The Balbiani ring 3 gene in Chironomus tentans has a diverged repetitive structure split by many introns.

A set of approximately 15 secretory proteins is synthesized by the salivary gland cells in the midge Chironomus tentans. These proteins are secreted but do not form insoluble fibers until they are transported out of the gland lumen. A Balbiani ring (BR) gene family consisting of four genes (BR1, BR2.1, BR2.2 and BR6) have previously been shown to encode four of these proteins, sp-I a to d, with relative molecular weights of 1 x 10(6). Each BR gene contains an uninterrupted block in which about 100 repeats are tandemly arranged. The repeats are virtually identical and efficient homogenization mechanisms must operate within each block. Here we describe a new BR gene, the BR3 gene, which according to structural similarities may belong to the BR gene family, but at the same time exhibits a strikingly different structure. The gene encodes a 10.9 kb transcript that contains 38 introns and is spliced into a 5.5 kb mRNA. The mRNA is translated into a cysteine-rich 185 kDa major component of the gland secretion. The coding sequence in the gene is built from diverged repeats in which mainly the cysteine codons are preserved and the sequence is split by the introns into 17 to 678-bp long exons. The introns are located at defined positions in relation to the repeat structure. In sharp contrast to the uninterrupted array of identical repeats in the BR1-BR6 genes, the repeats in the BR3 gene are not efficiently homogenized and have diverged extensively from each other. We propose that the splitting of the repeat structure into variable sized exons prevents homogenizations dependent on unequal aligning of homologous sequences.