Isodenaturation of avian and mammalian DNA of Mr > 2.5 x 108 in 85% (vol/vol) formamide led to the observation by electron microscopy of A+T-rich zones ("AT-rich linkers") of 300-3000 base pairs which aresdistributed over the entire genome in a characteristic pattern. Linkers of mean length 800 base pairs are found either isolated or in clusters of about four to six linkers of more heterogeneous size separated on average by 2500 base pairs (in duck DNA). In between clusters, single linkers segment the DNA at distances of 10 to more than 100 kilobase pairs, with a majority in the range of (3,4) chromosomes of higher organisms.The significance of the large excess of transcribed and untranscribed genomic DNA, compared to the complexity of the actual translated mRNA, is enigmatic, although the final confirmation of the existence of precursors to mRNA allows at least a tentative interpretation of the existence of transcribed nongenic DNA within the eukaryotic transcriptons (see discussion and references in ref. 5). Light satellite DNA (6-10) is an example of nontranscribed (9), probably structural, DNA. Further A+T-rich segments disseminated in eukaryotic DNA (11) might play an important role in DNA structure and function (see model for precursor mRNA in ref. 12) beyond their enrichment at the centromere (10) of metaphase chromosomes. A+T-rich DNA sequences have been found in bacterial promoter/operator regions (13, 14) and in gene-proximal eukaryotic spacers (15). Such sequences are possible targets of A+T-specific drugs such as distamycin or bromodeoxyuridine, which interfere, respectively, with transcription (16) and specific phases of differentiation (17).We thus set out to study the distribution and characteristics of A+T-rich regions in segments of eukaryotic DNA longer than the putative transcriptional units defined by the size of the transcription products (for discussion, see ref. 5).METHODS DNA Preparation. Erythroblasts were selected from the blood of anemic Pekin ducks and were washed in 10 mM Tris-HCl, pH 6.9/10 mM KCV3 mM MgCl2/1 mM MnCl2/ 0.25 M sucrose. The cells were lysed with 0.4% Cemusol NP-6 in washing buffer (pH 7.4) in a precision-bore, hydraulically driven Dounce homogenizer. Nuclei were washed with a double detergent solution (0.5% Tween/0.15% deoxycholate) for 10 sec, diluted with 10 vol of washing buffer, resuspended in 10 mM Tris-HCl, pH 7.4/150 mM NaCl, lysed by 0.5% NaDodSO4, gently transferred to a dialysis bag, and incubated overnight at 18°C with proteinase K (50 pLg/ml) in 5mM EDTA. Protein was removed by extraction in slow-motion roller bottles (18) first with saturated phenol, then with phenoVchloroform/ isoamyl alcohol, 1:1:0.02 (vol/vol), and finally with chloroform. DNA was then dialyzed against 10 mM NaCl/50 mM Tris HCV 1 mM WEDTA, treated with pancreatic RNase (1 mg/i50 ml for 3 hr at 20°C), extracted with satured phenol, and dialyzed exhaustively against Tris-HCl, pH 7.8/100 mM NaCVl1 mM EDTA. Rat and mouse liver cell suspensions were treated similarly.Electron M...