LONDON - The discovery of sections of DNA that have no apparent function inside bacterial genes suggests a novel hypothesis to explain how new genes could evolve at random.

Non-coding sequences of DNA - also called "junk DNA" or "selfish DNA" - are known to be common in the genomes of most organisms, but have hitherto been found only in those regions that do not code for proteins. Now, however, researchers in France have found that a repeated sequence of DNA can behave as a molecular parasite and insert itself within protein-coding genes of several Rickettsia species. Rickettsia are obligate intracellular bacteria, from which mitochondria, the "power plant" of all eukaryotic cells, are thought to originate.

Jean-Michel Claverie, director of research at the Centre National de la Recherche Scientifique (CNRS) in Marseille, France, told BioWorld International: "This is the first time it has been shown that these types of repeats can not only shuffle things around, but can also generate new protein sequences. This points toward a mechanism for the continuous creation of new genes."

Rickettsia conorii is an intracellular bacterium that causes Mediterranean spotted fever, also known as boutonneuse fever. R. conorii is closely related to Rickettsia prowazekii, which is the cause of typhus, and is transmitted by ticks and fleas. The genome of R. prowazekii recently has been reported.

Claverie's team, together with researchers from the WHO Reference Centre for Rickettsia in Marseille and the French National Center for Sequencing in Evry, decided to sequence the genome of R. conorii in order to study the evolution of the Rickettsias, and understand the molecular foundation for the different lifestyles of these bacteria and the diseases they cause. Their results are published in the Oct. 13, 2000, Science in an article titled "Selfish DNA in Protein-Coding Genes of Rickettsia." The first author is Hiroyuki Ogata, also from the Structural and Genetic Information Laboratory, a joint venture between Aventis Pharma and CNRS in Marseille.

To the team's surprise, it found that a certain type of repeat was found in 19 open reading frames (protein coding regions) of R. conorii, as well as in 25 different parts of the bacterium's genome that were not thought to encode proteins. The same repeat was identified a further 11 times within genes of other Rickettsia species. These repeats were of a type known as palindromic: They read the same from left to right as from right to left. As these sections occur on a single strand of the DNA, they pair with themselves, forming hairpin-like structures. The result is a short stretch of double helix, additional to the normal genome, which is able to move around the genome and insert itself elsewhere in the DNA.

Claverie said, "We don't know at the moment if this type of situation is only going to be found in the Rickettsia. It might turn out to be a trademark of this very specific type of bacterium. Or it might be a more general phenomenon. The latter seems likely as we know that bacteria can exchange genes, and there is no reason why a repeat that exists in Rickettsia should not have been passed along to other bacteria."

Interestingly, the insertion of the palindromic repeats into the proteins does not seem to affect their function. "But it is possible," Claverie said, "that in a different organism, the insertion of such sequences could turn innocuous proteins into harmful ones, and hence be linked to bacterial pathogenicity."

One of the team's next goals is to prove that the repeats identified are indeed mobile elements. "We are also doing protein structural work," Claverie said. "We have also cloned and expressed several examples of proteins which have this repeat inserted in them, to see what might be special about the repeat when it is inserted into a protein, and try to guess its possible function."

The team will use X-ray crystallography and nuclear magnetic resonance imaging in future structural genomics studies.