Psoralen-crosslinking of DNA as a probe for the structure of active nucleolar chromatin

Trimethylpsoralen was used to crosslink the extrachromosomal ribosomal DNA in nucleoli or nuclei of growing Dictyostelium discoideum cells. The DNA was extracted and was examined by spreading under denaturing conditions for electron microscopy. Intact 95,000 base ribosomal DNA molecules were seen, showing regularly spaced, single-stranded bubbles of about 200 to 400 bases in size, interrupted twice by 11,000 base heavily crosslinked stretches, which correspond to the known positions of the coding regions. The bubbles on the nontranscribed regions indicate the presence of nucleosomes during crosslinking. The DNA was digested with restriction enzymes and analysed by gel electrophoresis in parallel with DNA not treated with psoralen. Fragments from the non-coding region had the same mobility as untreated DNA, while those from the coding region had a markedly lower mobility, though not as low as that of crosslinked pure DNA. This shifting of the bands, specific to the coding region, was also seen when whole cells were treated with psoralen. Treatment of nucleoli with 2 m-NaCl (which is known to dissociate histones) before addition of psoralen led to strong crosslinking all along the ribosomal DNA, resulting in a decreased electrophoretic mobility of bands from the non-coding region, but no further retardation of those from the coding region. In differentiating Dictyostelium cells, slugs, where ribosomal RNA synthesis is very much reduced, the extent of psoralen-crosslinking in the coding region was reduced, but not completely to the level of that of the non-transcribed spacer. In order to test whether psoralen itself alters chromatin structure, crosslinked and non-crosslinked nucleoli from growing cells were lysed with heparin and spread for electron microscopy. There was no difference in the appearance or the frequency of the transcription units seen. Digestion of crosslinked nuclei with micrococcal nuclease indicated an undisturbed structure for bulk chromatin, as well as for the chromatin in the non-transcribed spacer of the ribosomal DNA. Thus psoralen-crosslinking does not lead to extensive disruption or distortion of the structure of either inactive or active chromatin. We conclude, taking the results presented in the Appendix into account, that the extent of psoralen-crosslinking in chromatin DNA is diagnostic for the structure of undistorted chromatin.(ABSTRACT TRUNCATED AT 400 WORDS)