proach to the same question has been to study the 
molecular functions of the homeotic proteins, 
which has led to the recognition that they are tran- 
scription factors. 
Many of the regulatory genes that control divi- 
sion of the Drosophila embryo into segments and 
many of the homeotic genes that control what 
structures will develop in different regions of the 
embryo have been found to contain homeoboxes, 
183 bp DNA sequences that encode evolutionarily 
conserved 6l-amino acid protein domains. Similar 
homeodomain proteins are found in Drosophila 
and in most other animal species that have been 
examined for the sequences, including humans. In 
all cases the homeodomains are similar in structure 
to bacterial DNA-binding proteins, and in several 
cases they have been shown to bind to DNA in a se- 
quence-specific manner. These results suggested 
that homeodomain-containing proteins might regu- 
late a largely unknown set of target genes by bind- 
ing to their DNA and affecting their transcription. 
Gary Winslow and Dr. Shigeo Hayashi have used a 
cultured cell system to test this hypothesis and 
have shown that three of the Drosophila homeo- 
domain proteins, the products of the Antp and Scr 
homeotic genes and the/fz segmentation gene, are 
capable of activating transcription of a variety of 
synthetic target genes. The target genes used were 
constructed using sequences that Dr. Hayashi had 
shown are bound in vitro by the protein encoded 
by the Antp homeotic gene. Therefore these 
homeodomain proteins are capable of controlling 
transcription, and the task now is to understand 
what determines their specificity of action. The Scr 
gene is required for the proper development of the 
posterior head and anterior thorax, whereas the 
Antp gene is required for all of the thoracic seg- 
ments to develop thoracic structures rather than 
more anterior structures. The ftz gene is required 
much earlier and is necessary to direct the division 
of the embryo into the correct number of segments. 
The homeodomains of the proteins are very similar, 
suggesting that the specificity of action may be due 
to the other parts of the proteins, which are quite 
distinct. Modified forms of the proteins are being 
constructed and expressed in cultured cells and in 
Drosophila to see the effects of the modifications 
on their functions. The exact structure of the target 
gene can also affect the activity of the homeo- 
domain proteins, and variations on target gene 
structure are being tested in cultured cells to deter- 
mine what aspects of the cis-acting sequences are 
important for function. 
III. Newly Characterized Homeotic Genes. 
Dr. John Tamkun (now at the University of Cali- 
fornia at Santa Cruz) has been collaborating with 
Dr. James Kennison (National Institutes of Health) 
in a study of new homeotic genes that were discov- 
ered because their alleles act as enhancers or sup- 
pressors of known homeotic mutations. The new 
genes that have been most studied, hrahma {brm) 
and kismet (kis), have different properties. The 
brm gene is required both during oogenesis and in 
the embryo. When brm function is drastically re- 
duced, there are severe defects in embryogenesis, 
suggesting that brm is required for the proper func- 
tion of other regulatory genes in the early embryo. 
In contrast, kis, alleles of which can cause ho- 
meotic transformations in adult Drosophila, may be 
a gene that is regulated by homeotic genes such as 
Antp or may encode a cofactor necessary for the 
function of other homeotic proteins. The brm gene 
has been cloned and sequenced; the protein se- 
quence has no relation to other reported proteins. 
The brm transcripts are expressed ubiquitously. 
IV A Transmembrane Protein Encoded by a Seg- 
mentation Gene. 
Not all of the segmentation genes encode tran- 
scription factors. Dr. Hooper has been investigating 
the gene patched (ptc), which acts in every seg- 
ment of the developing embryo to control in- 
trasegmental pattern formation. In the absence of 
ptc function the central part of each segment's pat- 
tern develops structures that are a mirror image 
of the anterior part of the segment. Cells in the 
center of the segment appear to be confused about 
where they are and which direction is anterior. The 
gene was cloned and found to encode a protein 
that is predicted to have between 7 and 12 trans- 
membrane domains. Therefore it is likely that the 
ptc protein is involved in cell-cell communication 
processes that inform cells of their positions within 
the segment primordium and of the orientation 
of the cell in relation to the embryo as a whole. 
The expression pattern of ptc is elaborate. Tran- 
scripts are first nearly everywhere, then disappear 
from the posterior of each segment primordium, 
and then disappear from the center of the anterior 
of each segment. The final pattern of 30 transverse 
stripes is one of the most complex known for a 
segmentation gene. In the nervous system, ptc is 
also required for proper patterning of cells. To- 
gether with the segmentation gene products that 
Continued 
290 
