This endosymbiosis, or merging of two cells, enabled the evolution of a highly stable and successful organism with the capacity to use energy from sunlight via photosynthesis.The Nature article is a tad clearer in terms of why this is a necessary inference. The conclusion reads, in part:
Further evolution led to photosynthetic organisms producing oxygen as a byproduct. The resulting oxygenation of Earth's atmosphere profoundly affected the evolution of life, leading to more complex organisms that consumed oxygen, which were the ancestors of modern oxygen-breathing creatures including humans. "Higher life would not have happened without this event," Lake said. "These are very important organisms. At the time these two early prokaryotes were evolving, there was no oxygen in the Earth's atmosphere. Humans could not live. No oxygen-breathing organisms could live."
We cannot say exactly how much time this symbiosis required, but it definitely did not happen in the past two billion years. We know this because the cyanobacterial double-membrane prokaryotes are responsible for producing the Earth’s oxygen atmosphere. This implies that their diversification, and hence that of the double-membrane clade, started before the rise of oxygen in the atmosphere, approximately 2.4 billion years ago48,49 or possibly earlier (2.7 billion years ago48,50). Because the double-membrane prokaryotes are descended from Actinobacteria and Clostridia, these two donor clades must trace their beginnings back to even earlier times. Thus there seem to be no obvious physical time constraints on how long the initial phase of the double-membrane symbiosis lasted. In the future, the double-membrane clade may provide a useful reference taxon for calibrating molecular clock studies.1The problem with a study like this is that folks like the DI will respond that there is no reason to infer evolution rather than special creation for something like this, despite clear evidence that the genome has evolved in other areas. This is not a smoking gun, it is a well-reasoned inference from the data.
1Lake, James E. (2009)Evidence for an early prokaryotic endosymbiosis. Nature. Vol. 460, 20 August, p. 967-971. doi:10.1038/nature08183