Apolipoprotein E (apoE) is believed to be closely involved in the pathogenesis of Alzheimer's disease (AD) because of its ability to bind to beta-amyloid (Abeta), the primary component of senile plaques. The presence of cystein residues in apoE2 and apoE3 allows these isoforms to form disulfide-linked complexes, such as apo(E-AII) complex and apo(AII-E-AII) complex. A 50-kDa complex [which corresponded to apo(E-AII)-Abeta, because it reacted with any of the three antibodies, anti-apoE, anti-apoAII, or anti-Abeta] was detected by immunoblot analysis in native cerebrospinal fluid (CSF) obtained from nondementia patients with the apoE phenotype E3/E3. However, a band considered to represent apoE-Abeta was not observed. The dissociation constant (Kd) values obtained for the specific binding of recombinant apoE2, apoE3, and apoE4 to Abeta(1-42) were 48.1 +/- 2.2 nM, 63.7 +/- 2.1 nM, and 75.9 +/- 1.8 nM, respectively. In contrast, the binding affinity of the partially purified apo(E3-AII) complex to Abeta(1-42) was very high, the Kd being 5.5 +/- 0.5 nM. No basic difference was observed between lipidated and nonlipidated apoE in terms of the characteristics of the binding of apoE isoforms to Abeta(1-42); however, lipidation reduced the binding capacity of each isoform in a dose-dependent manner. These findings seem consistent with the generally accepted idea that apoE4 is a risk factor for AD, insofar as only apoE4 is unable to form a complex with apoAII owing to its lack of a cystein residue. In addition, it is possible that apoE3 monomer (and possibly apoE2 monomer), like apoE4 but unlike apo(E-AII) complex, can act as a risk factor in the pathogenesis of AD. Copyright 1999 Wiley-Liss, Inc.