D88A mutant of cytochrome P450nor provides kinetic evidence for direct complex formation with electron donor NADH Mariko Umemura 1 , Fei Su 1 , Naoki Takaya 1 , Yoshitsugu Shiro 2 and Hirofumi Shoun 3 1 Institute of Applied Biochemistry, University of Tsukuba, Japan; 2 The Institute of Physical and Chemical Research Institute, RIKEN Harima Institute, Mikazuki-Cho, Sayo, Japan; 3 Department of Biotechnology, Graduate School of Agricultural and Life Sciences, The University of Tokyo, Japan The haem-distal pocket of nitric oxide reductase cyto- chrome P450 contains many Arg and Lys residues that are clustered to form a putative access channel for NADH. Asp88 is the sole negatively charged amino acid in this positive charge cluster, and thus it would b e interesting to know its functional r ole. Here we found the intriguing phenomenon that mutation at this site of P450nor (D88A or D88V) c onsiderably d ecreased t he overall nitric oxide reductase activity without blocking the reducing half reac- tion in which the ferric enzyme–NO complex is reduced with NADH to yield a specific intermediate (I). The results indicate that the catalytic turnover s ubsequent to the I formation was blocked by such m utation. This property of the mutants made it possible to perform kinetic a nalysis of the reduction step, which is impossible with the wild-type P450nor. These results a re the first kinetic evidence for direct complex formation between P450nor and an electron donor (NADH or N ADPH). The k inetic analysis also showed that the inhibition by chloride ions (Cl – )iscom- petitive with respect to NAD(P)H, which highlights the importance of the binding site for Cl – (the anion hole) in the interaction with NAD(P)H. We also characterized another mutant (D393A) of P450nor. The results demon- strated that both A sp residues p lay important roles in the interaction with NADH, whereas the role of Asp88 is unique in that it must be essential for the release of NAD + rather than binding to NADH. Keywords: cytochrome P450nor; P450nor; NADH. Cytochrome P450 is the t erm used for a g roup of haem proteins that widely exist i n life from b acteria to higher organisms such as mammals [1,2]. P450 usually catalyses a monooxygenase reaction, whereas its molecular and functional diversity is so remarkable that some P450 species catalyse dehydration, isomerization, reduction, C–C bond cleavage, and s o on [3]. P450nor is one of such diverse P450 species and is involved in fungal denitrification [4–6]. I t functions as a n itric oxide ( NO) reductase (NOR) t o reduce NO t o nitrous oxide (N 2 O), with NADH or NADPH (NAD(P)H) as the electron donor [4]. P450nor can complete this reaction without the aid o f other p roteinaceous components such a s P450 reductase and thus receives electrons directly from NAD(P)H (Eqn 1). 2NO þ NAD(P)H þ H þ ! N 2 O þ NAD(P) þ þ H 2 O ð1Þ Fe 3 þ þ NO ! Fe 3 þ ÀNO ð2Þ Fe 3 þ ÀNO þ NAD(P)H ! I þ NAD(P) þ ð3Þ I þ NO ðþ H þ Þ!Fe 3 þ þ N 2 O þ H 2 O ð4Þ The overall NOR reaction can be divided into three partial reactions [7]: first, t he resting enzyme with ferric haem (Fe 3 + ) binds the substrate NO to form a ferric enzyme–NO complex (Fe 3 + –NO) (Eqn 2). Fe 3 + –NO is then reduced by NAD(P)H to yield a specific intermediate (I) exhibiting a Soret absorption peak at 444 nm ( Eqn 3), and finally I reacts with a second NO to form the product N 2 O(Eqn4).We assume I to be the two-electron reduced product of Fe 3 + – NO, formally the (Fe 3 + –NO) 2– state [7]. Several lines of evidence support hydride (H – ) t ransfer from N AD(P)H to the Fe 3 + –NO complex to form I in Eqn 3. For example, I is formed upon reduction of the Fe 3 + –NO complex with sodium borohydride, and a kinetic isotope effect has been observed in the reduction step for the proR hydrogen of NADH [8]. This mean s that the eq uivalent of two e lectrons and one proton is provided by NADH. The chemical form o f I thus depends on when the second proton is provided. I wouldbeinthe(Fe 3 + –NO)2H + ) 2– state, which is e quivalent to a ferric-hydroxylamine radical complex, if the second proton is also provided in the reduction step (Eqn 3) [8]. On Correspondence to H. Shoun, Department of Biotechnology, Gradu- ate School of Agricultural and Life Sciences, The University of Tokyo, Bunkyo-ku, Tokyo 113-8657, Japan. Tel.:/Fax: +81 35841 5 148, E-mail: ahshoun@mail.ecc.u-tokyo.ac.jp Abbreviations: DA, difference in absorbance; k dec , rate constant of decomposition; k obs , observed (apparent) first order r ate constant; k red , rate constant of reduction; NOR, nitric oxide reductase; P450, cytochrome P450. (Received 10 November 2003, revised 6 April 2004, accepted 6 May 2004) Eur. J. Biochem. 