Alzheimer's disease (AD) beta-amyloid peptide (Abeta and betaA4) is derived from the amyloid precursor protein (APP) by the subsequent action of the so-far unidentified beta- and gamma-secretases. gamma-secretase, which generates the C-terminus of Abeta, cleaves within the transmembrane domain of APP, preferentially after Abeta-residue 40 (Abeta 40) but also after residue 42 (Abeta 42). This Abeta 42 represents the major subunit of the plaques in AD. Since the position of gamma-secretase cleavage is crucial for understanding the pathogenic pathway, we investigated the effect of different point mutations at Thr43 on gamma-secretase specificity in SPA4CT (SPC99)-expressing COS7 cells. These constructs only require gamma-cleavage for Abeta release. We observed that all Thr43 mutations altered the specificity of gamma-secretase. Small hydrophobic residues favored the generation of Abeta 42, leading to an increase in the 42/40 ratio of Abeta (1.6-2.8-fold). The increase was even stronger (5.6-5.8-fold) when combined with the familial mutation Val46Phe. Thus, these constructs might be highly valuable for the generation of animal models for AD. Processing of full-length APP or SPA4CT yielded the same 42/40 ratio of Abeta (4. 7%). Both constructs, bearing the familial AD mutation Val46Phe, led to a similar increase in the 42/40 ratio (3.3- versus 3.6-fold). The p3 fragment, produced by alpha- and gamma-secretase, showed 42/40 ratios similar to Abeta when derived from wild-type and mutant proteins. These results suggest that the different Abeta- and p3-species are generated by gamma-cleavage activities with a similar enzymatic mechanism.