ONE QUESTION in evolutionary biology not too many years ago was where new genes came from. One researcher who was influential in this area was Ohno, who described a variety of ways in which this could occur. One of the major ones is gene duplication followed by mutation, selection, and the eventual emergence of a new function for one of the gene copies. Certainly gene duplication unequivocally occurs fairly frequently–gene products that are required in large amounts frequently have multiple gene copies, and variations in gene copy numbers in humans leads to subpopulations with different phenotypes, such as faster metabolism of certain drugs.

Ohno originally thought that when a gene duplication occurs the second copy of the gene is under reduced selection since one copy is all that is needed by the organism. The second copy is neither beneficial nor detrimental–it is neutral. The problem is that while this neutrality will not apply a selective force upon the gene to retain the same sequence, it also renders the gene free of purifying selection to prevent deactivating mutations. In order to survive long enough to gain a new function the neutral gene must be able to drift to a high frequency in the population and survive long enough to acquire a new function by mutation. The problem is that under neutral selection this isn’t very likely to happen. The gene copy is more likely to remain if it is under positive selection, but then it is not going to be free to acquire new activity. So how does this occur?

A new paper in Proceedings of the National Academy of Sciences proposes a new model that involves positive selection at all stages. The authors call this the IAD model, for innovation, amplification, and divergence.

Before duplication, the gene has a primary function and a variety of secondary functions that are neither beneficial nor detrimental. When a change in the environment occurs, one trace activity that before was negligible now becomes valuable. The gene is under stronger positive selection, preserving it. This is the innovation step. Next, the gene is amplified. Chance duplications produce more copies of the gene. Organisms with more copies are under stronger positive selection because although the gene product isn’t very good at this now-important minor role, having multiple copies of the gene and being able to make a lot of the gene product is beneficial. The final step is divergence. Having multiple copies of the gene provides more chances for mutation to occur–if you attempt to shoot a single fish in a pond you’re likely to miss, but if you stick a bunch in a barrel your chances greatly improve. A chance mutation in one of the gene copies can increase its activity. This mutated copy will then be under even stronger selection than its non-mutated sisters, and will be selected for another round of amplification and divergence. The ultimate product is a gene that descended from the original gene but has a different function.

The authors propose this model as a solution for the puzzling Cairne phenomenon. In this experiment, Cairn discovered a strange occurrence when growing E. coli bacteria with a broken LacZ gene. This gene produces a protein that allows the bacteria to digest lactose. When the bacteria are given ample food the gene can be “repaired” by chance mutation that converts it back to its original sequence. This back-mutation occurs at a known rate, about 10^-8 per cell per division. However, when 10^8 bacteria are plated onto lactose-containing media, something strange occurs. Initially, there is no apparent growth, then after several days about a hundred different colonies appear with the LacZ gene reactivated by back-mutation.

This is puzzling because the back-mutation appears to occur without cell division and at a much higher rate than we would expect (if every cell went through one division, we would expect a single colony). Some people have said that it is as if the bacteria know what mutation they need to generate in order to reactivate the LacZ gene! For this reason this phenomenon has also been called adaptive mutation.

So how does the IAD model explain the Cairn phenomenon? The authors suggest that the broken LacZ gene still has slight lactase activity, so that would be the minor secondary activity in the IAD model. The cells can grow, but very slowly. This places strong selection on cells that have multiple copies of the LacZ gene, and strong selection on further amplification of the gene. With further cell division the result is cells with many copies of the gene, and each copy is vulnerable to mutation that might lead to reversion to the active gene. Once a revertant is generated, that cell can begin to grow rapidly and quickly overtakes the bacteria that still have the broken gene. Soon a visible colony appears.

In fact, the authors repeated the Cairn experiment and looked at the genomes of the bacteria involved (for the full story, see the open access article). Out of the hundred million cells plated, about ten thousand had a large section of their genome duplicated that contained copies of the genes. Another thousand had small segments containing the gene duplicated. Both populations were able to grow slowly on the lactose media.

The cells with the large duplication grew especially slowly because these were prone to having the duplication spliced out, and even if that didn’t happen copying all of the extra DNA imposed an additional burden on the starving cells. If a back-mutation occurred, the mutated cells would rapidly reproduce to form a colony with many back-mutated cells and a few of the original clone with the long repeat. As an alternative route, some cells with long repeats went through segmental deletions, converting the long repeat to short repeats.

Cells beginning with multiple short repeats were under strong selection for further amplification, with some cells ending up with as many as a hundred copies of the broken LacZ gene. These cells accumulated so many LacZ copies that they were able to grow fast enough to produce visible colonies. Eventually, a cell might back-revert, and the final colony could end up with some cells with a functioning LacZ gene and others with so many copies of the broken gene that they were able to grow reasonably competitively.

So the initially puzzling Cairn phenomenon turns out to be an illustration of how new functions emerge in gene copies, leading to the evolution of new genes.

Bergthorsson, Ulfar; Andersson, Dan I.; Roth, John R. “Ohno’s dilemma: Evolution of new genes under continuous selection.” Proceedings of the National Academy of Sciences, USA 2007, 104, 17004-17009.
Kugelbert, Elisabeth; Kofoid, Eric; Reams, Andrew B.; Andersson, Dan I.; Roth, John R. “Multiple pathways of selective gene amplification during adaptive mutation.” Proceedings of the National Academy of Sciences, USA 2006, 103, 17319-17324.