A fundamental problem in biology concerns how the genomic information present in fertilized eggs can give rise to the full spectrum of stably differentiated cell types required to form vertebrates and invertebrates. In the 1930s, C.H. Waddington’s largely observational mammalian embryology studies, which defined this problem, were central to establishing the field of epigenetics. It is now well known that there are master regulatory genes that must be kept on to specify a given cell lineage and off in the many other cell lineages that make up the body.
The problem of keeping these genes in the off state when required has received considerable attention, in large part due to the landmark genetic studies initiated by Pam and Ed Lewis in the 1940s that identified a set of genes required for this repression. This family of genes is called the Polycomb Group (PcG) because the visual phenotype of a heterozygous null allele in these genes is duplication on the second and third legs of the sex combs that wild-type male Drosophila flies have on their front legs. It turns out that PcG proteins repress key developmental master regulatory genes in organisms from plants to humans. The PcG is responsible for a diversity of important biological events, from why plants flower only in the spring (and not in a December warm spell) to how mammals form the correct body tissues in the correct locations.
Polycomb and histone methylation
Study of the PcG has converged with another active area related to epigenetics, the study of covalent modification of histones. The four core histones wrap DNA around them to form a nucleosome, and every gene is bound by nucleosomes, usually at one nucleosome for every ~147 base pairs of DNA. A physical mark on histones is one possibility for how regulatory information might be transmitted from an already differentiated cell to a daughter cell. For example, a covalent mark specifying repression might be found on the histones that coat a specific gene in a liver cell, and these histones might retain the repressive covalent mark when new nucleosomes are formed after DNA replication. One of the key protein complexes formed by the PcG gene products, Polycomb repressive complex 2 (PRC2), methylates lysine 27 of histone H3 (H3K27). This mark is widely believed to be an important component of epigenetic mechanisms and is believed to function by creating a binding pocket for another PcG complex (PRC1) that effects repression.
The fact that a hypothesis makes sense does not eliminate the need to test it as rigorously as possible.
Yet many who work on gene regulation are skeptical that methylation of lysine 27 confers epigenetic information—in this instance meaning information that is heritable and transmitted from mother to daughter cell to specify that a master regulatory gene be kept off. Significant issues with the model include whether the marked histones are faithfully replaced on the gene following replication, whether the energy created by a binding pocket constituted by a methyl group is sufficient to do the repressive job, and whether placement of the methyl mark can be accomplished with sufficient accuracy to effect defined regulation.
In comparison, gene-specific DNA binding proteins, known to play a role in epigenetic regulation, bind to their sites with energies considerably more formidable than can be created by a methyl mark on a histone. These proteins recognize specific DNA sequences that are lengthy enough to be unique within the genome, and one can easily imagine that the proteins will rebind accurately to those sequences following replication and cell division.
Mutating histones in large eukaryotes
There is a simple experiment that would go a long way toward addressing the concerns about the role played by covalent histone modification in heritable repression: mutate lysine 27 of histone H3 to arginine or to alanine. These substitutions would each prevent methylation by PRC2 and would be strongly predicted to impair development of an organism as observed either by gross phenotype or by a molecular phenotype such as RNA expression pattern. Put differently, if methylation of lysine 27 is a central mechanism for PcG-based epigenetic regulation, eliminating the ability of an organism to perform this function should derail the PcG and thereby derail appropriate epigenetic repression of master regulatory genes.
This experiment is somewhat complicated to design at the conceptual level. There are many mechanisms that can contribute to epigenetic regulation, including sequence-specific binding factors, DNA methylation, and noncoding RNAs, and it is anticipated that each different master regulatory gene will use a different combination of these mechanisms—and that many of these mechanisms will be redundant. Thus it is unclear how many genes, and which ones, might rely upon methylation of H3K27. It is even possible, though unlikely given the multitude of mechanisms involved in regulating development, that H3K27 mutation might be lethal at such an early stage that study is difficult. A further complicating issue is that this mark might be required in acute repressive settings in the developing organism (e.g., in the rapid response to a precise developmental cue), and thus there might be impacts on development that have to do with this type of acute response rather than with epigenetic memory of the repressed state. Thus, one would ideally like to follow any affected gene in time across cell divisions to determine to what extent there is a defect in memory of a repressed state.
These conceptual issues can be tackled, but there are two other major roadblocks to doing this experiment—one technical and one motivational. The technical issue is that the organisms that would be most suitable for doing this experiment—Drosophila and mice—have numerous genes for histone H3 (more than twenty in each organism). It is very hard, therefore, to mutate all the endogenous genes, and technologies with which one might eliminate the endogenous genes via large deletions or RNAi have the issue of expressing a mutated replacement gene at the correct level. Gene manipulation techniques are continually expanding, however, and recent advances make this experiment feasible in flies (EMBO Reports 11:772-76, 2010).
The motivational issue is worrisome—on two levels. Even in flies, where the experiment is feasible, it will take a lot of work. But more to the point is the experience of one investigator who recently recounted that he was asked repeatedly why he would waste time doing an experiment that is so hard when everyone already knows the answer. The received wisdom is that of course methylation of lysine 27 is critical for epigenetic regulation: Isn’t it usually called an epigenetic mark? This reliance upon what is essentially an act of faith—methylation of histones makes sense as a mark that might be epigenetic, therefore it must be—is dangerous to progress. The fact that a hypothesis makes sense does not eliminate the need to test it as rigorously as possible. Hopefully, mammalian technologies will advance so that point mutation of residues perceived to be epigenetic can in fact be performed, because the spectrum of mechanisms that govern these issues is not the same in flies and mammals. Until such definitive experiments are performed, skeptics will have free run, and the field will continue to spin its wheels.
Robert Kingston is Professor and Vice Chair of Genetics at Harvard Medical School and Chief of Molecular Biology at Massachusetts General Hospital. His research group at MGH studies the function of Polycomb Group complexes using biochemistry and cellular approaches. They focus on PRC1, PRC2 and the ATP-dependent nucleosome remodeling complexes that oppose the action of these repressive complexes.