Horizontal gene transfer (HGT) - direct transfer of a gene from one organism to another - is rampant within bacteria, so they may gain new function this way. However, that does not explain how the gene came to be in the first place.
Neofunctionalization: If the function is carried out by the original, the copy is free to evolve a new function by point mutations (etc.). However, such copies are much more likely to degrade by those mutations and lose the original function, thereby becoming a pseudogene.
If the gene is pleiotropic, i.e. it has more than one function (expressed in more than one trait or cell-type or at different times), then the gene and its new copy can turn off gene expression differentially such that they share the set of functions. However this doesn't allow either much chance for evolving new function by mutation.
So what to do?
Näsvall et al. gave me this present for my 40th: Real-Time Evolution of New Genes by Innovation, Amplification, and Divergence.
They describe a new model/mechanism by which duplicated genes can retain the selection pressure to not succumb to deleterious mutations. They call it the innovation-amplification-divergene model (IAD).
IAD works like this: A gene initially has one function only (A). Then some genetic changes makes it also have a new function, b, which at first is not of too great importance. Then some environmental change favors the gene variants with the minor b-function (the innovation stage). This is then followed by duplication of the gene, such that there are now more than one copy that carries out A and b (the amplification stage). At this stage there is selection for more b, and at some point genetic changes in one of the copies results in a gene that is better at the new function, B. At this point, selection for the genes that do both A and b is relaxed, because the new gene (blue) carries out the new function. The original gene then loses the b function, and we are left with two distinct genes. Viola!In other words, the green gene first becomes pleiotropic, is copied, followed by divergence, and then loss of pleiotropy. (How they could fail to mention pleiotropy in the article is beyond me.) The crucial feature is that at no point is the gene or any of the copes under no selection; there is always selection for them to be retained, so gene loss never occurs (pseudogenes are not created).
The researchers then look at a preexisting parental gene in Salmonella enterica that has low levels of two distinct activities that allows them to grow without the amino acids histidine and tryptophan, respectively.
Multiple evolutionary trajectories recovered through IAD. The x-axis indicates the HisA activity (assayed as growth rate in minimal glycerol medium with added tryptophan); the y axis indicates the TrpF activity (assayed as growth rate in minimal glycerol medium with added histidine). (A) Evolution of specialist enzymes (yellow) in which one activity is improved at the expense of the other. (B) Evolution of specialist enzymes (yellow) after initial evolution of a generalist enzyme (blue).
The figures here show how the generalist gene evolved to become specialists genes with increased function, doing better without both amino acids.
This is a model that explains how a gene with two functions can evolve to become two genes with distinct function under continued selection. It is this last part about selection that makes it novel, but it relies on the idea that the original gene had already evolved two distinct function - that it was pleiotropic.
Pleiotropy comes from the Greek πλείων pleion, meaning "more", and τρέπειν trepein, meaning "to turn, to convert". It designates the occurrence of a single gene affecting multiple traits, and is a hugely important concept in evolutionary biology.
