Dr. Desplan and his colleagues have also changed 
the specificity of Paired to that of Bicoid, by replac- 
ing the same serine with a lysine. Other site-di- 
rected mutations in the helix-turn-helix motif do 
not further alter the specificity. A second binding 
activity in Paired has also been defined. This activity 
is independent of the recognition helix, since major 
disruption of this motif leaves the binding unaf- 
fected. To address the specific function of each 
homeodomain, Treisman is now testing the in vivo 
activity of chimeric genes carrying a modified Paired 
homeodomain embedded in a ftz gene. 
III. Transcriptional Functions of the Homeodomain 
Proteins Fushi tarazu and Engrailed. 
The question of the transcriptional role of the 
homeodomain-containing proteins is addressed by 
using knowledge of the DNA binding of HD pro- 
teins and testing their function in an in vitro tran- 
scription system. The goal is not to dissect any spe- 
cific promoter but to analyze the mechanisms by 
which combinations of HD and zinc finger proteins 
bound to their respective sites act on the transcrip- 
tional machinery. Yoshiaki Okhuma (then an HHMI 
postdoctoral fellow in the laboratory) has purified 
to homogeneity the Engrailed and Ftz proteins 
from E. coli overexpressing cells. In collaboration 
with the laboratory of Dr. Robert Roeder (The 
Rockefeller University), he has analyzed the tran- 
scriptional role of these proteins in a fractionated 
in vitro transcription system from mammalian cells. 
The use of a heterologous transcription system is 
justified by the recent discovery that well-known 
transcription factors working in such systems do 
contain a homeodomain and also by the observa- 
tion of cross-species activities of the yeast and mam- 
PUBLICATIONS 
malian transcription machineries. Okhuma's first 
observation was that En could repress transcription 
of basal promoters, even in the absence of strong 
En-binding sites. He has shown that in this case. 
En was acting by binding to the TATA box and com- 
peting with the TATA-box binding factor TFIID. For- 
mation of a committed complex by preincubation 
of the promoter with TFIID prevented the repres- 
sion by En as well as its binding to the TATA box. 
The in vivo relevance of this observation, both in 
cell culture and in vivo, is being investigated by 
adding a TATA box to the engrailed promoter that 
is positively regulated by En. The TATA box contain- 
ing promoter of a construction negatively regulated 
by En in vitro and in cell culture is also being 
replaced by a different TATA box promoter and by 
the engrailed promoter that does not contain a 
TATA box. 
Recently Okhuma has shown that the Ftz protein 
could activate in vitro transcription in a binding 
site-dependent manner. This activation can be pre- 
vented by En. In this case. En seems to act by com- 
peting with Ftz for binding to the HD-binding sites. 
When more En protein is added, a further repres- 
sion appears to be mediated by the binding to the 
TATA box. 
This analysis leads to an investigation of both the 
general transcription functions carried out by the 
HD and zinc finger proteins and the particular 
properties of each gene product. Attention is now 
focused on the role of combinations of such factors 
in vitro and in vivo, and insight into the control 
mechanisms of the coordinate expression of the de- 
velopmental genes is expected. 
Dr. Desplan is also Assistant Professor and A. 
Meyer Fellow at The Rockefeller University 
Article 
Desplan, C, Theis, J., and O'Farrell, PH. 1988. The sequence specificity of homeodomain-DNA interaction. 
Ce// 54:1081-1090. 
190 
