Update on Mechanism and Catalytic Regulation in the NO Synthases*

Nitric-oxide synthases (NOSs, EC 1.14.13.39) 1 oxidize L -ar-ginine to nitric oxide (NO) and are interesting for several reasons. They are present in many life forms (1, 2), their gene regulation is complex (3), they are the only flavoheme enzymes that utilize tetrahydrobiopterin (H 4 B) as a redox cofactor, and their electron transfer reactions are regulated by a Ca 2 (cid:1) -bind-ing protein (calmodulin). In the past 5 years, crystal structures of NOS heme (oxygenase) domains and bacterial NOS-like proteins have shown how Arg, heme, and H 4 B bind in the active site (4, 5). Reviews are available on NOS biochemistry (6), regulation (7, 8), protein-protein interactions (9), and post-translational modifications (10). This minireview updates the NO biosynthetic mechanism and describes a global catalytic model that highlights the role of NO as an intrinsic regulator.

Nitric-oxide synthases (NOSs, EC 1.14.13.39) 1 oxidize L-arginine to nitric oxide (NO) and are interesting for several reasons. They are present in many life forms (1,2), their gene regulation is complex (3), they are the only flavoheme enzymes that utilize tetrahydrobiopterin (H 4 B) as a redox cofactor, and their electron transfer reactions are regulated by a Ca 2ϩ -binding protein (calmodulin). In the past 5 years, crystal structures of NOS heme (oxygenase) domains and bacterial NOS-like proteins have shown how Arg, heme, and H 4 B bind in the active site (4,5). Reviews are available on NOS biochemistry (6), regulation (7,8), protein-protein interactions (9), and posttranslational modifications (10). This minireview updates the NO biosynthetic mechanism and describes a global catalytic model that highlights the role of NO as an intrinsic regulator.

Mechanism of NO Biosynthesis
NOS is one of few heme-containing enzymes that make NO (11,12). NOS hydroxylates a guanidino nitrogen of Arg and then oxidizes the N -hydroxy-L-arginine intermediate (NOHA) to NO and L-citrulline (Scheme 1). The NOS flavoprotein domain first provides an electron (derived from NADPH) to the ferric heme (Fig. 1). This is the slowest step of the biosynthetic reaction and enables formation of a ferric heme-superoxy species (I) in the Arg or NOHA reactions (13,14). Species I is not reactive toward Arg but may (15) or may not (13) be reactive toward NOHA. Rates for many of the individual binding and electron transfer steps are known (16,17). Species I can receive an electron from H 4 B (18) or from the flavoprotein domain when H 4 B is absent (19). The H 4 B electron transfer is the second slowest step in the biosynthetic reaction. Its kinetics is influenced by surrounding protein residues and by the pterin structure itself (18,20). Timely electron transfer from H 4 B prevents superoxide release (Fig. 1). H 4 B may also donate an electron in the NOHA reaction (13,21), and in that case the radical is reduced back to H 4 B by a downstream reaction intermediate. Further discussion of H 4 B redox function in NOS is available (22,23). The heme-peroxo species (II) has only been observed in NOS at cryogenic temperature (24). Species II may become protonated and lose water to form a heme iron-oxo species (III) that hydroxylates Arg or may react directly with NOHA (Fig. 1). The reactivity of NOS species III or a related species has been studied (25). Importantly, the first observed product of NOS catalysis is a ferric heme-NO complex and not free NO (13,26,27) (Fig. 1).

