Interactions between the Isolated Oxygenase and Reductase Domains of Neuronal Nitric-oxide Synthase

Nitric-oxide synthase (NOS) is a fusion protein composed of an oxygenase domain with a heme-active site and a reductase domain with an NADPH binding site and requires Ca2+/calmodulin (CaM) for NO formation activity. We studied NO formation activity in reconstituted systems consisting of the isolated oxygenase and reductase domains of neuronal NOS with and without the CaM binding site. Reductase domains with 33-amino acid C-terminal truncations were also examined. These were shown to have faster cytochrome c reduction rates in the absence of CaM.NG -hydroxy-l-Arg, an intermediate in the physiological NO synthesis reaction, was found to be a viable substrate. Turnover rates forNG -hydroxy-l-Arg in the absence of Ca2+/CaM in most of the reconstituted systems were 2.3–3.1 min−1. Surprisingly, the NO formation activities with CaM binding sites on either reductase or oxygenase domains were decreased dramatically on addition of Ca2+/CaM. However, NADPH oxidation and cytochrome c reduction rates were increased by the same procedure. Activation of the reductase domains by CaM addition or by C-terminal deletion failed to increase the rate of NO synthesis. Therefore, both mechanisms appear to be less important than the domain-domain interaction, which is controlled by CaM binding in wild-type neuronal NOS, but not in the reconstituted systems.

The interaction between the oxygenase domain and the reductase domain is not clearly understood, despite its importance in controlling intermolecular/intersubunit electron transfer in the NOS enzyme system. The catalytic properties of the oxygenase domains (not including the CaM binding sites) of inducible NOS (iNOS) and endothelial NOS (eNOS) in reconstituted systems with the isolated recombinant reductase domain (including the CaM-binding site) have been reported (21,22). Hitherto, no study of the neuronal NOS (nNOS) oxygenase domain in a similar reconstituted system has been reported. In addition, it is worthwhile examining how the CaM-binding site, CaM binding itself, and the C-terminal truncation of the isolated reductase domain affect the reconstituted system. Another interesting issue is whether or not native CPR purified from rat liver microsomes is able to support NO synthesis in the reconstituted system containing oxygenase domain of NOS.
In the present study, we examined the interaction between the oxygenase domain of rat nNOS with and without the CaM binding site and the reductase domain with and without the CaM binding site. Reductase domains truncated by 33 amino acids at the C terminus (18,19) and native CPR purified from rat liver microsomes were also examined in conjunction with the nNOS oxygenase domains. NO formation activity was observed with the substrate NHA in a reconstituted system composed of the oxygenase and reductase domains with and without the CaM binding sites. Surprisingly, addition of Ca 2ϩ /CaM to the system markedly inhibited the NO formation activity of the reconstituted systems in direct contrast to its effect on the wild-type enzyme.

Materials
The complete cDNA for rat nNOS (GenBank accession no. X59949) in pBluescript SK(Ϫ) was kindly provided by Dr. S. Snyder (The Johns Hopkins University School of Medicine, Baltimore, MD). The complete cDNA of GroESL was kindly provided by Dr. A. A. Gatenby (DuPont Central Research and Development, Wilmington, DE). An E. coli expression vector, pCWori ϩ , was provided by Dr. M. R. Waterman (Vanderbilt University School of Medicine, Nashville). E. coli strains JM109 and BL21 used in this study were purchased from Takara Shuzo Co. (Kyoto, Japan). Restriction enzymes and other DNA modifying enzymes were purchased from Takara Shuzo Co., Roche Molecular Biochemicals (Mannheim, Germany), and New England Biolabs (Beverly, MA), Toyobo Co. (Osaka, Japan), and Invitrogen Oriental (Tokyo, Japan). H4B was purchased from Schircks Laboratories (Jona, Switzerland). Oligonucleotides for mutagenesis were purchased from Sawady Technology (Tokyo, Japan). PCR kits were obtained from Takara Shuzo. 2Ј,5Ј-ADP-Sepharose and CaM-Sepharose were products of Amersham Biosciences AB (Uppsala, Sweden). Nickel-nitrilotriacetic acid-agarose was a product of Qiagen Inc. (Valencia, CA). Bio-Spin 30 Tris column was from Bio-Rad. Other reagents of the highest grade available were obtained from Sigma or Wako Pure Chemicals (Osaka, Japan).
