The Drosophila melanogaster heat shock locus at 93D encodes at least three overlapping transcripts, 10-12 kilobases (kb), 1.9 kb, and 1.2 kb. The abundance of the three transcripts is significantly increased during heat shock; however, all are also found in non-heat-shocked cells. The 1.2-kb transcript is found in the cytoplasm. Sequence analysis of a 1.1-kb cDNA clone representing sequences within the 1.2-kb transcript and comparison to genomic sequences indicate that it is spliced; 700 base pairs of sequence found in genomic DNA are removed from the middle of the transcript. Sequence analysis further suggests that this RNA does not encode a heat shock protein. The largest open reading frame beginning with a methionine codon would encode a polypeptide of 34 amino acids. We have not been able to detect a heat shock-induced polypeptide of this size. A DNA clone from the analogous heat shock puff of Drosophila hydei has been analyzed by hybridization with the small subclones used to sequence the D. melanogaster cDNA plus a genomic fragment containing the 700-base-pair intron. Results of this hybridization indicated strong homology of the intron fragment. Weaker homology was detected with the two small fragments flanking the intron. Other fragments of the D. melanogaster cDNA showed no hybridization to the cloned D. hydei puff DNA.Most of the major heat shock puffs of Drosophila melanogaster have been shown to code for the prominent heat shock proteins (hsps); however, one of the largest puffs, 93D, does not encode any ofthe known hsps. The 93D puffhas a number of additional characteristics that distinguish it from the other heat shock puffs. Early experiments analyzing heat shock RNA by hybridization to polytene chromosomes indicated that, in contrast to the other puffs, a significant portion of RNA from 93D remains in the nucleus (1). Other in situ hybridization experiments showed cross-hybridization of sequences encoding hsps in D. melanogaster with heat shock puffs of Drosophila hydei but there was no hybridization to the large D. hydei heat shock puff 2-48B. Furthermore, total heat shock RNA from D. hydei did not cross-hybridize with puffs thought to be homologous to 2-48B in more closely related species (2). These experiments suggest that, if 2-48B of D. hydei is related to 93D of D. melanogaster, the sequence ofthis puff is evolving faster than are the sequences encoding the major hsps. Several other unusual features of 93D are shared by 2-48B. Each is the only puffin the genome that is also induced by benzamide and by colchicine (3, 4). During heat shock each of these puffs contains distinctive large ribonucleoprotein particles that share antigenic determinants not seen on other heat shock puffs (5). The unusual characteristics shared by these two puffs strongly suggest that the function of the locus is conserved in spite of the divergence at the nucleotide level.In this paper we report other features that distinguish 93D from previously studied heat shock puffs. There are multiple transcripts of pa...