The molecular basis for length heterogeneity in ribosomal DNA from Xenopus laevis

Abstract

The restriction endonuclease EcoRI cleaves Xenopus laevis ribosomal DNA twice in each repeating unit to yield two classes of fragments. One class is homogeneous in length and contains only gene sequences; the other class is heterogeneous in length and contains all of the non-transcribed spacer and some of the gene regions (see model in the Introduction). Four spacer-containing fragments of different length have been cloned in Escherichia coli. They have been compared by optical melting and by homoduplex and heteroduplex mapping in the electron microscope. The results show that two regions within the nontranscribed spacer account for the length heterogeneity. The region which varies most in length is adjacent to the transcription unit at its 5′ end, while the other variable-length region is near but not adjacent to the 3′ end of the transcription unit. Each of these variable-length spacer regions consists of internally repetitious simple sequences (“subrepeats”), which are probably shorter than 50 base-pairs in length. The length heterogeneity of rDNA repeating units is due to more or less copies of these subrepeats. These two regions of variable length are separated by a constant-length region. Another constant-length DNA region separates the 3′ end of the transcription unit from one of the variable spacer sequences. There is no evidence for subrepeats within these latter two spacer regions. It is proposed that the two variable-length regions within the spacer participate in and perhaps enhance the correction mechanisms which permit parallel evolution of tandem genes.