A test of the coordinated expression hypothesis for the origin and maintenance of the GAL cluster in yeast.
Lang GI and Botstein D. 2011. PLoS ONE. Sep 22; 10.1371/journal.pone.0025290. [PDF]
[Motivation]
What is the evolutionary force that favors the clustering of genes functioning in the same pathway?
[Hypothesis]
- coordinated expression, perhaps to facilitate “channeling” of toxic intermediates in metabolic pathways
- to utilize the genetic linkage to preserve co-adapted genes.
- reflect a history with frequent horizontal transfers.
- It seems to me that the results are quite inconclusive, despite the question being highly motivating. Their test for the first hypothesis, namely the clustering is favored because it contributes to coordinated expression, is achieved by tagging two genes in the pathway by different fluorescence proteins. Importantly, in a diploid cell, they only tag one copy of the gene and as a result, there are two ways of tagging: tagging the two genes on the same chromosome or two that are on different chromosomes. The hypothesis, therefore, can be tested by reading the expression correlation between the two genes tagged either in cis (same chromosome) or trans (on different chromosome).
While this is all good, I think there is a possibility that one or both of the genes might harbor some self-regulating abilities. Let’s look at the trans case. Although the two genes whose expression are being monitored are on different chromosomes, each of them does have another copy non-tagged. If their expression are partially self-regulated, then having one untagged copy in cis could potentially coordinate the expression of the two copies. While this may sound absurd, given that no explicit mechanism for coordinated expression is being offered and tested here, we could envision any possible mechanism.
To circumvent this problem, I would like to see this experiment redesigned with some elements borrowed from the second experiment, namely knocking out one copy (the untagged) and leave just one. This should eliminate the self-regulation concern. Given that they had done the second experiment, the additional experiments should be feasible.
- Concerning the test for the second hypothesis, my concern is that they don’t see a fitness difference not because there is not or it is too small for them to see in their experiments, but that the experiment design has fault that prevents them from seeing the effect. My reasoning is that the fitness difference, if any, should result from incompatibility between gene pairs that have each co-adapted with their native partners and when brought together behave less fit than either wildtype. The assumption here is that there has been co-evolution. In this case, I think it is crucial to pick two strains that are reasonably diverged — either two closely related species that can still make hybrids, or two distant strains within a species that harbor different ecological niches with regard to the function of the GAL pathway. In absence of that (the authors mentioned that all strains used in this experiment are derived from the same strain), I don’t see how this experiment can test the hypothesis.
I have to defend for this experimental design, because we did something virtually identical (although a bit slower). In the first experiment, trans pattern was used to mimic if two genes were unlinked. The second experiment was used to quantify the fitness effect of non-coordinated-expression. The second experiments did not provide evidence for the Hypothesis II. Rather, it provided evidence against hypothesis I (the sensitivity of the experiment becomes a problem though). I believe both experiments were designed correctly in general.
But the paper itself was not very well written, so that it is not so easy to understand the logic underlining the experimental design.
Comment by windwhere — October 11, 2011 @ 1:10 pm |
Qian, I actually realized that the problems I raised about using trans to mimic unlinked loci is a bit far-fetched. In fact, considering the hypothesis it is designed to test, namely that clustering is maintained in order to facilitate coordinated expression, the trans configuration is an appropriate and in fact quite genius way to prove or disapprove it. The self-regulation I’m concerned about doesn’t depend on, neither is it required for the clustering to cause coordinated expression. It could act as a location-independent mechanism for synchronizing the expression from trans-locus, such as two copies on sister chromosomes.
About test II, I still feel their design is not appropriate for the hypothesis they want to test. I think their hypothesis is that clustering offers fitness advantage by keeping co-adapted genes together. Then the key factor is genetic differentiation between orthologous genes in different alleles / haplotypes, because this will lead to incompatibility when one substitutes a cis-located gene in a cluster with a trans ortholog. But it seems to me that the two strains of yeast they used were identical or nearly identical, therefore no genetic differences. Then how could you still test the fitness advantage hypothesis?
Comment by hebin — October 12, 2011 @ 9:26 am |