Skip to main content

Analytical Chemistry

Epigenetics Is Not What You Would Call a Settled Field

Everyone knows the canonical bases of the nucleic acids. Well, OK, not every single person, but a whole of lot of people do, and I’m willing to bet that if you stopped a bunch of random strangers, you’d get more “A, T, C, G” answers than you might think, thanks to movies and popular culture. Maybe not so much uracil/uridine; the average passer-by cares more about DNA than RNA, I’m willing to bet.

But there are other bases, found in small quantities, such as 5-methylcytidine (and its 2-deoxy analog, m5dC).  Here’s a 2012 review from the Carrell group at Munich on the various noncanonical bases, as they were understood at the time. When you come across these little noncanonical additions in biochemistry, there are several ways you can think about them, which boil down to “Sure isn’t much of that, might not be that important” and “Sure isn’t much of that, might be really powerful stuff”. And because you can’t rule that second one out, it’s definitely worth investigating – at the very least, you may find out something about the unusual pathways that are using the weird structures, and it’s likely to be something new.

The 5-methylcytosine derivatives seem to be important as epigenetic markers, a way to modify transcription and translation outside of the straight sequence of nucleobases. This field has been humungous (technically speaking) for some years now, because it’s always been clear that the mechanisms for handling gene transcription are (and have to be) extremely complex, detailed, and highly regulated. Consider how DNA is wrapped up and wrapped up again for storage in the nucleus, but at the same time is available for wildly varied and complex programs of transcription during development and just normal cellular processes. There’s a nearly ungraspable amount of complex 3-D ballet going on down there constantly, and understanding it is going to be extremely important for a lot of disease states.

Epigenetic markers on histone proteins and on the nucleobases themselves are a big part of this process, and despite all the excitement, we’re only barely beginning to get a handle on them. A few years ago, there was a big burst of activity in the drug industry on epigenetics, which has since died down some among a general atmosphere of “Hmm, that was harder than we thought and we already thought it would be hard”. But we just need to catch our breath a bit, and learn more about what’s going on. The histone deacetylases were among the first enzymes in this area to get attention, but the various other lysine-modifying enzyme classes have since come in for a lot of exploration, with progress that has to be described as “highly varied”.

Attacking the modified nucleobase mechanisms is really a frontier area, because if you think we don’t understand histone acetylation very well, you should see these. To give you an idea of just how much we don’t know what’s going on, consider this new paper in Angewandte Chemie. It’s looking at not just m5dC, but at the 5-hydroxymethyl derivative, the 5-formyl, and the 5-carboxyl, all of which have been reported to be used in vivo. (Readers who have dealt with metabolizing enzymes will note that there’s a familiar pattern of oxidation going on).

It’s the same group that published that review linked above that have come out with this new paper, and it’s going to cause some consternation. They’re also looking at two even newer noncanonical bases, N6-methyldeoxyadenosine (m6dA) and N4-methyldeoxycytidine (m4dC), which were first reported in bacteria and the like but have recently been described in higher organisms (despite some earlier reports that failed to find, for example, N6-methyladenine in mouse DNA). The Munich group as developed a very sensitive LC/MS assay for these unusual nucleobases, and they can’t find either m6dA or m4dC in mice, either – not in brain tissue, not in the liver, and not even in stem cells, which is one of the last resorts for odd transcriptional mechanisms.

They were able to quantify the other four noncanonical bases; they’re definitely in there, but the conclude that the latest two may be a bridge too far. The people who have reported either of these (more here) in mammalian DNA, then, are going to be pretty interested to hear about this dry-well exercise, and I think we can expect to see some lively exchanges. Those working on the possible applications of all this to human disease would be well advised to wait for the dust to settle!

17 comments on “Epigenetics Is Not What You Would Call a Settled Field”

  1. luysii says:

    An oxidation product ( 8 – oxo – guanine) has been proposed as an epigenetic marker — e.g. it is useful — rather than toxic. Moreover production of it increases transcription of some genes (as repairing it opens up compacted DNA) This appears to be true for vascular endothelial growth factor (VEGF), whose transcription is increased in response to hypoxia (how hypoxia leads to the formation of an oxidative modification is far from clear to me).

    There is some delicious chemistry involved as the 8 oxo guanine occurs in an elegant DNA structure called a G quadruplex — something undreamed of by Watson and Crick. This may be why anti-oxidant prophylactic therapy hasn’t worked (in fact the reverse in one case).

    For details see —

    1. zero says:

      (how hypoxia leads to the formation of an oxidative modification is far from clear to me)

      Overshoot? Perhaps (parts of) the body responds to low oxygen states with an over-reaction that leads to oxidative stress?

