A bacterium that can grow by using arsenic instead of phosphorus.

Arsenic and an old lake: in this paper, Wolfe-Simon and colleagues claim to have weaned from phosphorous a microbe that (at least based on its ribosomal DNA [rDNA] sequence) shares the same last common ancestor as ourselves and other bacteria.

The fact that it's still on the terrestrial tree of life as we know it but can tolerate arsenic is consistent with an ancient adaptation to growth in the arsenic-rich mud where it was isolated. The real surprise is that it still grows in a medium containing arsenic and no added phosphorous, hence the title. While it does grow faster in the presence of added phosphate, the difference is not several orders of magnitude. How this organism uses or detoxifies arsenic should lead to important discoveries, possibly as fundamental as whether arsenic can replace phosphate in the backbone of DNA, as the authors propose but is highly debated, or possibly more practical, such as dealing with arsenic toxicity.

Based on Figure 2a, there appears to be limited RNA production in this cell, however, under strict arsenic growth conditions (unless RNA with an "arsenodiester" backbone stains less efficiently than conventional RNA, or a more mundane explanation would just be RNA degradation in this sample). A real decrease in RNA production would pose a challenge for many reasons, just one of many is the need for ribosomal RNA to make proteins, and this would contribute to the slower growth rate. Perhaps the microbe's metabolism is able to hold on to and recycle the limited phosphate resources its population started with, to supplement growth on arsenic under phosphorous-starvation conditions.

In contrast to the authors and evaluator of this paper, I believe it important to note that this article does not substantiate its key claim that the studied bacterium, Halomonas sp. strain GFAJ-1, "can grow by using arsenic instead of phosphorus". However, despite its many flaws, this article is important because it demonstrates the ability of a bacterial species to tolerate very high environmental levels of arsenate.

The authors' statement "our data show evidence for arsenate in macromolecules that normally contain phosphate, most notably nucleic acids and proteins" is poorly supported by the presented data and has been criticized by several scientists, based on (i) inherent instability of arsenate esters, which quickly hydrolyze in water; (ii) the altered geometry of putative DNA with an arsenate backbone, which would affect base stacking; (iii) apparent presence in the arsenate-grown cells of sufficient phosphate to account for most, if not all, biological functions; (iv) numerous experimental pitfalls that would normally prevent publication of such a paper in a reputable molecular biology journal (see, for example, the critique by Rosie Redfield {1} and Alex Bradley {2}, as well as a selection of comments by other scientists {3}). One could add to this list the undescriptive legends to Figs. 2A and 3A, which do not allow the reader to evaluate the data. Further, a comparison of the Figs. 1E and 3B suggests that, although the cells did contain arsenate, it was primarily accumulated in the intracellular vacuoles. The spectrum fit shown in Fig. 3A is consistent with the spectrum of unbound sodium arsenate determined in a very similar synchrotron experiment {4}.

Despite the limitations of the study, the ability of the new Halomonas GFAJ-1 isolate to grow, albeit very slowly, at pH 9.8 in 40mM sodium arsenate with only micromolar levels of phosphate, is truly remarkable, particularly given that its natural habitat in Mono Lake, California, contains more phosphate (0.4mM) than arsenate (0.2mM) {5}. This raises an important question on the mechanisms of survival in toxic habitats. So far, organisms isolated from extreme environments have demonstrated extreme resilience in maintaining their intracellular milieu to be the same as in their mesophilic counterparts. Their most common adaptation is the use of membrane pumps to import the missing chemicals and to export the toxic compounds. The recently sequenced genome {6} of a closely related Halomonas elongata strain DSM 2581 (with 94% identity on the rRNA level) encodes homologs of phosphate transporters of both low affinity-high velocity Pit (HELO_1616) and high affinity-low velocity Pst (HELO_4423-4420) types {7}. Additionally, H. elongata encodes several proteins (HELO_1499-1497, HELO_2612, HELO_2734-2733 and, potentially, HELO_2732) implicated in arsenate resistance (reviewed in {8}). It would be interesting to see whether these genes are also present in the GFAJ-1 strain and whether adaptation to arsenate has affected expression of these - and any other - genes. Another possible adaptation is the decreased usage of the limiting compound, such as the switch from phospholipids to sulfolipids, described in the phosphate-starved cyanobacteria {9}. It would be worthwhile to check whether the GFAJ-1 strain uses a similar trick. Yet another way for cells to get rid of arsenate would be sequestration in poorly soluble salts of divalent cations {10}, e.g. in the vacuoles seen in Figure 1E. Additionally, future work might still show that arsenate does, in fact, have a cellular role in GFAJ-1.

