Although nitric oxide (NO) 1 may not modulate all cellular functions and may not be present in all mammalian cells, the sheer volume of publications on the subject might lead one to conclude that this diminutive molecule is both omnipotent and omnipresent in human biology. This series of four Perspective articles, in the current and following issues of The Journal of Clinical Investigation, will explore the limits of this assertion. 2 NO is synthesized in mammalian cells by a family of three NO synthases (NOS). It is not known whether additional mammalian NOS isoforms exist, but the failure of homologybased molecular cloning approaches to identify novel NOS cDNAs makes it unlikely that newly discovered members of the mammalian NOS gene family will bear significant structural similarity to the current trio of isoforms (1). As for any newly described gene family, an accepted nomenclature of the NOS isoforms has evolved only as novel information becomes more generally established. The initial NOS nomenclature reflected the early observations that NO synthesis was not characteristic of unactivated inflammatory cells, but could become induced upon immunoactivation, hence the term iNOS. This prototypic “inducible” iNOS was contrasted to a “cNOS” activity that was constitutively expressed in certain characteristic cell types (neuronal, endothelial). However, it is now known that the levels of gene expression of both eNOS and nNOS may also be induced under different physiological conditions (eg, hemodynamic shear stress or nerve injury), and, conversely, that iNOS may function as a “constitutive” enzyme under physiological conditions in some cells (2). Thus, the designation of a NOS isoform’s being constitutive versus inducible NOS is misleading, and should be supplanted by a nomenclature that clearly identifies the specific enzyme isoform. A widely accepted nomenclature (3), which will be used in these Perspective articles, identifies the three mammalian enzyme isoforms as nNOS, iNOS, and eNOS, reflecting the tissues of origin for the original protein and cDNA isolates. As denoted by its prefix, nNOS was originally purified and cloned from neuronal tissues. However, nNOS is now known to be much more widely distributed, with an important level of expression in skeletal muscle. iNOS, originally purified and cloned from an immunoactivated macrophage cell line, has since been identified in myriad mammalian tissues, and iNOS expression has been studied in cells as diverse as cardiac myocytes, glial cells, and vascular smooth muscle cells (to name only a few). eNOS, the last of the three mammalian NOS isoforms to be isolated, was originally purified and cloned from vascular endothelium, but has since been discovered in cardiac myocytes, blood platelets, brain (hippocampus), and elsewhere. To add to the confusion, the human genes for the NOS isoforms are officially categorized in the order of their isolation and characterization; thus, the human genes encoding nNOS, iNOS, and eNOS are termed NOS1, NOS2, and NOS3, respectively.

Clearly, the same NOS isoform may play entirely distinct biological roles when expressed in different tissues, and it must not be assumed that pathways outlined in one tissue necessarily pertain when the same isoform is expressed in a different cell. For example, differential tissue-specific splicing of nNOS mRNA generates structurally distinct protein molecules when the enzyme is expressed in neurons versus skeletal muscle (4). Another example of tissue-specific regulation is reflected in the association of eNOS with different caveolin isoforms in endothelial cells versus cardiac myocytes (5). Furthermore, the …