Expression plasmids for the nNOS reductase domain with (designated RedCaM) or without (designated Red) CaM-binding site (residues 721-1429 and 746 -1429, respectively) and their C-terminal truncated domains (residues 721-1395 for RedCaM⌬33 and residues 746 -1395 for Red⌬33, respectively) were generated as previously described (16 -20). To clone RedCaM, a forward primer used for PCR was 5Ј-GGAAT-TCCATATGGGGACCCCCACGAAG-3Ј. In the case of Red, a forward primer was 5Ј-GGAATTCCATATG(CAC) 6 GGGCAGGCCA-3Ј, in which His 6 tag was attached to the N terminus of Red. Both forward primers introduced ATG in a NdeI site at the N-terminal end. In both cases, wild-type nNOS cDNA as a template and 5Ј-TCCCCCTCCCT-CATCTT-3Ј as a backward primer were used. The products were isolated in an agarose gel electrophoresis and recovered by band excision after digestion, with SphI and EcoRI enzymes, and ligated to pUC19 vector. The desired products were confirmed by sequencing using an automatic sequencer (DSQ-2000L; Shimadzu Co., Kyoto, Japan). The NdeI/SphI fragments from the desired plasmids and the SphI/XbaI fragment from the wild-type nNOS cDNA were ligated into NdeI/XbaI sites of pCWori ϩ .
Expression plasmids for the C-terminal truncated reductase domain with (designated RedCaM⌬33) and without (designated Red⌬33) CaMbinding site were also constructed in the present study. The synthetic oligonucleotide 5Ј-GGAATTCCATATGGGGACCCCCACGAAG-3Ј was used as a forward primer. nNOS cDNA and 5Ј-GGTCTAGATTAAAA-GATGTCCTCG-3Ј were used as a template and backward primer for PCR. Amplified product was isolated in an agarose gel electrophoresis, recovered by band excision after digestion with XbaI and EcoRI enzymes, and ligated to pBSK II vector. The product was confirmed by sequencing. The SphI/XbaI fragment of the plasmid was ligated to SphI/XbaI fragments of plasmids of RedCaM and Red to construct the expression plasmids for RedCaM⌬33 and Red⌬33.

Expression of nNOS Proteins
Full-length nNOS of wild type and the nNOS oxygenase domains, OxCaM and Ox, were expressed in E. coli cell line, BL21, which contains another plasmid, pGroESL, for expression of chaperone proteins as described previously (12,23,24,28).
Red-The E. coli cells were resuspended in buffer A. The cells were crushed on ice with sonication. After centrifugation at 35,000 rpm for 35 min at 4°C, ammonium sulfate was added to the resulting supernatant up to 40% saturation. The precipitate was collected and then dissolved in buffer C (50 mM sodium phosphate buffer (pH 7.8), 5 M H4B, 1 mM ␤-mercaptoethanol, 0.1 mM EDTA, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, 2 g/ml aprotinin, 2 g/ml leupeptin, 2 g/ml pepstatin) containing 10 mM imidazole. The solution was passed through a Sephadex G-25 column (4 ϫ 20 cm) pre-equilibrated with the same buffer. The eluted solution was applied to a nickel-nitrilotriacetic acidagarose column pre-equilibrated with buffer C containing 10 mM imidazole. The column was washed sequentially with buffer C containing 10 mM and 20 mM imidazole. The Red protein was eluted with buffer C containing 100 mM imidazole. The protein fractions were pooled and concentrated. After concentration the reductase domains were quickly frozen in liquid nitrogen and stored at Ϫ80°C.