271, 2887–2894 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04206.x the other hand, however, if the second proton is provided in the subsequent step, as indicated a bove (Eqn 4), I should be in the ( Fe 3 + –NOH + ) 2– state. The three-dimensional structure of P450nor [9] exhibits unique features, although overall structural similarity to other P450s is conserved. That is, there is an open space in the haem-distal pocket that has a hydrophilic environment including many hydrophilic amino acid residues and water molecules [10]. This unique feature suggests that the P450nor molecule has become d iversifiedsoastointeract with the hydrophilic electron donor NAD(P)H, which is distinct from the molecular evolution of u sual monooxy- genase P450s that generally employ hydr ophobic organic substances as substrates. We thus expect that this big haem- distal pocket forms an access c hannel for NAD(P)H. We are c urrently characterizing the distal-haem pocket o f P450nor by means of s ite-directed mutagenesis s tudies in order t o prove the working hypothesis that t he pocket forms an access channel for NAD(P)H. There are many Arg and Lys residues i nside and outside the haem-pocket of P450nor. We have shown t hat this positive charge cluster plays a crucial role i n a ttracting a nd binding to the n egatively charged NAD(P)H molecule [11]. We have shown also that the specificity of P 450nor for electron donors (NADH and NADPH) is mainly determined by a few amino a cid residues in the B¢-helix [12]. We have further shown that s ome NADH analogues cause a specific spectral c hange of t he bound haem upon mixing with P450nor, indicating that these NADH analogues can bind to P450nor [12]. The above r esults are h ighly indicative that t he reduction step (Eqn 3) comprises two steps obeying a normal enzymatic reaction, i.e. reversible c omplex (Michaelis inter- mediate) formation between P450nor and NADH, and a subsequent catalytic (ele ctron transfer) r eaction (Eqn 5 ). On the other hand, the possibility cannot b e ruled out yet that the electron transfer f rom NADH to P450nor is carried out in a manner like in a chemical reaction (one-step reaction; Eqn 6), as the reaction of P450nor is too rapid ( more than 1000 s )1 at 10 °C) [7] for an enzymatic reaction in which electron transfer f rom or t o NAD is involved. To discrim- inate enzymatic and chemical r eactions, classical k inetic analysis is still effective for determining whethe r or not the rate of an e nzymatic r eaction is s aturated as to th e s ubstrate concentration. However, the electron transfer from NADH to P450nor (Eqn 3) is too rapid to be followed with a high NADH concentration, even with a rapid reaction analyser. Thus, such k inetic analysis has not been performed yet. E þ NADH $ E-NADH ! I þ NAD þ ðE : Fe 3 þ À NOÞ ð5Þ E þ NADH ! I þ NAD þ ð6Þ There are two negatively charged amino acid r esidues, Asp88 and Asp393, in the haem-distal pocket o f P450nor in addition to the positive charge cluster. Here we c onstructed mutant protein of P450nor of Fusarium oxysporum in which thenegativechargeofeachoftheseresidueswascancelled by replacing it with a neutral residue. Some of the mutant proteins were shown to exhibit intriguing properties, providing kinetic evidence for the direct complex formation of P450nor with NADH. Materials and methods Mutagenesis and expression plasmids The construction of each mutant of P450nor of F. oxyspo- rum was carried out according to a standard method [13]. Each recombinant protein was produced using an expres- sion vector for P450nor (pT7-nor) [6]. Site-directed muta- genesis w as performed by m eans of PCR [ 14] u sing template pfp(450)-20 [15], which consists of the P450nor cDNA of F. oxysporum and the pUC18 vector. The p rimers used were M13-47 and M13RV ( Takara, Otsu, Japan), which are specific for vector pUC18. The primers used to construct the mutant proteins were as follows (mutated sites are under- lined): D88A, 5¢-ACATTTGTC GCCATGGATCC-3¢; D88V, 5¢-GCCGACATTTGTC GTCATGGATCC-3¢; D88N, 5¢-GCCGACATTTGTC AACATGGATCC-3¢; and D393A, 5¢-CTGAACCGA GCTGTCGGAAT-3¢. The resulting PCR product was inserted into pGEM-T (Promega), and then the mutation was confirmed by sequencing o f the inserted nucleotide f ragment. Al l p lasmids expressing mutant proteins were constructed by replacing the BssHII–PstIfragment containing P450nor cDNA [15] in pT7-nor with the corresponding portion of the mutant cDNA. The