On a more general note, this paper asks an important question of whether the C, H, N, O, P, and S-based chemical structure of all living cells is uniquely suited for life. In some enzymes, sulfur is successfully replaced by selenium; strontium can substitute for calcium, rubidium for potassium, and so on. With the exception of selenium {8}, these substitutions typically prove toxic or even lethal. Although GFAJ-1 does not seem to deviate from these general rules, finding organisms that do would be an extremely interesting development.


References:
{1} Redfield R, "Arsenic-associated bacteria (NASA's claims)" in "RRResearch" [blog], 4 Dec 2010, www.rrresearch.blogspot.com (Accessed 12 Dec 2010).
{2} Bradley A, "Arsenate-based DNA: a big idea with big holes" in "We, Beasties" [blog], 5 Dec 2010, www.scienceblogs.com (Accessed 12 Dec 2010).
{3} Zimmer C, "Of arsenic and aliens: What the critics said" in "The Loom", 8 Dec 2010, www.blogs.discovermagazine.com (Accessed 12 Dec 2010).
{4} Waychunas et al. Geochim Cosmochim Acta 1993, 57:2251-69 [DOI:10.1016/0016-7037(93)90567-G].
{5} Oremland et al. FEMS Microbiol Ecol 2004, 48:15-27 [PMID:19712427].
{6} Schwibbert et al. Envir Microbiol 2010, Sep 16, Epub ahead of print [PMID:20849449].
{7} Willsky and Malamy, J Bacteriol 1980, 144:356-65 [PMID:6998957].
{8} Stolz et al. Annu Rev Microbiol 2006, 60:107-30 [PMID:16704340].
{9} Van Mooy et al. Nature 2009, 458:69-72 [PMID:19182781].
{10} Magalhães MCF, Pure Appl Chem 2002, 74:1843-50 [DOI:10.1351/pac200274101843].

Wolfe-Simon and colleagues have enriched and isolated a Halomonas strain (gamma-Proteobacteria) from the hypersaline and alkaline Mono Lake in a medium without added phosphate but high concentrations of arsenate (up to 40mM). The high tolerance of arsenate in itself is remarkable, but there is no evidence for the claim in the title that it "can grow by using arsenic (As) instead of phosphorous (P)".

The medium without added P contained 3.1µM P, which is 1000-fold higher than the P concentration in the North Pacific Subtropical Gyre {1}. Prochlorococcus has evolved under these conditions by reducing its dependence on P, and has even replaced phospholipids with sulfolipids {1}. Reduced P requirements would be an alternative, and I think more parsimonious, explanation than claiming that As substitutes P.

The P content of the isolate GFAJ-1 in arsenate medium without added P was 0.019% of the dry weight; considering that the As grown cells were filled to about 90% with droplets that resemble polyhydroxyalkanoate, the cytosolic concentration of P would be about 0.2% of dry weight. This gives a ratio of 1P:646C, which is similar to the 1:500 ratio of Prochlorococcus {1}.

The arsenic:carbon ratio of gel-purified DNA of GFAJ-1 grown in As medium without added P is 1.34E-5. Considering that the P:C ratio of DNA is 1:7, the measured ratio of As:C in the DNA is 10000-fold (9.1E-5) lower than what you would expect if As replaced P in the DNA. An excellent exposition of further flaws in this article can be found on Rosie Redfield's blog, which everyone should read {2}.

References:
{1} Van Mooy et al. PNAS 2006, 103:8607-12 [PMID:16731626].
{2} Redfield R, "Arsenic-associated bacteria (NASA's claims)" in "RRResearch" [blog], 4 Dec 2010, www.rrresearch.blogspot.com (Accessed 12 Dec 2010).

This highly provocative, but ultimately disappointing, paper purports to describe a bacterium that grows with arsenic (As) instead of phosphorous (P).

