There has been much discussion, here and elsewhere, on ‘epigenetics’, broadly understood as the control of gene expression. People who cling to ‘classical’ models are portrayed, by revolutionaries and their cheerleaders, as dinosaurs standing in the way of progress.
I could perhaps explain, to any interested bystander, my own rationale for my position, since I’ve requested that of others.
Those of us taught molecular biology between the 1960’s and the noughties assimilated a model of gene regulation that started with the classic work of Jacob and Monod on the lac operon.
The structural genes – the parts coding for actual protein, by transcription into mRNA and translation – are flanked by regions to which transcription factors bind. Transcription proceeds in one direction only (a DNA strand has 2 directions due to the asymmetry of the ribose links). ‘Upstream’ of the protein regions, factors (proteins or RNAs) bind and block or promote the activity of transcription enzymes. In the case of lac, binding represses. So if there is no lactose, the enzymes are not produced. Lactose causes the disassociation of the repressor and the enzymes for its metabolism are produced. Neat, huh?
Now, the classical operon is a prokaryotic feature. Prokaryotes have no histones, and they appear to lack DNA methylation mechanisms, so there can be no involvement of these extra layers. Analogies to ‘memory’ are rather strained. The bound lac repressor is a ‘memory’ the way someone hanging on to one’s ankles is a memory of that person!
Modern eukaryotes spool their genomes on histones. This causes an extra problem/opportunity, beyond basic DNA management, in getting transcription turned on or off. If there is no transcription to do, the DNA is tightly wound on the spools, which involves their methylation (methyl groups are hydrophobic, so cause tightening in the presence of water). If transcription is required, conversely, the methyl groups need removing, and further relaxation can be provided by acetylation. And what causes these changes? Ultimately, it’s those transcription factors. Extra elements are introduced into the cascade, but basically the same kind of promoter/repressor system as in prokaryotes initiates the extra work that needs to be done to expose the reading frame to transcription.
DNA methylation again causes the helix to tighten. One sees something very similar in the thymine-uridine distinction between DNA and RNA. Thymine has exactly the same molecular relation to uridine as methylcytosine has to cytosine, the difference being that thymine’s methyl group is permanent and not neighbour-dependent. A-T base pairs in double-stranded DNA cause the helix to be more tightly wound than dsRNA; methylated cytosines means that C-G pairs reinforce this effect, but reversibly.
These mechanisms provide an additional means of control compared to prokarotes – but the much stronger claim is that they are a separate means of control – that control is by such modification instead of transcription factor binding.
Now if, parallel to all this, one buys the notion that evolution happens, and that eukaryotes derive from prokaryotes, then clearly the basic system would be expected to be one of promoter/repressor binding, subsequently amended to extend to histone modifications and the additional, novel (and by no means universal) mechanism of DNA methylation. It seems unlikely that a completely separate system of regulation would arise driven by histone changes, when histone changes must occur anyway.
If, conversely, one thinks evolution does not happen, and one is further seduced by superficial resemblance and analogy, the idea that the histone and DNA changes constitute a separate level of control seems to hold considerable appeal. There’s no problem with a Designer choosing one mechanism in prokaryotes and another in euks. But the inconvenient fact is that the cascade, as far as has been elucidated, starts with TFs. So had Mukherjee given more (and appropriate) weight to transcription factors, he would pretty much have been writing an article on 60 year old molecular biology. Instead, he’s describing a revolution that hasn’t happened yet.
Histone/DNA methylation codes are not impossible, they just seem very unlikely to people with a particular background. One can invoke old saws about Kuhn, and revolutions proceeding death by death, but one also needs hard data.
From the OP:
There are many definitions of epigenetics, this one from Nature is close to yours, and it is one I like the most.
I actually like that definition.
I used Felsenfeld’s definition on exams, but I prefer the one which the journal Nature uses.
http://www.nature.com/subjects/epigenetics
That said, however, note the emphasis on “nuclear organization, DNA methylation, histone modification and RNA transcription” not transcription factors (the dinosaur view).
Felsenfeld’s definition gives much stronger emphasis on development and evolution:
The reason I like the definition from Nature is that it doesn’t get into quasi-philosophical issues of when a biochemical change affects development or evolution — that tends to go into speculative territory. Changes in gene expression are much more measurable and above reproach, and the mechanisms involved often do affect development and sometimes evolution.