Before analysis, the buffer of protein was changed to a Tris buffer (50 mM Tris-HCl (pH 7.5), 0.1 M KCl, 0.1 mM EDTA, 10% glycerol, 5 M H4B, and 20 M DTT) using a Sephadex G-25 column. All purified nNOS enzymes were more than 95% pure, as judged by sodium dodecyl sulfate-polyacrylamide gel electrophoresis stained with Coomassie Blue R-250. The concentrations of the nNOS oxygenase domains were determined optically from the [CO-Fe(II)] Ϫ [Fe(II)] difference spectrum using ⌬⑀ 444 -467 nm ϭ 55 mM Ϫ1 cm Ϫ1 . The ⌬⑀ value was estimated by the pyridine hemochromogen method (25), assuming that one heme is bound to one subunit of the protein. Concentrations of the reductase domains were determined from the absorption spectrum, ⑀ 457 nm ϭ 22.9 mM Ϫ1 cm Ϫ1 (18) for both domains with or without CaM-binding site.

Catalytic Activities
NO concentrations generated in the reconstituted system were fluorometrically determined as NO 2 by using the NO 2 /NO 3 Assay Kit-F (2,3-diaminonaphthalene kit) of Dojindo (Kumamoto, Japan). Unless otherwise indicated, catalytic assays were carried out at 25°C in a reaction mixture containing 50 mM Tris-HCl (pH 7.5), 0.1 M heme domain, 0.5 M reductase domain or CPR, 0.1 mM NADPH, 0.5 mM NHA, 5 M H4B, 20 M DTT, 10 units/ml SOD, and 100 units/ml catalase in the presence or absence of 1 mM CaCl 2 and 10 g/ml CaM. The reaction mixture without the heme domain was preincubated at 25°C for 5 min, after which 0.1 M heme domain was added to start the reaction. The reaction mixture was incubated at 25°C for 30 min. The reaction was terminated by removing the enzyme with a membrane filter, Ultrafree-MC UFC3LGC (Millipore Japan Ltd., Tokyo, Japan), with centrifugation at 7,000 rpm for 10 min.
NO concentrations generated in the H 2 O 2 -dependent system were spectrophotometrically determined as NO 2 Ϫ by using the NO 2 /NO 3 Assay Kit (Griess method) of Cayman Chemical Co. (Cayman, MI). Unless otherwise indicated, catalytic assays were carried out at 25°C in a reaction mixture containing 50 mM Tris-HCl (pH 7.5), 0.1 M NOS, 1 mM NHA, 5 M H4B, 20 M DTT, and 10 units/ml SOD in the presence or absence of 1 mM CaCl 2 and 10 g/ml CaM. The reaction mixture without H 2 O 2 was preincubated at 25°C for 5 min, and then 30 mM H 2 O 2 was added to start the reaction. The reaction mixture was incubated at 25°C for 10 min. The reaction was terminated by adding 13,000 units/ml catalase.
The NADPH oxidation rate was determined spectrophotometrically as an absorbance decrease at 340 nm, using the extinction coefficient of 6.22 mM Ϫ1 cm Ϫ1 (25)(26)(27)(28). Cytochrome c reductase activity was determined by monitoring the absorbance at 550 nm using a ⌬⑀ red-ox ϭ 21 mM Ϫ1 cm Ϫ1 .

Optical Absorption and Fluorescence Spectra
Spectral experiments under aerobic conditions were carried out on a Shimadzu UV-2500 spectrophotometer maintained at 25°C by a temperature controller (25)(26)(27)(28). Anaerobic spectral experiments were conducted on a Shimadzu UV-160A spectrophotometer maintained at 15°C in a glove box under a nitrogen atmosphere with an O 2 concentration of less than 50 ppm (25,28). Fluorescence spectra were obtained with Shimadzu RF-500 and RF-5300PC spectrofluorophotometers. To ensure that the temperature of the solution was appropriate, the cell was incubated for 10 min prior to spectroscopic measurements. Titration experiments were repeated at least three times for each complex. Regression analyses were performed, and lines giving an optimum correlation coefficient were selected (31). Experimental errors were less than 20%.