introduced mutations were again confirmed by sequencing of the full-length exchanged c DNA. Expression The pT7-nor plasmid and derivatives of it were introduced into Escherichia coli JM109 (DE3). The transformed cells were cultured overnight at 30 °Cin LA broth [ 1% (w/v) t ryptone, 0.5% ( w/v) yeast e xtract, 0.5% (w/v) NaCl, 25 lgÆmL )1 ampicillin] supplemented with 0.5% (w/v) g lucose (preculture). P recultures (20 m L) were inoculated into 2 L LA broth in 5 L Erlenmeyer flasks w ith b affles, and i ncubated a t 3 0 °C with rotation at 120 r.p.m. for 6–7 h. Cells were induced by overnight i ncubation with 1 m M isopropyl thio-b- D - galactoside. Purification of mutant P450nor Each mutant P450nor protein was purified from a cell extract as re ported [12]. The transformed cells w ere harvested and suspended in Tris buffer [20 m M Tris/ HCl, pH 8.0, 0.1 m M dithiothreitol, 0.1 m M EDTA, 10% (v/v) glycerol] and then sonicated (200 watts, 10 min). The suspension was centrifuged at 10 000 g at 4 °Cfor 30 min. The supernatant was dialyzsd against Tris buffer and centrifuged at 10 000 g for 30 min, and then applied to a DEAE–cellulose (DE52, Whatman) column (bed volume, 30 mL) equilibrated with Tris buffer and e luted with a 0 –0.4 M KCl gradient. The P450nor fraction was concentrated by dehydration with polyethylene glycol and then dialysed against Tris buffer. The dialysate was applied to a Mono Q HR 5/5 column (Amersham Pharmacia Biotech) equilibrated with the same buffer andelutedwitha0–0.4 M KCl gradient. The P450nor fraction was concentrated, d ialysed, and s tored at 4 °C until further analyses. 2888 M. Umemura et al. (Eur. J. Biochem. 271) Ó FEBS 2004 Stopped-flow rapid scan analysis We analysed the P450nor reducing half-reaction by follow- ing the appearance of the intermediate (I) at 444 nm upon reduction of the Fe 3 –NO complex with NAD(P)H (Eqn 3) at 1 0 °C u sing a Unisoku rapid scan a nalyser ( Osaka, Japan) in 100 m M potassium phosphate buffer pH 7.2 as reported [7,11]. T he P450nor enzyme (final concentration 5 l M )intheFe 3 –NO c omplex was mixed with an equal volume of NADH or NADPH (final concentration 0.1 m M unless otherwise stated) anaerobically, and then the spec- trum of the mixture (200 lL volume) was recorded. The concentration of N O was kept almost equal t o that of P450nor to avoid catalytic turnover [7]. The gate time was set at 1 ms, and the rate o f I formation (k obs ) was calculated as described [7,11]. Titration of chloride ions (Cl – ) We investigated Cl – binding to P450nor by means of spectrophotometiric titration [11,12]. P450nor (10 l M )in 100 m M potassium p hosp hate buffer pH 7.2 was m ixed with an equal v olume (100 lL) of each concentration of potassium chloride and then the spectrum of the mixture (200-lL volume) was recorded. The dissociation constant (K d ) was calculated from the plot of Cl – concentration vs. the difference in absorbance (DA) at 413 nm from that at 395 nm caused by Cl – binding. Analytical methods The Nor activity of P450nor was assayed as reported [4]. P450nor (6 n M ) was incubated anaerobically with NO (55 l M ) in t he presence of 1.0 m M NADH in 100 m M potassium phosphate buffer p H 7.2 at 30 °C. The activity was determined by m easuring the i nitial rate of pr oduct (N 2 O) formation. N 2 O was determined by gas chromato- graphy [4]. Spectrophotometric analyses were carried out with a Beckman DU-7500 spectrophotometer. The amount of P450nor was determined from the CO-difference spec- trum using the value of 86.3 m M )1 Æcm )1 [16] for the molecular e xtinction c oefficient of the diffe rence between at 448 nm (CO-bound form) and 490 nm (dithionite- reduced form) [17]. Results Negative charges in the haem-distal pocket The location of the negatively charged amino acid residues in the haem-distal pocket of P450nor is shown in Fig. 1. T he B¢-helix and F,G-loop form the entrance of the pocket. There is an intriguing tendency in the distribution of charged amino acid residues that form the putative access channel for NADH. Most of the charged residues (Arg64, Asp88, Lys62, Lys291, a nd Arg2 92; fr om t he t op t o the bottom of t he pocket) are located on one side of the pocket beneath t he B ¢-helix, w hereas only one charged residue (Arg174) is present o n the other side beneath the F ,G-loop (Fig. 1 ) [ 9]. A sp88 is located in the m iddle of the a ccess channel, between the po sitive charges of Arg64, Lys291, and Lys62. Asp393 occupies a s ite away f rom the access c hannel across the haem. Here we constructed mutant protein of P450nor of F. oxysporum