Tuesday, 7 December 2010

Aliens in California! ...or not.

You’ve probably heard about it by now. In fact I’m probably late to the party, so to speak, in terms of blogging about this... but never mind. Here it goes... NASA discovers arsenic-loving life! 

But wait, what? It’s from California? Not Mars? And it’s a bacterium? That’s not as exciting as it sounded...

My initial reaction to the headlines was scepticism - I’ve got used to being sceptical about pretty much every science story I see on a mainstream news site. I guess it’s no surprise that the media are prone to hyperbole. It would be pretty easy, at a glance, to think an entirely new branch of life had been discovered, something fundamentally different from us and other life on Earth. As it turns out from a closer reading, that’s not the case; however, it’s still pretty exciting! Or I think so, anyway. 

For those who don’t have access to the paper, first, here’s a summary of what the researchers discovered, and how.

The researchers took “lake sediments” (otherwise known as mud) from Mono Lake in eastern California, which is much more alkaline, much saltier, and, crucially for this study, contains much higher levels of arsenic than other lakes. These odd conditions are due to the fact that it has no outlet, so salts cannot flow out of the lake and instead accumulate.

They placed samples of this mud in an artificial bacterial growth medium that contained all the necessary nutrients for life – except phosphorus. Phosphorus is pretty crucial for life – phosphate (PO43-) ions are part of the DNA and RNA backbones, as well as being important in signalling pathways. However, they did add arsenate (AsO43-) ions, which have a very similar structure and properties. Their hypothesis was that some bacteria in the sediment would be able to substitute arsenate for phosphorus and so reproduce without additional phosphorus.

So, what did they find?

They picked out one of the colonies that grew on the arsenate medium for further study, and this is now maintained at 40mM arsenate concentrations (which is significantly higher than the average arsenate concentration in Mono Lake, 200 uM).  They measured its growth rate when provided with arsenate, phosphate, or neither, as shown in this graph:

Figure 1: growth of GFAJ-1 when provided with phosphate (solid circles), arsenate (solid squares), or neither (open triangles). [from ref below].

The graph shows that although the bacteria could reproduce on arsenate medium, it could reproduce more (higher final level), and faster (steeper initial slope) when grown on phosphate. The control stays at about the same level during the whole experiment, at least for OD600, (cells/ml actually increases, although not as much as in the other samples) showing that the bacteria can’t reproduce (much) without at least one of the ions. However, there was a background level of ~3uM phosphate in the media due to trace contamination of other components.

They also measured the levels of arsenic and phosphorus inside the bacteria. In cells grown on arsenate, the mean level of As was 0.19 (± 0.25) % by dry weight, while the cells contained only 0.02 (± 0.01) % P by dry weight. When grown on phosphate, however, the cells on average contained 0.54 (± 0.21) % P and only 0.001 (± 0.0005) % As. The typical % of phosphorus required for growth (as stated by the authors) is 1-3% - so a lot higher than their findings of 0.19% for this bacterium.

I have issues with this figure of 0.19 ± 0.25. Firstly, the legend for the table doesn’t state what the ± value represents – I’d guess standard deviation, as that’s what they use in other places (e.g. the graph above). Secondly, those values lead to a graph that looks like this:

Figure 2: made by me.

In other words, their margin of error for As % is larger than the value itself, across eight replicates. That’s not data I’d be happy with. They’ve provided raw values in the supplementary material (Table S1 – “Cells”), which show that the eight replicates are made up of four repeats of each of two samples – one sample has consistently higher values (0.1-0.6%), one consistently lower (0.009-0.011%) and comparable to the levels of P found in those same samples (0.011-0.014%). If I was them, I’d have wanted to replicate those higher values to be sure... but it seems they were unable to. The authors suggest the variability may be “a result of collection during stationary phase and losses during the repeated centrifugations and washing cycles due to the potential instability of the cellular structures given their swollen state”.

Anyway. Carrying on:

The researchers found that the bacteria grown on arsenate were larger, and contained large intracellular compartments that presumably cause this increase in volume.

