Locations and contexts of sequences that hybridize to poly d(G-T) d(C-A) in mammalian ribosomal DNAs and two X-linked genes

Sequences located several kilobases both 5′ and 3′ of the stably transcribed portion of several genes hybridize to radio-labeled pure fragments of the alternating sequence poly (dG-dT)(dC-dA) [poly(GT)]. The genes include the ribosornal DNA of mouse, rat, and human, and also human glucose-6-phosphate dehydrogenase (G6PD) and mouse hypoxanthine-guanine phosphoribosyl transferase (HPRT). HPRT has additional hybridizing sequences in introns. Fragments that include the hybridizing sequences and up to 300 bp of adjoining DNA show perfect runs of poly(GT) (>30bp) in all but the human 5′ region of rDNA, which shows a somewhat different alternating purine:pyrimidine sequence, poly(GTAT) (36bp). Within 150 bp of these sequences in various instances are found a number of other sequences reported to affect DNA conformation in model systems. Most marked is an enhancement of sequences matching at least 67% to the consensus binding sequence for topoisomerase II. Two to ten-fold less of such sequences were found in other sequenced portions of the nontranscribed spacer or in the transcribed portion of rDNA. The conservation of the locations of tracts of alternating purlne:pyrimidine between evolutionarily diverse species is consistent with a possible functional role for these sequences.