Response to Dagan and Martin
Dear Sir, We read with interest the letter penned by Dagan and Martin (1) in response to our recent Problems and Paradigms article on the origin of eukaryotes. (2) Among the hypotheses that we critically evaluate-and disfavour-stands their own, (3) which proposes, without precedent or mechanism, that the ancestor to mitochondria was 'engulfed' by an archaeon. It therefore comes as little surprise that Martin and Dagan are opposed to our conclusion that, on current evidence, the only model that explains the genomic and phylogenetic data suggests phagocytosis evolved prior to the incorporation of the mitochondrion. (4) Their first criticism is that we deny that horizontal gene transfer (HGT) occurs and that, in doing so, we have used an inappropriate framework for evaluating the various hypotheses for the origin of the eukaryote cell. First, a minor clarification. We did not cite the Thermotoga example (5) to suggest that there is no HGT in prokaryotes, we cite it to point out that genomic fusion signals are indeed observed. However, such fusion does not obviously lead to formation of new domains (as proposed in model i in our paper)-if that was not clear, we apologise. We are perfectly at ease with considering significant levels of HGT in microbial evolution and have never argued otherwise.
The real issue is thus, does HGT invalidate our argument that a phylogenetic test should be employed in order to test between different theories? Our point is that, even in the presence of appreciable rates of HGT (for some, but demonstrably not all, genes), there is nevertheless consensus that a phylogenetic signal from a significant number of genes for the bacterial origin of mitochondria is seen. It is important to stress here that where reliable data are available the result is unequivocal: eukaryote genes with bacterial nearest neighbours group within bacteria, a sister group relationship is not observed. (6,7) For any given eukaryotic gene, it is conceivable that it may have independently entered the eukaryote lineage via a transfer event from some other bacterial lineage, thereby complicating the evolutionary history of the eukaryotes. (6) However, no one argues that the a-proteobacterial origin of mitochondria is obscured because 'HGT is so extensive that the evolutionary history of the host lineage is obscured'. If that really were so, we might as well all go home because we could not use phylogenetics to examine the evolutionary affinity between mitochondria and a-proteobacteria. We are sure that Dagan and Martin agree that this is as preposterous as the other extreme, which they seem intent on pinning on us-the 'genome of Eden'.
Yes, HGT complicates the picture, but it is not so extreme that it invalidates the use of phylogenetics. Nor is our original