A short time ago, a study describing the discovery of a bacterium that could apparently utilize arsenate in place of phosphate was published in Science. This report has since generated a lot of discussion and debate in the blog-o-sphere, most of it rather uncomplementary. I count myself among those who are skeptical of the more daring of the authors’ claims.
Something I found curious was the gel in Fig. 2A, an analysis of nucleic acids isolated from bacteria grown in arsenate-rich (middle lane) or phosphate-rich (right lane) media:
Something is missing in the middle lane – the ribosomal RNA that is so abundant in the right-hand lane. Possibly related to this curious result, the authors make this statement that is hidden away in the paper:
“Since these cells were harvested in stationary phase (11), the fraction of P associated with RNA is likely small (14).”
These authors seem to be implying that cells grown in arsenate have little RNA, far less than DNA. This seems very unlikely to me. Even slowly-growing cells should have as much or more ribosomal RNA as DNA. Indeed, in contrast to what the authors are claiming here, reference 14 mentions that total RNA levels in non-growing bacterial cells do not change very much.
Reference 11 is the methods, and reference 14 is: J. Mandelstam, Bactierol Rev. 24, 289 (1960). The authors of the Science paper seem to have mixed up a discussion of RNA stability with a consideration of overall RNA quantities. On p.297, Mandelstam states in this regard:
“In experiments with a leucine auxotroph (26) cells were labeled with P32-orthophosphate and then starved of leucine in an unlabeled medium. About 20 per cent of the p32 of the RNA was lost in 90 min and at least half of this was recovered from the incubation medium. There must have been a good deal of resynthesis of RNA from unlabeled phosphorus, because the total RNA decreased only 0 to 12 per cent, which is far less than the decrease in label.”
Obviously, the study Mandelstam cites does not claim that total RNA levels decrease dramatically (let alone to the extent that would have to be the case in the arsenate-grown cells) in non-growing cells.
I don’t have any ready explanation for this figure, but I am not comfortable with the authors’ explanations. There would seem to be more to these cells than meets the eye.
These cells may not be growing at all. If they’re the same cells in Figure 2A, they stopped dividing about 15 days before harvesting, probably because the ~3 µM phosphate in the medium was used up, and even stopped accumulating polyhydroxybutyrate several days before harvesting.
A very small amount of rRNA is still visible in lane 2. I suggested (somewhere in my several blog posts and responses to comments about this paper, at http://rrresearch.blogspot.com (scroll down to see the arsenic posts)) that once the cells ran out of phosphate in the medium, they scavenged the phosphate from almost all their ribosomes. I don’t know if anyone has ever measured the relative amounts of DNA and RNA in cells that are not adapted to phosphate starvation but have been starved for several weeks.
The similar intensities of the DNA bands may also be misleading. The wad of rRNA in lane 3 will have absorbed all the EthBr in the agarose it passed through, so the much greater amount of DNA in this lane has been stained only at its edges.
Given the many other errors in the authors’ methods, this anomaly could have a number of other explanations.
Thanks for the comment.
A quick perusal of some rather old* studies (dating to the 70’s and 80’s) reveals that E. coli that have been starved to the point of loss of viability (several logs worth) do in fact start to lose ribosomes. Judging from those studies, I would guess that the “arsenate-eating” bacteria are not really eating arsenate, but rather that they have starved themselves to the point of being dead. So the “cells” analyzed in this figure are long past the point of viability, and it is rather unlikely that they are really growing slowly on arsenate. They may not be cells at all, but rather vesicles of some sort.
An odd thing about Fig. 1A is the way these authors counted cells. Typically, one would just take a sample, do some serial dilutions, and plate out in rich media to count viable cells. This group stained and counted everything and called them cells. It would have been nice to see some viable cell counts during the “stationary phase” that is seen in Fig. 1A.
* – I can’t believe I’m calling papers from the 70’s and 80’s old….
If the cells grown in arsenate-rich medium incorporated arsenate into its DNA, wouldn’t the DNA run higher on the gel because it’s heavier?
The gel that I showed here doesn’t have the resolution to show the difference you are expecting.