in which either Asp88 or Asp393 was replaced with a neutral amino acid. Each mutant protein was expressed i n E. coli and purified. All of the purified proteins exhibited the same spectral properties as those of the native protein. A representative spectrum (D88A) is shown in Fig. 2. D88A exhibited the same characteristics as those of the wild-type P450nor, i.e. its ferric resting form (Fe 3+ ) comprises a mixture of high- and low-spin states, a nd the CO-bound form gives a Soret band at 448 nm. These spectra show that the environment of the haem was not modulated very much by the m utation, Fig. 1. Stereoview of the haem-distal poc ket of P450nor. Negative charge residues (D88 and D393) are depicted in r ed, haem in magenta, h aem-iron in grey, a nd positive charge residues (K62, R6 4, R174, and K291) in blue. Data from PDB 1 CL6 [9]. Ó FEBS 2004 Asp88 in haem-distal pocket of P450nor (Eur. J. Biochem. 271) 2889 indicating that each mutant protein was properly folded in the heterologous host c ells. Mutations at Asp88 All mutations at Asp88 decreased the overall NOR activity of P450nor to a considerable extent, as shown i n Table 1. We further examined the partial reaction (reduction of the Fe 3 + –NO complex with NADH t o y ield the intermediate I; Eqn 3) of the mutant proteins. T his reducing half reaction can be observed as an isolated process under specific conditions by following the time-dependent accumulation of I (444 nm species; F ig. 3B and Fig. 4) with a rapid scan apparatus [7]. Such c onditions can b e attained by adding a similar a mount of NO to that of the e nzyme (P450nor), with which I cannot react further with a second NO because n o free NO remains; thus I could be i n a quasi-stable state suitable for accumulation (cf. Figs 3 and 4). The apparent first-order rate constant (k obs ) for the reduction (I forma- tion) can be obtained from the time-dependent decrease during t he process in t he absorbance at 427 nm, which is the isosbestic point of the spectrum of I and the resting enzyme Fe 3+ (cf. legend t o Fig. 3) [7]. The k obs value was obt ained for each mutant protein and is presented in Table 1. It is intriguing that the k obs due to NADH did not change or was even enhanced by the mutation replacing Asp88 with a hydrophobic amino acid residue (D88A or D88V), w hereas the m utation r eplacing Asp88 with a hydrophilic residue (D88N) d ecreased the k obs . T he extent of the decrease in k obs (36%; 16 s )1 as compared with 45 s )1 for the wild-type enzyme) agreed well with that in the overall N OR activity (36.5%) of the D88N mutant, showing that the inactivation caused by the mutation arose from blocking of the reduction step. P450nor of F. oxysporum shows electron donor specific- ity towards NADH [4,5,12]. The accumulation of I is not observed when the wild-type enzyme is reduced by NADPH, a less effective electron donor (Fig. 3A). This suggests that a higher formation rate is required for the accumulation of I. I must be highly re active with free NO to complete the overall reaction (Eqn 4), and Fe 3 + –NO complex formation (Eqn 2) should be in rapid equilibrium between association and dissociation [7]. Thus, during the I-forming process, previously formed I has a chance t o further react with NO even under the conditions used (with no excess NO) by taking it from the remaining Fe 3 + –NO complex i f the I formationrateismuchslowerthanthe following step (Eqn 4), which results in n o accumulation of I. This should b e the case for the reduction of wild-type P450nor by NADPH (Fig. 3A and Table 1). The I forma- tion step, the rate-limiting step of the overall reaction [7], must not be much slower than the subsequent s teps for I to accumulate. Therefore, the accumulation of I even after slow reduction of the D 88A or D88V mutant with NADPH (Fig. 3 B and Table 1) m eans that the mutation decelerated the subsequent steps, so that the reaction rate b ecame comparable to or even lower t han the slow reduction. Mutation at Asp393 We previously showed that a hydrogen bond network including Asp393, Ser286, and a few water molecules is formed upo n Fe 3 + –NO c omplex formation, and that the network should play a key role in providing a proton that i s required for intermediate formation [9,10]. As shown previously, mutation at Asp393 (D393A) greatly blocked the reduction step as well as the overall activity (Table 1). However, when the reduction was examined with a higher NADH concentration (0.5 m M ), we could observe the I formation o f the mutant protein ( Fig. 4), suggesting that the Fig. 