (24) suggested that higher levels of expres- Table II. Comparison of Cyt c Reductase Activity of Recombinant sion may be obtained in an enriched medium containing Cyt c Reductase and Maize Leaf NADH:NR in the Presence of glucose. Others using the pET vector system for expressing NADPH or Antibodies heterologous proteins in E. coli found more protein ofinterest Purified Cyt c reductase and NADH:NR were adjusted to the same in the soluble fraction when a lower growth temperature was NADH:Cyt c reductase activity (5 nmol Cyt c reduced/min/mL) in the utilized (20). The highest specific activity was found in the standard assay. enriched LBM9 medium when cells were grown at 37°C, as Relative Cyt c compared to growth in LB medium. Growth at 32°C in either Reductase Activity the enriched LBM9 or LB media did not enhance the pro- Additive Final [IgG] duction ofthe recombinant Cyt c reductase activity. A second reductase NR transformant of E. coli containing an inframe construct, mg/mL % pZCR2, was also analyzed for Cyt c reductase activity and None was found to have a similar level to the pZCRl transformant. 0 100 100 NADPHa 0 7 4 A a transformant harboring construct with the insert DNA in Anti-NR 0.02 33 17 the opposite orientation, pZCR3, was used as a negative Anti-FD 0.48 58 70 control in these analyses. The factors limiting the expression Zm2(69) 0.48 21 26 of the recombinant Cyt c reductase activity remain to be 8 NADH was omitted from this assay and replaced by 0.1 mm established. NADPH.

Purification of Recombinant Cyt c Reductase

When an E. coli lysate with expressed Cyt c reductase Sepharose column revealed that both of these high specific activity was mixed with blue Sepharose and the gel washed to activity Cyt c reductase fractions contained a series of poly- remove nonbinding proteins, the activity was completely peptides ranging in size from 58 to 43 kD with the largest bound. However, NADH did not elute the Cyt c reductase amount appearing to be 43 kD, especially in fraction 8 (Fig. activity, but it was eluted with 1 M NaCl (data not shown). 2A, lanes 4 and 5). For comparison, the recombinant FAD This procedure yielded a low specific activity Cyt c reductase, domain of maize NR and maize leaf NADH:NR, 31 and which contained a large number of protein bands on SDS- approximately 100 kD (2, 1 1), respectively, were also included PAGE. An effective purification was achieved using anti-NR in this gel (Fig. 2A, lanes 3 and 6). When a similar gel was monoclonal antibody Zm2(69) Sepharose with elution at pH electrophoresed with these same proteins and transferred to 11, which is a method for purification of maize leaf nitrocellulose, polyclonal anti-NR bound to all of the poly- NADH:NR (12). Cyt c reductase activity was eluted from the peptides in fractions 7 and 8 (Fig. 2B, lanes 2 and 6, respec- Zm2(69) Sepharose in a sharp peak, which absorbed light at tively), as well as the FAD domain (Fig. 2B, lane 1) and holo- 413 nm, a characteristic ofthe heme-Fe cofactor ofNR (Table NR (Fig. 2B, lanes 3 and 5). These results indicate that the I). The specific activity of the Cyt c reductase activity was full size 58-kD polypeptide predicted to be made from the increased 800-fold in the peak fraction (fraction 7) of the target sequence in pZCRl is expressed, but this polypeptide Zm2(69) Sepharose pH 11 elution, in which 60% of the is unstable in the cells and/or lysate. It appears to be degraded activity bound to the column was recovered (Table I). When in steps with at least two intermediate polypeptides and a NADPH was substituted for NADH in the Cyt c reductase stable 43-kD polypeptide accumulating as the final product. assay, the purified recombinant enzyme was slightly more Inclusion of PMSF in the lysis buffer did not prevent the active than maize leaf NADH:NR with this substrate but degradation of the 58-kD polypeptide (data not shown). Al- clearly had a strong preference for NADH (Table II). though this size heterogeneity of the recombinant Cyt c re- SDS-PAGE analysis of fractions 7 and 8 from the Zm2(69) ductase complicates its biochemical analysis, the impact of

