Marine biology: Flea market

A newly discovered virus may be the most abundant organism on the planet

What is the commonest living thing on Earth? Until now, those in the know would probably have answered Pelagibacter ubique, the most successful member of a group of bacteria, called SAR11, that jointly constitute about a third of the single-celled organisms in the ocean. But this is not P. ubique’s only claim to fame, for unlike almost every other known cellular creature, it and its relatives have seemed to be untroubled by viruses.

As Jonathan Swift put it in a much-misquoted poem, “So, naturalists observe, a flea/Hath smaller fleas that on him prey”. Parasites, in other words, are everywhere. They are also, usually, more abundant than their hosts. An astute observer might therefore have suspected that the actual most-common species on Earth would be a “flea” that parasitised P. ubique, rather than the bacterium itself. The absence of such fleas (in the form of viruses called bacteriophages, that attack bacteria) has puzzled virologists since 1990, when the SAR11 group was identified. Some thought the advantage this absence conferred explained the group’s abundance. But no. As they report in this week’s Nature, Stephen Giovannoni of Oregon State University and his colleagues have discovered the elusive phages. Swift’s wisdom, it seems, still holds good.

Tracking down a particular virus in the ocean makes finding a needle in a haystack look a trivial task. A litre of seawater has billions of viruses in it. Modern genetic techniques can obtain DNA sequences from these viruses, but that cannot tie a particular virus to a particular host.

To do so, Dr Giovannoni (pictured) borrowed a technique from homeopathy: he diluted some seawater to such an extent that, statistically speaking, he expected a 100-microlitre-sized aliquot to contain only one or two viruses. The difference between his approach and a homeopath’s was that what homeopathy dilutes almost to nothing are chemicals, and thus cannot breed. A virus can, given a suitable host. So he mixed each of several hundred aliquots into tubes of water containing P. ubique. Then he waited.

The race is to the Swift

After 60 hours, he looked to see what had happened. In most cases the bacteria had thrived. In a few, though, they had been killed by what looked like viral infection. It was these samples that he ran through the DNA-sequencing machine, in the knowledge that the only viral DNA present would be from whatever it was had killed the bacteria.

His reward was to find not one, but four viruses that parasitise P. ubique. He then compared their DNA with databases of DNA found in seawater from around the world, to find out how abundant each is. The upshot was that a virus dubbed HTVC010P was the commonest. It thus displaces its host as the likely winner of the most-common-living-thing prize.

That does depend, of course, on your definition of “living thing”. Some biologists count viruses as organisms. Some do not. The reason is that a virus relies for its growth and reproduction on the metabolic processes of the cell it infects. This means viruses themselves are hard to parasitise, since they do no work on which another organism can free-ride. Which is why the next two lines of Swift’s poem, “And these have smaller fleas to bite ’em/And so proceed ad infinitum”, are wrong—and why, because HTVC010P itself can have no parasites, it probably really is the commonest organism on the planet.

First published in The Economist.  Also available in audio here.


  1. Zhao et al., Abundant SAR11 viruses in the ocean, Nature2013.
  2. Brown et al., Global biogeography of SAR11 marine bacteria, Mol Syst Biol2012.
  3. Swift, Poetry: A Rhapsody, 1733.

Image credit: Lynn Ketchum

Drug development: Teaching old pills new tricks

Exploding research costs and falling sales: there seems to be no cure for the pharma industry’s two big afflictions. But it may have found a way to both cut costs and open up new markets: repurposing drugs already approved for treatment of one disease or those that failed to gain approval in the late stages of development. Alas, this is not as easy as it sounds—mostly for legal reasons.

Finding new uses for old or failed drugs is on average 40% cheaper than inventing a new drug from scratch: it allows to skip the early stages of development. Since coming up with a new drug can cost more than $1 billion, such savings are nothing to sneeze at. Repurposing also trims the risk of failure because new drugs hit a dead end mostly during the early stages of development.

In 2007, a report in Nature, a science journal, counted 41 drugs that have found new uses. But there should be many more, experts say. This is why America’s National Institutes of Health, the country’s biggest government agency financing drug research, and the Medical Research Council, its British counterpart, each have launched new grant programmes. Worth $20m and £10m ($15m) respectively, they are meant to allow university researchers analyse failed drugs from big pharma firms such as Pfizer, AstraZeneca and Eli Lilly and see whether they can be repurposed.