2. Absorption s pectra of the D88 A mutant of P45 0nor. The spectra are for the ferric resting (solid line), dithionite-reduced (dotted line), and CO-bound (broken line ) forms, respectively, of P450nor (5.0 l M ) in 10 m M potassium phosphate buffer pH 7.2. Table 1. Kinetic parameters for the reduction step f or P4 50nor w ild-type an d mutant enzymes. k obs , Ob served fi rst-order rate constant for reduction (I formation ); k dec , first-order rate constant for spontaneous decomposition of I; ND, not determined. P450nor Overall activity (%) NADH (m M ) NADPH (m M ) 0.1 0.5 0.1 1.0 k obs (s )1 ) k dec (s )1 ) k obs (s )1 ) k obs (s )1 ) k dec (s )1 ) k obs (s )1 ) k dec (s )1 ) Wild-type 100 45 ± 5 0.027 ND ND ND ND D88A 25.5 ± 1.5 41 ± 4 0.10 1.6 ± 0.2 0.10 7.6 ± 0.8 0.10 D88V 22.5 ± 0.5 99 ± 6 0.089 7.3 ± 0.5 0.19 22 ± 2 0.11 D88N 36.5 ± 1.5 16 ± 1 0.077 ND ND 2.9 ± 0.5 ND D393A 10.0 ND ND 8.7 2890 M. Umemura et al. (Eur. J. Biochem. 271) Ó FEBS 2004 hydrogen bond network containing Asp393 is essential for the binding of NADH rather than the electron transfer to form I. Confusion r egarding the isosbestic point (at 440 nm) in the l atter stage of the process ( Fig. 4) is due t o the appearance of the 413 nm species (resting Fe 3 + ). This means that the k obs is not sufficiently high for conversion of all of the 431 nm (Fe 3 + –NO) species to I. This situation is intermediate between the results in Fig. 3A and B. Thus, if the r eduction of the D393A mutant is performed with NADH at a higher con centration, the formation of I should be more complete, affording a clearer isosbestic point. Saturation kinetics of the intermediate formation by the D88A or D88V mutant The D 88A or D88V mutation e nables the intermediate I to accumulate even after slow reduction by NADPH. This suggests that the k obs can be d etermined in a wide range of NADPH concentrations for kinetic analysis when these mutant proteins are utilized, which is impossible with the wild-type P450nor, a s noted above. As shown in Fig. 5, the k obs for the reduction step for the D88A mutant showed saturation kinetics in terms o f the NADPH c oncentration, affording V max (k red ; first order reduction rate) and K m values (10 °C) for NADPH of 12.7 s )1 and 0.64 m M , Fig. 3. Spectral changes during reduction with NADPH of the Fe 3+ – NO complex of the wild-type (A) and D88A mutant (B) of P450nor. Each spectrum was recorded with a rapid scan apparatus at t he indi- cated time after mixing the Fe 3 + –NO complex solution with NADPH solution (fina l c oncentration, 0 .1 m M ), as des cribed in M aterials and methods. In (A) the Fe 3 + –NO complex (431 nm species) was con- verted to the resting (Fe 3 + ) state (413 nm species) due to catalytic turnover without a ccu mulation of the intermed iate I.In(B)Fe 3 + –NO was converted to I (444 nm species) upon reduction with NADPH. The k obs for I formation is usually obtained from the time-dependent decrease in the absorbance at 427 nm, at which the isosbestic point between the spectra of I (444 nm species) and the Fe 3 + state (413 nm species) exists (cf. Fig. 6). Thus, a time-dep endent trace of I formation can avoid the interf erence due to the s pontan eous dec omposition of I that follows its formation, although the rate of decomposition is m uch slower than that of I formation (cf. Figure 6). Fig. 5. Saturation kinetics observed o n the reduction (I formation) of th e Fe 3 + –NO complex of the D 88A m utant with N ADPH. The k obs for the reductionwasobtainedateachNADPHconcentrationasdescribedin the l egend to F ig. 3. The mean value fo r two to four experiments w ith each NADPH concen tratio n was used for each plot. The data we re fitted with KALEIDAGRAPH (Abelbeck Software), w hich gave V max and K m values of 12.7 ± 0.55 s )1 and 0.64 ± 0.077 m M , respectively. Fig. 4. Formation of the spectral inter mediate ( I ) observed upon reduction of the Fe 3 + –NO complex of the D393A mutant with a higher concentration o f NADH. Each sp ectrum was obtained as i n Fig. 3 after mixing th e Fe 3 + –NO c omplex of the D 393 mutant with N ADH (final, 0.5 m M ). Ó FEBS 2004 Asp88 in haem-distal pocket of P450nor (Eur. J. Biochem. 