Table I. Purification of Recombinant Cyt c Reductase by Monoclonal Antibody-Based Immunoaffinity Chromatography Using Anti-NR Zm2(69) Sepharose See "Materials and Methods" for details of the procedure. Eighty percent of the Cyt c reductase activity in the extract bound to Zm2(69) Sepharose, which was determined by subtracting the residual activity after binding from the original activity in the extract. Cyt c Specific Fold Fraction Volume Reductase A413rm Protein Activity PurificationYield Activitytciiy Prfcto mL units/mL mg/mL units/mg % Extract 920 0.145 NDb 0.172 0.84 1 100 Zm2(69) Sepharose Fraction 7 1.1 60 0.18 0.09 670 800 50 Fraction 8 1.1 30 0.11 0.055 550 650 25 a NADH:Cyt c reductase activity units = ,gmol Cyt c reduced/min. b ND, Not determined. EXPRESSION OF CYT c REDUCTASE OF MAIZE NITRATE REDUCTASE 697

1 2 3 4 5 6 FAD domain of maize NR (Table II). All three antibodies A inhibited the Cyt c reductase activity of the recombinant enzyme and maize leaf NADH:NR. Anti-NR was by far the (-9 most inhibitory with a greater effectiveness for the holoen- zyme than the recombinant fragment, whereas anti-FD and Zm2(69) were slightly more effective inhibitors of the frag- *15 ment (Table II). Anti-FD is a weak inhibitor and adding increasing amounts of it did not enhance its effectiveness, ._w :3 I whereas the degree of inhibition produced by anti-NR and Zm2(69) was highly concentration dependent (data not shown). SDS-PAGE gels of the proteins in lysates of E. coli 1 2 3 4 5 6 expressing the FAD domain and Cyt c reductase and the B purified FAD domain and Cyt c reductase were transferred 9 , to nitrocellulose and developed with anti-FD. Although there were many proteins present in the lysates (Fig. 3A, lanes 1 -15 and 2), the anti-FD bound only to a 30-kD protein in a lysate :II from cells expressing the FAD domain, which establishes the specificity of this antibody (Fig. 3B, lane 1). However, al- though the anti-FD bound to purified Cyt c reductase poly- peptides (Fig. 3B, lane 4), no antibody binding was detected Figure 2. SDS-PAGE analysis of recombinant Cyt c reductase and protein blotting with anti-NR. A, A 10% acrylamide SDS-PAGE gel in the lysate of cells expressing them, which may be due to was used for separation of the polypeptides and stained with Coo- their low concentration in the lysate (Fig. 3B, lane 2). Anti- massie blue. Lane 1, High molecular mass standard proteins of 200, FD also binds to purified holo-NR from maize leaves in 116.25, 97.4, 66.2, and 45 kD (Bio-Rad); lane 2, low molecular mass protein blots (Fig. 3B, lane 5). These tests leave little doubt standard proteins of 97.4, 66.2, 45 and 31 kD; lane 3, purified recombinant FAD domain of NR (11); lane 4, fraction 7 of the immunoaffinity-purified recombinant Cyt c reductase (Table I); lane 5, 1 2 3 4 5 B fraction 8 (Table I); lane 6, purified maize leaf NADH:NR (12). B, A 12% acrylamide SDS-PAGE gel was used to separate the polypep- A - :..:.2'Yw~7 tides, which were electrophoretically transferred to supported nitro- cellulose membrane (OPTIBIND, Schleicher and Schuell) and devel- 1:- oped with anti-NR (4) followed by binding of antibody for rabbit IgG conjugated to alkaline phosphatase