One of the defining characteristics of life is responding to stuff that the environment throws at it. At the level of cells, such responses are often accomplished by what we call signalling pathways. These are chains of interacting proteins that detect a stimulus (chemicals, voltage differences, pressure, light, etc.) on one end, and affect gene regulation or modify the activity of cellular components on the other end. One of the most common way of passing a message from one protein to another is phosphorylation – an enzyme called a kinase attaches phosphate groups to another protein, changing its behaviour. Kinases that phosphorylate proteins are unsurprisingly called protein kinases. (Their families are named after their favourite amino acid to phosphorylate, so we have tyrosine kinases, histidine kinases, etc.)
There are shitloads of protein kinasess. Legend has it that the acronym JAK, which officially refers to the “two-faced” Janus kinases, originally stood for “Just Another Kinase”. (I guess “Just Another Damned Kinase” didn’t abbreviate so well.) Every cell encounters many different stimuli, each of which may require a different response, and a diversity of signalling pathways can provide a more sophisticated ability to handle all conceivable circumstances. And sometimes, it’s best if such pathways keep to themselves.
Capra et al. (2012) investigate a curious property of a simple signalling pathway in bacteria. This pathway reacts to a shortage of phosphate, and consists only of the histidine kinase PhoR, and the regulatory protein it phosphorylates (PhoB). (Presumably there is still enough phosphate for the enzyme to work when the reaction kicks in…) The PhoR-PhoB pathway is found in all sorts of bacteria. In each major group, the handful of amino acids that determine the specificity of the interaction are strongly conserved. However, these “specificity residues” sometimes differ markedly between groups. Their conservation within groups suggests that changing them has dire consequences. So how and, most importantly, why were they changed anyway?
The study focused on three groups of bacteria: the alpha, beta and gamma classes of proteobacteria, which include familiar bugs like E. coli. In fact, E. coli (a gamma-proteobacterium) was one of the two main experimental species, the other one being the alpha-proteobacterium Caulobacter crescentus. The alpha bugs have an odd set of PhoR specificity residues compared to other proteobacteria, and the researchers hypothesised that this isn’t accidental. Instead, they thought, it might prevent PhoR from meddling with another signalling pathway that gamma-proteobacteria like E. coli lack.
The differences certainly aren’t without consequence. PhoR from E. coli can barely phosphorylate PhoB from alpha-proteobacteria, while it works quite happily on the same protein from other gamma-proteobacteria. It also does reasonably well on PhoB from the beta class, in accordance with the greater similarity of their specificity residues. C. crescentus PhoR only really works on PhoB from its own class.
How about that hypothesised other pathway? Well, when E. coli and C. crescentus PhoR are tested on the regulatory proteins from all similar pathways in C. crescentus, one particular molecule stands out. NtrX is the member of a pathway that has been duplicated in alpha- but not gamma-proteobacteria – and E. coli PhoR phosphorylates it! Is this duplication the reason why PhoR took a strange direction in this class of bacteria?
Multiple lines of evidence indicate that the researchers’ hunch was right. Replacing just one of the three altered specificity residues in C. crescentus PhoR to match the sequence in the other classes causes it to start interacting with NtrX at the expense of its normal function. C. crescentus with such “gamma-like” PhoR grows just as lousily in a phosphate-poor environment as C. crescentus with no PhoR at all, but only if NtrX is also present – delete the ntrx gene from the genome, and the disadvantage almost completely disappears. (NtrX isn’t disposable, though – under normal circumstances, it’s NtrX-deficient bacteria who perform badly.) A gamma-like PhoR can still interact normally with its correct target*, but it’ll simply ignore poor PhoB when NtrX is also around.
[*Which suggests to me that more than those few amino acids are involved in the PhoR-PhoB interaction, since a C. crescentus PhoR with specificity residues completely identical to those of E. coli still phosphorylates C. crescentus PhoB much better than PhoR from E. coli. However, those three do seem to be the main culprits in the NtrX mix-up.]
(In an interesting twist, it turns out that beta-proteobacteria also possess the NtrX pathway, but they tweaked NtrX instead of PhoR. The result is the same – each protein minds its own business, peace and prosperity and mad procreation ensue.)
The authors hypothesise that the Ntr pathways must have duplicated and diverged at a time when phosphate limitation didn’t come up often, given how much of a nuisance NtrX becomes to old-fashioned PhoR when phosphate is scarce. When phosphate did become a problem, the bugs were stuck with an already established NtrX pathway that they couldn’t just boot out of their genomes. Under those circumstances, any mutation getting NtrX out of PhoR’s way would have been the definition of beneficial.
Avoiding crosstalk seems to be a general feature of this kind of pathway: when you compare the specificity residues of all the signalling kinases and kinase targets from the same kind of bacterium, it’s as though they’re all doing their darnedest to be as different as possible. Capra et al. note that signalling pathways relying on a small set of amino acids to ensure specificity are very common in all life forms. They also often proliferate by gene duplication, which would make the crosstalk-avoidance issue a huge force in protein evolution. Good thing that so few mutations are needed, then – where would the complexity of the living world be if duplicated pathways all died, stuck between being redundant and screwing the organism?
Capra EJ et al. (2012) Adaptive mutations that prevent crosstalk enable the expansion of paralogous signaling protein families. Cell 150:222-232