Mike the Mad Biologist

Or maybe wild-ass speculation. As the data continue to come in about the E. coli outbreak in Germany caused by E. coli O104:H4 HUSEC041 (well, everything but the public health and epidemiological data which are a contradictory, incoherent mess), it appears that one of the things that has made this strain so dangerous is its ability to make Shiga-like toxin–a compound that stops cells from making proteins cells–the mechanism is similar to that of the biowarfare agent ricin (Note: there are several different Shiga-like toxins; for clarity, since it’s not germane to this post, I’ll lump them together and refer to them as Shiga-like toxin).

What makes Shiga-like toxin interesting (to the extent that pissing and shitting blood can be called “interesting”) is that the gene that encodes Shiga-like toxin, stx, is often carried on the genomes of phage (bacterial viruses). To explain what that means (and I promise, I’ll get to the E. coli disease argument), we have to briefly review the biology of phage. Phage have two different strategies. One is lysis, where the phage injects its DNA (or RNA) into the bacterium, takes over the cell, turns the cell into a phage factory, and then lyses (pops the cell, killing it) allowing the phage to be released into the environment, where they can find new, uninfected bacteria.

The other strategy is lysogeny. Here, the phage DNA integrates into the host’s genome*, essentially becoming another set of genes that the bacterium possesses. When the bacterium divides and replicates, so does the virus–as part of the bacterial genome. And as part of the genome, some genes, such as stx, will be expressed, just as other genes in the bacterial genome can be. But in those bacterial cells that are stressed (e.g., nutrient deprived or whacked with antibiotics), the phage will enter the lytic phage–it will turn on the genes that allow it to take over the bacterium and produce more phage. In other words, when times get bad, lysogenic phage jump ship.

Ok, enough of the phage biology primer.

The reason I went through all of that is that, whenever there’s an outbreak caused by E. coli that produce Shiga-like toxin (often called ‘STEC’), there’s often an implicit assumption that the bacterium ‘wants’ to do this–that Stx production is advantageous. But what the phage primer should indicate is that phage are bacterial parasites. We focus on the toxin production for obvious medical reasons, but, from the bacterial perspective (not that they think or anything), the toxin can be harmful to the E. coli carrying it, not in a biochemical sense, but in an ecological and evolutionary sense (no idea regarding the creationist sense…).

Here’s the problem for the hapless E. coli infected with Stx phage: you’ve just lost the ability to exist as a harmless commensal in six billion people, calves, and possibly other organisms. It’s not entirely clear to me that becoming an STEC is a good thing: yes, that new strain can sweep through a person, but is that really a better long-term strategy than being a commensal? Something else to consider is that EHEC strains infected with Stx, such as O157:H7, appear to have evolved their improved binding, colonization, and external survival abilities first, and then acquired Stx phage (two such phage in the case of E. coli O157:H7). In other words, the ‘pre-Stx’ O157:H7 was one of the few strains that wouldn’t rapidly become extinct when it was forced to become a human and calf pathogen.

Now, I won’t deny there might be situations where Stx phage are helpful. Under certain circumstances, the STEC strategy might be better than a commensal strategy. There’s also some evidence that STEC E. coli are more likely to survive protozoal predation. But that doesn’t mean E. coli lineages (the descendants of an infected bacterium) aren’t slowly stumbling towards extinction–it just can take a very long time in large populations. After all, if the effect of Stx phage across the range of environments the host E. coli will experience is mild, and if the Stx phage can infect other bacteria (and remember that it’s more likely to do so when the host isn’t doing well), then the phage could persist, perhaps even spread, even though in the long-run, it’s harmful to any particular E. coli host.

So maybe Stx phage aren’t just bad for us, but are also bad for E. coli. Admittedly, this is just speculation, but it would be great if someone tried to figure this out.

This perspective also has consequences for understanding the evolution of STEC-caused disease. I’ve been asked by a couple of journalists if I think the German outbreak strain, which carries an Stx phage, is a truly bizarre occurrence (a bacterial ‘Black Swan’) or if it’s something we’re going to encounter more and more. I don’t have a clue as to what will happen, but if the Stx phages evolve to infect new bacterial hosts more readily**, we could be in a lot of trouble.

*Some phage won’t enter into the genome directly, but will, instead, form a plasmid, which is a mini-chromosome. These phage ‘plasmids’ replicate when the cell does, and have mechanisms that kill the host bacterium if the ‘daughter’ cells don’t contain a phage plasmid.

**Mind you, this could occur either through increased infection rates, or through lessening any cost to the host.


  1. #2 Kevin
    June 9, 2011

    I love virulence theory.

    But actually working out whether one strategy or the other (commensal vs pathogen) is “better,” seems like a nonsense question. Pathogens and commensals have different niches, so unless you could show that a bug that was strictly commensal became a pathogen and then that lineage got wiped out, I imagine you would just be running around in semantic circles.

  2. #3 Mike the Mad Biologist
    June 9, 2011


    I really disagree. With the exception of the shigellic E. coli (where a reservoir host isn’t known), most E. coli are opportunistic pathogens (although some are more pathogenic than others). This includes most of the non-shigella Stx producers. I think it remains an open question whether the population size of these E. coli is decreasing.

    That’s what matters here–does a pathogenic life history strategy that affects the commensal strategy ultimately decrease absolute fitness in both habitats?

    I’m going to have to write a post about this…

  3. #4 Lab Rat
    June 10, 2011

    Really interesting post. I didn’t realise that it was the Shiga phage toxin that had got into this strain of E. coli.

    And while it might be a potentially ‘bad’ evolutionary strategy for the bacteria, you must admit it’s quite a good one for the phage gene in question, which can now be passed around in an entirely new bacterial environment.

    If the phage really is stopping the bacteria from spreading it would be interesting to see if an E. coli strain with a suppressed version of the toxin gene arises. Which would be useful from a public health perspective and fascinating for the microbiologists.