Change in perspective for Natural Selection

[jim mcnamara]

https://phys.org/news/2020-01-fossil-upend-basic-tenet-evolutionary.html
Sandra Catania et al.
Evolutionary Persistence of DNA Methylation for Millions of Years after Ancient Loss of a De Novo Methyltransferase,
Cell (2020). DOI: 10.1016/j.cell.2019.12.012

Using pathogenic fungus species, the researchers have found that a methylated DNA segment has apparently been selected for. This is important because it demonstrates that Natural Selection may not always operate solely on genes in a DNA sequence. This operates on the presence of a methylation mark in just one place in the DNA sequence. In other words this is epigenetic selection at work. The researchers posit that this methylation mark is in a sequence that should not be altered by transposons -- i.e., jumping genes inserting themselves into a place in the DNA sequence that would kill the cell by rendering the gene inactive. The statistical models used indicate that this methylation mark has persisted for millions of years.

@Ygggdrasil may be aware of more aspects of the report. I find it interesting in that it may be a precursor for new studies on other epigenetic changes that persist due to selection pressure. ... if any exist.

 

[BillTre]

This seems to be an interesting finding:

• that some specific methylation pattern has been maintained millions of years, something that strongly implies a selection to maintain it in the genome for so long.

However, the claims that this is the leading edge of a change in perspective on how (or, on what) natural selection operates, seems inappropriate to me.
They basically claim that the methylation pattern they found is not a gene so its a new way of considering natural selection. Taking a "gene" as some molecularly well defined entity like a transcription unit makes non-transcriptional molecular feature (methylation in this case) not count as a gene.

However, older gene definitions (predating molecular biology) refer simply to an inheritied scorable trait (transcription and translation were unknown). Mapping and genetic tests of inheritance are key to identifying genes by this approach. What the map-able traits are, as long as they were reliably score-able, does not matter.
That both molecular and more abstract genetic features can all be put on the same genetic map indicates they are all encoded on the same piece of DNA.

This mapping relationship is the basis of how the genes (such as ether-a-go-go, a fly mutation that causes the flies to shake when they are etherized (common way to reversably knock out flies to look at they and sort them for setting up crosses, thus easy to find mutations), which might not have an obvious basis in genetics were mapped to locations in the genome (along the various DNA molecules).
The DNA sequence corresponding tot he map location and its molecular functions were then studied in detail to reveal details of its molecular functioning.

A whole series of different molecular functions that can be a changed (resulting in a sore-able phenotype) by changing how the cell's molecular functions, resulting in different phenotypes.
There are a huge number of different genetically encoded molecular processes, that can be mutated, to produce a distinct phenotype. Here are some:

• The basic: protein encoding gene, using processes of transcription, translation, possibly followed by processes for proper cellular localization
• genes transcribing an RNA product: tRNAs, rRNAs, various RNA transcribed sequences that seem to have control functions.
• DNA binding sites that are based on the DNA sequence. The basis of many described molecular control systems.
• mutations affecting gene splicing, which can make changes in the protein produced.
• changes in cell localization.
• Changes in teleomeres or centromeres that could have big lethal effects in chromosome structure or inheritance or if affecting only a small part of a large repetitive genomic sequence (structure)
• Sites of genome sequence with as yet unknown functions.

 

[Ygggdrasil]

I disagree here with @BillTre. Most research on evolution assumes that any new traits come about due to changes in the DNA sequence of an organism. Here, the authors provide data supporting an alternative way certain organisms can transmit information across generations that does not rely on changes to DNA sequences, but to changes in DNA methylation that are heritable across generations.

Now, we have known for a while that DNA methylation can be faithfully copied during cell division. We know this because the enzymes that methylate DNA come in two different flavors: 1)de novo methyltransferases that can install DNA methylation at specific sites, and 2) maintenance methyltransferases that are involved in copying DNA methylation during DNA replication. Classic molecular biology experiments have shown that removal of the de novo methyltransferases does not result in any change in DNA methylation; cells containing only the maintenance methyltransferases are perfectly capable of maintaining the correct levels of DNA methylation (at least over the short term). After removal of both types of methyltransferase, however, cells will eventually lose their methylation. Re-expressing the maintenance methyltransferases in cells that have lost methylation does nothing to restore DNA methylation; only re-expression of the de novo methyltransferases helps to restore the original levels of DNA methylation.

This paper shows that the experiments above have already been performed in nature by the yeast species Cryptococcus neoformans. The ancestor to this yeast contained a de novo methyltransferase and a maintenance methyltransferase, but aroudn 50-150 million years ago, the lineage leading to C. neoformans lost the de novo methyltransferase. The researchers show that the maintenance methyltransferase does a decent job of accurately copying DNA methylation, though after many generations they do see the occasional loss of some methylation sites and the occasional gain of some methylation sites. Because they see overall much more loss of DNA methylation than gain of DNA methylation, these observations would lead one to think that C. neoformans should have lost all of its DNA methylation in the 50-150 million years since the loss of its de novo methyltransferase. The fact that DNA methylation persists in C. neoformans strongly suggests that the DNA methylation is under evolutionary selection.

While the paper makes a strong case that the epigenetic inheritance of DNA methylation can contribute to natural selection, they don't actually demonstrate this. The researchers can create a strain of C. neoformans that completely lacks DNA methylation by temporarily disabling the yeast's maintenance methyltransferase. It would have been a nice experiment to show that two strains that have the same genetic sequence, but only differ in the presence or absence of DNA methylation, have different evolutionary fitnesses in laboratory evolution experiments.

Do these finding have relevance to evolution in the other kingdoms of life? The study suggests that is likely to be the case in yeasts and many other single celled organisms as we have a good understaning of how DNA methylation can be transmitted across generations in organisms that reproduce via cell division. Yeast and other single celled organisms also show other mechanisms for transgenerational epigenetic inheritance, for example, through prion proteins. The existence of transgenerational epigenetic inheritance in multicelular organisms, however, is less clear. In plants, we have maybe a dozen examples of epialleles (alleles where changes in DNA methylation in response to environmental stimuli can cause alterations in phenotypes that can persist over generation), though it is unclear whether these changes are stable enough over evolutionary timescales to contribute to evolution. Invertebrates (such as the nematode worm C. elegans) may show transgenerational epigenetic inheritance of small regulatory RNA molecules. In mammals, while there is limited observational evidence suggesting the possibility of transgenerational epigenetic inheritance, most people who study the subject are skeptical that DNA methylation can be inherited epigenetically because DNA methylation gets nearly completely erased and re-programmed during gamete formation and embryonic development.