Position-Dependent Gene Expression in Drosophila 
Tulle I. Hazelrigg, Ph.D. — Assistant Investigator 
Dr. Hazelrigg is also Assistant Professor of Biology at the University of Utah. She received her B.A. degree 
from Oberlin College and her Ph.D. degree in genetics from Indiana University, where she worked with 
Thomas Kaufman on the Drosophila Antennapedia gene complex. She did postdoctoral work in the Carnegie 
Institution's Department of Embryology and in the Biochemistry Department at the University of Califor- 
nia, Berkeley. In both places she worked with Gerald Rubin, analyzing DNA elements that regulate the 
expression of the Drosophila white gene. 
WE are studying two problems in the genetic 
regulation of pattern formation in Dro- 
sophila. One is the question of how determinants 
come to be localized within a cell (for instance, 
the egg) where they can initiate correct develop- 
mental fates. The second is the action of DNA reg- 
ulatory sequences that respond to positional cues 
within a tissue and confine the expression of 
genes to particular cells. 
During early development of the Drosophila 
embryo, maternal-effect genes express products 
involved in establishing the basic body plan. 
Among this class of genes is exuperantia {exu), 
which is needed for the correct determination of 
anterior embryonic structures. The product of 
the exu gene is needed during the development 
of the oocyte for anterior localization of the RNA 
of another gene, bicoid (bed). This localization 
leads, during early embryogenesis, to a steep an- 
teroposterior concentration gradient of bed pro- 
tein, which acts as an embryonic anterior determi- 
nant. We are interested in how the exu gene 
functions to bring about this subcellular localiza- 
tion of the bed RNA. 
We have analyzed the effects of exu mutations 
on spermatogenesis. During a long period of 
growth and development, primary spermato- 
cytes, which have not yet undergone meiosis, 
appear normal in exu mutants; defects can be ob- 
served during meiosis and subsequent differen- 
tiation of spermatids. Fully differentiated, motile 
sperm are never produced. One exu allele is fe- 
male-specific, showing only the maternal -effect 
phenotype, and the other allele is male-specific. 
Both alleles have been informative about the 
functioning of exu in gametogenesis in the two 
sexes. 
We have cloned the exu gene and found that it 
encodes overlapping male and female tran- 
scripts, which differ in size in the two sexes. The 
two sex-specific transcripts differ also in splicing 
and polyadenylation patterns, but sequence anal- 
ysis has shown that they both encode the same 
predicted 58-kDa polypeptide. 
We have determined the sequence of the fe- 
male-specific exu mutation. It changes a single 
amino acid in the protein, possibly identifying a 
region more critical for its female than its male 
function. 
The male transcript is longer than the female in 
its 3'-untranslated tail. For part of the tail, the 
male-specific exu mutation is deleted. We are 
currently attempting to understand the function 
of the male-specific tail. We have also found that 
the tra-2 (transformer- 2) gene, which functions 
in Drosophila sex determination, is needed for 
efficient processing of the exu transcript in the 
male mode. 
Together these results suggest that sex-specific 
processing of the exu transcript in the germline is 
a biologically important event. Why is the exu 
gene needed maternally for early development of 
the embryo and also during spermatogenesis? 
Since exu mutants disturb the anterior localiza- 
tion of bed RNA in eggs, one model is that the 
gene functions in both cases to localize RNAs in 
the germ cells. Since there is no known effect of 
bed mutations on spermatogenesis, different do- 
mains in the exu protein could act to recognize 
different RNAs. 
Alternatively, the exu product could serve a 
more general role in germ cell cytoarchitecture, 
and the disruptive effect of exu mutations on bed 
RNA localization could be a pleiotropic effect of 
disturbing this structure. We do not believe, how- 
ever, that the latter scenario is correct. We have 
determined the distribution of the exu protein 
during oogenesis and early development with the 
use of antibodies raised against the protein. We 
find it to be present in the oocyte in a sharp con- 
centration gradient at times in oogenesis when 
the bed^NA is localized, with highest concentra- 
tions at the anterior ends of developing oocytes. 
But the protein appears to be degraded after it 
enters oocytes (it is produced by the nurse cells, 
which are attached to the oocyte and nourish it 
during its development) and is not present in the 
mature oocyte or early embryo. These results sug- 
gest that the exu protein's role in bed RNA local- 
ization may be transient. 
Perhaps the exu protein modifies another pro- 
tein that the bed RNA interacts with, or performs 
187 