A Global Mechanism for NOS Catalysis
As noted, practically all NO binds to the NOS ferric heme before exiting the enzyme. In isolation, this is just an example of how product release can limit enzyme catalysis. However, the attached flavoprotein domain of NOS provides an alternative path for the ferric heme-NO complex to return to the initial ferric state ( Fig. 2A). The flavoprotein can reduce the ferric heme-NO complex to the ferrous heme-NO species, which releases NO very slowly (17,27) and so reacts instead with O 2 to regenerate ferric enzyme. Consequently after NO biosynthesis is finished two different cycles compete; NO dissociation from the ferric heme (k d ) is part of a "productive cycle" that releases NO and is essential for NOS bioactivity. Conversely, reduction of the ferric heme-NO complex (k r Ј) channels the enzyme into a "futile cycle" that ultimately generates nitrate in place of NO. NOS futile cycling is also influenced by the rate at which O 2 reacts with the ferrous heme-NO species (k ox in Fig. 2). Together, the productive and futile cycles create a global kinetic mechanism for NOS catalysis. Thus, to synthesize NO is good but not sufficient; a NOS must also control partitioning between both cycles by balancing heme reduction (k r , k r Ј) 2 and NO dissociation (k d ) if it is to release the NO that it makes.
Computer simulations of a more detailed global kinetic mechanism 3 have been run using individual rate measurements available in the literature (16,17), and these accurately model the pre-steady-state and steady-state behaviors of the mammalian endothelial (eNOS), neuronal (nNOS), and inducible (iNOS) NOS isozymes. Some fundamental concepts are described below.
Each NOS Distributes Differently during Catalysis-Values of the three kinetic parameters (k r , k ox , k d ) differ significantly among NOSs (Table I), and this causes each NOS to distribute differently during steady-state NO synthesis. The enzyme distributions in Fig. 2B were derived from computer simulations of the global kinetic mechanism and mimic distributions estimated from actual experiments (17,28,29). NOSs distribute into five main forms during steady-state NO synthesis, namely the ferric, ferrous, ferrous-O 2 (or ferric-superoxy), ferric-NO, and ferrous-NO forms. For nNOS, a fast k r , k r Ј relative to k d and k ox causes it to exist predominantly as a ferrous-NO species. For eNOS the situation is reversed; a slow k r relative to k d or k ox causes it to exist predominantly as a ferric species. For iNOS, its moderately fast k r and fast k ox create an enzyme distribution that is between the two extremes.
Knowing the distribution pattern helps to understand NOS catalytic behavior. For example, the specific activities of the three mammalian NOSs are of rank order iNOS Ն nNOS Ͼ eNOS, with nNOS activity being 4 times that of eNOS. However, if one considers their actual rates of NO biosynthesis (the speed at which each NOS makes one NO) it is clear from their k r values ( Table I) that nNOS is at least twice as fast as iNOS and about 30 times faster than eNOS. This discrepancy is explained by the global kinetic model and the different enzyme distribution pattern of each NOS in the steady state (17).
NOSs Have an Optimal Rate of Heme Reduction-With sufficient substrates the rate of NO biosynthesis is simply limited by k r (Fig. 1). However, the relationship between k r and the rate of NO release is modified by ferric heme-NO complex formation at the end of each catalytic event. The global kinetic model predicts that NOS will increase its NO release as a function of k r , but at a certain point the NO release will reach a maximum and then fall despite the enzyme working faster and faster (Fig. 3). The bell-shaped curves can be rationalized by considering how k r impacts the ferric heme-NO product complex. Increasing k r speeds its formation but also partitions more of it into the futile cycle (k r Ј), which diminishes NO release via the productive pathway and reciprocally increases futile cycling and nitrate production (Fig. 3). The relationship between k r and the rate of products release is also influenced by the k d and k ox values of each NOS. The approximate location of each NOS on its curve is indicated by an arrow in Fig. 3. Note that nNOS has evolved a near optimal k r , whereas eNOS and iNOS have sub-optimal rates.
The Effect of External NO-Binding solution NO to the NOS heme becomes significant at low micromolar NO concentrations. This equilibrium binding event differs from the reaction of newly generated NO with the ferric heme in the heme pocket. Solution NO binding is particularly important when the ferric form of a NOS predominates during steady-state catalysis (as for eNOS and iNOS). In this circumstance, solution NO binding increases the concentration of the ferric heme-NO species and alters the enzyme distribution pattern (Fig. 4A). It leads to greater production of the ferrous heme-NO species and so shunts more NOS through the futile cycle, which consequently lowers the NO release rate. Thus, NOS activity can differ considerably when it is measured in the presence or absence of an NO scavenger. The effect of solution NO has been documented for the three NOS (28 -30), and interesting circumstances can result. For example, at about 1 M NO concentration iNOS becomes equally efficient at generating and scavenging NO (Fig. 4B). This enables iNOS to maintain a steady NO concentration in the environment even while it supports almost no net release of NO (17,29). At higher NO concentrations, the fast k ox of iNOS enables it to be an NADPH-dependent NO oxygenase at the expense of its NO biosynthesis activity (Fig. 4B).

Interplay of Three Kinetic Parameters Determines Activity
NOS activity depends on interplay of k r , k ox , and k d . With this in mind, we provide two examples of NOS catalytic behavior that could only be understood in the context of the global kinetic model and the three kinetic parameters.
More Is Less: the Case of S1412D nNOS-Both eNOS and nNOS contain a consensus sequence for Akt-dependent Ser phosphorylation in their C-terminal regulatory elements (8,9). Phosphorylation at this Ser (or point mutation to Asp) caused a 3-fold increase in eNOS activity. However, identical point mutation in nNOS (S1412D) lowered its NO release rate by 30% (31). Analysis showed that S1412D nNOS actually has a faster k r than wild type (Table I) and therefore a faster NO biosynthesis. Calmodulin mutants were used to slow down k r in S1412D nNOS. This slowed down its NO biosynthesis but increased its NO release rate (31). This occurs because the S1412D mutation increased k r beyond optimal in nNOS. Conversely, increasing k r in eNOS by the same mechanism should increase its NO release because its basal k r is below optimal (Fig. 3).
Less Is More: the Case of W409F nNOS-Most NOSs contain

FIG. 2. Global kinetic model for NOS (A) and NOS distribution patterns during steady-state NO synthesis (B).
A, ferric enzyme reduction (k r ) is rate-limiting for the biosynthetic reactions (central linear portion). k cat1 and k cat2 are the conversion rates of the Fe II O 2 species to products in the Arg and NOHA reactions, respectively. The ferric heme-NO product complex (Fe III NO) can either release NO (k d ) or become reduced (k r Ј) to a ferrous heme-NO complex (Fe II NO), which reacts with O 2 (k ox ) to regenerate ferric enzyme.

