Science, 291, 252-253 (2001) [in Letters] -- "Response to Dr. Dwyer's letter"

Selfish DNA and the Origin of Genes

Hiroyuki Ogata, Stéphane Audic, Jean-Michel Claverie*

Information Génétique & Structural, CNRS-AVENTIS UMR 1889,
31 Chemin Joseph Aiguier, 13402 Marseille Cedex 20, France
* To whom correspondence should be addressed. E-mail: jean-michel.claverie@igs.cnrs-mrs.fr

Structural and functional modularity of proteins is well established. Occurrences of homologous domains in otherwise different proteins suggest the recurrent use of modular units in evolution. The combinatorial advantage of modular units to design diverse proteins is obvious, but the precise relation between evolutionarily successful modules and mobile sequence units is not yet clear. The "trexon" hypothesis proposed by Dwyer (1) and the palindromic element (RPE) that we discovered in several Rickettsia species (2) provide an interesting alternative to the "exon shuffling theory," in which the mobile element precisely coincides with the limits of existing coding exons, thus restricting the evolutionary game to some sort of "card shuffling." The finding of the RPEs suggests a greater flexibility in the evolution of genes.

First, the insertions of RPEs realize a flow of genetic material across the boundary between noncoding and protein-coding sequences. In addition, we recently noticed that one of the RPEs (rpe22 of R. conorii) previously annotated as "intergenic" is located within the gene coding for tmRNA (the transfer/messenger RNA molecule used to rescue stalled ribosomes and to clear the cell of incomplete polypeptides) (3). Thus, the RPE appears capable of parasitizing both protein and RNA structures. The generality of such an influx of genetic material from noncoding to coding sequences deserves further study.

Second, the host proteins targeted by the RPE are different among the species in the same genus. This indicates that RPE insertions occurred after the divergence of those Rickettsia species and that the RPE proliferation might be continuing.

Third, the insertion of the RPE at sites that code for a part of the protein that is on the surface, but not necessarily in between domains or within the constraints of exon boundaries, argues for the possibility of a significant evolution of preexisting protein domains and/or coding exons. For example, the RPE found in the DNA polymerase I of R. helvetica and R. felis is located on the surface of the exonuclease domain (4). Domain insertions within other domains have been described for other proteins, for example, the cat muscle pyruvate kinase, which consists of four different domains. One of the domains forming a beta barrel is located within one of the loops of the other alpha/beta barrel domain (1PKM of Protein Data Bank). The structural and functional consequences of such iterative insertions of domains within domains (a "Russian doll" evolutionary model) remain to be analyzed to better understand the flexibility of genes and genome, as well as the evolutionary modularity of genetic material.

References and Notes

  1. D. S. Dwyer, J. Theor. Biol. 194, 11 (1998).
  2. H. Ogata et al., Science 290, 347 (2000).
  3. As suggested to us by K. P. Williams [see K. C. Keiler et al., Proc. Natl. Acad. Sci. U.S.A. 97, 7778 (2000)].
  4. See the supplemental figure associated with our report (2) available at Science Online at www.sciencemag.org/feature/data/1051142.shl