Extraordinary claims require extraordinary evidence, as the great Carl Sagan once said. And they don’t come much more extraordinary than Wolfe-Simon et al’s highly publicized claim that bacteria can incorporate arsenic into their DNA instead of phosphate, and hence that the six fundamental elements of life (carbon, hydrogen, nitrogen, oxygen, sulphur and phosphorus) are not so fundamental after all.
The claim was met with a resounding outcry by the wider scientific community, with extensive methodological criticism highlighting many errors which could have introduced the illusion of arsenate incorporation. However, Rosie Redfield and colleagues went the whole hog by replicating the study in its entirety, modifying it to include rigorous checks for contamination, the absence of which caused such a negative reaction to the original paper. The results of this replication were published in Science Express on July 8th.
Wolfe-Simon et al had claimed that their growth medium contained no phosphate, and hence that growth when arsenic was added was a result of arsenate incorporation into the DNA of the bacteria (which was a strain called GFAJ-1). However, Redfield et al refute this claim by pointing to the large body of literature cataloguing bacterial growth at the level of phosphate (3-4 uM) remaining in this ‘phosphate-free’ medium. Conclusion: growth of GFAJ-1 when arsenate was added could have been caused by trace levels of phosphate in Wolfe-Simon et al’s medium.
It was also reported by Wolfe-Simon et al that GFAJ-1 cells grew very slowly in the medium, but grew faster when arsenic was added. This was refuted by Redfield et al, who replicated the level of phosphate in Wolfe-Simon et al’s ‘phosphate-free’ medium, and observed significant growth. Conclusion: GFAJ-1 doesn’t need arsenic to grow quickly at low levels of phosphate.
The most contentious claim in the original paper was that as much as 4% of the phosphate in the DNA backbone of GFAJ-1 was replaced by arsenate (Wolfe-Simon et al, 2011). Redfield et al checked for the presence of arsenate bonds after three serial washes of the DNA with distilled water, and found arsenate present at a level of 5 x 10-8 M, 50-fold lower than the 4% claimed by Wolfe-Simon et al. Incidentally, they found a similar level in their water blank, suggesting that even this low level is a result of remaining contamination. Conclusion: high arsenate levels in the original paper were most likely a result of contamination, introduced by insufficient washing of DNA.
Logically, one would not expect arsenate bonds to exist, because they have previously been reported to be unstable, quickly breaking down by hydrolysis. Wolfe-Simon et al had claimed that internal proteins or compartmentalization may protect arsenate bonds from this hydrolysis. Unfortunately, this claim was also refuted by Redfield et al, who showed by gel migration that GFAJ-1 DNA is not associated with hydrolysis-protecting proteins. They also compared the size of DNA fragments seen before and after removal of any potential hydrolysis-protecting proteins, and found no difference in the fragment size. Conclusion: there is no hydrolysis-protection mechanism active in GFAJ-1 DNA, and hence arsenate bonds are not even logically possible in this bacterium.
So it would appear that there is no evidence (extraordinary or not) for arsenic incorporation. I don’t think I can improve on the conclusion by Redfield:
“The end result is that the fundamental biopolymers conserved across all forms of life remain, in terms of chemical backbone, invariant.”
I suppose there are two ways to look at this. The first is positive: after all, Redfield et al have ruthlessly employed the scientific method not only to highlight an erroneous claim, but to specify and quantify the source of the error. But one can’t shake the negative perspective: valuable time has been spent disproving something that should never have been published in the first place.
Wolfe-Simon, F. et al (2011) A bacterium that can grow by using arsenic instead of phosphorus. Science 332: 1163-1166.
Reaves, M.L. et al (2012) Absence of detectable arsenate in DNA from arsenate-grown GFAJ-1 cells. Sciencexpress 8 July: 1-4