RESULTS
The OxCaM-Fe(III) complex had a Soret peak around 405 nm and the addition of L-Arg moved the peak to 395 nm, indicating that the substrate binding site is not altered in this domain on elimination of the reductase domain. The Ox-Fe(III) complex had a peak at around 415 nm ascribed to a low spin complex, suggesting that in the absence of the CaM binding site, the properties of the heme are altered. However, the Soret absorption band was moved to 400 nm on addition of L-Arg, confirming that the substrate binding site was preserved in this mutant. To confirm that the substrate and H4B binding sites were preserved, we observed spectral changes with the reconstituted system composed of Ox and RedCaM (Fig. 1). On addition of the substrate, NHA, to the solution, the Soret spectrum altered so as to indicate an increase in the high spin content, but the spectra were still a mixture of the low and high spin complexes. Further addition of H4B resulted in the formation of the high spin complex. Therefore, it is suggested that the NHA and H4B binding sites of the oxygenase domain in the reconstituted system were not altered by isolating the domain from the full-length wild-type protein. The Fe(II)-CO complex did not show any absorption at ϳ420 nm (shown later), suggesting that the thiol coordination to the heme was preserved and no denatured form was detected in the reconstituted systems. These spectral findings are very similar to those observed for the reconstituted system composed of OxCaM and Red, and also to the isolated oxygenase domains per se (1-10).
Red and RedCaM had absorption spectral peaks at ϳ387 and 457 nm with a shoulder at ϳ479 nm and a broad absorption centered at ϳ590 nm (data not shown). These spectral features are very similar to those reported previously (16 -18). As summarized in Table I, the cytochrome c reduction rate with Red was 480 min Ϫ1 , which is similar to that (380 min Ϫ1 ) reported previously (16 -18). The rate with RedCaM in the absence of Ca 2ϩ /CaM was 510 min Ϫ1 , and addition of Ca 2ϩ /CaM increased the rate up to 2,550 min Ϫ1 , as reported previously (16 -18). The rate of cytochrome c reduction with RedCaM⌬33 in the presence of Ca 2ϩ /CaM was higher than that in its absence, unlike full-length nNOS⌬33, which has been shown to be decreased slightly in the presence of Ca 2ϩ /CaM. However, overall, the reductase domains behave as expected.
We detected a small amount of NO formation activity with L-Arg (less than 0.22 min Ϫ1 ) in the reconstituted systems composed of nNOS oxygenase and reductase domains. Ca 2ϩ /CaM addition had little effect on the activity. With CPR, on the other hand, we did not detect any NO formation activity using either the fluorometric method or the Griess method. The H 2 O 2 -supported system was unable to support NO formation activity from L-Arg.
NO formation from NHA was observed for all reconstituted systems composed of the oxygenase domains and the reductase domains or CPR (Table II). By increasing the RedCaM concentration, the activity of OxCaM increased linearly with up to a ratio of 1:5 heme:reductase (Fig. 2). This optimal ratio of the oxygenase to the reductase domain was similar to that ob- served for the iNOS system (21) and was used throughout these experiments. The NO formation activity of the reconstituted system consisting of 0.1 M OxCaM and 0.5 M RedCaM in the absence of Ca 2ϩ /CaM was maximal at 3.11 min Ϫ1 . This was 26% of that (12.0 min Ϫ1 ) obtained for full-length nNOS in the presence of Ca 2ϩ /CaM. Note that the NO formation activity obtained by the fluorometric method is lower than that obtained by the oxyhemoglobin method, because they are calculated after 30-min turnover rather than by the initial rate method. Surprisingly, addition of Ca 2ϩ /CaM to the reconstituted systems containing OxCaM or RedCaM markedly decreased the NO formation rate (Table II). Similar decreases in activity were observed for systems containing Red⌬33 or Red-CaM⌬33 despite the lack of CaM dependence in these systems. With the reconstituted system containing CPR, the maximal activity (0.72 min Ϫ1 ) was observed with an OxCaM:CPR ratio of 1:5, similar to the system containing the nNOS reductase domains. Addition of Ca 2ϩ /CaM to this system also showed marked decrease in activity. The NO formation activity in the Ox/CPR system was very low at 0.09 min Ϫ1 and was largely unaffected by addition of Ca 2ϩ /CaM.
