Arsenic-permissive bacteria – implications for arsenical cancer chemotherapy

Like many of you, I waited this week for details on the NASA press conference and Science paper on a major discovery – painted as an “alien lifeform” by some news outlets. The truth did not live up to the hype but it was an impressive biological finding: a group led by Felisa Wolfe-Simon discovered a bacterium in California’s Mono Lake that could still grow when phosphorous was completely replaced with arsenic. The bacterium, strain GFAJ-1 of the Halomonadaceae family of Gammaproteobacteria, appears to use arsenic in place of phosphorus in molecules where phosphate is used – shown most conclusively here for DNA.

The story is perhaps best told by science writer extraordinaire, Ed Yong, at Not Exactly Rocket Science, and biologist PZ Myers is to be commended for representing us well to our chemistry colleagues by actually breaking out the periodic table in his excellent teaching post.

Addendum (4 Dec, 2:46 pm): An excellent chemistry-flavored post at The Curious Wavefunction also came to my attention – he/she cited this great 1987 Science paper, “Why Nature Chose Phosphates (PDF),” from the late Harvard chemist Frank Westheimer which discusses, among other things, the difficulties in overcoming the lability of arsenate esters in biomolecules.

The bacterium doesn’t exactly thrive on arsenic, to be sure. It’s home, Mono Lake, contains arsenic in the form of arsenate at a concentration of about 200 μM. The research team isolated and propagated the bacterial strain from lake sediment and observed that when grown in a defined culture medium, it normally incorporated a small amount of arsenic into biomolecules (as determined by ICP-MS from dry weight calculations) relative to phosphorus at a ratio of 0.002 to 1.

But to truly allow significant arsenic incorporation, the bacterial culture had to be grown stepwise in arsenate (AsO43-) at concentrations from 100 μM to 5 mM. Then, a single colony was isolated and grown and maintained at 40 mM arsenate in growth medium. Under these conditions, the bacteria didn’t seem particularly happy as they grew more slowly with an oblong morphology and large vacuoles inside the cells, but they did multiply by 20-fold after six days in the highest concentration. Here, the bacteria preferentially incorporated arsenic at a ratio of 7.3 to phosphorus in dry weight. So, the bacteria are still able to use some phosphate from the trace amount in the medium (about 3.1 μM). But the remarkable feat is that the bacteria can be maintained on 40 mM arsenate. That is the impressive discovery.

I began to think about this in the context of arsenic chemotherapy. Although arsenic is toxic to most cells, Paul Ehrlich’s group had pioneered the use of an arsenical compound to treat syphylis in the early 1900s before penicillin was available. In a post we wrote in 2007, we describe the revisiting of arsenic as medicine when the chemotherapeutic drug, arsenic trioxide (Trisenox, Cell Therapeutics), was approved by the US FDA in 2000 for the second line treatment of acute promyelocytic leukemia.

Lawrence Boise, PhD. Emory University and Winship Cancer Institute

Thinking back on this, I wondered if the chemotherapeutic effect of arsenic trioxide might be due to arsenic incorporation into cancer cell DNA, RNA, or protein and the cell’s inability to tolerate this substitution. Such a scenario would require that the arsenate be incorporated into ribo- and deoxyribonucleotide triphosphates and then used in nucleic acid polymerization or as a phosphate-mimietic source for protein kinases.

Such a mechanism is not as far-fetched as one might think. Some of our first chemotherapeutics took advantage of this “counterfeit incorporation” mechanism such as the antimetabolite nucleotide analogs, 6-mercaptopurine, or the antiviral drugs, acyclovir or valacyclovir. These agents resemble normal precursors enough to be used by biosynthetic enzymes but then prevent the catalysis of subsequent steps, leading to cell death.

So, I was fortunate to be able to reach Dr. Lawrence Boise, Professor of Hematology and Medical Oncology and Georgia Cancer Coalition Distinguished Cancer Scholar at Emory University’s Winship Cancer Institute. Dr. Boise is well-known for his work on arsenic-containing compounds in leukemia and multiple myeloma and was a pioneer in understanding tumor cell apoptosis, or programmed cell death.

When my e-mail caught up with Dr. Boise on his way to the American Society of Hematology meeting, he responded to my musings as follows:

In the case of APL where it is FDA approved and most effective the mechanism appears to be primarily related to arsenic binding to the PML portion of the PML-RARα fusion protein, resulting in its degradation, so any role for DNA incorporation where it is most effective as a therapeutic would be minor at best.

Of course, APL is an unusual leukemia that develops from the novel fusion protein Dr. Boise describes. But what about in other tumor types? Could arsenic incorporation into nucleic acids lead to DNA damage or other effects contributing to tumor cell death?

Again, Boise:

Arsenicals have been tested in other diseases and the mechanism(s) by which it induces apoptosis are still not completely clear. In our gene expression profile studies in myeloma cell lines we did not see evidence of a DNA damage response (paper was published in JBC in 2009 although we focused anti-oxidant response since it was the primary observation).

So, although my bubble is burst, I still wouldn’t be surprised if eukaryotic biologists begin to investigate whether arsenic can indeed be incorporated into DNA, perhaps not causing DNA strand breaks but instead interfering with DNA replication or transcription. Moreover, I’d be interested to know if proteins can be modified by arsenylation from protein kinases using adenosine triarsenate instead of ATP. Or what about arsenylation of inositol to disrupt second messenger signaling?

Just imagine, a whole new field of arsenomics.

Update (5 Dec, 11:07 am): University of British Columbia microbiologist, Rosie Redfield, has posted a very detailed technical critique of the Science paper that questions some of the findings.

Author: David Kroll

Share This Post On


  1. I think that the concentration of As in Mono Lake would depend on whether one is considering the main body of water or one of the more isolated sloughs. Where were the bacteria samples taken?

  2. Glassed in the forehead? Sounds like a bad night out in Glasgow/Sydney


  1. Quick Links | A Blog Around The Clock - [...] bacteria (NASA’s claims) and Arsenic-permissive bacteria – implications for arsenical cancer chemotherapy and Are there viruses of arsenic-utilizing bacteria? …
  2. Arsenic Bacteria link-dump | A Blog Around The Clock - [...] bacteria (NASA’s claims) and Arsenic-permissive bacteria – implications for arsenical cancer chemotherapy and Are there viruses of arsenic-utilizing bacteria? …
  3. Post-publication peer review in public: poison or progress? | Terra Sigillata - [...] stemming from NASA’s embargoed press conference and paper in Science (somewhat ballyhooed here by yours truly), is accumulating criticism …
  4. Arsenic bacteria – a post-mortem, a review, and some navel-gazing | Not Exactly Rocket Science - [...] practicing scientists took the findings on face value, speculating about everything from the implications for chemotherapy to whether the …