The engrailed (en) gene functions throughout Drosophila development and is expressed in a succession of intricate spatial patterns as development proceeds. Normal en function relies on an extremely large cis-acting regulatory region (70 kilobases). We are using evolutionary conservation to help identify en sequences important in regulating patterned expression. Sequence comparison of 2.6 kilobases upstream of the en coding region of D. melanogaster and D. virilis (estimated divergence time, 60 million years) showed that 30% of this DNA occurs in islands of near perfect sequence conservation. One of these conserved islands contains binding sites for homeodomain-containing proteins. It has been shown genetically that homeodomain-containing proteins regulate en expression. Our data suggested that this regulation may be direct. The remaining conserved islands may contain binding sites for other regulatory proteins.Many of the genes that control development in Drosophila have been identified by mutation and analyzed genetically and molecularly. Studies of the action and interaction of these genes have led to their classification into a hierarchy (for reviews, see references 26 and 49). The patterns of expression of developmentally important genes in Drosophila can be extremely complex. For example, en is first expressed in a series of stripes which transect the anteriorposterior axis of the embryo (11,19,33). Later in development, en is expressed in particular cells in the head, terminalia, central and peripheral nervous systems (39), and imaginal disks (8, 33). Molecular mapping of mutations of the en gene and other developmental loci having intricate spatial patterns of expression has shown that very large cis-acting regulatory regions are associated with these genes. For example, the en gene extends across 70 kilobases (kb) but has only a 4.5-kb transcription unit (15,34). Using in vitro DNA constructs and P-element transformation, we have begun to functionally dissect the en cis-acting regulatory regions (13; J. A. Kassis, manuscript in preparation). Because of the large size of the cis-regulatory region, we are guiding our search for important regulatory sequences by identifying sequences that are evolutionarily conserved.We isolated the en gene from a distantly related Drosophila species, D. virilis, and compared it with the D. melanogaster gene (28,29). Importantly, these two species are sufficiently separated such that only functional sequences should be conserved (estimated divergence time, 60 million years [2]). Also important is that en expression is similar in the two species (J. A. Kassis, unpublished data). Here 3.1 kb from D. virilis. We found that about 30% of this DNA is conserved between the two species. The conservation is best described as islands of near perfect conservation within largely divergent regions. We showed that some of the conserved DNA sequences contain in vitro binding sites for homeodomain (HD)-containing proteins. Mutations in genes which encode HD-containing proteins alter en e...