Experiments with mutated versions of epigenetic enzymes have shown that the changes in histone modifications at the FLC gene are critically important in controlling the flowering response. For example, there is a gene called SDG27 which adds methyl groups to the lysine amino acid at position 4 on histone H[276], so it is an epigenetic writer. This methylation is associated with active gene expression. The SDG27 gene can be mutated experimentally, so that it no longer encodes an active protein. Plants with this mutation have less of this active histone modification at the FLC gene promoter. They produce less FLC protein, and so aren’t so good at repressing the genes that trigger flowering. The SDG27 mutants flower earlier than the normal plants3. This demonstrates that the epigenetic modifications at the FLC promoter don’t simply reflect the activity levels of the gene, they actually alter the expression. The modifications do actually cause the change in expression.
Cold weather induces a protein in plant cells called VIN3. This protein can bind to the FLC promoter. VIN3 is a type of protein called a chromatin remodeller. It can change how tightly chromatin is wound up. When VIN3 binds to the FLC promoter, it alters the local structure of the chromatin, making it more accessible to other proteins. Often, opening up chromatin leads to an increase in gene expression. However, in this case, VIN3 attracts yet another enzyme that can add methyl groups to histone proteins. However, this particular enzyme adds methyl groups to the lysine amino acid at position 27 on histone H3. This modification represses gene expression and is one of the most important methods that the plant cell uses to switch off the FLC gene[277][278].
This still raises the question of how cold weather results in epigenetic changes to the FLC gene specifically. What is the targeting mechanism? We still don’t know all the details, but one of the stages has been elucidated. Following cold weather, the cells in Arabidopsis thaliana produce a long RNA, which doesn’t code for protein. This RNA is called COLDAIR. The COLDAIR non-coding RNA is localised specifically at the FLC gene. When localised, it binds to the enzyme complex that creates the important repressive mark at position 27 on histone H3. COLDAIR therefore acts as a targeting mechanism for the enzyme complex[279].
When Arabidopsis thaliana produces new seeds, the repressive histone marks at the FLC gene are removed. They are replaced by activating chromatin modifications. This ensures that when the seeds germinate the FLC gene will be switched on, and repress flowering until the new plants have grown through winter.
From these data we can see that flowering plants clearly use some of the same epigenetic machinery as many animal cells. These include modifications of histone proteins, and the use of long non-coding RNAs to target these modifications. True, animal and plant cells use these tools for different end-points – remember the orthopaedic surgeon and the carpenter from the previous chapter – but this is strong evidence for common ancestry and one basic set of tools.
The epigenetic similarities between plants and animals don’t end here either. Just like animals, plants also produce thousands of different small RNA molecules. These don’t code for proteins, instead they silence genes. It was scientists working with plants who first realised that these very small RNA molecules can move from one cell to another, silencing gene expression as they go[280][281]. This spreads the epigenetic response to a stimulus from one initial location to distant parts of the organism.
Research in Arabidopsis thaliana has shown that plants use epigenetic modifications to regulate thousands of genes[282]. This regulation probably serves the same purposes as in animal cells. It helps cells to maintain appropriate but short-term responses to environmental stimuli, and it also locks differentiated cells in permanent patterns of specific gene expression. Because of epigenetic mechanisms we humans don’t have teeth in our eyeballs, and plants don’t have leaves growing out of their roots.
Flowering plants share a characteristic epigenetic phenomenon with mammals, and with no other members of the animal kingdom. Flowering plants are the only organisms we know of besides placental mammals in which genes are imprinted. Imprinting is the process we examined in Chapter 8, where the expression pattern of a gene is dependent on whether it was inherited from the mother or father.
At first glance, this similarity between flowering plants and mammals seems positively bizarre. But there’s an interesting parallel between us and our floral relations. In all higher mammals, the fertilised zygote is the source of both the embryo and the placenta. The placenta nourishes the developing embryo, but doesn’t ultimately form part of the new individual. Something rather similar happens when fertilisation occurs in flowering plants. The process is slightly more complicated, but the final fertilised seed contains the embryo and an accessory tissue called the endosperm, shown in Figure 15.2.
Figure 15.2 The major anatomical components of a seed. The relatively small embryo that will give rise to the new plant is nourished by the endosperm, in a manner somewhat analogous to the nourishment of mammalian embryos by the placenta.
Just like the placenta in mammalian development, the endosperm nourishes the embryo. It promotes development and germination but it doesn’t contribute genetically to the next generation. The presence of any accessory tissues during development, be this a placenta or an endosperm, seems to favour the generation of imprinted control of the expression of a select group of genes.
In fact, something very sophisticated happens in the endosperm of seeds. Just like most animal genomes, the genomes of flowering plants contain retrotransposons. These are usually referred to as TEs – transposable elements. These are the repetitive elements that don’t encode proteins, but can cause havoc if they are activated. This is especially because they can move around in the genome and disrupt gene expression.
Normally such TEs are tightly repressed, but in the endosperm these sequences are switched on. The cells of the endosperm create small RNA molecules from these TEs. These small RNAs travel out from the endosperm into the embryo. They find the TEs in the embryo’s genome that have the same sequence as themselves. These TE small RNA molecules then seem to recruit the machinery that permanently inactivates these potentially dangerous genomic elements. The risk to the endosperm genome through re-activation of the TEs is high. But because the endosperm doesn’t contribute to the next generation genetically, it can undertake this suicide mission, for the greater good[283][284][285][286].
Although mammals and flowering plants both carry out imprinting, they seem to use slightly different mechanisms. Mammals inactivate the appropriate copy of the imprinted gene by using DNA methylation. In plants, the paternally-derived copy of the gene is always the one that carries the DNA methylation. However, it’s not always this methylated copy of the gene that is inactivated[287]. In plant imprinting, therefore, DNA methylation tells the cell how a gene was inherited, not how the gene should be expressed.