271) 2891 respectively. Similarly, the D88V mutant exhibited satura- tion kinetics (data not shown ) with k red and K m values of 29.4 s )1 and 0.2 9 m M , respectively. The results de monstrate that reduction of the P 450nor–NO complex by NADPH proceeds in an enzymatic man ner (Eqn 5) and not in a chemical reaction manner (Eqn 6), and thus the ternary complex of P450nor, NO, and NADPH should be formed prior to the electron transfer from NADPH to form I.This mechanism should b e ascribed to the reduction step due to NADH (Eqn 3). Spontaneous decomposition of intermediate I The intermediate I is so stable that its a ccumulation can be observedwitharapidscanapparatus,whereasI slowly decomposes after completion of the reduction ( I formation) to give th e resting form ( Fe 3 + ) (Fig. 6). The decomposition process comprises single exponential decay, and t he rate constant for the decomposition (k dec ) can be obtained by following the c hange in a bsorbance at 440 nm, which is the isosbestic point between the Fe 3 + –NO state and I [7]. We observed this process in addition to the I-forming process with each mutant protein, and the obtained k dec (I decomposition) values a re listed i n Table 1. The k dec value f or each mutant (0.10–0.19 s )1 ) w as increased by several fold as c ompared with that (0.027 s )1 ) for the wild- type enzyme, indicating that I became more unstable with mutation. The k dec value was a lso shown to be i ndependent of the N AD(P)H concentration used for the I- forming process, as previously observed for the wild-type enzyme [7]. Competitive inhibition by chloride ions Halogen ions such as chloride and b romide are reverse type I ligands for P450nor, a nd inhibit its enzymatic r eaction [4,11]. T wo binding sites for bromide were revealed by X-ray crystallography [11], one of which was located near the haem and termed the anion hole. Chloride ions (Cl – ) caused spectral perturbation of reverse type I i n the bound haem of the D88A mutant (Fig. 7A), which is similar to t hat observed on its binding to wild-type P450nor [11,18]. The K d for the complex was determined by spectrophotometric titration (Fig. 7B) to be 0 .69 M for the mutant, which was almost equal to the value for the wild-type enzyme (data not shown). The results indicated that the environment around the anion hole was not modulated by the mutation . It has now become possible to k inetically analyse inhibition by Cl – utilizing the I-forming process due to NADPH of the D88A mutant. As expected, Cl – inhibited the process in a manner competitive w ith N ADPH, the K i being 0 .70 M (Fig. 8 ). The excellent agreement of the K d (K i ) values obtained with the different methods (Figs 7 and 8) strongly supports the conclusion above that the ternary complex between the Fe 3 – NO binary complex and NADPH is formed prior to the electron transfer from NADPH to the b inary complex. Discussion Here we found an intriguing phenomenon concerning the properties of mutants D88A and D88V. The reducing half reaction of these mutants yielding I was not blocked although the overall N OR activity was decreased to a Fig. 6. Spontaneous decomposition of intermediate I. The Fe 3 + –NO complex o f th e D8 8A m utant wa s r educ ed wit h 0 .1 m M NADH as in Fig. 3, and each spectrum was recorded at the indicated time (after mixing), it being shown that I is in a quasi-stable state, i.e. it i s decomposing slowly. Fig. 7. Spectral perturbation upon binding of Cl – to the D88A m utan t. (A) Absolute spectra of the ferric res ting (solid line ) and Cl – –bound (broken line) forms of the mutant protein. KCl, 1.0 M (B) Titration with KCl. The difference between the spectra with and without Cl – was recorded at each KCl concentration, in 10 m M potassium phosphate buffer pH 7 .2. 2892 M. Umemura et al. (Eur. J. Biochem. 271) Ó FEBS 2004 considerable extent (Table 1). Accumulation of I could be observed even after slow reduction of the mutants by NADPH (Fig. 3B and Table 1) in cont rast to the little accumulation on the reduction of wild-type P450nor under the same conditions (Fig. 3A). It is evident that this accumulation of I did not arise from the stabilization of I as r egards spontaneous decomposition, as the decomposition rate (k dec ) increased with the mutation (Table 1). As noted above, the I formation (Eqn 3) and the following reaction of I with the second NO (Eqn 4) compete with each other for free NO under the conditions used, and the accumulation of I means that the former reaction (I formation) overcomes the competition. Acquisition of the ability by the mutant proteins to accumulate I after slow reduction indicates that the rate-limiting step in the NADPH-dependent overall reaction changes with the mutation, and that the new rate-limiting step should be the process subsequent to the formation of I. Two events must occur following I formation during catalytic t urnover, i.e. dissociation of NAD(P) + from the p rotein and subse- quent reaction of