Recently, the media has been discussing the micro-evolution of a new antibiotic resistant strain of bacteria, dubbed the “NDM-1 superbug.” This seems to be a very sad case of one of those things that evolution is pretty good at doing — making small, incremental improvements upon an enzyme through a step-by step process. That, plus the tendency of bacteria to collect multiple antibiotic resistances, makes this gene a real problem. However, it by no means provides evidence for the ability of evolutionary processes to produce new functions within a cell.
The problem is that antibiotics are frequently used — and abused. Beta-lactamases, the enzymes that degrade penicillin and penicillin-like antibiotics (they are all characterized by a “beta-lactam” ring) were around before the anitbiotic era; microbial samples from the 1930s have been found that contained penicillin-resistant bacteria. The enormous populations of bacteria are then subjected to selection. Those bacteria that possessed the beta-lactamase genes survived, while those that didn’t died. This resulted in many bacteria becoming resistant to penicillin. The resistance did not have to arise in every type of bacteria; those that possessed beta-lactamases could transfer the genes to those that did not, resulting in further spread of penicillin resistance. We then developed variants of penicillin, and each variant initially was successful, but eventually new variants of these beta-lactamases also arose. These variants were the result of small changes in the DNA sequence that resulted in incremental improvements in these beta-lactamases. If a variant improved survival of a microbe in an environment in which antibiotic was present (such as within a person being treated with the antibiotic for an infection) then that variant survived and spread through the microbial population.
This process can be easily duplicated in the laboratory. If you put billions of bacteria on an agar plate containing 50 ug/ml ampicillin (a type of antibiotic), the weak beta lactamases in E. coli will be overwhelmed; you will get no survivors. This happens hundreds of times every day, since ampicllin is a very common antibiotic used in research laboratories. Evolution to high levels of ampicillin resistance is not seen because multiple steps would all be required for it to occur. But if you start with only 2 ug/ml ampicillin, you’ll get a few survivors; they will be mutants that have an improved beta lactamase. Continuing to do this will eventually result in a strain resistant to high levels of ampicillin. It speaks of the ability of incremental steps to improve a function, but nothing about evolution’s ability to create a function to begin with. For mutation and selection to be successful, a gene has to either be already functioning, or within 1 (at most 2) base changes from working. In my lab, we deliberately inactivated one of the genes for tryptophan biosynthesis by putting in two base changes to one of the genes needed to make tryptophan. Evolution has been helpless to re-evolve the ability to make tryptophan in this population of E. coli. In spite of only needing to produce two changes (with both being required), we have yet to see this evolutionary event occur; trillions of cells have been tested, and we’ve grown them for over 10,000 generations in an environment that readily selects for bacteria capable of synthesizing tryptophan.
The other part of this problem is our bad habit of introducing antibiotics one at a time, a policy that is tailor-made to produce resistant bacteria. Unfortunately, microbes also will transfer resistance genes from one bacteria to another, through transmissible plasmids. Some of these, by very standard and well-known pathways, will then collect antibiotic resistance genes. This has been the case with NDM-1. The microbes where it has been found are already resistant to other antibiotics; with the addition of NDM-1, they become “superbugs,” resistant to all of the standard antibiotics used. (The report that NDM-1 by itself makes a microbe resistant to multiple antibiotics isn’t true; it simply makes it resistant to essentially all beta-lactamases).
Can anything be done about this situation? Unfortunately, we become the victims of our own bad policies. If, early in the antibiotic era, a set of multiple antibiotics had been introduced, all with different ways of killing bacteria, the situation today would be quite different. Unfortunately, to repeat this today would require the discovery of two or three completely unique antibiotics and holding them back from use until they both could be introduced. Unfortunately, that will probably not happen, and we will have to live in a world where bacteria pose a larger risk to us than they did to our parents or grandparents.