Minireview: NO Synthase Mechanism and Catalytic Regulation 36168
a Trp residue whose side chain forms a hydrogen bond with the heme thiolate ligand (Fig. 4C). Substituting Phe or Tyr for Trp-409 in nNOS created mutant enzymes whose NO release rates were 2-5 times greater than wild-type nNOS (32). The Trp-409 mutants actually have a decreased k r compared with wild type but have a faster k ox (Table I). Although this makes their NO biosynthesis slower than wild type, the increased k ox causes less enzyme to populate the futile cycle during the steady state (Fig. 4C). This "positive" effect on enzyme distribution overcomes their slower k r and enables the Trp-409 mutants to have greater NO release rates than wild type.

Some Implications
Why Do NOSs Have Slow Heme Reduction Rates?-Other flavoheme enzymes have faster k r values compared with the NOSs (Table I), suggesting that NOSs are under a unique selective pressure. Their slow k r is understandable, because increasing k r beyond optimal will partition more NOS into the futile cycle and ultimately convert it into an NO oxygenase (Fig. 3). However, the slow k r makes it difficult for NOS to couple oxygen activation to substrate oxidation (19,33). This is because delivery of the second electron to the heme must be rapid enough for the enzyme to generate the heme-oxy species that will react with Arg or NOHA before autoxidation of ferricsuperoxy species I occurs (Fig. 1). NOS solves this dilemma by utilizing H 4 B as a source of the second electron. H 4 B delivers the second electron about 3-30 times faster than can the NOS flavoprotein, and this difference is sufficient to minimize superoxide release from the heme and so enable coupled oxygen activation (20,23). Thus, heme-NO binding imposes a kinetic constraint on NOS heme reduction that impacts its coupled oxygen activation. NOS overcomes this problem by using two kinetically distinct sources of electrons: a slow electron transfer from the flavoprotein to minimize ferrous heme-NO formation and futile cycling, and a fast reduction by H 4 B at the kinetically sensitive step in its oxygen activation cascade.
The Unusual O 2 Response of NOSs-Apparent K m O 2 values differ among the NOSs and in some cases are much higher than those reported for related monooxygenases (Table I). This is because NOS interacts with O 2 in two ways; the ferrous enzyme binds O 2 during NO biosynthesis, and the ferrous heme-NO species reacts with O 2 in the futile cycle (k ox ). Both  3. How heme reduction rate (k r ) alters NOS product release rates in the steady state. Simulations were performed under the conditions described in Fig. 2. The arrows locate the approximate heme reduction rates of each NOS measured experimentally at 10°C (16,17). B) and enzyme distribution pattern of W409F nNOS during the steady state (C). A and B, simulations were done as described in Fig.  2 but factored in the presence of solution NO at the indicated concentrations. C, the structure was adapted from the crystal structure of mouse iNOS oxygenase (1DWX) and highlights a hydrogen bond between Trp-409 and the heme-thiolate ligand (Cys-415) in nNOS. The enzyme distribution shown in the pie graph was calculated by simulation using kinetic parameters from Table I (34,35), suggesting it is intrinsic to the enzyme. Solution NO binding to the ferric heme also increases the apparent K m O 2 of eNOS and iNOS (28,29). A low K m O 2 may be required for eNOS to increase blood supply under hypoxic conditions, whereas a higher apparent K m O 2 of iNOS explains how it may regulate ventilation-perfusion matching in the lung (36). Different Strategies to Increase NO Release-NOSs are likely to increase their activities by different mechanisms. For example, increasing k r only makes sense for eNOS and iNOS because they have a suboptimal rate. Indeed, eNOS activators like phosphorylation, HSP 90, and dynamin (9, 10) could function this way. For nNOS a better strategy would be to increase its k ox , as demonstrated by the Trp-409 nNOS mutants. Whether mechanisms exist to do so in cells can now be investigated.

Concluding Remarks
There are gaps in our understanding of NO biosynthesis. For example, the heme-oxy species that react with Arg or NOHA have not been demonstrated. This will require that the rate of the second electron transfer be increased and/or the reactive heme-oxy species be stabilized, perhaps using strategies developed for other monooxygenases (37,38). The global kinetic mechanism prompts its own questions. How do values for the three kinetic parameters differ among NOSs throughout the animal kingdom? Are the kinetic parameters of some NOSs set to favor futile cycling instead of NO release? Bacterial NOSs may be configured this way (39). What is the physical basis of the three kinetic parameters? Presently, we only know that the NOS flavoprotein determines k r (40,41) whereas the oxygenase domain determines k ox and k d (17). The wide variation in k ox and k d values among heme proteins (Table I) suggests that interesting structure-function relationships remain to be explored. NOS function in signaling and cytotoxicity may also be linked to its productive and futile catalytic cycles. Although written as nitrate, the identity of the futile cycle product depends on the oxidation mechanism of the ferrous heme-NO complex and could be the cytotoxic molecule peroxynitrite. It will be interesting to see how NOS product release, which differs according to the environment and kinetic parameters of a NOS, correlates with its physiologic roles.