We have determined the complete nucleotide sequence (4712 nucleotides) of the mouse 28S rRNA gene. Comparison with all other homologs indicates that the potential for major variations in size during the evolution has been restricted to a unique set of a few sites within a largely conserved secondary structure core. The D (divergent) domains, responsible for the large increase in size of the molecule from procaryotes to higher eukaryotes, represent half the mouse 28S rRNA length. They show a clear potential to form self-contained secondary structures. Their high GC content in vertebrates is correlated with the folding of very long stable stems. Their comparison with the two other vertebrates, xenopus and rat, reveals an history of repeated insertions and deletions. During the evolution of vertebrates, insertion or deletion of new sequence tracts preferentially takes place in the subareas of D domains where the more recently fixed insertions/deletions were located in the ancestor sequence. These D domains appear closely related to the transcribed spacers of rRNA precursor but a sizable fraction displays a much slower rate of sequence variation.

RNA polymerase I terminates transcription of mouse rDNA 565 bp downstream of the 3' end of mature 28S rRNA. This specific termination event can be duplicated in a nuclear extract system. RNA molecules with authentic 3' ends are transcribed from ribosomal minigene constructs provided the templates retain a minimal length of downstream spacer sequences. The nucleotide sequence of the region of transcription termination contains a set of repetitive structural elements consisting of 18 bp conserved nucleotides surrounded by stretches of pyrimidines. Termination in vivo occurs within the first element. This site is preferentially used in vitro at low template concentrations. At increasing DNA concentrations a termination site within the second repetitive element is used. Competition experiments with defined 3'-terminal fragments suggest that transcription termination by RNA polymerase I requires interaction of some factor (or factors) with the repetitive structural elements in the 3' nontranscribed spacer.

We present a secondary structure model for the entire sequence of mouse 28S rRNA (1) which is based on an extensive comparative analysis of the available eukaryotic sequences, i.e. yeast (2, 3), Physarum polycephalum (4), Xenopus laevis (5) and rat (6). It has been derived with close reference to the models previously proposed for yeast 26S rRNA (2) and for prokaryotic 23S rRNA (7-9). Examination of the recently published eukaryotic sequences confirms that all pro- and eukaryotic large rRNAs share a largely conserved secondary structure core, as already apparent from the previous analysis of yeast 26S rRNA (2). These new comparative data confirm most features of the yeast model (2). They also provide the basis for a few modifications and for new proposals which extend the boundaries of the common structural core (now representing about 85% of E. coli 23S rRNA length) and bring new insights for tracing the structural evolution, in higher eukaryotes, of the domains which have no prokaryotic equivalent and are inserted at specific locations within the common structural core of the large subunit rRNA.

We report the sequence of the 4006-nucleotide 5' external transcribed spacer (5'ETS) of the mouse ribosomal primary transcript. These data complete the sequence of the 13.4-kb mouse rRNA gene, thus providing a mammalian rRNA gene structure, in addition to yeast and Xenopus. The mouse 5'ETS displays a highly biased base content (very high in GC and particularly low in A), closely similar to the other transcribed spacers of the mouse ribosomal gene. This region seems to have accumulated sequence variation relatively rapidly during vertebrate evolution, with the possible insertion in rodents of sequences structurally similar to retroposons. About half the length of the mouse 5'ETS can fold into a giant and highly stable secondary structure, which is probably evolutionarily conserved in mammals and which could play an important role in the higher-order organization of mammalian pre-ribosomes.