and staining for enzyme activity 0 - :31 bound to the polypeptides (9). Lane 1, Purified recombinant FAD domain of NR (11); lane 2, fraction 7 (Table I); lanes 3 and 5, purified maize leaf NADH:NR (12); lane 4, RAINBOW protein molecular mass markers (Amersham) used to calibrate blot but not visible in photo- graph; lane 6, fraction 8 (Table I). The protein blot contained artifacts 1 2 3 4 5 of unknown origin at 60 and 70 kD in all lanes, including the RAINBOW B 9 t,rM"; standards. The purified maize leaf NADH:NR contained a small amount of 85-kD protein, which bound anti-NR and is presumed to be a degradation product of purified 1 00-kD NR polypeptide. . 15 - :31 this complexity was minimized by using spectra to evaluate the heme-Fe content, which was taken to represent the frac- Figure 3. SDS-PAGE analysis of E. coli lysates and recombinant Cyt tion of the protein mixture containing a functional enzyme c reductase and protein blotting with anti-FD. A, A 12% acrylamide form. SDS-PAGE gel was used for separation of the polypeptides and silver stained (Bio-Rad). Lane 1, Lysate of E. coli cells expressing recom- Immunochemical Analysis of Recombinant Cyt c binant FAD domain of NR (1 1); lane 2, lysate of E. coli cells expressing Reductase recombinant Cyt c reductase; lane 3, purified recombinant FAD domain (11); lane 4, purified recombinant Cyt c reductase from a As described above, recombinant Cyt c reductase binds to different Zm2(69) Sepharose column than described in Table I; lane monoclonal antibody Zm2(69), and polyclonal anti-NR binds 5, purified maize leaf NADH:NR (12); lane 6, combined high and low to its denatured polypeptides. Because these two antibody molecular mass standard proteins of 116.25, 97.4, 66.2, 45, 31, and have been shown to be for NR 21.5 kD (Bio-Rad). B, Identical gel to A transferred to nitrocellulose preparations specific (4, 12), and developed with anti-FD as in Figure 2B; lane assignments are as the Cyt c reductase activity expressed was encoded by pZCRl in A, except standards were omitted. The blot for lane 5 containing and has epitopes in common with maize NADH:NR. The maize leaf NADH:NR was from a separate gel and appeared to immunochemical character of recombinant Cyt c reductase contain degradation products of holo-NR at 43 and 30 kD, which was further delineated with these two antibodies and a poly- bound anti-FD. RAINBOW protein markers were used to calibrate clonal antibody prepared against the denatured recombinant the blots. 698 CAMPBELL Plant Physiol. Vol. 99, 1992 that recombinant Cyt c reductase is immunochemically re- Table l1l. Comparison of the Kinetic Properties of NADH:NR and its lated to NR and its recombinant FAD domain. This is to be Recombinant Fragments expected because these proteins are predicted to have >95% Ferricyanide reductase activity was assayed in 25 mm Mops (pH identity in their common amino acid sequence portions. 7.0) with 10 AM NADH and the apparent Km for ferricyanide is shown. Cyt c reductase activity was assayed in 30 mm Pi (pH 7.5) with 100 Spectral and Kinetic Analyses of Recombinant Cyt c ,M NADH and the apparent Km for Cyt c is shown. Apparent Vmax Reductase values are given in units of two electrons transferred/min per nmol heme for NR and recombinant Cyt c reductase and per nmol FAD for Spectra were taken for the oxidized and NADH-reduced the recombinant FAD domain. Cyt c reductase and compared with those of purified maize Ferricyanide Cyt c Reductase leaf NADH:NR (Fig. 4). In the visible range, both proteins Enzyme Form Reductase Activity Activity have major peaks at 413 nm in the oxidized form and 557 Km V.ax Km Vmax and 424 nm in the