To quote Shakespeare, “a rose is a rose by any other name.” I’m really averse to over emphasis on definitions and names and nomenclature. Histone modifications are important to gene expression and development, I see little to gain in arguments over definitions and amorphous concepts and trying to reduce biology to a miniscule set of fundamental concepts.
Complex systems have a large set of fundamental components, many of them life critical. Pointless to argue which one has primacy. The life of a chicken was critically dependent on egg shells, does that therefore make egg shells primacy in defining chickenhood? Pointless arguments. Same goes over arguments as to which epigenetic mechanism has primacy.
In multicellular Eukaryotes, histone modifications, DNA methyl marks, transcription factors — all are life critical. Pointless to argue that one has primacy over the other, HOWEVER, histone modifications and the supporting machinery are unique to eukaryotes and especially multicellular eukaryotes. I think it’s a unwise to try to explain multicellular gene regulation in terms of bacterial gene regulation — it’s a force fit that doesn’t work.
And back to the question of what initiates cascade of epigenetic mechanisms. Well if an organism loses an organ or suffers injury, epigenetic changes have to happen. No one on the planet really understands how the healing cascade works exactly. I don’t see how transcription factors alone can sense injury and then recruit healing response. I think it must be an integrated decentralized system as a matter of principle to effect self-healing. But that system in all its details is still a black box.
Felsenfeld alluded to it:
stcordova,
I have two main questions here. First, why do you favor a definition of “epigenetic” that specifically excludes (or at least ignores) transcription factors? Do you think they aren’t important in control of gene expression?
Second, why do you think epigenetic changes are relevant to evolution? Do you know of a case in which evolution occurred without genetic change?
stcordova,
It would be better if you sided with it because that was where the data pointed, don’t you think?
I don’t think that’s quite right. It is more that evolutionists have an expectation of the possibility of continuity. There does not have to be continuity. It just so happens, on investigation, that there is. There is no fundamental problem for evolutionary theory if a planet had two completely different life forms with two completely independent origins, and two completely different ways of doing everything.
But yes, people with a decent biological background who do not experience a kind of retinal detachment with regard to evolutionary theory, can see why molecular factors interacting with DNA would be expected to be conserved as a mechanism. Histones both get in the way and provide an additional level of control, so those factors have to get past it. But there’s no point histone opening up if there is nothing to let through. Histones don’t have the variation to permit ‘recruitment’ of a transcription factor – especially when the same TF has a genome-wide effect. Some TFs are involved in 15% of the genome. Clearly, it makes more sense to suppose that the TFs have a key than that histones are a complex system of intelligent doors.
True. I don’t want a derail on sex, but I agree that bacterial mechanisms that people call ‘sex’ are not really the same thing. There is some continuity because they both incorporate DSB repair mechanisms (whose sequence homology indicates relationship), but they are mechanistically and functionally unrelated IMO.
Which is of interest. I argue for an expectation of continuity in the matter of regulation, where the driver is some kind of molecular factor that ultimately needs to get to the DNA, by hook or by crook. But I argue against continuity in the matter of sex.
So perhaps you err in your rationalisation of why ‘evolutionists’ argue against histone primacy. It’s not because it goes against evolutionary theory – if it did, then so would sex, or any other prokaryote/eukaryote discontinuity you care to name.
Of course, when we look at mitochondria and chloroplasts, we see a very interesting continuity. At plastid level, eukaryotes are prokaryotes – including their mechanisms of gene regulation, and reproduction.
I’m still kind of waiting on Sal to comment on how epigenetic changes wormhole their way into to heritable traits.
It would be nice to see Sal comment on the resemblance of the Thompson Tone Discriminator to his network diagrams, and what this might say about the evolvability of networks.
And maybe a bit on royal jelly, not cribbed from quack medical sites.
Petrushka,
I just saw your comment prior to loggin in. I don’t see them when I’m logged in now since you are on my ignore list. So your questions are likely not be get answers. Nothing personal, Allan Miller has higher priority because of his relevant background.
I’m not sure anyone completely understands transcription regulation, but it seems to me there is a natural tendency (confirmed by Larry Moran) to spuriously transcribe, and thus a lot of regulation is about repression and silencing.