Yet such schemes are not enough, as work by Grant Churchill, a researcher at Oxford University, shows. In a recent paper in Nature Communications, another science journal, he describes how he and his colleagues looked for a drug to treat bipolar disorder, which causes uncontrollable mood swings. Instead of developing a new compound, they tested a library of known ones and found that ebselen, a drug first developed to treat stroke, was a candidate. Their claim, based on animal tests, is that ebselen is as good as and much safer than lithium, currently considered the best treatment for bipolar disorder.

But this was where things hit a hurdle that is hard to overcome. Universities do not have the money to further develop promising drug candidates that need to be tested on a large scale. Expensive human trials are usually carried out by pharma firms, which own the patent for a drug and thus can hope to make their money back. But in the case of many repurposed drugs, like ebselen, the patent has expired. Filing for a new one, which is possible, is not of much help: patients could simply buy versions of the drug which are already available from other makers.

One way of solving this problem would be to change the patent system, for instance by extending the length of patent protection, but this could hamper innovation in other ways. A better solution, argues Benjamin Roin, a law professor at Harvard University, is to have regulators grant the drugmaker that has repurposed the drug some exclusivity and thus time to recover research costs: it is rare that a drug is used in the same form and the same dosage for two different diseases; regulators could wait a few years before they allow other firms to offer the drug for the new purpose. If old drugs can learn new tricks, regulators should do so, too.

First published on


  1. Singh et al., A safe lithium mimetic for bipolar disorder, Nature Communications2013.
  2. DiMasi et al., The price of innovation: new estimates of drug development costs, Journal of Health Education2003.
  3. Chong & Sullivan, New uses for old drugs, Nature2007.
  4. Roin, Unpatentable Drugs and the Standards of Patentability, Texas Law Review2009.

Image credit: The Economist

Cancer drugs: Refusing to die

Suicide is a part of life. Whenever any of the 100 trillion or so cells that make up the human body malfunction, which happens all the time even in healthy tissue, they are programmed to provoke their own death. The mechanism hinges on a protein called TRAIL, which is produced by the damaged cell and binds to receptors on its surface, causing inflammation. That is a signal for the immune system to sweep in and, through a process called apoptosis, break down the damaged cell and recycle its parts to feed healthy ones. If this self-destruct is subverted, however, the result is a tumour.

When TRAIL’s tumour-suppressing ability was first discovered in 1995 researchers hoped that by discriminating between cancer cells and healthy ones, TRAIL would do away with the debilitating side-effects associated with traditional treatments like radio- and chemotherapy. These are good at destroying tumours but also cause lots of collateral damage. Unfortunately, it turned out that simply injecting a synthetic version of the molecule into the patient’s body provoked only a limited immune response in a handful of cancers.

That, says Joshua Allen from the Pennsylvania State Cancer Institute, was because people assumed that cancer’s subversion of TRAIL consisted merely in halting the molecule’s production within the cell. It turns out, however, that cancerous cells also suppress their TRAIL receptors, so no amount of synthetic TRAIL sloshing about would ever be enough. What you need, Dr Allen reasoned, is something to reboot the TRAIL-producing pathway within cells as well as to unblock their TRAIL receptors. Only then would the immune system be spurred into action.

So he and his colleagues sifted through a library of molecules maintained by America’s National Cancer Institute and found a molecule, called TIC10, whose biochemistry seemed to fit the bill. When enough of these molecules accumulate in a cancer cell, they activate a protein called FOXO3a. This binds to DNA and flips on many biological pathways, including those involved in the TRAIL mechanism that lead to the immune-system alerting inflammation.

As Dr Allen and his colleagues report in Science Translational Medicine, tests in mice with brain tumours confirmed the biochemical hunch. Murine subject given TIC10 lived twice as long as those that received no treatment. The drug also worked for lymphoma, as well as breast, colon and lung cancers. And it did not seem to cause the wasting side-effects typically associated with chemotherapy, suggesting that it can indeed tell cancer cells from healthy ones. As an added bonus, TIC10 is small compared to TRAIL, and cheaper to concoct than the complex protein is.

Last year Dr Allen secured a $1.3m grant from Pennsylvania’s department of health to begin clinical trials. These will be carried out in collaboration with Oncoceutics, a drug company. Nine out of ten promising molecules which work in mice fail in humans, so “Cure for cancer” headlines must wait. If TIC10 does live up to its promise, though, it would make one killer app.

First published on

Image from here