protein is the Drosophila homologue of the mam- 
mahan SHPTP2 protein, with which it shares 62% 
overall identity. 
To elucidate the function of the D-ra/and csw 
proteins, Dr. Perrimon's laboratory has undertaken a 
structure-function analysis of these proteins. The 
D-raf protein contains three highly conserved do- 
mains: a cysteine-rich region, speculated to play a 
role in protein-protein interactions; a serine/ 
threonine-rich region of unknown function; and the 
catalytic kinase domain. To gain insights into the 
functions of each of these domains, the laboratory 
will introduce a number of modifications into the 
£>-m/protein. These engineered mutants will be as- 
sayed in vivo for their ability to rescue all or parts of 
the D-ra/mutant phenotypes. Similarly, a number of 
csw derivatives have been constructed and are 
currently being assayed for their ability to rescue the 
csw mutant phenotypes. These include modifica- 
tions in the two SH2 domains and the tyrosine phos- 
phatase catalytic region. 
A number of approaches are being used to identify 
and characterize additional components of this sig- 
nal transduction pathway. First, using mutations 
with residual activities in D-ra/and csw, Dr. Perri- 
mon and his colleagues have identified genetically a 
number of second-site interacting loci. Future work 
on these genes should identify either novel proteins 
or previously known proteins that function in torso 
signaling. Second, using the "two-hybrid" tech- 
nique developed in the laboratories of Drs. Stanley 
Fields and Roger Brent, which allows the cloning of 
unknown proteins that associate with a known pro- 
tein, Dr. Perrimon and his colleagues will attempt to 
isolate proteins that physically interact with D-raf. 
The wingless Segment Polarity Pathway 
Intrasegmental patterning in the Drosophila em- 
bryo is regulated by cell-cell communication. One 
signaling pathway that specifies positional informa- 
tion throughout the segment is mediated by the se- 
creted wingless {wg) protein, which shows signifi- 
cant homology to the mammalian Wnt-1 protein. 
The early role of wg is to stabilize the expression of 
the engrailed (en) homeodomain protein by initiat- 
ing a phase of en autoregulation in the adjacent 
cells. Dr. Perrimon and his colleagues are studying 
the role of three genes: dishevelled (dsh), porcu- 
pine (pore), and zeste-white-3 (zw3) in li^g signal- 
ing. Their working hypothesis is that pore is re- 
quired for the secretion of the wg protein and dsh 
and zw3 are involved in transducing the wg signal. 
Dr. Perrimon and his colleagues have shown that 
zw3 acts as a repressor of en autoregulation, and 
genetic epistasis experiments indicate that wg sig- 
naling operates by inactivating the zw3 repression 
of en autoactivation. At least three different serine/ 
threonine protein kinases that diverge at their amino 
termini are encoded by zw3. The differences among 
the three proteins most likely involve the regulation 
of zw3 proteins rather than their target specificities, 
since constitutive expression throughout develop- 
ment of a single zw3 protein is able to rescue all 
aspects of the zw3 mutant phenotype. zw3 is 85% 
identical to the mammalian protein glycogen syn- 
thase kinase- 3 (GSK3), a known repressor of tran- 
scription factors. Dr. Perrimon and his colleagues 
have demonstrated that zw3 encodes the Drosoph- 
ila homologue of mammalian serine/threonine ki- 
nase GSK3- (A grant from the March of Dimes pro- 
vided support for the project described above.) 
Genetic interactions between the segment polar- 
ity genes indicate that the effect of wg on zw3 is 
mediated by dsh, which encodes a novel protein 
whose structure and sequence have failed to reveal 
any clues about its function. Using the Drosophila 
dsh gene as a probe, the laboratory, in collaboration 
with Dr. Daniel Sussman (Lake Placid, New York), 
has isolated vertebrate homologues, suggesting that 
dsh function has been conserved during evolution. 
(A grant from the National Institutes of Health pro- 
vided support for the project described above.) 
The dsh, pore, and zw3 segment polarity loci 
were identified as the outcome of screens designed 
to analyze the maternal-effect phenotypes of zygotic 
lethal mutations. These screens were made possible 
by the use of the dominant female sterile (DFS) tech- 
nique that so far has remained limited to the X chro- 
mosome (one-fifth of the Drosophila genome) . To 
isolate novel autosomal segment polarity loci. Dr. 
Perrimon and his colleagues have recently extended 
the original DFS technique to the autosomes. In ad- 
dition, the efficiency of the original technique has 
been improved using the site-specific recombina- 
tion system developed by Dr. Kent Golic and Dr. 
Susan Lindquist (HHMI, University of Chicago). 
Many genes have been identified that when mu- 
tated generate a segment polarity phenotype. Their 
gene products range from transcription factors to 
structural proteins, kinases, transmembrane pro- 
teins, and secreted factors. In most cases the rela- 
tionships between the segment polarity genes have 
been difficult to analyze, because more than one 
signaling pathway likely operates during segmenta- 
tion. In addition, analyses of the epistatic relation- 
ships between the segment polarity genes have been 
difficult because of the lack of dominant mutations. 
To generate embryonic dominant phenotypes asso- 
ciated with the segment polarity genes. Dr. Perri- 
mon and his colleagues will use the Gal4 system of 
GENETICS 251 
