The role of presenilins in gamma-secretase activity: catalyst or cofactor?

Abstract

Over the last decade, there has been considerable progress in our understanding of the biochemical pathogenesis of Alzheimer's disease. The amyloid b-protein (Ab) was ®rst reported by Glenner and Wong in 1984 and subsequently extracted from amyloid plaques by Masters and coworkers (Glenner and Wong 1984; Masters et al. 1985). Today, the accumulation of Ab in the brain is considered to be a major factor contributing to a cascade of events causing neuronal dysfunction and Alzheimer's disease. Ab is derived from the amyloid b-protein precursor (APP) (Kang et al. 1987) through the action of proteases termed band g-secretases, which cleave on the Nand C-terminal ends of the Ab sequence in APP (see Nunan and Small 2000 for a recent review). Aggregation of Ab to form neurotoxic ®brillar or proto®brillar oligomers is likely to be necessary for the neurotoxic effects of Ab (Small and McLean 1999). It is understandable that the mechanisms of Ab production, aggregation and clearance from the brain have been the subject of considerable interest. The band g-secretases are major targets for drug development and, already, a number of inhibitors of these enzymes have been developed. The identi®cation, puri®cation and sequencing of these enzymes have been a top priority, as an accurate three-dimensional structure based on X-ray crystal data, NMR or homology modeling would provide a major step towards rational drug design (De Strooper and KoEnig 1999). Pharmacological and biochemical studies have shown that both band g-secretases are likely to belong to the family of aspartyl (acid) proteases, which utilize two aspartate residues at the active site (Haass et al. 1993; Knops et al. 1995; Vassar et al. 1999; Wolfe et al. 1999a). This conclusion is consistent with the view that both band g-secretase cleavage of APP occurs in secretory compartments that have an acidic pH, as many aspartyl proteases have optimum activity at acidic pH. The aspartyl proteases are only one of a group of six wellestablished protease families that have been identi®ed (Table 1). Members of each of these families share similarities in amino acid sequence and exhibit common secondary and tertiary folding, particularly in the region of the active site (Neurath 1984). Recently, an aspartyl protease, termed the b-site APP cleaving enzyme (BACE), has been identi®ed that has many of the requirements of a b-secretase (Hussain et al. 1999; Sinha et al. 1999; Vassar et al. 1999; Yan et al. 1999). This protease possesses approximately 30% homology with other members of the aspartyl protease family and its crystal structure shows very similar folding to that seen in other aspartyl proteases (Hong et al. 2000). Studies on the localization and speci®city of BACE support the view that it is a b-secretase. Unlike the b-secretase, the g-secretase has not been unequivocally identi®ed. There have been several candidates proposed (Evin et al. 1995; Klafki et al. 1995; Figueiredo-Pereira et al. 1999) and some studies have suggested that there may be several g-secretases (Murphy et al. 2000). Previously it had been demonstrated that the knock-out of the genes encoding proteins called presenilins also knocks-out g-secretase activity (Herreman et al. 2000). This ®nding clearly established the intimate association between g-secretase and the presenilins. In 1999, Wolfe and coworkers proposed that the presenilins were the g-secretase and that two transmembrane aspartate residues in the presenilin sequence were analogous to the catalytic aspartates of aspartyl proteases (Wolfe et al. 1999b). This hypothesis has been viewed as controversial (according to a recent poll conducted on 206 researchers for the Alzheimer Research Forum, http://www.alzforum.org). First, the presenilins did not seem to ®t into any known family of proteases. Did they emerge via a mechanism of convergent evolution and if so, how? Are they the ®rst