We designed a screen of random GFP::cDNA fusion proteins for several reasons:

• Traditional analysis of cellular structure in fixed and stained thin sections is subject to several limitations: it is time consuming, it is subject to fixation artifacts, it is often limited by the availability of specific markers such as antibodies, and this method is severely limited in its ability to yield information about dynamic processes. Expression of GFP fusion proteins offers an opportunity to introduce specific markers for subcellular structures that can be directly visualized in live cells. These markers are introduced genetically, obviating the need for microinjection, they are renewable by continued gene expression, and their specificity can be exacting and biologically informative, as it is conferred by peptide structure.

• A large scale screen for fusion proteins that have specific localization patterns may prove an efficient means of creating a collection of useful and varied subcellular markers.

• A generalized and random approach to generating cellular markers offered the possibility of revealing novel cellular structures or domains.

• This method can be used as a means of gene discovery based on localization pattern. Gene discovery may also be possible based on dynamic patterns of localization. For example, proteins whose localization pattern changes in response to an applied signal such as a hormone, wounding, or pathogen interaction.

• A sufficiently large collection of localization tags can be used to gain insight into protein targeting motifs.

• Live cell structural markers offer the possibililty of designing microscopic genetic screens for changes in subcellular structure. These screens would be much more efficient and rapid than those based on antibody detection of structures in which the screened material is fixed and processed for immunostaining.