To clarify that the heme active site is involved in the NO formation activity with the reconstituted system, we examined the effect of an inhibitor, N-nitro-L-Arg methyl ester (NAME), on the activity. As shown in Fig. 3, addition of NAME markedly suppressed the NO formation activity in the reconstituted system with OxCaM/RedCaM, confirming that the heme participates in the NO formation from NHA. An inhibitor for the reductase reaction, dipenyleneiodonium chloride, and a weak heme ligand, KCN, also substantially inhibited the reaction. In all the assays conducted, we always added SOD and catalase to avoid any nonenzymatic reactions involving superoxide or H 2 O 2 for NHA and L-Arg. In the absence of SOD and catalase, inhibitory effect by these inhibitors was much less than that in their presence, indicating that most of the NO formation activity described in the present study is derived from the enzymatic reaction in the heme active site. It also should be noticed that addition of H4B is required for NO formation activity in all of the reconstituted systems, confirming that they follow the conventional catalytic mechanism.
To determine the correlation of catalytic NO synthesis with heme reduction in the reconstituted system, we monitored the rate of heme reduction under anaerobic conditions by addition of excess NADPH. Fig. 4A shows how the Fe(II)-CO complexes are formed by adding excess NADPH to the OxCaM/RedCaM system in the presence of Ca 2ϩ /CaM. Almost 80 -90% heme was reduced by adding excess NADPH in the OxCaM/RedCaM system. On addition of Ca 2ϩ /CaM, the heme reduction rate of the OxCaM/RedCaM system was essentially not affected (Fig. 4B). The other systems also showed a heme reduction pattern similar to that observed for the OxCaM/RedCaM system (data not shown). Taken together, the heme reduction rates observed for the reconstituted system appear not to be tightly associated with the NO formation rates in these systems. Note that heme reduction in the iNOS reconstituted system on addition of NADPH was not observed (21), whereas that in eNOS was observed (22). Table III shows the NADPH oxidation rates under various conditions. NADPH oxidation is an alternative measure of the rate of electron transfer to the heme in the wild-type nNOS system, but does not correlate well with the rate of NO formation in this case. For all of the systems with the CaM-binding site on the reductase domain, addition of Ca 2ϩ /CaM increased the NADPH oxidation rate, suggesting that electrons were used in CaM-dependent fashion. A small decrease in the NADPH oxidation rate was observed for the Ox/Red, OxCaM/   Red⌬33, and OxCaM/CPR systems. There is no correlation between NO formation and NADPH consumption. NADPH consumption is much faster than NO formation in all cases (Ͼ10fold), indicating severe uncoupling in the reconstituted systems. This may originate primarily from the reaction of O 2 with the reductase domains alone (which are present in excess).
The reaction of NHA with H 2 O 2 or superoxide anion to form NO and L-citrulline is catalyzed by the NOS oxygenase domain (Ref. 4 and references therein). We examined the H 2 O 2 -dependent (shunt reaction) NO formation activity using NHA as the substrate. Both OxCaM and Ox provided high NO formation activities (4.0 -6.0 min Ϫ1 ) similar to those observed for the full-length nNOS. The NO formation activities observed for OxCaM and full-length nNOS were only slightly (10%) enhanced by Ca 2ϩ /CaM. It appears, therefore, that the mechanism occurring in the shunt reaction is fairly different from that observed in the reconstituted system.
It has been reported that dimer formation is required for NO formation in full-length nNOS (1-7). Both oxygenase domains, OxCaM and Ox, studied here were mainly dimeric both in the presence and absence of H4B in terms of gel-filtration column chromatography (data not shown).

DISCUSSION
Previous studies on the constitutive NOS enzymes have yielded a number of interesting facts. 1) The intramolecular electron transfer from FAD to FMN in the reductase domain is facilitated by CaM binding (1-8, 32, 33). 2) The interdomain electron transfer from the reductase domain to the heme domain is facilitated by CaM binding (1-8, 32, 33). 3) Interdomain electron transfer is conducted in a cross-wise manner from the reductase domain of a subunit to the oxygenase domain of the other subunit (5,9,10). 4) There is an autoinhibitory loop in the FMN-binding subdomain of the reductase domain in nNOS and eNOS not present in iNOS (18, 27, 34 -36). 5) The C terminus of the reductase domain inhibits reduction of FAD (19,20). The present paper explores additional interesting characteristics of CaM-binding and the CaM-binding site.
