Proteoglycans are associated with senile plaques in Alzheimer's disease and may be involved in A beta fibril formation and plaque formation. In vitro, glycosaminoglycans have been shown to inhibit the proteolysis of A beta fibrils, accelerate formation and maintain their stability. To model their interaction, we investigated the binding of a sulfated proteoglycan derived from a natural source; marine sponge Microciona prolifera aggregation factor (MAF). This species-specific re-aggregation of sponge cells has two functional properties, a Ca2+ independent cell binding activity and a Ca2+ dependent self-aggregation. It has been shown that a novel sulfated disaccharide and a pyruvylated trisaccharide are important in the Ca(2+)-dependent MAF aggregation. Aggregation demonstrated by homophilic binding of MAF subunits may be chemically distinct from other heterotypic binding effects. We investigated A beta-MAF interactions and show that MAF induces a structural transition in A beta 40 and A beta 42 from random to beta-structure as detected by circular dichroism spectroscopy. Electron microscopy revealed that the structural transition correlated with an increase in the number of A beta 40 and A beta 42 aggregated that have a truncated fibrillar morphology. Finally, MAF increased A beta-induced toxicity of nerve growth factor (NGF)-differentiated PC-12 cells in the absence of Ca2+. The addition of Ca2+ to MAF-A beta incubations resulted in a moderate attenuation of toxicity possibly due to a reduction in A beta-cell interactions caused by extensive lateral aggregation of the MAF-A beta complexes. Our results indicate that A beta is generally susceptible to proteoglycan-mediated aggregation and fibril formation. We also propose that the MAF model system may be useful in delineating these interactions and represent a means to develop and examine potential inhibitors of the proteoglycan effects.