Histones, DNA methylation and interfering non-coding RNAs regulate by blocking and interfering and repressing, not so much by encouraging, if at all.
Some genes are always “on”, other genes specific to development are shut down. Development is for multicellular creatures, not prokaryotes.
Silencing and repression are important. We wouldn’t want bone specific genes expressing too much in neuron cell lest someone become a bone head.
If there were not targeted repression and silencing, but a constant amount of gene expression in each cell, the notion of epigenetics in development of somatic cell lines would be meaningless.
As I said, in network topologies, and gene regulation is well-modeled as a regulatory network, arguments about what’s at the top of hierarchy are pointless and misplaced.
In addition to chromatin-based epigenetic mechanisms, there are likely glycomic and cytoplasmic epigenetic mechanisms, but we don’t have the sequencing techniques and technology to really understand how the glycome and cytoplasm really work.
And last but not least, there is now talk at the NIH about the EPITRANSCRIPTOME! We have:
Genome (DNA)
Transcriptome (RNA)
Proteome (amino acids)
We now have
Epigenome
Epitranscriptome
Epiprotoeome
Glycome
All these are involved in heritable changes in the somatic cell lines. Whether one wants to call it “epigenetic” or whatever is like calling a rose by another name.
The problem might be that we might have to refer to the chromatin as having “epigenetic traits” (ala Shelly Berger), since now there is an Epitrancriptome on the horizon.
This NIH document details a proposed quest for the Epitranscriptome:
https://dpcpsi.nih.gov/sites/default/files/council%20jan%2030%202015%20Pres%20E4.pdf
It may be worth noting there is structural inheritance:
And the one under everyone’s nose is the heritable properties of a zygote’s cytoplasm!
A blood cell or neuron cell doesn’t have the sort of cytoplasm that will (even with the right DNA) make a human. Hence cytoplasm is a template for other cytoplasm networks. DNA controls the sequence of proteins in the cytoplasm, it is necessary but not sufficient to make an copy of the cytoplasm.
You did say that. And I pointed out that you were wrong (citing SRY in therians and and MAT in yeast). So you are still wrong. Repeating it makes you look dumb.
stcordova,
This has been done to death. Find a biological molecule or state not generated by a recent copy of the DNA molecule and I’ll buy you a beer.
You have an eccentric view of heredity, from a biological perspective, informed more by human affairs than anything particularly relevant to biology. Naturally, when a cell splits in two, its ‘estate’ is divided. This is not what is meant by heredity.
Find a biological molecule or state not generated by a recent copy of the DNA molecule and I’ll buy you a beer.
The direction that paramecium parts its hair does not count.
stcordova,
You don’t need a fancy histone code to repress at that level, so it hardly helps your case. From a histone pov, you just need to shut the door.
I think I spy the death of Darwin, undone by his marriage to Mendel. Once we get Mendel out the way (paramecium haircuts and handedness of shell spirality are a Foot In The Door, the rest will surely follow), Darwin must die. Even though he was unaware of Mendel, and was rather a taken with Lamarck.
Ironic that Sal, in his haste to highlight the importance of epigenetics to multicellular eukaryotic development, should cite paramecium as an example.
In related news, did he actually raise the handedness of snail shells, or was that you Allan?
It’s mainly fuzzy thinking, aided and abetted by fuzzy use of language.
Don’t confuse your misrepresentation of the argument put forward as evidence the argument is wrong. Eggshells are necessary for chickens to reproduce, it doesn’t mean eggshells are therefore eggshell genes and transcription factors are a the top of the developmental hierarchy.
It’s a pointless argument to try to say which component is more essential than another when so many are essential.
Prions. I’m looking forward to that beer. 🙂
Are you suggesting that proteins make themselves? I’m referring both to the misshapen one that causes disease, and the “normal” ones that are converted.
Sal sez:
DNA_J cites SRY and MAT.
Sal sez:
DNA_J notes Sal’s repeated error.
Sal sez:
It’s not that your argument is wrong, Sal. You don’t even have an argument, merely a semi-random array of allusions “I pointed to this….I pointed to that…”
It’s the facts that you have wrong. SRY in therians and MAT in yeast sit, quite unambiguously, at the top of a regulatory hierarchy.
