The question

Several neurogenetic disorders are caused by trinucleotide repeat expansions, which disrupt gene function1. Repeat expansions in noncoding regions do not affect the encoded protein sequence but often cause gene silencing; how they do this is still unclear in diseases such as Friedreich’s ataxia. Repeat expansions in noncoding regions are also often extensive and difficult to engineer, and thus are difficult to study. We previously discovered a repeat expansion in Arabidopsis, and in the past decade we developed this system as an in vivo model to study triplet-repeat-expansion-associated genetic defects2,3. The Bur-0 Arabidopsis strain harbours an intronic repeat expansion in ISOPROPYL MALATE ISOMERASE SUBUNIT 1 (IIL1), which leads to epigenetic silencing of IIL1 and, in turn, results in a temperature-dependent growth defect2. This clear phenotype enables us to investigate the pathways and mechanisms that underlie this gene silencing by conducting genetic screens in Arabidopsis.

The discovery

We previously showed that the irregularly impaired leaves (iil) phenotype (narrower, twisted leaves) of Bur-0 when grown at 27 °C is associated with the production of 24-nucleotide small-interfering RNAs (siRNAs), which mediate silencing of IIL1 by the RNA-directed DNA methylation (RdDM) pathway3. To identify factors that are involved in this gene silencing, we performed ethyl methanesulfonate (EMS) mutagenesis in Bur-0 and screened for genetic suppressors of the iil phenotype (that is, plants with normal leaves at 27 °C). Genome sequencing revealed that most of the suppressors (eight of nine) contained mutations in RdDM component genes. However, in one mutant (4b), no mutations were detected in RdDM pathway components. Mapping and cloning revealed that 4b contains a mutation in FOURTH ULP GENE CLASS 1 (FUG1), which encodes an uncharacterized deSUMOylase (a protease that removes small ubiquitin-related modifier (SUMO) from target proteins). Catalytic site mutations confirmed that FUG1 protease activity is essential for epigenetic silencing caused by the repeat expansions.

To find FUG1 targets, we carried out a yeast-two hybrid screen and identified AL3 — a member of the Alfin-like gene family, which encodes readers of trimethylation of histone H3 at lysine 4 (H3K4me3; a mark of accessible chromatin). Knockdown of AL3 suppresses the iil phenotype and results in increased IIL1 expression. Interestingly, trimethylation of histone H3 at lysine 27 (H3K27me3; a repressive mark) is increased at the IIL1 locus in Bur-0, an increase that is suppressed by AL3 mutation. To ascertain how AL3 (a H3K4me3 reader) leads to H3K27me3 at IIL1, we tested for AL3 interactors by yeast two hybrid assays, which yielded LHP1 (a chromodomain protein that is required for the spread of H3K27me3 marks) (Fig. 1).

Fig. 1: Repeat expansion-induced gene silencing requires FUG1, AL3 and LHP1.
figure 1

a–c, Schematic depicting a model of epigenetic silencing at IL3 by an intronic triplet repeat expansion (a). In the Bur-0 strain (b), interactions among FUG1, AL3 and LHP1 result in increased repressive H3K27me3 marks that silence the IIL1 locus, which is associated with the iil phenotype. In the absence of any of these components, active marks of H3K4me3 or H3K36me3 can be seen at the IIL1 locus and the iil phenotype is suppressed (c depicts the FUG1-mutant strain 4b). © 2024, Sureshkumar, S. et al.

Full size image

Interestingly, the FUG1 protease domain is very similar to that of human sentrin-like protease 7 (SENP7), which interacts with the chromodomain-containing protein heterochromatin protein 1 (HP1), an interaction that is essential for heterochromatinization. In plants, it seems that Alfin-like proteins act as mediators between SUMO proteases and chromodomain proteins to induce epigenetic silencing.

The implications

Post-translational modifiers are involved in the responses of plants to various environmental stimuli4. Our study links post-translational modifiers such as SUMO proteases to chromatin remodelling associated with repeat-expansion-induced gene silencing in plants. It would be interesting to test whether the same holds true in human diseases such as Friedreich’s ataxia. SUMOylation–deSUMOylation cycles are well known in the stress responses of plants4, which raises the question of whether these cycles are associated with chromatin remodelling in plant stress responses. Additionally, global efforts are underway to identify ways of reversing epigenetic silencing in triplet-repeat-expansion diseases, and post-translational modifiers would represent attractive therapeutic targets if similar mechanisms operate in these diseases.

Several key questions remain unanswered in our study. First, we do not know how potential SUMOylation or deSUMOylation of AL3 affects its function. Second, the interaction between AL3 (a reader of active H3K4me3 marks) and LHP1 (a reader of inactive H3K27me3 marks) suggests a chromatin state switch, but the exact mechanism of this switch remains to be deciphered.

Our work clearly shows that interactions among post-translational modifiers, histone readers and chromodomain proteins are crucial for epigenetic silencing caused by repeat expansions. In the future, it will be interesting to examine the effects of perturbing these pathways in Friedreich’s ataxia or other human genetic diseases that are caused by noncoding repeat expansions.

Sridevi Sureshkumar & Sureshkumar Balasubramanian

Monash University, Clayton, Victoria, Australia.

Expert opinion

“This work investigates the mechanism of repeat-expansion-induced epigenetic silencing in Arabidopsis thaliana. On the basis of their data, the authors propose a model in which the SUMO protease FUG1 deSUMOylates the PhD finger domain protein AL3, thereby changing its ability to interact with LHP1 and enabling a switch from H3K4me3 to H3K27me3 and silencing of the repeats. Linking two new players, FUG1 and AL3 to repeat-induced gene silencing is an interesting discovery.” An anonymous reviewer.

Behind the paper

As a geneticist, when you have such a clear phenotype it is almost imperative to do a genetic suppressor screen. We were prompted to carry out a screen from often being asked whenever we presented our findings on siRNAs in Bur-0 whether we picked up RdDM pathway components in genetic screens. We had mixed feelings when the first suppressor proved to have a mutation in a RdDM pathway gene: joy from confirming our previous results, but also trepidation about studying novel genes. In the case of FUG1, we were at a loss to explain its mechanism of action, but fruitful collaborations with dedicated researchers with varied expertise and the contributions of students from diverse backgrounds helped to link the varied components. The project was also facilitated by funding from the Australian Research Council and the National Health and Medical Research Council, Australia over the years. S.S. & S.B.

From the editor

“The mechanism of repeat-expansion-induced epigenetic silencing remains unclear. This study stands out because it reports three new regulators involved in this silencing. Moreover, the finding that one of the SUMO proteases, which normally function at the post-translational level, also functions in this process is a big surprise.” Jun Lyu, Senior Editor, Nature Plants.