Epigenetics has been defined as the study of heritable traits that do not involve changes in the DNA sequence. This view has been broadened by an avalanche of biochemical evidence revealing a complex and versatile array of molecular mechanisms that regulate gene expression without changing DNA sequences. These include chemical modifications of DNA and RNA molecules, as well as post-translational modifications of histones — the proteins around which DNA coils to form chromatin strands. Post-translational modifications of histones include acetylation, phosphorylation and methylation (addition of acetyl, phosphate and methyl groups, respectively) at specific amino-acid residues of these proteins1. Writing in Nature, Farrelly et al.2 report that histones can also be modified by the addition of serotonin, a molecule with essential roles in regulating neuron activity.
Serotonin is generated from the metabolism of the amino acid tryptophan. It functions as a neurotransmitter — a molecule that acts as a signal between neurons — and as a trophic factor that helps neurons to grow, survive and differentiate. Psychiatric disorders such as schizophrenia, depression and autism spectrum disorder have been linked to serotonin-dependent signalling during key periods of the development of the nervous system3.
Signalling through various serotonin receptors leads to chromatin remodelling, whereby the conformation of chromatin changes in a way that permits gene expression4 (Fig. 1a). Serotonin-receptor-dependent chromatin remodelling is mediated by signals that include histone post-translational modifications4. How serotonin-driven signals are integrated with other molecular signals that affect chromatin architecture remains poorly explored.
The study by Farrelly and colleagues reveals that serotonin can act on chromatin in a receptor-independent manner, by directly targeting histones through a post-translational modification called serotonylation. This chemical modification has been known for more than a decade on some non-nuclear proteins. Some small enzymes belonging to the GTPase family are serotonylated at specific glutamine amino-acid residues by other enzymes called calcium-dependent transglutaminases5–7 (Fig. 1b). This process has a role in the induction of cell division in smooth-muscle cells5, the regulation of insulin secretion by β-cells in the pancreas6 and the internalization of the protein that transports serotonin from blood into platelets7.
Although serotonylation in the nucleus had not previously been described, some intriguing hints supported its existence. A small fraction of the total amount of the enzyme transglutaminase 2 (TGM2) in cells8, as well as a portion of the cells’ serotonin content9, were found in the nucleus. These observations suggested that serotonylation could target nuclear proteins, and thereby influence gene expression independently of serotonin receptors and their signalling pathways. Farrelly and colleagues detected serotonylation on the glutamine residue at position 5 of histone H3 (the H3Q5 position). As with many other targets of post-translational modifications of histones, this residue is located on the protein’s amino-terminal region. No other serotonylation sites seem to exist on histones H3, H2A, H2B or H4, which highlights the remarkable specificity of this modification.
The N-terminal tail of histone H3 is the best-characterized region of all histones. Many modifications that have functional significance (alone or in combination) have been described in this protein region over the past decade10. The trimethylation (addition of three methyl groups) of the lysine amino-acid residue at position 4 of histone H3 (H3K4) is considered the most reliable mark for identifying parts of the genome that are in a state that enables transcription. Notably, the enzymes responsible for transferring the first, second and third methyl groups to this lysine residue are unique11.
Farrelly and colleagues report that TGM2 serotonylates H3Q5 when H3K4 is trimethylated (Fig. 1c). The combination of these two post-translational modifications is called H3K4me3Q5ser. Given that the modified lysine and glutamine residues are adjacent, the stability (or half-life) of the two modifications might be co-dependent. This proximity might also aid the recruitment of specialized chromatin-remodelling protein complexes. Indeed, the authors’ findings suggest that H3K4me3Q5ser might help the function of the transcription factor TFIID, which acts on chromatin to promote transcription.
These findings raise other compelling questions. Does TGM2 have a role in the function of the enzymes that methylate H3K4, such as MLL1? If so, future studies should try to clarify the functional interplay between these enzymes. Does serotonylation of H3Q5 influence other post-translational modifications, in a similar way to how the trimethylation of H3K4 and the acetylation of lysine residues at positions 9 and 14 of histone H3 influence each other12? Are the intracellular pools of serotonin replenished in different ways depending on how serotonin is being used in various cellular compartments at any given time? Does extra-nuclear serotonin influence the serotonylation of histones by being transported into the nucleus on demand?
Serotonylation of histones and its potential influence on transcription might be only the tip of the iceberg in an ever-expanding scenario of post-translational modifications associated with chromatin changes. Histaminylation and dopaminylation (addition of histamine, an amino acid, and dopamine, a neurotransmitter, respectively) are likely to join the party, which could complicate the task of deciphering the language of histone modifications. However, an exciting road to discovery seems to lie ahead.