I with the s econd NO (Eqn 4). Because Asp88 i s located rather far away from the bound haem, blocking of the release of NAD(P) + is more probable than that of Eqn 4 (which must involve the haem) as the cau se of the inactivation o f P450nor by the mutation. It is also intriguing that a m utation to also replace Asp88 with a hydrophilic amino a cid (D88N) had an inhibitory effect on I formation, in contrast with other m utations (D88A and D88V). On t he other h and, the accumulation of I was still observed even in the case of slow reduction of the D88N mutant with a higher c oncentration of N ADPH (1.0 m M ; Table 1), suggesting that the subsequent process was a lso b locked in this mutant. It therefore seems that the replacement of Asp88 with a hydrophilic re sidu e (D88N) blocked both I formation and the subsequent step. The opposite effects on the I formation of these mutations are rather difficult to explain, while it would appear that the hydropathy of the amino acid residue at the 88th site would affect the hydrogen bond network c ontaining many water molecules, which would be important for the reduction step [10]. Determination of the properties of the D88A mutant made it possible f or the first time to perform kinetic analyses of the reduction step and the inhibition by Cl – . The kinetic analysis of the competitive inhibition (Fig. 8) was based on the assumption that enzyme–subst rate complex (Michaelis complex) and enzyme–inhibitor complex formation are both in rapid equilibrium as compared with the following catalytic process ( electron t ran sfer from N ADPH in this case). The excellent agreement between the K iand K d values, respectively, obtained by kinetic (Fig. 8) and spectrophoto- metric (Fig. 7 ) analyses means that this assumption is valid. It can thus be concluded that the reduction step (Eqn 3) progresses in an enzymatic manner (Eqn 5), that is, reversible complex formation between P450nor and NADPH (or NADH) precedes the electron transfer from NAD(P)H t o the Fe 3 –NO c omplex to yield I. Thus, the present results are the first kinetic evidence supporting our assumption that P450nor directly binds to NAD(P)H [11,12], although such direct binding of NADH is unpre- cedented for a P450. The kinetic analyses (Fig s 7 and 8 ) a lso provided the first evidence that C l – binds to P450nor in a manner c ompetitive in terms of NADH (or N ADPH). T he competitive inhibition by Cl – highlights the key role of the anion hole (the Br – binding site near haem) [ 11] in the interaction with NAD(P)H. Now, many amino a cid residues located inside the h aem- distal pocket h ave b een identified a s b eing important for the interaction with NAD(P)H. They a re Lys62, Arg64, Arg174, Lys291, Arg292 [11], S er286 [9,10], Thr243 [ 19], Asp393 [9,10] (present study), and Asp88 (present study). All o f t hese charged or hydrophilic amino acid residues are conserved among P450nor isozymes [6,20]. It i s noteworthy that many of these c harged amino acid r esidues (Lys62, Arg64, Asp88, Arg174, Lys2 91, and Arg292) are concen- trated in a rather narrow area in the pocket (Fig. 1), suggesting that these charged residues form an access channel f or NADH. It is a lso noteworthy that A sp88 is exceptional among these amino acid residues in that its mutation (D88A or D 88V) decreased t he overall NOR activity without blocking the I formation step. This phenomenon could be attributed to blocking by the mutation of the steps subsequent to I formation, as noted above. On the other hand, immediate dissociation of NAD + is also essential f or attaining the e xtremely high catalytic turnover of the P450nor reaction. It would therefore appear that the inclusion of a negative charge (Asp88) in the positive charge cluster is important for releasing NAD + , leading to a charge balance in the access channel. This charge balance would b e important for both binding to NADH and release of NAD + . It is evident from our present and previous results that P450nor has evolved so as to interact directly with NAD(P)H by having many charged and hydrophilic amino acid residues in its distal pocket. This unique molecular evolution of P450nor is in sharp contrast w ith that o f other members of the P450 superfamily that have evolved a hydrophobic haem-distal pocket. Acknowledgements This study was supported by a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science 14104005 (to H.S.). Fig. 8. Inhibition by Cl – of the NADPH-dependent I formation of t he D88A mutant. The k obs was obtained as in Fig. 5 at each NADPH concentration in the presence of the indicated amount of KCl (0, 0.1 or 0.5 M ). Ó FEBS 2004 Asp88 in haem-distal pocket of P450nor (Eur. J. Biochem. 