Interesting. Sal does not appear to note that his new claim that the most essential parts of inheritance are cytoplasmic would result in uniparental inheritance. People should always look like their mothers, not their fathers. How does that prediction hold up?
Contrast with:
and
Guess you don’t know as much as you pretend to know.
As you put it: “Don’t confuse your misrepresentation of the argument put forward as evidence the argument is wrong. ”
SRY and MAT sit at the top of a regulatory hierarchy. Lots of fun stuff, including 16 other transcription factors, affect the timing of SRY expression. Doesn’t make your idiocy any less wrong.
What about MAT? (Yeah, I know spare copies of MAT are suppressed via HDACs — we’ll get you to discuss Zhang et al 2014 yet! — but again that doesn’t make you any less wrong.)
Yikes.
DNA_Jock,
It was in the Wiki quote. Sal seems to be arguing (I spy an OP likely pursuing that very theme) that the addition of DNA-generated molecules to an existing scaffold, in which the scaffold influences placement, constitutes ‘inheritance’. Yep, that’ll do for Darwin alright!
stcordova,
Where do prions come from? I don’t just give it away you know!
Allan Miller,
Well, that highlights the risk of quoting stuff you don’t understand. Snails (Lymnaea peregra) were the original textbook example of a maternal effect gene, and similarly Sal’s / wikipedia’s Tetrahymena citation notes:
Oh well.
Just to drive home of the conceptual difficulties of DNA_Jock’s assertion:
But if SRY is controlled by something else, the claim it is unambiguously at the top of the regulatory hierarchy is falsified!
I pointed out this paper:
Oops. No retraction by DNA_Jock.
It’s pointless to argue which character of a password is the most important. In comparable manner, at this stage in our knowledge it’s pointless to argue what’s at the top of the regulatory hierarchy, especially if the regulatory system has a network topology over time and hence hierarchy is not the correct way to describe it’s construction.
In any case, I provided the experimental evidence refuting DNA_Jock’s now falsified claim that SRY is unambiguously at the top of the regulatory hierarchy.
Sal,
You evidently have a great deal of difficulty in following even the simplest of arguments. I would like to encourage you to slow down and read very carefully. You display a history of mis-representing what others have said, clearly due to your failure to actually understand what they have said.
You made a claim (well, an over-wrought analogy, in fact, but another problem with your style of argumentation is your belief that argument-by-analogy is a valid form of argument. Analogies are pedagogical tools…) that
I pointed out that this is wrong, citing SRY and MAT as counter-examples (of centralized control).
You repeated your “arguments about what’s at the top of hierarchy are pointless and misplaced” claim.
I remind you that “SRY in therians and MAT in yeast sit, quite unambiguously, at the top of a regulatory hierarchy.” (Remember now, the topic is centralized control)
You cite some (decades-old) literature that describes the potential influence of DNA methylation on the timing of SRY expression. The abstract that you quote actually includes the statement “The Sry gene encodes a master protein for initiating testis differentiation in mammals” FFS!
I point out that lots of stuff, including 16 TFs, affect the timing of SRY expression., and this in no way detracts from the statement “SRY sits at the top of a regulatory hierarchy” (heck, it doesn’t refute the statement “MAT sits at the top of a regulatory hierarchy” either! BTW, I do enjoy the way you are avoiding that counter-example!)
So you just repeat your claim, with a rather sad modification, viz:
Really? A network topology over time?
So Sal, how would you describe the control of Mating Type in S. cerevisiae, if not a hierarchy?
With MAT at the top?
The fact Sry gene controls something else doesn’t negate the fact it is regulated by chromatin. You make a non-sequitur.
Neither does the fact that the experiment was conducted 2004. Only goes to show you ignore information that’s already been available for a decade.
So your response makes a non-sequitur and shows you failed to account for experimental results available for a decade.
Ho boy. So your conclusion from reading Nishino et al. 2004 is that it is a fact that Sry “is regulated by chromatin.”
Let’s see if we can disabuse you of that error.
Would you care to describe, in your own words, including the specifics of the experimental technique used, what led Nishino et al. 2004 to their (albeit very tentative) conclusion that “expression of the Sry gene is under the control of an epigenetic mechanism mediated by DNA methylation”.
There will be a follow-up question.