They fractionated the cells and observed radiolabelled arsenic in the protein, lipid, metabolite and nucleic acid (DNA/RNA) fractions – the last of these fractions contained about a tenth of the total radiolabelled arsenate. This is roughly similar to their estimate that in a (stationary phase, with one complete average-sized genome) bacterial cell, 4% of phosphate would be found in the DNA (and only a small amount in RNA). These two values are consistent, supporting the idea that AsO43- could substitute for PO43-.

The presence of As in the DNA fraction was confirmed by gel extracting DNA and analysing it; this showed consistently higher As:P ratios in the cells grown on arsenate. 

Figure 3: agarose gel with standard (1), DNA from cells grown on arsenate (2) and DNA from cells grown on phosphate (3). [from ref below].

The difference in positions of the two bands also indicates a difference in the DNA, although I would probably expect the +As band to be higher if As is incorporated into DNA, as As is a heavier atom than P, so As-containing DNA would be heavier and therefore slower-moving... however, differences in electronegativity could cause different migration speeds, I suppose. The large lower band in lane 3 is also a bit concerning – it indicates that the cells sampled, or the treatment of the cells, differed in some way other than the growth media.

Then they did various chemistry-y things that I don’t quite understand (even though I used to want to be a chemist), but which indicate that the environment, i.e. bonding partners, of the arsenic in the As+ cells is consistent with arsenate replacing phosphate in the DNA backbone, and also in proteins and small metabolites.

So, to sum up, the researchers showed: 

  • That GFAJ-1 can grow without (much) phosphate, if arsenate is present.
  • That in these conditions, arsenate is taken up into the cell...
  •  ...and that it is present in protein, lipid, metabolite and nucleic acid fractions, and in chemical environments consistent with being incorporated into these molecules.

Why is this exciting?

Well, it’s not because NASA have discovered a distinct lifeform that has replaced phosphorus with arsenic. Sorry newspapers. GFAJ-1 is in fact a strain of bacteria with recognised close relatives among the Halomonadaceae, and grows significantly better when supplied with phosphorus than it does with arsenic alone.

It has previously been suggested that alien life might be based on silicon instead of carbon, or otherwise substitute elements that are key in terrestrial life.  Arsenic has a similar structure to phosphorus, but behaves differently in water – arsenates are unstable in water at physiological pH. This similarity but instability can explain why arsenic is toxic to us (and others) – it can be initially incorporated into early steps in pathways, but the molecules disintegrate and the pathways cannot be successfully completed. How the GFAJ-1 bacteria get around this problem, if they do incorporate arsenate into crucial molecules, is unknown. If the paper is confirmed by future investigation, GFAJ-1 will offer interesting insights into evolution and into the chemistry of life.

However, this paper doesn’t conclusively prove that arsenic is functionally incorporated into DNA and proteins, nor do the researchers succeed in eliminating all phosphorus from the cells, so the conclusion that this bacterium can simply substitute arsenic for phosphorus and survive doesn’t really hold up. It’ll be interesting to see what further research reveals about GFAJ-1 – and whether what it does is as widely reported.


So, from my sceptical but very much non-expert overview, to some significantly more expert opinions on why the paper is almost certainly wrong:

A Scienceblogs guest post by one Alex Bradley makes some very good points about the half life of arsenate-backbone DNA.

RRResearch has a quantitative analysis of phosphate requirements and concentrations.

And, a link to The Curious Wavefunction, who ponders the chemistry of arsenate-using life and also provided me with the links to the two posts above :)

Original paper: 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. (2010). A Bacterium That Can Grow by Using Arsenic Instead of Phosphorus Science DOI: 10.1126/science.1197258


  1. Great post. I was thinking about doing something similar, especially along the lines of your little column graph :)

  2. Thanks! Heh, I think everyone on the internet has had / will have their say on this topic, it's just a matter of putting your own take on it, and, well, I missed a lot of stuff... seeing other people's views is always good, though, and doing your own post on this is definitely worth it :)

    (also I really recommend the RRResearch post if you haven't seen it already. It's an awesome takedown.)