reduced form. However, they differ in the UV with Cyt c reductase peaking at 265 nm, and NR peaking JAM units/nmol JM units/nmol at 275 nm and having a broader absorbance, which may be FAD domain 10 41.6 -a due to the presence of the Mo-pterin cofactor. These spectra Cyt c reductase 12 41.6 19 38.9 are very similar to those of squash NADH:NR and Chlorella NR 11 42.1 14 44.5 NADH:NR and its recombinant Cyt b domain (5, 17). The a The FAD domain does not reduce Cyt c (11). major differences are in the ratios of the UV absorbance to the Soret peaks (413 and 424 nm), with the ratios for maize leaf NADH:NR (RI = A28onm:A4I3nm = 2.10; R2 = A280 nm: A424 nm = 1.67) being greater than those for squash (R1 = 1.79; R2= 1.02). The small differences in the spectral properties of R2 = 1.26) or Chlorella NADH:NR (RI = 1.84; R2 = 1.45); the NADH:NR forms are probably due to differences in the recombinant Cyt c reductase has lower ratios (RI = 1.32; amino acid sequence and, in the case of the maize leaf NR, due to the pH 11 treatment during purification, which results in decreased NADH:NR activity (12). The much lower ratios for recombinant Cyt c reductase are likely due to the absence of the Mo-pterin cofactor. The ratios of the peaks in the 0.2 visible portion of the spectra, which can be attributed to the heme-Fe of their Cyt b domain, are similar for all four proteins, indicating that the environment of the heme-Fe is similar in each of these proteins. The kinetic properties of the recombinant Cyt c reductase 0.1 were compared to those of maize leaf NADH:NR and the recombinant FAD domain (Table III). Because the FAD domain contains only FAD as an internal cofactor and cannot reduce Cyt c (1 1), the kinetics of ferricyanide reduction were Q compared for these three enzyme forms and found to be ~0 0.0 virtually identical. The recombinant Cyt c reductase has a C) 2.5 slightly higher Km (Cyt c) and lower Vmax for Cyt c reduction than maize leaf NADH:NR (Table III). These kinetic con- stants for partial reactions of NR are very similar to those for 2.0 Chlorella and spinach NADH:NR (1, 14, 22).

1.5 SUMMARY 1.0 Zeduced A new cDNA clone for maize NADH:NR was used to express a protein in E. coli with Cyt c reductase activity. The 0.5 557 protein was isolated in a partially degraded form but displayed spectral and kinetic properties consistent with it having a 0.0 stable combination of the Cyt b and FAD domains of 300 400 500 600 withNADH:NR.native andThedenaturedrecombinantmaizeCytleafc reductaseNADH:NR,sharesasepitopeswell as Wavelength I(nm) the recombinant FAD domain of this enzyme. Although the recombinant Cyt c reductase was not expressed at a high level, Figure 4. Spectra of recombinant Cyt c reduct;ase (CCR) and maize the system will be useful for determining essential amino acid leaf NADH:NR. The proteins were purified by pH 11 elution from residues of the Cyt b and FAD domains of NR via site- Zm2(69) Sepharose. Oxidized (... .) and reduce(d --- -) spectra were directed mutagenesis, especially when studied in combination taken at 250C as described in "Materials andJ Methods." Proteins with the recombinant FAD domain which is expressed at a were reduced with solid NADH. high level ( 11). EXPRESSION OF CYT c REDUCTASE OF MAIZE NITRATE REDUCTASE 699

