Evolution - Gene Duplication

Gain of Function

• Multicopy genes in Chemostats - When cells are fed with xylose in place of glucose there is a gene amplification effect, multiple copies of glucose metabolising genes arise.
• Regulation of multicopy genes - What happens to regulation when genes are present in multiple copies?
• Difference in Activation Pattern - With multicopy genes, activation patterns will differ from gene to gene depending on their effectiveness, even though these genes are regulated by copies of the same mechanism.
• Preferential Gene Loss - Gene activation differences provide an opportunity for preferential loss of defective genes.

Multicopy genes in Chemostats

One of the interesting experiments concerned depriving cells which normally required glucose of glucose and providing them instead with another sugar, xylose.

Cells from the chemostat were analysed and found to have gained multiple copies of genes responsible for an early stage in glucose metabolism. These additional genes occured as tandem repeats, a section of DNA repeated a number of times over in sequence.

In this situation multiple copies were advantageous because the gene responsible for glucose break down was not 100% specific for glucose. The enzyme had a weak side specificity for xylose. By amplifying the gene, that is having multiple copies, enough of the enzyme was produced to metabolise xylose.

The repetition of a section of DNA is believed to occur through an error in copying DNA. A loop can form from a stretch of one strand of DNA and rather than copying this loop once as it should, DNA polymerase may traverse this loop two or more times.

Multiple copies also have an indirect advantage. They increase rapidity of subsequent evolution. With multiple copies:

• The genome makes more experiments with changes to that gene per generation.
• Mutations that damage one copy only of the duplicated gene are not lethal.
• It makes possible evolution of a new option. In this example it makes possible the evolution of a xylose metabolism option without destruction of a previous option, glucose metabolism.

Regulation of Multicopy Genes

If there are two copies of the gene for the enzyme, regulation should still act correctly. I say 'should' because as far as I know regulation of multicopy genes has not been studied experimentally. My expectations are based on what we know about the mechanism of feedback control of single regulated genes.

For duplicated genes with duplicated regulatory regions, one can draw an analogy with a house which has two independent central heating systems controlled by two thermostats. Both boilers will switch on together when it gets too cold. The house should stay in the right temperature range because both boilers will switch off when the right temperature is reached.

However, the boiler analogy hides the stochastic random nature of interactions at the molecular scale. There are random effects. If the two copies of the gene are labelled as A and B, sometimes A will be 'switched on' first, sometimes B.

• If A gets switched on first and rapidly restores the right levels in the cell, then B won't get switched on. Similarly...
• If B gets switched on first and rapidly restores the levels in the cell, then A won't get switched on.

Difference in Activation Pattern

Although not essential to the argument, the effect is amplified if cells operate a Most Recently Used (MRU) strategy. If genes which have recently been activated are more likely to be reactivated again then we could see a very marked difference in production of A and B.

This MRU mechanism is both:

• Likely from known biochemistry - inactive genes tend to become tightly coiled and so less accesible to reactivation.
• Advantageous - For it means that with multicopy genes the most appropriate variant of the gene for the substrates available will be manufactured in greatest amounts.

Preferential Gene Loss

Early on in evolution duplicated genes like A and B might well have been equally likely to be lost.

Suppose though that at some point in evolution a cell arose more likely to lose the least recently used gene of the gene pair A, B. Because of the regulatory effect described earlier losing the least recently used gene means it would be more likely to lose a defective copy and keep the working version.

In fact we do not even have to suppose that the genes A and B have an elaborate regulatory mechanism such as that for betagalactosidase. All we need is that A and B can be switched on independently and that the cell makes some distinction between them on the basis of which was most recently used. If B is defective, the cell is less likely to be in a state ready for cell division after using B than it is after using A. Thus even without elaborate machinery, by passing on the most recently used gene of a repeated pair the cell is more likely to be passing on the working copy.

Given the advantage of the mechanism, given that it does not posit extraordinary new properties for DNA and given the frequency of tandem gene reduplications, far from being a purely theoretical construct I posit that a mechanism like the one described has arisen and that it is a basic system to all single cell organisms.