Eating can be bad for your health. Oh, and don’t forget the phages.

Sure, we have obesity problems in this country, but we also have more direct food safety problems. Summer has brought with it news of the bungled tomato-Salmonella affair, and now, from the Midwest, contaminated beef.

One of our local supermarket chains has been forced to recall hamburger meat because of over a dozen cases of E. coli-related disease. These cases have occurred over a wide area, and the bacteria are genetically linked, indicating a likely common source.

erv handed me some legitimate criticism regarding my very brief post on diarrhea. It was certainly not my intention to give you the more comprehensive talk that I give my students and residents, but even if I had I would likely have left out the fascinating role played by bacteriophages.

E. coli can be a nasty bug. We all have plenty of E. coli living in our colon peacefully, but some strains aren’t such good guests. E. coli can lead to several different types of intestinal disease, depending on the strain.

E. coli can cause human disease by directly invading the colon wall, causing pain, fever, and bloody diarrhea. Some strains simply secrete a toxin causing your gut wall to gush fluid into the lumen leading to a cholera-like illness.

Then there’s O157:H7 (and occasionally other strains, known collectively as EHEC—clearly needs a cooler name). This little guy acquired the ability to produce a Shiga-like toxin, which causes no end of havoc. Before I get into the details that will get me in trouble with erv, let me tell you a little about hemolytic-uremic syndrome (HUS). This illness has received a decent amount of press over the years, but it’s a pretty complicated disease. First of all, HUS very closely resembles another disease called TTP. Despite the outward similarities, TTP and HUS have some important differences.

Both illnesses are characterized by a combination of fever, a certain type of anemia, kidney failure, confusion, and low platelet count. Confusion is more common with TTP; HUS more commonly affects children; and HUS is often associated with diarrhea caused by E. coli. Still it can be very hard to distinguish the two diseases, which is unfortunate as treatment is radically different. TTP is (usually) an autoimmune disease and is treated as such. HUS is not.

HUS affects only a small number of people with EHEC disease, and most recover, but even those who do survive often require prolonged hospitalization, and can be left with some serious medical problems.

As an internist, these two diseases are fascinating for a number of reasons: they represent a diagnostic challenge; they are associated with interesting laboratory and microscopic findings; they teach you a lot about human biology; and, at least in the case of TTP, we have rational and effective ways of treating the illness.

But one of the coolest facts about HUS isn’t medical but microbiologic (and here’s where I’ll get into trouble from erv).

Most E. coli does not produce the toxin (usually called “Shiga toxin”) that causes HUS. In the strains that do produce the toxin, the toxin is not encoded by E. coli‘s intrinsic DNA but by a virus inhabiting the bacteria. Let me ‘splain—there is too much, let me sum up.

A human eats a rare burger. That meat is contaminated by a bacteria (E. coli), and the human becomes infected with the bacteria. The bacteria is itself hosting a viral “infection”. This virus (or “phage”), rather than harming the bacteria, sits with its DNA peacefully integrated into the bacterial genome, allowing itself to be replicated along with the bacteria, only reproducing in large numbers and harming the bacteria when the bacterium is under stress. The viral DNA, in addition to encoding any proteins necessary to the virus happens to encode for Shiga toxin. This is most unfortunate, as the phage infecting the bacteria infecting the human now allows for the production of the toxin responsible for HUS.

I’ll leave it to others to explain how the Shiga-toxin genes became part of the phage, but perhaps a phage infected a Shigella bacterium (one that normally produces the toxin), and when it was snipped out during replication, it took an extra gene with it. Later, when infecting an E. coli, this gene was integrated into the newest host’s genome, and voila, nasty HUS-casing strain.

If you made it this far, you won’t mind me summing up a bit. We started with a food recall prompted by contaminated meat. This meat has caused infections in humans. These infections can (but apparently haven’t) lead to a serious disease. The disease illustrates some important biological principles, such as a type of gene transfer (transduction).

Most lay-people haven’t heard of phages, and most physicians probably don’t think about them very often, but phages represent an enormously important part of biology and medicine. They cause disease, they can be a vehicle for certain therapies, and ultimately, they represent a large part of the biosphere’s genetic information. Pretty cool, eh?

8 thoughts on “Eating can be bad for your health. Oh, and don’t forget the phages.”

  1. Not to diminish the Awesome Power of Phage ™, but I seemed to remember the description of E. coli acquiring Shiga toxin via plasmids (conjugation) rather than phages (transduction). Could it be that different strains of EHEC acquired it at different times in different fashions?
    Do you have a citation verifying O157 getting it via phage?

    Just making sure the HGT stories are straight.

  2. I had the same recollection, and was corrected when I did the research for the piece.

    See, for example,

    Interesting note (to me at least) was the one of the earliest refs I found. It’s from Nature, 1984, by O’Brien. She characterized the prophage homology (correct word?) in the pre-PCR era. She basically used RFLP and protein characterization to compare prophage genes in E coli and Shigella—no real DNA sequencing was available. Very cool stuff.

  3. Becca,

    I thought I remembered the same thing. But after a bit of searching, I realized I was misremembering the hemolysin gene, which is plasmid-encoded and is another virulence factor found in EHEC (link).

  4. Cool, thanks! And determining homology without being able to sequence? I think that just about trumps most of the old-skool macho points stories I’ve heard lately.

  5. Back in the dark ages, when I took my first molecular genetics class, I had a professor who insisted on teaching old-school methodology for virtually everything. If nothing else, it gives you an appreciation for how easy things are these days (who needs a maxi-prep kit when you can just pour a CsCl gradient, add enough ethidium bromide to mutate a horse, and ultracentrifuge it for hours?)

    Anyway, I seem to recall that you can get a rough estimate of DNA homology by annealing non-identical strands and then raising the temperature while measuring A260. Then there’s some equation that correlates melting point differences with homology. Or am I remembering that incorrectly?

Comments are closed.