I have to say that I had never seen Eric Olson's seminar before, and it was awesome. Lately the Olson's lab has been looking at HDACs, i.e. histone deacetylases.
Don't forget that in the nucleus DNA is wrapped around nucleosomes that are composed of proteins called histones (see image, right). One important idea in biology today is that histones can get modified and this changes their regulates how they bind to the DNA. Thus by altering the histones you can regulate how accessible the DNA is for transcription (i.e. the conversion of DNA into RNA).
Histones are modified in many ways. Histones can gain or lose acetyl, methyl, phospho or other small chemical groups. In other cases histones can be modified by being attached to big proteins like SUMO or ubiquitin. It is thought that the pattern of modifications along the length of the DNA can be inherited from mother cell to daughter cells. and this EXTRA inheritance often said to contain extra information. the fact that a cell not only passes on DNA information but also histone modification patterns has fed the "epigenetics" fad (epigenetics=information beyond what is coded in the DNA). Thus histone modifications are described with mystical delight ... like all fads, this one has been oversold. A subclass of the hisones are regulated by external cues. If cellular signals are on, these HDACs are phosphorylated and are sequestered into the cytoplasm by 14-3-3 proteins. If signals are off the HDACs remain unphosphorylated and can then bind to certain DNA sequences via the DNA binding Mef2 family of proteins. HDACs that are loaded onto parts of the genome can deacetylate all the nearby histones and turn off the transcription of nearby genes into RNA.
But Olson demonstrated that when you knock out certain HDACs from mice you get very clean, STRIKING, phenotypes.
When hearts under go stress they expand and accumulate fibrous connective tissue. This leads to heart failure and death. Knocking out HDAC5 or 9 ... you get an increase in heart hypertrophy.
During development cartilage is converted to bone by chondrocytes (cells of the cartilage). In this development process chondrocytes will expand (i.e. hypertrophy) and die (by apoptosis) as the cartilage mineralizes. Once the catilage is replaced by calcium and magnesium salts ... you have bone. Knockout HDAC4 ... you get a mouse that has massive ossification of cartilaginous tissues. Then if you overexpress Mef2C in chondrocytes, you have a mouse whose non-spongy bones are all cartilaginous. So almost all bones (except for the scull) are now converted to cartilage. Yes, a rubber mouse.
Probably because you blocked the chodrocytes from hypertrophy-ing and thus stopped bone development.
Finally knockout HDAC7 ... you get a mouse that is fine until embryonic day 8.5, at which point all the blood vessels burst. The poor critter dies of a body-wide aneurysm.
HDAC7 regulates aspect of endothelial cell development. Endothelial cells form the lining of all you blood vessel walls.
In each case an HDAC regulates some very precise step in development.
So very interesting clues ... HDAC activity correlates with gross morphological changes at the cellular level, usually cell hypertrophy. The specificity and the clean phenotypes are very much against conventional wisdom with regards to HDACs and epigenetics ...
I am interested by your last comment about Olsen's studies challenging conventional wisdom with regards to HDACs and epigenetics. Could you elaborate on what you feel this conventional wisdom is?
I am only 6 months into a PhD so I may have missed the hype period surrounding epigenetics but from my reading I had got the impression that epigenetic changes were important features during development, contributing to 'switches' in expression patterns. Also in a seminar I recently attended, Paul Marks, the guy who developed SAHA, a HDAC inhibitor for cancer treatment, said that the Class II HDACs are lineage specific which fits in well with the specificity seen in Olsen's studies.