Vreeland et al. (2000) reported the successful isolation of a spore-forming bacteria, Bacillus strain 2-9-3, from a brine inclusion within a halite crystal recovered from a presumably 250 Ma Permian Salado Formation in Carlsbad, New Mexico. The authors claim the bacterium isolate to be as old as the formation itself, thereby nominating it to be the oldest viable organism ever recovered. Since then, both the age of the brine inclusion and the bacteria isolate itself have been questioned (Hazen and Roedder, 2001; Graur and Pupko, 2001; Nickle et al., 2002; Willerslev et al., 2004a; Hebsgaard et al., 2005). Recently, Satterfield et al. (2005) presented evidence that brine inclusions from the same layer of salt that housed Bacillus strain 2-9-3 are composed of evaporated Late Permian seawater. The authors find these results to be in strong support of the old age of strain 2-9-3. We strongly disagree with this conclusion. Although the age of the brine from where strain 2-9-3 was isolated might, indeed, be 250 Ma, it does not necessarily mean that Bacillus strain 2-9-3 is of similar age. In fact, there are a number of both theoretical and empirical reasons to suggest otherwise.
1. The use of nonspecific media for the culturing of Bacillus strain 2-9-3 makes the risk of false positive results extremely high. Although Satterfield et al. (2005, p. 265) state that Bacillus strain 2-9-3 “has received significant publicity because of the extreme sterilization techniques used to avoid contamination by modern microorganisms,” Vreeland et al. (2000) did not follow some of the most basic authentication criteria, such as replication of their result in an independent laboratory. This criterion is of great importance if laboratory based contamination is to be excluded (i.e., all types of contamination related to the laboratory), as it's unlikely that different laboratories would obtain the exact same result due to a common lab contaminant (Willerslev and Cooper, 2005). Intriguingly, no claims of geologically ancient cultures or DNA sequences published to date (i.e., claims >1 Ma) have followed this simple criterion of authentication (Hebsgaard et al., 2005).
2. DNA is a relatively unstable molecule compared to other cellular components such as lignin and cutine, and will degrade with time if not repaired. The rate of degradation is known to be highly dependent on the environment, particularly the temperature (Smith et al., 2001; Willerslev et al., 2004b; Willerslev and Cooper, 2005). For example, calculations have shown that free DNA experiencing depurination damage will break down to <100 base-pair (bp) fragments in less than 10 k.y. under warm and humid conditions, and in less than 100 k.y. under cold conditions (Poinar et al., 1996; Smith et al., 2001; Willerslev et al., 2004b). No metabolic activity has yet been measured from endospores like Bacillus strain 2-9-3, excluding the possibility of active DNA repair prior to germination. Although endospores have special adaptations such as DNA binding α/β-type small acid soluble proteins to reduce the rate of genomic modification, it still remains unlikely that they should be able to germinate after hundreds of millions of years of dormancy, particular under nonfrozen conditions. In support of this, a recent study of bacterial DNA in permafrost—an environment considered the most promising for long-term DNA survival—has shown that DNA from endospore-forming bacteria >600 bp in size cannot be obtained from samples older than 0.5 Ma, and not even 120-bp DNA fragments can be reproducibly obtained from samples ≥2 Ma (Willerslev et al., 2004a).
3. Relative rate tests conducted on 16S ribosomal DNA and protein coding gene (recA and splB) sequences obtained from Bacillus strain 2-9-3 strongly suggest that it is not geologically ancient (Graur and Pupko, 2001; Nickle et al., 2002; Maughan et al., 2002; Hebsgaard et al., 2005). This test investigates the relative genetic distance from an outgroup to the postulated ancient organism and its closest contemporary relatives. One assumption of the relative rate test is that the mutation rate is similar and constant in the DNA sequences for both the ancient organism and its contemporary relatives. If Bacillus strain 2-9-3 is ancient, then a significantly shorter genetic distance to the outgroup is expected as compared to its contemporary relatives, as the latter should have an additional 250 m.y. to accumulate substitutions. Although mutation rates are not always similar across organisms, it is striking that not a single claim of geologically ancient DNA, including that of Bacillus strain 2-9-3, has passed the rate test so far, strongly suggesting that contamination is involved (Hebsgaard et al., 2005).
Therefore, despite the recent results of Satterfield et al. (2005), it still remains highly controversial whether the Bacillus strain 2-9-3 is, indeed, millions of years old.