Back in December 2010, there was a press conference held by NASA to announce the discovery of a bacterium found in a high salt, high pH lake with high concentrations of arsenic that seemed to have substituted arsenic for phosphorus in the bacterium’s biomolecules. This set off a wave of response in the blogosphere regarding what Felisa Wolfe-Simon and her team did nor did not do to confirm arsenic was incorporated into DNA molecules. Controversy ranged from the ability of arsenic to form a stable compound to the types of experiments conducted to confirm incorporation of arsenic into molecules like DNA.

Wolfe-Simon et al. stated they would address all critiques of the Science paper if the critiques were also subject to peer review like their paper titled “A Bacterium That Can Grow by Using Arsenic Instead of Phosphorus”. On May 27, Science published eight Letters to the Editor and the Wolfe-Simon et al. response to the critical comments.

One of the main criticisms of Wolfe-Simon paper concerned the phosphate content of medium used to grow the bacterium GFAJ-1. The researchers had grown the bacterium in both phosphate+/arsenic– medium as well as phosphate–/arsenic+ medium. The mineral salt medium to which the phosphate or arsenic was added, had a background of ~3µM phosphorus. Some argued this could account for the successful growth seen in the arsenic+/phosphorus– medium. In their response to the critique, Wolfe-Simon et al. counters with the growth difference observed without any added arsenic compared to the growth of GFAJ-1 in the presence of arsenic must be explained by any alternative model since they both have the same amount of phosphorus in the medium.

Another possible explanation suggested for the bacterium growth in the presence of arsenic was the toxic element stimulated uptake of phosphorus into the bacterium. The counter argument: there would need to be arsenic detoxification products as the arsenate would need to be removed so the cell could use the phosphorus. The research team did not find any signs of arsenic reduction or methylation, known byproducts of detoxification. However, the team acknowledges this is an avenue they should explore.

Calculations on the amount of intracellular phosphorus of bacteria had suggested that 1.3% of dry weight in a cell would need to be P to create the needed RNA, DNA, ATP and lipids. These calculations were based on normal bacterial populations, not an extremeophile microbe like GFAJ-1. One scenario that was called “ultra-low P” allowed 0.3% P, still at least 5X higher than what Wolfe-Simon et al. measured in their experiments. Two scientists argued that 0.03% P content could be possible for bacteria adapted to depletion of P for populations of cells with an average of 0.5% of P by dry weight. While this debate may seem esoteric and is subject to calculated percent of P in biomolecules and the sensitivity of detection methods (ICP-MS in this case), Wolfe-Simon et al. say their data for their As–/P+ growth of GFAJ-1 agree with the P depletion scenario (they measured 0.33–0.66% of P in dry weight) proposed by the critique but the amount of P under As+/P– conditions is even smaller (0.036%).

The way the data for the intracellular arsenic and phosphorus content were averaged was also critiqued. Artifacts can be introduced when averaging separate experiments. The GFAJ-1 data were then presented from the two experiments separately rather than averaged together. The data does highlight some variability but Wolfe-Simon et al. point out the replicates of each experiment agreed with the exception of the As percentage for one of the As+/P– measurements. Stationary phase collection, sample preparation and possible loss of membrane integrity were suggested as explanations for the variability in the amount of As between the two experiments, but additional data would help clarify this matter. While the researchers acknowledged the greater disparity in the As measurements (0.37% and 0.01%), they argued the %P was lower for the As+/P– but only in the experiment that did not have possible sample preparation issues. Again, more experiments may clarify this issue.

Another criticism of the paper focused on the sample preparation methods for the genomic DNA, the biomolecule selected for analysis by the research group to detect the presence of arsenic incorporated into its structure. They explained how they purified the DNA, a method using a series of organic extractions followed by ethanol precipitation. Furthermore, they argued that the three washes of the bacterial pellet using the mineral salt medium (minus glucose, metals, vitamins, phosphate or arsenate) and the organic extractions would remove any elemental arsenic, and since arsenate and DNA are both negatively charged, it is unlikely to stick to the DNA pellet. The ⁷³AsO₃^– experiment had 11% of the radioactive label associated with the DNA/RNA pellet, and since Wolfe-Simon et al. believe that the As is not a contaminant due to washing and the negative charge, As must be associated with the extracted nucleic acid. In the original Science paper, X-ray analysis showed arsenic in structures with oxygen and carbon similar to the bond distances measured in crystalized DNA, suggesting that arsenic could substitute phosphorus in the DNA backbone.

Since the DNA was analyzed using NanoSIMS without extraction from the agarose gel, one concern was the amount of carbon detected came from the gel. However, any ¹²C^– was assumed to have come from the gel matrix and not used in any calculations. In addition, Wolfe-Simon et al. clarified that they did not mean to imply “wholesale substitution of As for P” with the ratios of As:C because it was not supported by the data. Furthermore, the team acknowledged purifying the DNA from the agarose gel would be an improved experiment by eliminating any factors introduced by having the gel matrix present during analysis and allow them to quantitate As:C and P:C ratios.

Finally, much argument came from the stability of arsenate substituted for phosphate in biomolecules. One critique stated the cellular environment was too reducing for arsenate compounds to be present. However, Wolfe-Simon et al. countered with the lack of redox compounds found in the cell per the arsenic-enhanced uptake of phosphorus discussion. Further critiques pointed out how rapidly arsenate esters hydrolyzed in small model compounds. However, the GFAJ-1 team suggest that large molecules with arsenate esters are more sterically hindered and less likely to hydrolyze than small compounds and cite literature that hint as the arsenate esters become more complex, the hydrolysis decreases. They also reiterated the hypothesis that GFAJ-1 may have evolved stabilizing structures as a coping mechanism.

Much of the controversy about arsenic substitution could be resolved by more data. Whether it comes from Wolfe-Simon and her colleagues or from other labs, reproducible data and more methods to cross check the data from an experiment would help move the field forward. Unfortunately, answers in science do not come swiftly. Hence, we put the “re” in research as I liked to joke when I was in the lab (with appropriate controls of course). Thus, there will continue to be debate about whether this bacterium is merely arsenic tolerant or actually uses arsenate in its biological molecules.

Reference
Wolfe-Simon, F., Blum, J., Kulp, T., Gordon, G., Hoeft, S., Pett-Ridge, J., Stolz, J., Webb, S., Weber, P., Davies, P., Anbar, A., & Oremland, R. (2011). Response to Comments on “A Bacterium That Can Grow Using Arsenic Instead of Phosphorus” Science DOI: 10.1126/science.1202098