Cell, Vol. 78, 927-930, September 23, 1994, Copyright 0 1994 by Cell Press Nitric Oxide Synthase: Aspects Concerning Structure and Catalysis Michael A. Marletta Interdepartmental Program in Medicinal Chemistry College of Pharmacy and Department of Biological Chemistry School of Medicine The University of Michigan Ann Arbor, Michigan 46109-l 065 The rapid development of our understanding of the biologi- cal actions of nitric oxide (NO) has, to a large degree, been paralleled by our understanding of the enzyme responsi- ble for the synthesis of NO, nitric oxide synthase (NOS). The relatively fast pace of advance in NOS enzymology is primarily due to the fact that structure/function questions have crossed over several well-established enzymatic problems. As will be outlined below, NOS is a complex enzyme involving several tightly bound redox cofactors that are apparently organized into discrete domains that can be associated with a particular activity. First, the en- zyme has significant homology to NADPH cytochrome P-450 reductase and has been shown to contain a cyto- chrome P-450-type heme and to carry out P-450 chemistry in the formation of NO. What then is the relationship of NOS to the large class of cytochrome P-450 isoenzymes? Second, constitutive NOS isoforms require Ca*+ and cal- modulin (CaM) while inducible NOS (INOS) shows no re- quirement, although these isoforms apparently have CaM as a tightly bound subunit. What is the nature of this im- portant difference in the recognition and binding of CaM? NOS also has a tightly associated reduced pterin that is very important for an enzyme activity whose function is still not known. Third, the product of the reaction, NO, typically is a strong heme ligand. How is it that the enzyme escapes self-inactivation during turnover? These ques- tions shall be the main the focus of this review. General Characteristics of the Reaction The reaction catalyzed by NOS is illustrated in Figure 1. As shown, the reaction requires molecular oxygen (02) and reducing equivalents in the form of NADPH (Marletta, 1993). Except for some minor structural modifications, all NOS isoforms specifically utilize L-arginine as the sub- strate. The products of the reaction are NO and citrulline (presumed to be the L-isomer). It is assumed, given the monooxygenase-like activity of NOS and the heavy iso- tope labeling studies, that H20 is the ultimate fate of the other oxygen atom. The initial NOS cDNA isolated was that of the neuronal isoform from rat cerebellum, where sequence analysis showed a significant homology of NADPH cytochrome P-450 reductase to the C-terminal sequence of this NOS isoform (Bredt et al., 1991). All clones isolated subsequently have demonstrated the same homology. Given that the normal function of this reductase is to supply reducing equivalents to cytochrome P-450, it has been assumed that this domain in NOS serves the same function. The findings that NOS also con- Review tains a P-450-type heme moiety (see references in Mar- letta, 1993) and that this heme functions in the chemistry of the reaction (White and Marletta, 1992; Pufahl and Mar- letta, 1993) strengthen this hypothesis. Relationship to Cytochrome P-450 Several laboratories have now shown with both inducible and constitutive isoforms that upon reduction and treat- ment with carbon monoxide, NOS shows a h,, of - 450 nm. This relatively rare spectral characteristic is relegated to the large family of cytochrome P-450 enzymes. Most of the P-450s fall into a well-described supergene family whose function involves oxidative metabolism (hydroxyla- tion) of endogenous and xenobiotic compounds (Nelson et al., 1993). In addition, with one notable exception, the P-450s require a flavoprotein reductase and sometimes an iron-sulfur protein to transfer electrons into the heme prosthetic group that is responsible for the oxidative catal- ysis. The exception to this rule is a fatty acid monooxygen- ase, P-450~~3, isolated from Bacillus megaterium, where the flavoprotein reductase and the heme are contained within a single polypeptide (Narhi and Fulco, 1966). This P-450 has been called a self-sufficient P-450 since the need for a separate reductase is no longer necessary. NOS would appear to be the first self-sufficient mamma- lian P-450. However, comparison of NOS to the P45Os does not reveal any significant homology. This remains true even when the analyses are focused on the N-termi- nal domain where the heme site is located. The signature P-450 spectrum is derived from the ligation of a cysteine thiolate to the iron of the heme, and a lo-residue con- sensus sequence that includes this cysteine residue (FXXGXXXCXG) has been found in most of the P-450s characterized. As several groups have noted (McMillan et al., 1992; Renaud et al., 1993; White and Marletta, 1993), this lo-residue sequence is not present in NOS. However, several cysteines are conserved among the cloned NOS isoforms; alignment of some of those conserved cysteines with the P-450 consensus sequence is shown in Table 1. P-450,,, is a camphor hydroxylase isolated from Pseu- domonas putida, and P-450eM3 is the self-sufficient P-450 fatty acid hydroxylase mentioned above. The conserved cysteine in NOS that provides the best alignment to this lo-residue P-450 sequence corresponds to neuronal NOS (nNOS) residues 409-417 and murine macrophage iNOS residues 166-196. In both cases, with a gap at position 7, the alignment of the amino acid side chains of known