One of the defining features of plants is a body plan based on the physical properties of cell walls. Structural analyses of the polysaccharide components, combined with high-resolution imaging, have provided the basis for much of the current understanding of cell walls. The application of genetic methods has begun to provide new insights into how walls are made, how they are controlled, and how they function. However, progress in integrating biophysical, developmental, and genetic information into a useful model will require a system-based approach.

We describe a general approach for identifying components of subcellular structures in a multicellular organism by exploiting the ability to generate thousands of independent transformants in Arabidopsis thaliana. A library of Arabidopsis cDNAs was constructed so that the cDNAs were inserted at the 3 end of the green fluorescent protein (GFP) coding sequence. The library was introduced en masse into Arabidopsis by Agrobacterium-mediated transformation. Fluorescence imaging of 5,700 transgenic plants indicated that Ϸ2% of lines expressed a fusion protein with a different subcellular distribution than that of soluble GFP. About half of the markers identified were targeted to peroxisomes or other subcellular destinations by non-native coding sequence (i.e., out-of-frame cDNAs). This observation suggests that some targeting signals are of sufficiently low information content that they can be generated frequently by chance. The potential of the approach for identifying markers with unique dynamic processes is demonstrated by the identification of a GFP fusion protein that displays a cell-cycle regulated change in subcellular distribution. Our results indicate that screening GFP-fusion protein libraries is a useful approach for identifying and visualizing components of subcellular structures and their associated dynamics in higher plant cells.