An expandable gene that encodes a Drosophila glue protein is not expressed in variants lacking remote upstream sequences.

The Drosophila melanogaster gene Sgs4 encodes one of the glue polypeptides, sgs-4, synthesized in the larval salivary gland. We have examined the structure and expression of Sgs4 in five strains that produce abundant amounts of sgs-4 and its mRNA and in four that do not. The nonproducers include three Japanese strains that accumulate trace amounts of mRNA and one strain, BER-1, that contains no detectable Sgs4 RNA. Sgs4 carries a tandem array of repeated 21 bp elements within its coding sequence. The number of elements per array varies, causing considerable differences in the lengths of Sgs4 and its mRNA among the strains. These differences in length are not correlated with differences in mRNA abundance; rather, the low or zero abundance in nonproducers correlates with the loss of DNA upstream from the gene. The Japanese nonproducers carry a 52 bp deletion 305 bp upstream from the 5' end of Sgs4, and BER-1 carries a 95 bp deletion 392 bp upstream. Curiously, each deletion encompasses one or more of the salivary-gland-specific DNAase I-hypersensitive sites which are known to flank the Sgs4 gene.