Mechanisms of Embryonic Induction in Vertebrates 
meobox sequences that are clearly novel are to 
obtain cDNA clones, to map them in the mouse 
genome, and to investigate their expression pat- 
terns by in situ hybridization in the embryonic 
kidney. A longer-term goal is to identify the target 
genes that these homeoboxes interact with in 
mouse kidney development. 
Approaches to Identifying Target Genes of 
Homeobox Proteins 
Since vertebrate development appears to pro- 
ceed by a sequential hierarchy of transcriptional 
activation (and repression) , and since homeobox 
genes are apparently important in mediating such 
events, we have sought to develop a technique 
for identifying genes that homeobox genes di- 
rectly regulate. One technique being explored 
makes use of a yeast expression system in which 
an individual yeast cell contains an expression 
plasmid for the homeoprotein whose binding site 
is being sought, fused to Gal4, a strong transcrip- 
tional activator. This homeobox-Gal4 fusion pro- 
tein is placed under the control of an inducible 
promoter, so that the synthesis of the protein can 
be turned on or off. 
In yeast, such a system has been shown capa- 
ble of activating a target plasmid containing a 
sequence-specific binding site upstream of a pro- 
moter controlling the expression of the bacterial 
gene lacZ, which thus serves as a reporter gene. 
Conveniently, /«cZ encodes an enzyme, jS-galac- 
tosidase, that will cleave a fluorescent substrate, 
resulting in a fluorescent yeast cell. Such cells 
can be separated with the fluorescence-activated 
cell sorter (FAGS). We hope to use this system to 
remove background from subgenomic libraries 
and then, after induction of homeobox expres- 
sion, to sort cells containing specific homeobox 
binding sites. 
A second general approach under consider- 
ation is the addition of labeled or retrievable ho- 
meodomains to chromatin, isolated from the de- 
velopmental stage and tissue of interest. This is 
patterned after an analogous approach used in 
Drosophila. However, a current limitation is that 
high-titer monoclonal antibodies are not avail- 
able for many mouse homeobox genes. 
296 