However, this is misleading and the media hype surrounding the announcement has only made matters worse. In my opinion, this work should not have been evaluated as there are many reasons why this paper does not live up to the claim in its title. The following points highlight only the most egregious errors and mis-statements: (i) the minus (added) P medium actually contains 3 micromolar P (a lot); (ii) the evidence that there is As incorporated into DNA is unconvincing and even the data that are presented, liberally interpreted, indicate a maximal possible substitution of a tiny fraction of P (many suspect there is in fact none); (iii) there is, in fact, no credible evidence that As is incorporated into any metabolites or macromolecules at all. Science magazine will, no doubt, receive letters to the editor that address these issues, and the paper has already been excoriated in numerous venues so I won't belabor the point. However, let me address a somewhat peripheral point: the authors claim that, "there are no prior reports of substitutions for any of the six major elements essential for life." This is patently untrue since there are very interesting examples of precisely this type of adaptation in many microbes. While perhaps not involving an entirely 'new' element to biology, two evolutionarily relevant examples are worth mentioning. First, it has been known from chemostat experiments in the late 1960s that Bacillus subtilis replaces an abundant P-glycerol polymer in its cell walls (teichoic acids) with a carboxylate-containing alternative (teichuronic acid) and thereby decreases the cell's P demand by a factor of ~2 ({1} and references therein). Second, in marine environments (many ocean surface waters are P-limited) cyanobacteria routinely replace phospholipids with sulfolipids {2}, thereby increasing competitiveness in a P-limited environment. Certainly, finding a new type of genetic material would be exciting, but this has not yet happened. Ultimately, this work will stimulate further experiments and will, incidentally, provide useful material for undergraduate and graduate student courses training the next generation how to critically develop, read, and evaluate data.

References:
{1} Lang et al. J Bacteriol 1982, 151:367-75 [PMID:6806244].
{2} Van Mooy et al. Proc Natl Acad Sci USA 2006, 103:8607-12 [PMID:16731626].

This paper is fascinating for two reasons – firstly, it demonstrates that the halophile GFAJ-1 can break the rules of biochemistry as we know them to incorporate arsenic into its DNA and possibly protein, and, secondly, it showcases the approach of looking for 'weird life' or extraterrestrial life on earth.

Searching for weird life on earth could answer a couple of compelling questions. First, since astrobiologists have long proposed that microbial life could survive space travel on meteors or rocks ejected from nearby planets like Mars {1}, we may be able to find extraterrestrial life or signatures of it faster by looking for it on Earth than we will waiting to receive an extraterrestrial visit or transmission. Also, in order to estimate the probability of life developing on other earth-like planets, it would be useful to determine if life evolved on our planet more than once. It has been proposed that a 'shadow biosphere' could exist on Earth containing life that evolved independently from life as we know it. One proposed strategy to look for such life is to make an educated guess about what kind of signature such life would present and then to look for the signature in microbial populations inhabiting extreme environments, the rationale being that microbial life represents the greatest number of terrestrial life forms and that 'weird' microbial life may still exist in extreme environments due to low competition {2}. In a 2008 hypothesis paper, the author of this paper proposed the incorporation of arsenic into biological macromolecules as a possible signature for weird life {3}. Arsenic has a very similar structure to phosphate but slightly different chemical properties, which make it toxic for most species but not perhaps for an organism that evolved in an arsenic-rich environment. In this paper, the authors describe such an arsenic signature in GFAJ-1.

The authors demonstrate that the halophile GFAJ-1 collected from Mono Lake, CA, (a hypersaline lake containing 200µM arsenic) can undergo arsenic-dependent growth and incorporate arsenic into DNA under low phosphate conditions. They used NanoSIMS and were able to detect arsenic in gel-purified DNA extracts. In addition, they were able to show in sub-cellular fractionation studies that roughly 75% of the radio-labeled arsenic in GFAJ-1 grown under low phosphate conditions was associated with a phenol-extracted 'protein' fraction. Although the authors do not have evidence yet for the mechanisms by which arsenic is being incorporated into these molecules, they propose that arsenic may be involved in the formation of DNA as well as small metabolites like NADH and ATP. They also propose that it may be substituting for phosphate in protein phosphorylation. It should be noted that there is considerable evidence that arsenic can bind many proteins in 'normal' organisms {4}. Although the evolution of GFAJ-1 is well rooted in our current Tree of Life, it will be interesting to find out the mechanisms and pathways that arsenic is used in by GFAJ-1 and also to what extent this happens in nature.