Abstract
Endosymbiosis is an idea that provided a remarkable amount of explanatory power about the origins of eukaryotic organelles. But it also promoted a number of assumptions that have also been influential, but are less well-examined. Here we look at two of these to see whether or not they fit current evidence. The assumption we first address is that endosymbiotic relationships such as nutritional symbioses and eukaryotic organelles are mutualisms. We argue instead that they are more one-sided associations that can be regarded as context-dependent power struggles like any other ecological interaction. The second assumption is that during endosymbiotic interactions (such as the origin of organelles), the host genomes will acquire a great many genes from endosymbionts that assume functions in host systems (as opposed to the well-documented genes whose products are simply targeted back to the endosymbiont or organelle). The idea that these genes exist in large numbers has been influential in a number of hypotheses about organelle evolution and distribution, but in the most carefully-examined systems no such mass migration of genes is evident. Overall, we argue that both the nature and impact of endosymbiosis need to be constantly re-evaluated to fully understand what roles it really plays in both cell biology and evolution.