ACKNOWLEDGMENTS 12. Hyde GE, Wilberding JL, Meyer AL, Campbell ER, Campbell WH (1989) Monoclonal antibody-based immunoaffinity chro- I thank Ellen R. Campbell and Yueru Zhang for technical assist- matography for purifying corn and squash NADH:nitrate re- ance and Dr. Robert C. Bugos for reading and commenting on this ductases. Plant Mol Biol 13: 233-246 manuscript. 13. Johnson JL, Bastian NR, Rajagopalan KV (1990) Molybdopterin guanine dinucleotide, a modified form of molybdopterin iden- LITERATURE CITED tified in the molybdenum cofactor of dimethyl sulfoxide re- ductase from Rhodobacter sphaeroidesforma specialis denitri- 1. Barber MJ, Notton BA (1990) Spinach nitrate reductase. Effect ficans. Proc Natl Acad Sci USA 87: 3190-3194 of ionic strength and pH on the full and partial enzyme 14. Kay CJ, Barber MJ (1986) Assimilatory nitrate reductase from activities. Plant Physiol 93: 537-540 Chlorella: effect of ionic strength and pH on catalytic activity. 2. Campbell WH (1989) Structure and regulation of nitrate reduc- J Biol Chem 261: 14125-14129 tase in higher plants. In JL Wray, JR Kinghorn, eds, Molecular and Genetic Aspects of Nitrate Assimilation. Oxford Science 15. Kubo Y, Ogura N, Nakagawa H (1988) Limited proteolysis of the nitrate reductase from spinach leaves. J Biol Chem Publications, Oxford, United Kingdom, pp 125-154 263: 3. Campbell WH, Kinghorn JR (1990) Functional domains of 19684-19689 assimilatory nitrate and nitrite reductases. Trends 16. Redinbaugh MG, Campbell WH (1985) Adaptation of the dye Biochem Sci 15: 315-319 binding protein assay to microtiter plates. Anal Biochem 147: 4. Campbell WH, Remmler JL (1986) Regulation of corn leaf 144-147 nitrate reductase. I. Immunochemical methods for analysis of 17. Redinbaugh MG, Campbell WH (1985) Quaternary structure the enzyme's protein component. Plant Physiol 80: 435-441 and composition of squash NADH:nitrate reductase. J Biol 5. Cannons AC, lida N, Solomonson LP (1991) Expression of a Chem 260: 3380-3385 cDNA clone encoding the heme binding domain of Chlorella 18. Sambrook J, Fritsch EF, Maniatis T (1989) Molecular Cloning. nitrate reductase. Biochem J 278: 203-209 A Laboratory Manual, Ed 2. Cold Spring Harbor Laboratory 6. Cheng CL, Dewdney J, Nam HG, den Boer BGW, Goodman Press, Cold Spring Harbor, NY HM (1988) A new locus (NIA 1) in Arabidopsis thaliana en- 19. Schnorr KM, Jurick M, Huang C, Culley D, Kleinhofs A (1991) coding nitrate reductase. EMBO J 7: 3309-3314 Analysis ofbarley nitrate reductase cDNA and genomic clones. 7. Choi HK, Kleinhofs A, An G (1989) Nucleotide sequence of rice Mol Gen Genet 227: 411-416 nitrate reductase genes. Plant Mol Biol 13: 731-733 20. Slice LW, Taylor SS (1989) Expression of the catalytic subunit 8. Crawford NM, Campbell WH (1991) Fertile fields. Plant Cell 2: of cAMP-dependent protein kinase in Escherichia coli. J Biol 829-835 Chem 264: 20940-20946 9. Gowri G, Campbell WH (1988) cDNA clones for corn leaf 21. Smarrelli J, WH NADH:nitrate reductase and chloroplast NAD(P)+: glyc- Campbell (1979) NADH dehydrogenase activity eraldehyde-3-phosphate dehydrogenase. Plant Physiol 90: of higher plant nitrate reductase (NADH). Plant Sci Lett 16: 792-798 139-147 10. Gowri G, Kenis J, Ingemarsson B, Campbell WH (1992) Nitrate 22. Solomonson LP, Barber MJ (1990) Assimilatory nitrate reduc- reductase transcript is expressed in the primary response of tase: Functional properties and regulation. Annu Rev Plant maize to environmental nitrate. Plant Mol Biol 18: 55-64 Physiol Plant Mol Biol 41: 225-253 11. Hyde GE, Campbell WH (1990) High-level expression in Esch- 23. Studier FW (1991) Use ofbacteriophage T7 lysozyme to improve erichia coli of the catalytically active flavin domain of corn an inducible T7 expression system. J Mol Biol 219: 37-44 leaf NADH:nitrate reductase and its comparison to human 24. Studier FW, Rosenburg AH, Dunn JJ, Dubendorff JW (1990) NADH: reductase. Biochem Biophys Res Com- Use of T7 RNA polymerase to direct expression of cloned mun 168: 1285-1291 genes. Methods Enzymol 185: 60-89