The importance of chromatin regulation is evidenced here:
There were differences as noted in the part of the wiki article not quoted, but the conclusion:
Proximally and partially but not totally and finally. You seem unwilling to view the nuances and want simplistic (and thus erroneous) descriptions.
Your error is the assumption regulation can be described ultimately by transcription factors. It’s a silly as saying the life of chicken is defined by the egg shell since egg shells are mission critical.
No, Sal, you have missed the point. I am highlighting the fact that you have conflated “under the control of an epigenetic mechanism mediated by DNA methylation” with “regulated by chromatin”. Hence my request that you describe, in your own words, the methods used in Nishino, 2004.
My broader point is that all you have offered to date is argument-by-analogy (thank you for repeating your lame egg-shell analogy, btw, truly TGTKOG), vague allusions to review articles, and fuzzy thinking enabled by fuzzy language. You have really not put forward anything that could be construed as an argument.
You have offered up a single primary source (Nishino, 2004) for your claims about SRY.
I would like to discuss the methods that these authors used, or discuss Zhang et al. 2014, which demonstrates that it is transcription factors that are driving the bus.
You, evidently, would rather not.
I understand your reluctance.
Methylation marks (or absence thereof) repress or allow gene expression for a variety of genes — not just for SRY but lots of genes (like the HOX clusters), that is well known. Therefore DNA methylation is a form of regulation. Regulation by chromatin is regulation by the components and mechanism of chromatin such as DNA methylation, histone modification, etc.
Your insistence that there is some sort of absolute hierarchy in gene regulation is not consistent with the viewpoint of regulation via networks not strict top down control. The strict hierarchical viewpoint is simplistic to the point of being erroneous, maybe useful as a pedagogical tool for beginners.
And in case you miss the mainstream viewpoint, it is represented here in Wiki:
Don’t confuse my disagreement with you as my lack of understanding. I understand what you’re saying, and I also understand why what you’re saying is overly simplistic to the point of distortion.
Here is a depiction of a genetic/epigenetic/microRNA regulatory network. As you can see, the notion of “hierarchy” or “top of the hierarchy” is not a good metaphor for describing the relationship of the parts of a regulatory network.
https://www.researchgate.net/figure/262194139_fig1_miRNA-and-EMT-TF-regulatory-networks-controlling-epithelial-plasticityNotes-EMT-TFs
The New Yorker highlighted the battle lines between the Old Guard and the New Guard, the transcription-factor centric vs. the chromatin centric view of epigenetics.
As I’ve said, I personally think there is more to the epigenome than chromatin, but a lot of the data on the epigenome is chromatin based, and hence I’ve tended to side with the views expressed by Allis, Felsenfeld and the New Yorker article with some qualification.
The glycome’s role in the epigenome has yet to be elucidated not to mention cortical inheritance and who knows what else.
In any case, here is a regulatory network diagram that put chromatin as a central feature, especially the histones (just like in the New Yorker article). Medical researchers who have less ideological agendas are at the forefront of paradigm shifts.
The diagram shows how a transcription factors behavior (in this case MyoD) can be blocked or recruited according to the chromatin/histone state.
The conflation continues. What Nishino 2004 did was demonstrate regulation by chromatin because Sal says so. LOL.
And here is a depiction of a regulatory hierarchy:
SRY+ –> male
SRY- –> apparently normal female who fails to go through puberty (presents with delayed puberty or with apparent ovarian cancer)
Here is a depiction of another regulatory hierarchy:
MATa –> cell produces a-factor, responds to alpha-factor, able to mate with alpha cells, can switch mating type. Cannot sporulate.
MATalpha –> cell produces alpha-factor and an alpha-factor antagonist, responds to a-factor, able to mate with a cells, can switch mating type. Cannot sporulate.
MATalpha/MATa –> cell cannot produce or respond to pheromones, cannot mate, cannot switch mating type. Can sporulate.
But “arguments about what’s at the top of hierarchy are pointless and misplaced”, because Sal says so
ROFL
From this paper:
I cited one of the types of chromatin modification earlier:
Epigenetics of sex determination and gonadogenesis by Francesc Piferrer*
So SRY is controlled epigenetically by a mechanism involving DNA methylation. DNA methylation is a chromatin modification.
