Most patients with Marfan's syndrome (> 95%) and 75% of patients with uncertain diagnosis can be classified into four groups (Aoyama et al, 1994, 1995) based on abnormal patterns of synthesis, intracellular transport, and/or matrix deposition of fibrillin-1 in fibroblast cultures. Herein we report a systematic study of fibrillin assembly in normal and Marfan's syndrome fibroblasts and correlations between pulse-chase, immunofluorescence, and immunoelectron microscopic data. Normal control fibroblasts were grown at confluent conditions from 2 to 10 days before passage and then maintained at hyperconfluent cell densities for an additional period of 1 to 6 days before assaying. Maximum deposition in the extracellular matrix of pulse-labeled fibrillin required at least 6 days of confluent and 4 to 5 days of hyperconfluent culture. This result is explained by immunofluorescence studies with fibrillin-1-specific antibodies, because 1 day after seeding cells at hyperconfluency, patches of regular immunostained structures were already present. Within these patches, fluorescence intensity and fibrillar material increased over 3 to 4 days, and after only 5 days, fibrillar networks extended throughout the culture. We propose that fibrillin-containing microfibrillar material is passaged together with the cells, newly synthesized fibrillin molecules are deposited onto preexisting microfibrillar assemblies, and several additional days of culture at high cell density are necessary for the cells to construct a sufficient microfibrillar network binding and detection of pulse-labeled fibrillin molecules in insoluble form during a 20-hour chase period. This fraction is decreased to a varying extent in fibroblast cultures of four biosynthetically distinct groups of Marfan's syndrome patients, but only Groups II and IV clearly showed reduction in immunostainable microfibrils. In long-term cultures, immunoelectron microscopy of the extracellular matrix with fibrillin antibodies also detected differences among these groups and in comparison to normal controls with respect to the arrangement of fibrillin-containing microfibrils, thickness of microfibrillar bundles, and the presence of amorphous material. The data support the idea of different pathogenetic mechanisms for each biosynthetically defined group of Marfan's syndrome, which depends on the nature of fibrillin-1 mutations.