5 November 2007
Michael J Behe
The latest issue of the journal Trends in Ecology and Evolution (TREE) carries a tediously disdainful review (1) of The Edge which revisits the blunders of previous reviews while adding new ones. This is the second of a three part series concerning the review.
Like other Darwinian reviewers, the one for TREE questions the number I specify of 1 in 1020 for the origin of chloroquine resistance, citing a recent interesting paper on the development of CQR in India, which showed different strains of malaria with various numbers of mutations in their pfcrt genes. (4) Yet such field studies, while very valuable, can be fraught with uncertainty. For example, another recent paper (cited by the first) on CQR in Cambodia (5) reported data showing that the mutations K76T and A220S were not necessarily associated with CQR, which is inconsistent with the great majority of other reports. The most reliable data we have on the independent occurrence of resistance is that which surveys not just the mutations in the pfcrt resistance gene itself, but looks at surrounding DNA sequences for sequence heterogeneity. If drug resistance arose many times, easily and frequently, DNA surrounding the resistance gene would be expected to be as heterogeneous as other DNA regions in the genome. On the other hand, if resistance arose rarely with difficulty, heterogeneity would be suppressed around the resistance gene because of something called “hitchhiking” with the selected DNA. In several thorough studies, DNA heterogeneity was seen to be quite suppressed around pfcrt(6,7) (the chloroquine resistance gene), meaning that the resistance gene arose rarely and swept through a population. Thus since a person sick with malaria can carry 1012 parasites, and since resistance arises rarely, perhaps once for every tens of millions of persons, the number for resistance events of 1 in 1020 seems to be a very good approximation, as the renowned malaria researcher Nicholas White wrote. (8)
(The same sort of data is available for the pyrimethamine resistance gene. (9) The authors of that paper were puzzled: “Why did the triple-mutant allele not arise independently in Africa? Assuming a mutation rate of 10-9 per base per generation, we would expect 10 to 1000 independent origins of triple-mutant parasites in every infection (1010-12 parasites) containing double-mutantdhfr alleles. The implication is that complex compensatory mutations are required to restore parasite fitness.”)
Perhaps even more remarkable than the rarity of malaria resistance to chloroquine — and even more of an indication of the weakness of Darwinian forces — is the puny final result. Truly in this instance the Darwinian elephant labored mightily and brought forth a gnat. After a hundred billion billion chances, we end up with a few measly point mutations in pfcrt. These results from malaria are our best evidence by far of what Darwinian processes can accomplish when given a huge number of chances and strong selective pressure. Behold the result.
The TREE reviewer goes on to complain about my noting that no protein-protein binding sites evolved in malaria in an astronomical number of opportunities: “He apparently thinks that evolutionary theory says anything should evolve a new binding site in response to any arbitrary situation.” (1) The reviewer’s complaint begs a large question, however: when does evolutionary theory say that a protein-protein binding site should evolve? What non-arbitrary situation would cause that? In fact, evolutionary theory says nothing about specifics of what should or shouldn’t evolve. Therefore, we need to get our ideas about what should or shouldn’t evolve not from evolutionary theory, but from evolutionary data. And what we see in our best set of data from malaria is that no such protein sites evolved by Darwinian means in an astronomical number of opportunities. Furthermore, mutations in only one protein,pfcrt, were really able to do much in the face of chloroquine, showing that the number of proteins that it may be helpful to evolve in any given situation might be extremely small: one, maybe none. Ditto for pyrimethamine resistance.
If only one protein could evolve to help malaria avoid chloroquine poisoning, why should we think that a cell will luckily have a dozen or score of proteins that happen to be able to evolve to make a molecular machine?