So SRY is controlled epigenetically, which means it is not unambiguously at the top of the developmental hierarchy, only a hierarchy where it is depicted at the top, but not in the complete sense when one shows (as I have) that something else regulates SRY.
I’ve cited relevant literature, DNA_jock just provides ridicule of arguments I’m not making.
So what’s DNA_jock excuse for ignoring the fact that SRY is controlled epigenetically by something else. I therefore refutes his claim SRY is unambiguously at the top of the regulatory hierarchy.
SRY is only at the top of a hierarchy if one doesn’t include what controls SRY in the depiction of the hierarchy.
You’re just laughing at your own misrepresentations and arguments I didn’t make. Kind of disgusting, but if you don’t have actual experimental evidence as I’ve provided, shouting and heckling is about all you can do.
I have been clear from the start that the timing of SRY expression is controlled by “Lots of fun stuff, including 16 other transcription factors”. You do not seem to be paying attention…
No. Logic fail.
Care to try the same line of reasoning with MAT? I mean, HO ‘controls’ the identity of MAT, so (by Sal-logic(tm)) MAT cannot sit at the top of a regulatory hierarchy either.
Sal-logic(tm) is fun.
As I wrote before:
When you cite literature, but refuse to discuss it, your citation becomes just another empty appeal to authority, on a par with all the things that you have “pointed to”, without ever having made a coherent argument.
If SRY expression is controlled by 16 other transcription factors, how does this imply it is unambiguously at the top of the regulatory hierarchy, unless you build the hierarchy to exclude the things that actually have precedence over SRY?
Your earlier diagram therefore is not complete:
16 factors + chromatin + other stuff control SRY, ergo SRY is not unambiguously at the top of the regulatory hierarchy unless you are only referring to hierarchies where SRY is defined at the top by excluding the mechanism regulating SRY.
Only shows your claim of “unambiguous” is falsified even by your own data.
“timing”, Sal.
Your unwillingness to discuss what Nishino actually showed, or discuss MAT, is noted.
😉
Here’s another research paper you ignored that shows SRY is regulated by chromatin state (specifically the DNA methylations):
Take that, man, you’re way behind the curve in terms of available literature.
http://www.nature.com/nature/journal/v447/n7143/full/nature05915.html
Still want to argue chromatin isn’t involved in gene regulation?
Regarding MAT:
So if MAT is regulated by something else, isn’t it kinda hard to insist it is unambiguously at the top of the regulatory hierarchy?
Especially damaging to your case is that part of the regulation of MAT is through DNA Methylations and histone modifications.
I’ve provided links to several experiments supporting my point.
So if MAT and SRY are regulated by something else, how can you insist they are unambiguously at the top of the regulatory heiarchy? Huh? Huh? Huh?
Some gene regulation involves feedback loops where
A regulates B which regulates C which regulates A.
It is hard then to state what is at the top of the hierarchy in such a scheme.
One example:
http://www.ncbi.nlm.nih.gov/pubmed/19221490
Remind me how old the earth is again Sal? Huh? Huh? Huh?
Our conversation has moved to Noyau.
However, I must point out yet another failure to comprehend on your part here.
When I introduced “SRY” and “MAT”, I wrote (page 2 of this thread):
MAT in yeast is completely and utterly unrelated to the human methionine adenosyltrasferases (the MAT isozymes) that you are attempting throw in my face.
Which fact you would have been aware of, if only you had read my second example of a hierarchy, instead of studiously ignoring it.
I wrote:
Kinda hard to come away with the impression that I am talking about humans…
When you conflate humans and yeast, you come across as ignorant and a little desperate.
(In related news, 10-HDA is unrelated to HDA-10 😉 )
Hey, I finally learned something from you. Thanks for the free-of-charge tutoring. It’s true, I don’t spend time reading you’re questions to me too carefully especially after I demonstrated your misleading insinuation that SRY is unambiguously at the top of the regulatory hierarchy.
But regarding the claim MAT is unambiguously at the top of the regulatory hierarchy, suffice to show MAT behavior is regulated by something else, the HO endonuclease. Knockout the HO gene and no MAT modification.
But HO is regulated also by MAT domains.
http://europepmc.org/articles/PMC545453
There is a loop apparently. As I said, hard to characterize things with a strict hierarchy.