September 1, 2010 § Leave a comment
Earlier this week, I commented in passing about the ubiquity of phosphorylation as a way of regulating protein activity in eukaryotes. It’s been estimated that somewhere in the region of 1/3 to 1/2 of proteins are — at some time, under some circumstances — phosphorylated on serine, threonine or tyrosine, causing changes in protein behavior that are often dramatic. Bacteria and archaebacteria use this form of regulation much less, and understanding this difference may help us understand the split between the three domains of life. A recent collaboration between Matthias Mann, Ewan Birney and our own Jeremy Gunawardena looked at the evolution of Ser/Thr/Tyr phosphorylation across prokaryotes and eukaryotes, and also in the eukaryotic mitochondrion (Gnad et al. 2010 Evolutionary constraints of phosphorylation in eukaryotes, prokaryotes and mitochondria. Mol Cell Proteomics. PMID: 20688971).
To be able to compare the phosphoproteomes of different species, you need a set of comparable databases of phosphorylation events in those species. This means that all the phosphoproteomic studies have to be done the same way, otherwise your analysis may be confounded by varying sensitivities or other differences between the methods. Fortunately quite a few species have already been characterized using a uniform method, but the mitochondrion has not. Gnad et al. therefore purified mitochondria from 4 different mouse cell lines and from mouse primary brown adipocytes and characterized the phosphorylations seen in all five cell types. They found 174 phosphorylation sites on 74 mitochondrial proteins, about half of which were novel. Now they can compare these both with phosphorylation datasets from prokaryotes, and the mitochondrial subsets of the phosphorylation datasets from various eukaryotes.
As mitochondria are believed to have evolved from endosymbiotic bacteria, the first question you might ask is how similar are their phosphorylation sites to those of bacteria? Not very. Although the proteins found in mitochondria have relatives in bacteria (homologs for 35% of mitochondrial proteins are found in prokaryotes, as compared to 13% of non-mitochondrial proteins), there is almost no overlap between the phosphorylation sites detected in mitochondria and those detected in the three different prokaryotes for which Gnad et al. have data.
Note, please, the word “detected” in that sentence (twice). One big problem with phosphoproteomics is that you don’t know what you don’t know. A protein might be low abundance in one species but not another, making it hard to detect the phosphorylated peptides in that species; or the protein might only be phosphorylated under conditions you didn’t use. Or a sequence change that has nothing to do with phosphorylation might reduce the “flyability” of the peptide. When we detect a phosphorylated peptide, we can be pretty sure that the relevant site on that peptide is phosphorylated, at least sometimes. When we don’t, all we know is that we didn’t detect it. But because Gnad et al. chose to analyze datasets that were all produced using the same method, it is at least possible to make fairly confident statements about the overall amount of Ser/Thr/Tyr phosphorylation: bacteria do very little of it, and eukaryotes do a lot. And mitochondria look much more like bacteria than like eukaryotes: of roughly 1100 proteins that are believed to be mitochondrial, 74 have phosphorylation sites. This is not all that different to what’s seen in E. coli, where 100 of 4,000 proteins are phosphorylated. And yet the identity of the proteins phosphorylated is not conserved.
There are only two proteins, the elongation factors EF-G and EF-Ts, that are phosphorylated both in mitochondria and in all three of the bacteria studied. Even for those proteins the phosphorylation sites are not the same between mitochondria and bacteria. The only protein in which a detected phosphorylation site was apparently conserved was a protein that rejoices in the name of “protein similar to nucleoside diphosphate kinase”, which is phosphorylated on Thr123 in mitochondria and Ser123 in B. subtilis. While one can’t rule out the possibility that there are undetected similarities in phosphorylation sites between mitochondria and bacteria, the findings are consistent with the idea that most of the phosphoregulation of mitochondrial processes evolved after the endosymbiotic event.
What about phosphoregulation in eukaryotes, is that conserved? Yes, to some extent, but there is an interesting break point. Let’s look at serines that are phosphorylated in yeast and see if they’re conserved in human (i.e. whether the sequence of the homolog still has a serine at that position); as a comparison, we’ll use the serines that aren’t (known to be) phosphorylated in yeast. In both cases, you see about 13% conservation, so there is no evidence that the phosphosites are under any special kind of selection. But comparing flies to humans you get 20% of phosphoserines conserved, as opposed to 16% of non-phosphorylated serines. Even in chimpanzee, where a whopping 92% of serines are conserved in human, phosphoserines are even more conserved at 97%. This is consistent with other evidence that suggests that many of the kinases that now exist in humans evolved after the lineage that led to humans diverged from the lineage that led to modern-day yeast. [Nematodes are outliers as well, but probably for a different reason.] Phosphosites on mitochondrial proteins tend to be more conserved than non-phosphorylated sites, too, suggesting that the job of regulating mitochondrial behavior has similarities across all species.
Stepping back from the details of these data, we see that the phosphoproteome has two major discontinuities. One is between bacteria and eukaryotes: based on current techniques, fewer than 5% of bacterial proteins, but 35-50% of eukaryotic proteins, are phosphorylated on Ser/Thr/Tyr. Mitochondria look a lot like prokaryotes from this point of view, even though they live in the eukaryotic cell and have access to its enormous repertoire of kinases. The other discontinuity is between yeast and other eukaryotes: although yeast isn’t so different from other eukaryotes in terms of the frequency of phosphorylation, it seems to use phosphorylation for different purposes. Not a shock, perhaps, since a unicellular organism has less need for several of the processes regulated by phosphorylation (such as differentiation), but it emphasizes the fact that comparison of phosphoproteomes can give you a new way to ask questions about how regulatory processes evolved.
Gnad F, Forner F, Zielinska DF, Birney E, Gunawardena J, & Mann M (2010). Evolutionary constraints of phosphorylation in eukaryotes, prokaryotes and mitochondria. Molecular & Cellular Proteomics : MCP PMID: 20688971