271) 2893 References 1. Nelson, D.R. (1999) C ytochrome P450 and the individuality of species. Arch. Biochem. Biophys. 369, 1–10. 2. Omura, T. (1999) Forty years of cytoc hrome P450. Biochem. Biophys. Res. Commun. 266, 690–698. 3. Sono, M., Roach , M .P., Cou lter, E.D . & Dawson, J.H. (1996) Heme-containing oxygenases. Ch em . Rev. 96, 2841–2887. 4. Nakahara, K., Tanimoto, T., Hatano, K., Usuda, K. & Shoun, H. (1993) Cytochrome P-450 55A1 (P-450dNIR) acts as nitric oxide reductase employing NAD H as the direct electron don or. J. Biol. Chem. 268, 835 0–8355. 5. Usuda, K., T oritsuka, N., Matsuo, Y., Kim, D H. & S houn, H. (1995) Denitrification by the fungus Cylindrocarpon tonkinense: Anaerobic cell g rowth and two isozyme forms o f cytochrome P-450nor. Appl. Environ. Microbiol. 61 , 883–889. 6. Zhang, L., Takaya, N., Kitazume, T., Kondo , T. & Shoun, H. (2001) Purification and cDNA cloning of nitric ox ide reductase cytochrome P450nor (CYP55A4) from Trichosporon cutaneum. Eur. J. Biochem. 268, 3198–3204. 7. Shiro, Y., Fujii, M., Iizuka, T., Adachi, S., Tsukamoto, K., Nakahara, K. & Shoun, H. (1995) Spectroscopic and kinetic stu- dies on re action of cytochrome P450nor with nitric oxide (NO): Implication for its NO reduction mechanism. J. Biol. Chem. 270, 1617–1623. 8. Daiber,A.,Nauser,T.,Takaya,N.,Kudo,T.,Weber,P.,Hults- chig, C., Shoun, H. & U llrich, V. (2002) Isotope effec ts and intermediates in the reduction of NO by P450NOR. J. Inorg. Biochem. 88, 3 43–352. 9. Park, S.Y., Shimizu, H., Adachi, S., Nakagawa, A., Tanaka, I., Nakahara, K., Shoun, H., Ob ayashi,E.,Nakamura,H.,Iizuka,T. & S hiro, Y. ( 1997) Crystal structure of n itric oxide reductase from denitrifying fungus Fusarium oxysporum. Nat. Struct. Biol. 4, 827– 832. 10. Shimizu, H., O bayashi, E., Gomi, Y., Arakawa, H., Park, S Y., Nakamura, H., Adachi, S., Shoun, H. & Shiro, Y. (2000) Proton delivery in NO reduction by fungal nitric-oxide red uctase: Cryo- genic crystallography, spectroscopy, and k inetics of ferric-NO complexes of wild-type and mutant enzymes. J. Biol. C hem. 275, 4816–4826. 11. Kudo, T., Takaya, N., P ark, S Y., Shiro , Y . & S hou n, H. (2001) A positively charged cluster formed in the heme-distal p ocket of cytochrome P450nor i s e ssential f or interaction w ith N ADH. J. Biol. Chem. 276, 5020–5026. 12. Zhang, L., Kudo, T., Takaya, N. & Shoun, H. (2002) The B’ helix determines cytochrome P450nor specificity for the electron donors NADH and NADPH. J. Biol. C hem. 277, 33842–33847. 13. Sambrook, J., Fritsch, E.F. & Maniatis, T. (1989) Molecular Cloning: a L aboratory Manual, 2 nd edn. Cold Spring Harbor Laboratory, Cold. Spring Harbor Press, New York. 14. Higuchi, R., Krummel, B. & Saiki, R.K . (1988) A general method of in vitro preparation and spec ific mutagen esis of DNA frag- ments: s tudy of protein and DNA interactions. Nucleic A cids Res. 16, 7351–7367. 15. Kizawa,H.,Tomura,D.,Oda,M.,Fukamizu,A.,Hoshino,T., Gotoh, O., Yasui, T. & S houn, H. (1991) Nu cle otide sequ ence of the unique nitrate/nitrite-inducible cytochrom e P-450 cDNA from Fusarium oxysporum. J. Biol. Chem. 26 6 , 10632–10637. 16. Nakahara, K. & Shoun, H. (1996) N-Terminal processing and amino acid sequences of two isofo rms of nitric oxide reductase cytochrome P450 nor from Fusarium oxysporum. J. Bioc hem. 120, 1082–1087. 17. Omura, T. & Sato, R. (1964) The carbon monoxide-binding pig- ment of liver microsomes. II. Solubilization, purification and properties. J. Biol. Chem. 239, 2379–2385. 18. Shoun, H., S udo, Y., Seto, Y. & Beppu, T. (1983) Purification a n d properties of a cytochrom e P-450 of a f ungus, Fusaium oxysporum. J. Biochem. 94, 1219–1229. 19. Okamoto, N., Imai, Y., Shoun, H . & Shiro, Y. (1998) Site- directed mutagenesis o f t he conse rved thr eonine (Th r243) o f the distal helix of fun gal cytochrome P4 50nor. Biochemistry 37, 8839– 8847. 20. Kudo,T.,Tomura,D.,Liu,D.,Dai,X.&Shoun,H.(1996)Two isozymes of P450nor of Cylindrocarpon tonkinense:molecular cloning of the cDNAs and genes, expression in the yeast, and the putative NAD(P)H-binding site. Biochimie 78, 792–799. 2894 M. Umemura et al. (Eur. J. Biochem. 271) Ó FEBS 2004 . D88A mutant of cytochrome P450nor provides kinetic evidence for direct complex formation with electron donor NADH Mariko Umemura 1 ,. properties, providing kinetic evidence for the direct complex formation of P450nor with NADH. Materials and methods Mutagenesis and expression plasmids The construction of