The recent discovery of a large number of histone demethylases suggests a central role for these enzymes in regulating histone methylation dynamics. Histone H3K27 trimethylation (H3K27me3) has been linked to polycomb-group-protein-mediated suppression of Hox genes and animal body patterning, X-chromosome inactivation and possibly maintenance of embryonic stem cell (ESC) identity. An imbalance of H3K27 methylation owing to overexpression of the methylase EZH2 has been implicated in metastatic prostate and aggressive breast cancers. Here we show that the JmjC-domain-containing related proteins UTX and JMJD3 catalyse demethylation of H3K27me3/2. UTX is enriched around the transcription start sites of many HOX genes in primary human fibroblasts, in which HOX genes are differentially expressed, but is selectively excluded from the HOX loci in ESCs, in which HOX genes are largely silent. Consistently, RNA interference inhibition of UTX led to increased H3K27me3 levels at some HOX gene promoters. Importantly, morpholino oligonucleotide inhibition of a zebrafish UTX homologue resulted in mis-regulation of hox genes and a striking posterior developmental defect, which was partially rescued by wild-type, but not by catalytically inactive, human UTX. Taken together, these findings identify a small family of H3K27 demethylases with important, evolutionarily conserved roles in H3K27 methylation regulation and in animal anterior-posterior development.

Histone methylation regulates chromatin structure and transcription. The recently identified histone demethylase lysine-specific demethylase 1 (LSD1) is chemically restricted to demethylation of only mono- and di- but not trimethylated histone H3 lysine 4 (H3K4me3). We show that the X-linked mental retardation (XLMR) gene SMCX (JARID1C), which encodes a JmjC-domain protein, reversed H3K4me3 to di- and mono- but not unmethylated products. Other SMCX family members, including SMCY, RBP2, and PLU-1, also demethylated H3K4me3. SMCX bound H3K9me3 via its N-terminal PHD (plant homeodomain) finger, which may help coordinate H3K4 demethylation and H3K9 methylation in transcriptional repression. Significantly, several XLMR-patient point mutations reduced SMCX demethylase activity and binding to H3K9me3 peptides, respectively. Importantly, studies in zebrafish and primary mammalian neurons demonstrated a role for SMCX in neuronal survival and dendritic development and a link to the demethylase activity. Our findings thus identify a family of H3K4me3 demethylases and uncover a critical link between histone modifications and XLMR.

There is an interesting overlap of function in a wide range of organisms between genes that modulate the stress responses and those that regulate aging phenotypes and, in some cases, lifespan. We have therefore screened mutagenized zebrafish embryos for the altered expression of a stress biomarker, senescence-associated β-galactosidase (SA-β-gal) in our current study. We validated the use of embryonic SA-β-gal production as a screening tool by analyzing a collection of retrovirus-insertional mutants. From a pool of 306 such mutants, we identified 11 candidates that showed higher embryonic SA-β-gal activity, two of which were selected for further study. One of these mutants is null for a homologue of Drosophila spinster, a gene known to regulate lifespan in flies, whereas the other harbors a mutation in a homologue of the human telomeric repeat binding factor 2 (terf2) gene, which plays roles in telomere protection and telomere-length regulation. Although the homozygous spinster and terf2 mutants are embryonic lethal, heterozygous adult fish are viable and show an accelerated appearance of aging symptoms including lipofuscin accumulation, which is another biomarker, and shorter lifespan. We next used the same SA-β-gal assay to screen chemically mutagenized zebrafish, each of which was heterozygous for lesions in multiple genes, under the sensitizing conditions of oxidative stress. We obtained eight additional mutants from this screen that, when bred to homozygosity, showed enhanced SA-β-gal activity even in the absence of stress, and further displayed embryonic neural and muscular degenerative phenotypes. Adult fish that are heterozygous for these mutations also showed the premature expression of aging biomarkers and the accelerated onset of aging phenotypes. Our current strategy of mutant screening for a senescence-associated biomarker in zebrafish embryos may thus prove to be a useful new tool for the genetic dissection of vertebrate stress response and senescence mechanisms.