Cleavage of amyloid precursor protein (APP) by the beta- and gamma-secretases generates the amino and carboxy termini, respectively, of the A beta amyloidogenic peptides A beta40 and A beta42--the major constituents of the amyloid plaques in the brain parenchyma of Alzheimer's disease patients. There is evidence that the polytopic membrane-spanning proteins, presenilin 1 and 2 (PS1 and PS2), are important determinants of gamma-secretase activity: mutations in PS1 and PS2 that are associated with early-onset familial Alzheimer's disease increase the production of A beta42 (refs 4-6), the more amyloidogenic peptide; gamma-secretase activity is reduced in neuronal cultures derived from PS1-deficient mouse embryos; and directed mutagenesis of two conserved aspartates in transmembrane segments of PS1 inactivates the ability of gamma-secretase to catalyse processing of APP within its transmembrane domain. It is unknown, however, whether PS1 (which has little or no homology to any known aspartyl protease) is itself a transmembrane aspartyl protease or a gamma-secretase cofactor, or helps to colocalize gamma-secretase and APP. Here we report photoaffinity labelling of PS1 (and PS2) by potent gamma-secretase inhibitors that were designed to function as transition state analogue inhibitors directed to the active site of an aspartyl protease. This observation indicates that PS1 (and PS2) may contain the active site of gamma-secretase. Interestingly, the intact, single-chain form of wild-type PS1 is not labelled by an active-site-directed photoaffinity probe, suggesting that intact wild-type PS1 may be an aspartyl protease zymogen.

The functional significance of conserved polar amino acids within the putative transmembrane region of the fi-adrenergic receptor (PiAR) was examined by oligonucleotide-directed mutagenesis of the hamster gene encoding PAR and expression of the mutant genes in COS-7 cells. Although a substitution of aspartate at position 113 with an asparagine residue did not affect expression or processing of the protein, the resulting mutant BAR did not show detectable binding toward the antagonist iodocyanopindolol. Replacement of the aspartate and asparagine residues at positions 79 and 318, respectively, had no effect on the affinity of the receptor toward antagonists but reduced the affinity of the receptor toward agonists by 1 order of magnitude. Furthermore, we observed that substitution of the proline at position 323 with a serine residue resulted in improper or incomplete processing of the .BAR, presumably reflecting a role for this residue in the folding of the receptor. Together with our previous results from deletion mutagenesis studies, these observations indicate that the ligand binding site involves the transmembrane region of the BAR. The recent cloning of the genes encoding the hamster (2) and human 32AR (3), the avian PAR (4), and the porcine Ml muscarinic cholinergic receptor (MAR) (5) and the deduction of the primary sequences of these proteins suggest a structural basis for the mechanistic similarities among the Gprotein-linked receptors. These hormone receptors share sequence homology with each other and with the visual opsins, which transduce their signals through the activation of the G-protein transducin (6). Most of the conservation in sequence among these proteins is found within seven hydrophobic domains of approximately 20-25 residues in length. Based on the model proposed for the opsins (7), these hydrophobic regions of the receptors, which are linked by more divergent hydrophilic regions of various lengths, would alternately traverse the membrane with the N terminus of the PAR exposed externally and the C terminus exposed intracellularly (2). 4384The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Presenilin-1 (PS1) and presenilin 2 (PS2) are proposed to be transmembrane aspartyl proteases that cleave amyloid precursor protein and Notch. PS1-and PS2-mediated activities were individually characterized using blastocyst-derived (BD) cells and membranes from PS1 ؉/؊ -PS2 ؊/؊ and PS1 PS2؉/؉ mice, respectively. The relative amounts of PS1 and PS2 in the various BD cells were determined from the intensities of the anti-PS1 and anti-PS2 immunoblot signals by comparison with standard curves using radiolabeled PS1 and PS2 standards produced by in vitro transcription and translation. Cellular membranes from wild type, PS1, and PS1 PS2؊/؊ BD cells generated the A␤40 and A␤42 products from the C100FLAG substrate. PS1-associated ␥-secretase displays considerably higher specific activity than PS2-associated ␥-secretase. Moreover, the PS1 PS2؊/؊ BD cells and corresponding membranes exhibited much higher ␥-secretase activity as compared with other BD cells and membranes. The PS1-mediated ␥-secretase activity correlated better with the amount of PS1 that is modifiable by a photoactivated active sitedirected ␥-secretase inhibitor rather than total PS1; hence, only a small portion (<14%) of the PS1 in wild-type membranes appears to be engaged in an active ␥-secretase complex. This finding suggests that PS1 may serve other biological functions in addition to that associated with its ␥-secretase activity. Furthermore, the PS1 ␥-secretase complex and the PS2 ␥-secretase complex activities can be discriminated on the basis of their susceptibility to inhibition by a potent ␥-secretase inhibitor. The distinct yet overlapping enzymatic properties of the PS1 ␥-secretase complex and the PS2 ␥-secretase complex imply that these two putative aspartyl class proteases may contribute to different biological processes.

On the basis of the homology between the amino acid sequences of the beta‐adrenergic receptor (beta AR) and the opsin proteins we have proposed that the ligand binding domain lies within the seven transmembrane hydrophobic regions of the protein, which are connected by hydrophilic regions alternatively exposed extracellularly and intracellularly. We have systematically examined the importance of each of these regions by making a sequential series of deletions in the gene for the hamster beta AR which encompass most of the protein coding region. The ability of the corresponding mutant receptors to be expressed, localized to the cell membrane, and bind beta‐adrenergic ligands has been analyzed, using transient expression in COS‐7 cells. The hydrophobic regions and the hydrophilic segments immediately adjacent to the membrane cannot be removed without affecting the processing and membrane localization of the beta AR. However, most of the hydrophilic regions appear to be dispensable for ligand binding. In addition, we observed that substitution of the conserved cysteine residues at positions 106 and 184 dramatically altered the ligand binding characteristics of the beta AR, suggesting the occurrence of a disulfide bond between these two residues in the native protein. These data are discussed in terms of the tertiary structure of the beta AR.