Localization of human and mouse N-ethylmaleimide-sensitive factor (NSF) gene: a two-domain member of the AAA family that is involved in membrane fusion

The process of vesicle targeting and fusion in the secretory and endocytic pathways has been described by the SNARE hypothesis (Rothman 1994). This proposes that vesicles dock with specific target membranes by binding to membrane-specific SNAREs (soluble N-ethylmaleimide-sensitive factors attachment protein receptors); targeting specificity is also affected by the Rabs, a group of small soluble GTPases (Simons and Zerial 1993; Fischer von Mollard et al. 1994). After the vesicle has bound to the target membrane, the SNARE multimer is joined by the soluble SNAP proteins and N-ethylmaleimide-sensitive factor (NSF); this large complex is thought to allow membrane fusion, and the ATPase activity of NSF appears to be essential for this process (Rothman 1994). While there are a variety of different SNARES, Rabs, and SNAPs involved in membrane fusion, there is only one NSF, and the SNARE hypothesis describes NSF-dependent fusion. Mammalian N-ethylmaleimide-sensitive protein was first identified as the protein that restored the ability of Golgi membranes that had been inactivated with the reagent N-ethylmaleimide, to re-engage in vesicular transport (Glick and Rothman 1987). NSF was subsequently cloned from Chinese hamster cells by Wilson and colleagues (Block et al. 1988; Wilson et al. 1989). NSF is a member of the AAA (ATPases Associated with diverse cellular Activities) gene family. The genes are most related throughout an approximately 200-amino acid domain (the AAA domain) that binds ATP; however, the family is notable not only for its conservation but also for the diverse functions of its proteins in eukaryotic cells. The family can be subdivided into those with either one or two ATP-binding domains. NSF is a two-domain member of the AAA family. VCP, the valosin-containing protein that is also involved in membrane fusion, is another two-AAA domain protein (Schmidt et al. 1985; Pleasure et al. 1993), as is PAS1, a Saccharomyces cerevisiae protein required for peroxisome biogenesis (Erdmann et al. 1991). Mammalian one-domain proteins include P26S4, subunit 4 of the 26S protease (Dubiel et al. 1992; Hoyle and Fisher 1996); MSS1 ($7) subunit 7 of the 26S protease (Shibuya et al. 1992); TBP1, the HIV Tat binding protein 1 (Nelbock et al. 1990); TRIP 1, the thyroid-hormone-receptor interacting protein (Lee et al. 1995). The AAA family is highly conserved between eukaryotes, and members are known in archaebacteria and eubacteria (Hershko and Ciechanover 1992; Confalonieri et al. 1994; Tomoyasu et al. 1995). A partial human NSF sequence has been described by Hong and colleagues (unpublished, accession no. U03985) that includes a portion of the 5' and 3' untranslated regions (UTR) and the complete coding region. However, from database searches performed with the Blast program, we determined that this human