Substrate Specificities-Optical absorption spectral changes caused by adding L-Arg suggest that the substrate binding site is well conserved in the both oxygenase domains, Ox and Ox-CaM. It is interesting to note, however, that L-Arg is not a viable substrate for the reconstituted system, despite its being the physiological substrate for the holoenzyme. A certain important conformation/configuration appears to be required for the monooxygenation of L-Arg. This important structural feature is a characteristic found only in the dimeric full-length nNOS protein, whereas the oxygenase domains, both OxCaM and Ox, must lack this feature. A previous report has suggested that a specific interaction between the oxygenase domain and the reductase domain may be lacking in the iNOS reconstituted system (21). If this is the root cause, the interaction must be influencing the conformation/configuration of the heme active site, which is responsible for catalytic monooxygenation of L-Arg.
The Isolated Reductase Domains-The cytochrome c reduction rates of RedCaM and RedCaM⌬33 were markedly increased by addition of Ca 2ϩ /CaM (Table I), suggesting that the function of the CaM-binding site in the reductase domain is well preserved in these domains. The cytochrome c reduction rates of full-length nNOS⌬33 in the absence of Ca 2ϩ /CaM are much higher than the wild-type nNOS under the same conditions (Table I) (20). Both Red⌬33 and RedCaM⌬33 in the absence of Ca 2ϩ /CaM also have higher rates than those of the corresponding domains, Red and RedCaM, in the absence of Ca 2ϩ /CaM, confirming that the effect of the ⌬33 mutation is preserved in the isolated domain.
The NADPH oxidation rates for systems containing RedCaM or RedCaM⌬33 in the presence of Ca 2ϩ /CaM were always more than 2-fold higher than in its absence (Table III). Therefore, it appears that the CaM-binding site of the isolated reductase domain binds CaM and facilitates electron transfer even in the reconstituted system. This behavior is also consistent with the effect on cytochrome c reduction, which is similar in the reductase domain of full-length nNOS (25-28).  a The NADPH oxidation rate was determined spectrophotometrically as an absorbance decrease at 340 nm, using an extinction coefficient of 6.22 mM Ϫ1 cm Ϫ1 (25)(26)(27)(28). Experimental conditions were the same as those in Table II.
In Either Domain, the Presence of the CaM-binding Site Appears to Assist NO Formation-The OxCaM systems were found to have slightly higher NO formation activity than the Ox systems in the absence of Ca 2ϩ /CaM (Table II). In particular, those of OxCaM/Red⌬33 and OxCaM/RedCaM⌬33 were 3 times higher than those of Ox/Red⌬33 and Ox/RedCaM⌬33, respectively. The CaM-binding site may itself be an important component of the heme domain, required for optimal NO formation catalysis. The position of the Soret peak of the Ox protein was at a higher wavelength than for the OxCaM protein after purification, suggesting that the heme is in some way different, even though the substrate binding sites are retained. The CaM binding site may therefore have a subtle effect on the enzyme's active site.
The highest NO formation activity was observed for the reconstituted system in which both domains contained the CaM-binding site in the absence of Ca 2ϩ /CaM (Table II). The CaM-binding site of RedCaM or RedCaM⌬33 may be important for the domain-domain interaction and must lie in the proximity of the domain-domain interface.
Suppression of Activities by Ca 2ϩ /CaM Binding-The NO formation rate in the reconstituted system was substantially suppressed on addition of Ca 2ϩ /CaM. This is the opposite effect of CaM binding on the full-length wild-type nNOS, with which CaM binding facilitates electron transfer and activates NO formation activity (1)(2)(3)(4)(5)(6)(7)(8)11). The bound CaM could easily impede the catalytically useful collisions between the domains in the reconstituted system. It is suggested that CaM binding activates full-length wild-type nNOS by causing a reorientation between the oxygenase and reductase domains, such that electron transfer becomes viable. When the domains are separate, this effect would be expected to have disappeared. CaM binding does not appear to active nNOS directly at the mechanistic active site.
