I always assumed that epigenetic processes were defined as any means of storing information that is somehow associated (by physical interaction) with a piece of DNA but don't involve changes in the DNA sequence, in other words addition of functional groups to the DNA bases or modification of histones.

http://www.haloscan.com/comments...- epigenetic.php

I am the author of the Epigenetics article you so nicely refer to. But I am puzzled: what do you intend to mean by saying my observations are "hard line"? Do you think they are true, false, or (evidently) somewhere in between...?

hi mark,

thanks for stopping by. we're always excited to see the original author commenting at gnxp.

in a previous post (link), i discussed some of the requirements for referring to a phenomenon or mechanism as 'epigenetic'. i suggested that it was dangerous to take the term for a phenomenon of heritable gene regulation and tie it to a specific set of molecular mechanisms such as histone modification and DNA methylation. it is especially problematic if the term 'epigenetic mechanism' becomes untethered from the heritability requirement and people just start referring to chromatin modification as epigenetics.

i took it for granted though, that histone modification was a legitimate 'epigenetic mechanism'. my understanding was that histone modifications are necessary for establishing a 'state' of gene silencing and expression which is passed from mother to daughter cells. most of the reviews in the recent review issues of Cell and Nature Reviews Genetics seem to take this for granted as well.

if i remember correctly (sorry i don't have access to the paper just this moment), you make the point that DNA methylation isn't even independent of primary sequence, much less histone modifications. in my reading, 'epigenetic' is commonly used to refer to the mechanism by which genes are heritably silenced, so to the extent that DNA methylation and histone modifications carry this task out (and it seems that many in the field agree that they do) they are 'epigenetic modifications'.

in my commentary, i was drawing a line at heritability, arguing that an epigenetic modifications can't occur in a postmitotic cell. i described your commentary as taking a 'hard line' because i already felt like i was being a little uptight about usage and you went beyond me.

i think your observations about biology are most likely true. i'm certainly not expert enough to refute them if i wanted to. as far as observations about definitions, i don't think a definition can be true or false. language will evolve. epigenetics meant the interaction between genes and the environment at one point, right? now it means something different. i do think it is worth making sure we define terms that are being used in the literature so we can communicate clearly, which is why i really liked your essay, but i'll have to think some more about where exactly i think the line should be drawn for 'epigenetic'.

The classical definition of an “epigenetic” change – eg in a protein modification, in a state of expression of a gene and so on – that is self-maintaining – ie, is inherited in the absence of the signal (or initial event) that initiated the change. This change must NOT involve a mutation. Prions, for example, can change conformational states and the new state can be self-perpetuating. To take a gene expression example: in many cases a specific DNA-binding transcriptional activator, once expressed, activates transcription of its own gene, and thus that state of expression can be self-maintaining (and the change “epigenetic”). Histone modifications might or might not be involved in turning on and off genes in any given case, but there is no evidence that any such modifications are self-perpetuating. To be precise: where examined, all such modifications require the continuing action of a RECRUITER, in this case a specific DNA binding protein that brings the modifying enzyme to specific regions of DNA. ( DNA methylation can be self-perpetuating, but see my article for further discussion of the role of recruiters in this modification.) I did NOT say that histone modifications CANNOT be epigenetic – and I even give a vaguely plausible way in which some such modification might be self-perpetuating – I am just noting that there is no evidence that is happens, and good evidence that, where looked for, it does not. Moreover I am arguing that assuming that histone modifications ARE “epigenetic” (self-perpetuating) provides a “solution” to a “problem” that probably does not exist (see the article). Much of regulation in biology involves imparting “specificity” to enzymes – RNA polymerase must be directed to the proper gene; the ubyquitylating enzymes directed to a proper protein substrate; a histone modifying enzyme to the proper nucleosome (s), and so on. At the heart of these regulatory reactions lie RECRUITERS. Thus for example a transcriptional activator is analogous the the F-box subunit of an E3ligase…each recruits (using simple binding reactions) an enzyme to a specific substrate. This is all laid out in the short book Genes and Signals by Ptashne and Gann, CSH Press. This conversation would be easier if you had read it…



In the meantime I am inspired by your self description: “We are united by a reverence for the exploration of the heterodox in an empirical and analytic fashion drained of excessive emotive enthusiasm or revulsion. We are part of the remnant!” Why so wishy-washy on this issue of self-perpetuation of histone modifications just because it seems to have bamboozled so many?

If I seem wishy-washy it is because I know the limits of my knowledge. Epigenetics is not my main focus. I did some reading in the area to examine claims made by a memory / synaptic plasticity researcher. Being outside the field, I have two choices: 1) rapidly absorb a very large and complex literature or 2) rely on consensus.

I do not contend that there is evidence that histone modifications are self-perpetuating. Because of that, they do not currently fit the classical definition that you have described. Your points are well taken and will affect my thinking when I consider transcriptional regulation, but it wouldn't do for me to be swayed by one author's interpretation of the literature without reading the primary literature myself. So I'll add your book to my reading list and when I feel comfortable enough with the material maybe I'll be able to hold a firmer opinion.

Great response!