Actin and cell movement.

The primary locomotory organelle of non-muscle, eukaryotic cells is the lamellipodium, a thin layer of cytoplasm that exhibits active protrusive activity. Earlier studies have implicated actin polymerization in the formation of lamellipodia, whereby actin monomers insert at the front and dissociate at the rear, in a treadmilling fashion. However, other models based on gel swelling and a breakdown of actin networks at the site of protrusion of the lamellipodium have also been proposed. By employing videomicroscopy and electron microscopy of the same cells, in this case mouse macrophages, it could be shown that lamellipodial protrusion is directly linked with the growth of dense actin meshworks. The gel swelling and cortical breakdown models are not supported by this data. Using rapidly locomoting fish keratocytes, Theriot and Mitchison (Nature 352, 126-131, 1991) recently obtained results that they interpreted as supporting a form of actin filament dynamics different from treadmilling. In their new "nucleation release model" the actin filaments in the lamellipodium are predicted as being very short (less than 0.5 micron) and randomly organized. We have now investigated the ultrastructure of the keratocyte cytoskeleton. Our results show that the actin filaments in these cells are very long and organized in dense and regular, more or less orthogonal networks. A gradient of filament density across the rear part of the lamellipodium suggests that the filaments are graded in length. These data support a treadmilling type model for the eratocyte.