    2. b says:

      I believe metabolic processes in hypoxic conditions ultimately end up generating more ROS:

      “Paradoxically, cellular hypoxia augments the rate of ROS generation at that site, leading to the production of H2O2 in the intermembrane space”

  2. SirWired says:

    Whenever somebody claims that gene sequencing can solve all our medical problems, I point them here:

    1. dstar says:

      Oh, but it will. Eventually. Barring anything that’s flat out not solvable, of course.

      Now, whether we’ll still be recognizable as Homo Sapiens by the time we figure it out, or whether we’ll have evolved into something with tentacles that can only be described using words like ‘squamous’, well… I’m not putting any money down on that one.

  3. MBP says:

    I’ve never heard of 8-oxo-dG being considered an epigenetic marker. Epigenetic markers are typically thought to be controlled via normal cellular processes. In contrast, 8-oxo-dG is formed via spontaneous reaction of reactive oxygen species (ROS) with DNA, so it is non-specific. It’s usually grouped as either a DNA lesion or as a biomarker of oxidative stress, not an epigenetic marker.

    1. luysii says:

      The authors of the paper discussed in the link above ( [ Proc. Natl. Acad. Sci. vol. 114 pp. 2788 – 2790, 2604 – 2609 ’17 ] ) call 8 oxo guanine an epigenetic factor. According to them when present 8 oxo guanine increases the transcription of Vascular Endothelial Growth Factor (VEGF).

  4. David Antonini says:

    I don’t really understand much about the biology side here, but it seems that pursuing costly drug/therapy development applications of such things way before the science/processes are at all well understood. Feels a bit akin to trying to build the ISS for-profit in the mid 1960s. Or alchemy efforts in the Middle Ages before much chemistry (or nuclear physics, regarding elemental transition) was understood.

    1. zero says:

      A drug discovery effort could turn up useful and surprising information that could inform academic efforts…

      1. David Antonini says:

        True. But I wonder if it’s a bit of an “eventually you pin the tail on the donkey” thing – at what cost efficiency?

        1. Patrick says:

          As a nonparticipant in the field itself, I have always been stunned by the willingness of drug discovery programs to run out ahead of the settled science. It’s wild, and often is extremely wasteful, people are so desperate for new drugs that we do it anyway.

          This is to say, it’s in fact pretty common for drug discovery efforts to start with the first inkling of the possibility of a way to influence a disease state.

          1. Hap says:

            I don’t know if there’s enough money directed at fundamental biology to find things out comprehensively through research – people mostly have to perform reconnaissance by fire to figure out what’s going on, especially if applied research (in this case, research towards specific compounds) is what’s getting funded.

            Also, the HepC scrum makes it look like that if you don’t get an effective drug out early, you may not be able to get enough market share for the development to pay off. By the time it’s clear that compounds acting by a mechanism can be used to treat a disease, you may not be able to make a drug that will pay enough to be worth it. In most cases, that means guessing before you actually know that hitting a mechanism will treat a disease.

    2. Morten G says:

      If working on drugs in areas with “settled” biology was a viable business model then it would happen. That’s capitalism for you (an area of study that inspired Darwin to come up with his theory of evolution – both very strong optimisation approaches).

  5. RBW says:

    Worth pointing out the early epigenetic drugs were discovered phenotypically before their targets were understood.
    This year actually marks the 50th anniversary of an epigenetic drug in man.

  6. Rishi Porecha says:

    This stuff is fascinating: 5-hmC, 5-fC, and 5-carboxyC are introduced into DNA by enzymes that site-specifically oxidize the methyl group on 5-mC using a radical mechanism. Not too much study on the mechanisms of these enzymes (hydroxylases, yes, but not on enzymes that act on DNA), but if the radical reacts with anything other than the H-atoms on the methyl on 5-mC, then it would damage the DNA. If the radical is lost to water, then it can form a hydroxy-radical that can react with DNA. Also, it could be possible that the transition to 5-carboxyC could lead to the eventual removal of DNA methylation marks – especially if they discover an enzyme that forms a Schiff’s base with the 6-position of 5-carboxyC. So much is still unknown – not to mention the histone modification and genome architecture stuff too.

  7. Design Monkey says:

    >Hmm, that was harder than we thought and we already thought it would be hard”

    getting first foot-in-the-door (SAHA, belinostat) wasn’t that hard, they are quite simple thingies (generic non specific crowbars thrown at the HDAC). But after those low hanging fruits, yep, anything selective ain’t going to easy. And even if the compounds themselves might be highly selective against specific epigenetically involved enzymes, that still would be throwing a crowbar at very subtly regulated machinery.

  8. Mikeb. says:

    How about the epitranscriptome? Drugging that might be a hell of a lot safer than messing around with dna.

Comments are closed.