Previous studies of reconstituted systems consisting of the isolated domains of iNOS (21) and of eNOS (22) did not report that CaM suppresses catalysis. Why are there differences between the present study and previous studies? In iNOS, CaM binds very tightly even in the absence of Ca 2ϩ (1)(2)(3)(4)(5)(6)(7)(8)(9)(10). It would be difficult, therefore, to observe CaM dependence. It is unclear why the study on eNOS indicated that the reconstituted system was deactivated in the absence of CaM, but this may be because of the lower activity of the eNOS reductase domain, which may limit electron transfer more severely in the absence of CaM. This effect also appears to limit overall turnover in full-length eNOS.
Heme Reduction Rates, NADPH Oxidation Rates, and Catalytic Activities-Differences in heme reduction rates between the systems were marginal in the presence and absence of Ca 2ϩ /CaM (Fig. 4) in comparison with the marked difference observed for the NO formation activities (Table I). Note that many of the NADPH-derived electrons are not well coupled with catalysis in the reconstituted system and serve to reduce molecular oxygen and produce superoxide anion and/or H 2 O 2 . Nevertheless, NADPH oxidation rates in the reconstituted systems containing RedCaM or RedCaM⌬33 were markedly increased by adding Ca 2ϩ /CaM (Table III). These are not in accordance with the catalytic findings in that addition of Ca 2ϩ / CaM suppressed the NO formation activities (Table I) in our reconstituted system. The fact that NADPH consumption is not coupled to NO synthesis is unsurprising, given the stoichiometry of 5 reductase domains to 1 oxygenase domain. Electron transfer to the oxygenase domain is clearly hampered on domain separation, but NADPH/oxygen consumption at the flavin sites of the reductase domain will be unaffected. This portion of the NADPH consumption will remain Ca 2ϩ /CaM-dependent and unaffected by the presence of the oxygenase domain.
The dependence of NO synthesis on electron transfer is complicated in nNOS; as explained by recent catalytic models, faster heme reduction does not necessarily lead to faster NO synthesis. The rate of turnover in these reconstituted systems may be affected by changes at any point in such catalytic models. The rate of NO synthesis must be balanced against the rate of superoxide/peroxide formation, the rate of formation/ decomposition of dead-end complexes such as the ferrous heme-NO complex, etc. However, several firm conclusions can be made. 1) The presence of the CaM binding site improves the rate of catalytic turnover and is therefore likely to be an important structural feature of both reductase and oxygenase domains.
2) The effect of CaM-binding on nNOS depends crucially on whether the two domains are linked. CaM binding to both the oxygenase and reductase fragments is detrimental to overall turnover. This reinforces the view that the primary role of CaM binding is to rearrange the two domains with respect to each other. This factor overrides any improvements in reductase activity observed on CaM binding. 3) The ⌬33 mutation, although activating the reductase domain, had little effect on NO synthesis. This correlates with the lack of CaM dependence observed, suggesting that NO synthesis is not controlled by factors within the reductase domain, but by interactions between the reductase and oxygenase domain.
Heme reduction in the reconstituted system was not affected by CaM, unlike the NO formation rate. For one cycle of NO formation, three electrons are required to be transferred from the reductase domain to the heme domain. Resent studies suggested that H4B might provide the second electron required for activation of the ferrous-dioxy intermediate during catalysis (13)(14)(15). In the absence of H4B, no NO formation activity was observed in the reconstituted system (data not shown), suggesting that it plays a critical catalytic role in these systems as well as in native NOS. It is possible that there is a successive or alternative electron transfer pathway via H4B, as has been suggested for the full-length NOS system (13)(14)(15). This electron transfer may be hampered by adding Ca 2ϩ /CaM, rather than electron transfer to the heme, explaining why NO synthesis, but not heme reduction, was impeded by Ca 2ϩ /CaM binding. Nevertheless, the heme remains the center of catalysis, because heme active site inhibitors, NAME and KCN, markedly inhibited catalysis.
A recently isolated bacterial NOS consisting only of the oxygenase domain appears to function similarly with only NHA acting as a viable substrate (37). A comparative study, therefore, may serve as a useful working hypothesis for the further investigation of NOS function.