The Ribosome as a Missing Link in Prebiotic Evolution III: Over-Representation of tRNA- and rRNA-Like Sequences and Plieofunctionality of Ribosome-Related Molecules Argues for the Evolution of Primitive Genomes from Ribosomal RNA Modules

Abstract

We propose that ribosomal RNA (rRNA) formed the basis of the first cellular genomes, and provide evidence from a review of relevant literature and proteonomic tests. We have proposed previously that the ribosome may represent the vestige of the first self-replicating entity in which rRNAs also functioned as genes that were transcribed into functional messenger RNAs (mRNAs) encoding ribosomal proteins. rRNAs also encoded polymerases to replicate itself and a full complement of the transfer RNAs (tRNAs) required to translate its genes. We explore here a further prediction of our "ribosome-first" theory: the ribosomal genome provided the basis for the first cellular genomes. Modern genomes should therefore contain an unexpectedly large percentage of tRNA- and rRNA-like modules derived from both sense and antisense reading frames, and these should encode non-ribosomal proteins, as well as ribosomal ones with key cell functions. Ribosomal proteins should also have been co-opted by cellular evolution to play extra-ribosomal functions. We review existing literature supporting these predictions. We provide additional, new data demonstrating that rRNA-like sequences occur at significantly higher frequencies than predicted on the basis of mRNA duplications or randomized RNA sequences. These data support our "ribosome-first" theory of cellular evolution.

Keywords: LINE; SINE; aminoacyl tRNA synthetase; extra-ribosomal; genome; mRNA; non-canonical; paralog; rRNA; ribosome; ribosome-binding protein; tRNA.

Figure 1

Locations of 5S rRNA, 16S rRNA, and 23S rRNA frame 1 homologies within two strains of E. coli K12 genomes. Note that the scale is non-linear. Homology locations are not identical between the 5S, 16S, and 23S rRNAs, demonstrating that the sequences probably have independent origins. Similarly, the homology locations are not identical even between two closely related E. coli K12 strains, suggesting that these sequences are subject to reorganization processes within these genomes.

Figure 2

Locations of the six frames of 5S rRNA homologies within two strains of E. coli K12. Note that the scale is non-linear. As in the previous Figure, these homology locations differ between the two strains indicating that they are subject to reorganization processes. Notably, the locations also differ considerably from one reading frame to another, demonstrating that the search algorithm is identifying non-overlapping sequences.

Figure 3

Comparison of 5S rRNA variables with matched protein mRNA and randomized RNA controls (original data in Tables above). Error bars show SD. While the mean length of the matches found in each group was very similar, the number of total matches, the number of identical amino acids found within those matches, and the number of exact matches found to the 5S rRNA was very significantly greater than those found for the proteins’ control mRNA and randomized RNA sequences tested.

Figure 4

Comparison of 16S rRNA variables with matched protein mRNA and randomized RNA controls (original data in Tables above). Error bars show SD. The mean number of matches found within the E. coli genome was greatest for the 16S rRNA, as was the mean number of perfect identities, length of the sequences with identical sequences and the number of identical amino acid matches found within each match was significantly greater for the 23S rRNA than for the mRNA of the proteins controls or the randomized RNA sequences.

Figure 5

Comparison of 23S rRNA variables with matched protein mRNA and randomized RNA controls (original data in Tables above). Error bars show SD. Although the mean number of matches found within the E. coli genome was greatest for the protein controls, the mean number of perfect identities, length of the sequences with identical sequences and the number of identical amino acid matches found within each match was significantly greater for the 23S rRNA than for the mRNA of the protein controls or the randomized RNA sequences.

Figure 6

A summary of the evidence for pleiofunctionality in rRNA-related molecules supporting the hypothesis of a self-replicating ribosome. rRNA = ribosomal ribonucleic acid (only one rRNA is shown though there are three in the ribosome, the 5S, 16S, and 23S). crRNA = complementary strand of rRNA posited to exist to make possible rRNA replication in self-replicating ribosome. The crRNA would also act as a messenger RNA (mRNA), or act itself as a source of transcription to produce mRNAs (see Root-Bernstein and Root-Bernstein, 2016; 2017). tRNA = transfer ribonucleic acid. aa-synthetase = amino acyl synthetase, which loads amino acids onto tRNA. Ribo-protein = ribosomal proteins (see Table 3, Table 4, Table 5 and Table 6). Boxes summarize the data from Table 1, Table 2, Table 3, Table 4, Table 5, Table 6 and Table 7 and the related text.

Figure 7

Schematic model of how RNAs perform various ribosomal functions and/or encoding various ribosome-related proteins may have become ligated to form the first genome. The visual formalisms are the same as in Figure 6 and illustrate the idea that “genes” that may have existed as overlapping sequences in a variety of reading frames within a self-replicating set or rRNAs may have been separated out into a single reading frame for better functional control. This schema does not identify whether the original primitive genome was constituted of RNA or DNA (though RNA seems likely if a ribosomal origin is correct) but given a common origin for RNA and DNA polymerases (see text) the existence of either one could have provided the basis for the other. It is also important to remember that SINEs and other insertion-facilitating RNAs most likely evolved from tRNAs or that they have common ancestors (see text), so that molecular mechanisms of genome growth would have been present from the outset in a ribosome-first scenario.