The Permian bacterium that isn't.

There is growing evidence for the presence of viable microorganisms in geological salt formations that are millions of years old. It is still not known, however, whether these bacteria are dormant organisms that are themselves millions of years old or whether the salt crystals merely provide a habitat in which contemporary microorganisms can grow, perhaps interspersed with relatively short periods of dormancy (McGenity et al. 2000). Vreeland, Rosenzweig and Powers (2000) have recently reported the isolation and growth of a halotolerant spore-forming Bacillus species from a brine inclusion within a 250-Myr-old salt crystal from the Permian Salado Formation in New Mexico. This bacterium, Bacillus strain 2-9-3, was informally christened Bacillus permians, and a 16S ribosomal RNA gene was sequenced and deposited in GenBank under the name B. permians (accession number AF166093). It has been claimed that B. permians was trapped inside the salt crystal 250 MYA and survived within the crystal until the present, most probably as a spore. Serious doubts have been raised concerning the possibility of spore survival for 250 Myr (Tomas Lindahl, personal communication), mostly because spores contain no active DNA repair enzymes, so the DNA is expected to decay into small fragments due to such factors as the natural radioactive radiation in the soil, and the bacterium is expected to lose its viability within at most several hundred years (Lindahl 1993). In this note, we apply theproof-of-the-pudding-is-in-the-eating principle to test whether the newly reported B. permians 16S ribosomal RNA gene sequence is ancient or not. There are several reasons to doubt the antiquity of B. permians. The first concerns the extraordinary similarity of its 16S rRNA gene sequence to that of Bacillus marismortui. Bacillus marismortui was described by Arahal et al. (1999) as a moderately halophilic species from the Dead Sea and was later renamed Salibacillus marismortui (Arahal et al. 2000). The B. permians sequence differs from that of S. marismortui by only one transition and one transversion out of the 1,555 aligned and unambiguously determined nucleotides. In comparison, the 16S rRNA gene from Staphylococcus succinus, which was claimed to be ‘‘25–35 million years old’’ (Lambert et al. 1998), differs from its homolog in its closest present-day relative (a urinary pathogen called Staphylococcus saprophyticus) by 19 substitutions out of 1,525 aligned nucleotides. Using Kimura’s (1980) two-parameter model, the difference between the B. permians and S. marismortui sequences translates into 1.3