INSECT PATTERN FORMATION 
NORBEKT Perrimon, Fh.D., Assistant Investigator 
Dr. Perrimon's laboratory is studying the molecu- 
lar mechanisms of cellular differentiation during 
pattern formation. The goal is to use genetic and 
molecular approaches to define networks of inter- 
acting genes that are involved in the elaboration of 
specific pattern elements. Two patterning systems, 
embryonic segmentation and neurogenesis, are 
being investigated, with the belief that similar gene 
networks may be operating in both systems. 
I. Role of l(l)pole hole in Signal Transduction. 
The function of a number of maternally encoded 
gene products is to direct the expression of zygotic 
genes. The molecular mechanisms underlying early 
determinative processes rely on the interactions be- 
tween maternally stored and zygotically activated 
gene products. One goal has been to study the role 
of the l(l)pole hole [l(l)ph] gene in establishing 
the spatial coordinates of the embryo. This labora- 
tory has shown that l(l)ph encodes a single 3.2 kb 
RNA that has homology to the mammalian raf 
proto-oncogene, which is a serine-threonine ki- 
nase. Embryos derived from females that lack the 
l(l)ph wild-type gene product are missing their 
most anterior and posterior structures. A similar 
embryonic phenotype is observed in embryos de- 
rived from females homozygous for the torso muta- 
tion. The torso gene has recently been cloned by 
Dr. Nusslein-Volhard's group and was found to 
share extensive amino acid homology with a mem- 
brane-bound tyrosine kinase. Genetic epistasis ex- 
periments suggest that l(l)ph acts downstream of 
torso. One current model suggests that the mam- 
malian c-raf acts as a "molecular bottleneck," serv- 
ing as a signal transducer for a variety of growth fac- 
tor receptors with tyrosine kinase activity. In 
response to extracellular signals, c-raf, which is lo- 
cated at the membrane, is phosphorylated on a 
subset of its tyrosine residues and is thereby acti- 
vated. After activation, c-raf moves to the nucleus, 
where it participates in the control of gene expres- 
sion. It is therefore postulated that the l(l)ph gene 
product is phosphorylated by the torso gene, there- 
by mediating the torso information. This hypothesis 
is being tested using polyclonal antibodies against 
the l(l)ph gene product and phosphorylation as- 
says. The combination of biochemical and genetic 
approaches should allow the understanding of the 
role of c-ra/[i.e., l(l)ph] in signal transduction. 
II. Mechanisms of Intrasegmental Patterning. 
Once a group of cells is assigned to a specific 
germ layer, both cell lineage and cell-cell interac- 
tions define the particular identities of single cells. 
The establishment of intrasegmental patterning 
provides an attractive system to study these mecha- 
nisms. This laboratory has identified three new seg- 
ment polarity loci, dishevelled (dsh), porcupine 
(pore), and zeste-white-3 , that are associated with 
maternal-effect lethal phenotypes that affect the 
fates of rows of cells within every segment. Another 
segment polarity gene, wingless (wg), has been 
studied and found by others to be the Drosophila 
homologue of the murine oncogene int-1, whose 
misexpression can result in mammary tumors. The 
wg gene has been shown to act nonautonomously 
and is believed to encode a short-range morpho- 
genetic signal. The receptor(s) for such a signal is 
likely to be required in most cells, and when the 
gene encoding the receptor is mutated the embry- 
onic phenotype should be identical to that of wg. 
Identified by genetic means, pore and dsh are two 
likely candidates for "receptor" gene. A battery of 
clonal analysis experiments has allowed the dem- 
onstration of the cellular autonomy of dsh. An in- 
vestigation has revealed that pore, like wg and un- 
like dsh, is not cell autonomous. The present 
model is that pore is required for production of the 
wg signal and that dsh is required to receive or 
transduce it. Immunological approaches are being 
used to examine (in collaboration with Dr. R. 
Nusse) the pattern of wg protein expression in 
both pore and dsh mutants. Molecular characteriza- 
tion of both genes is in progress and will likely cul- 
minate with the characterization of gene products 
involved in the D-intl transducing machinery. 
III. Ventral Midline as a Developmental System. 
Dr. Perrimon's laboratory is studying the determi- 
nation of a set of cells that are located along the 
ventral midline of the embryo. Affecting these mid- 
line cells are seven loci that appear to play impor- 
tant roles. Genetic and molecular analyses of three 
of these loci, orthodentiele {otd), midline, and 
spitz, are in progress. Within the central nervous 
system (CNS) these genes affect the development of 
both neural (midline neurons) and nonneural 
(midline glia) cells that play key roles in the forma- 
Continued 
283 
