Generation of a Drosophila Model of Intellectual Disability to Investigate the Transcriptional, Behavioral and Neuronal Defects Caused by Mutations in Histone Demethylase KDMS.
MetadataShow full item record
Mutations in lysine demethylase 5 (KDM5) family histone demethylases cause intellectual disability (ID) in humans. However, the mechanism underlying the cognitive impairment is unknown. KDM5 proteins are multi-domain transcriptional regulators, most well-known for their Jumonji C (JmjC) domain-encoded demethylase activity that specifically removes trimethylated lysine 4 of histone H3 (H3K4me3), a mark associated with active promoters. Current models of KDM5-induced ID suggest that it is the loss of demethylase activity that disrupts cognition. This is based primarily on a limited number of in vitro demethylase activity studies in which some, but not all, KDM5 missense mutations show reduced enzymatic activity of up to 2-fold. However, the in vivo effects of these mutations remain uncharacterized. In addition, the enzymatic defects, if any, of a majority of ID-associated KDM5 mutations proteins have also not been assessed in vitro or in vivo. It is also notable that no reported patient-associated mutations lie in catalytically critical residues within the Jumonji C domain. Indeed, most missense mutations lie outside of this domain. Thus, the causal link between KDMS's demethylase activity in neuronal function remains unclear. The role of KDM5 in neuronal function in general is evolutionarily conserved, with loss of KDM5 disrupting neurodevelopmental processes in model organisms including mice, worms and zebrafish. However, the in vivo defects of the specific reported KDM5 mutations on transcription, cognitive behaviors and neuronal development are not known. In addition, it is not currently clear whether missense mutations in distinct regions of KDM5 affect common downstream effectors through a demethylase-dependent mechanism or if each mutation has individual downstream consequences. Studies answering these questions are crucial to define the molecular mechanisms linking KDM5regulated transcription and cognition. Here, we establish Drosophila as a model system to understand this connection by generating seven fly strains harboring evolutionarily conserved missense mutations in KDM5 associated with ID. These mutations do not affect protein levels and produce viable, phenotypically normal flies. Significantly, only two of the seven mutations reduce histone demethylase activity levels in vivo, suggesting that reduced enzymatic activity may not be a necessary prerequisite for KDM5 mutations to cause ID. We carried out detailed studies including transcriptome analyses, cognitive assays and neuronal morphology on one ID-associated KDM5 mutation that affects demethylase activity, kdm5A512P. mRNA-seq from kdm5A512P mutant heads revealed a significant enrichment for genes required for ribosomal assembly and function. Comparison with ChIP-seq data available from whole adults revealed that a significant number of these affected genes are also direct targets of KDM5. Consistent with the downregulation of this class of gene, incorporation of the tRNA analog puromycin revealed a 2-fold reduction in translation specifically in head tissue but not in thoraces. Because translation is critical for neuronal function, we utilized association based appetitive learning and/or memory tests of kdm5A5I2P adults. This revealed both short-term and long-term memory defects compared to wildtype animals. KDM5A512P is likely to abolish enzymatic activity, since H3K4me3 levels are increased to a similar extent to the demethylase dead strain (kdm5JmjC* ). Indeed, consistent with the primary defect of kdm5AS12P being the loss of histone demethylase function, the transcriptional and behavioral changes in kdm5AS12P flies were indistinguishable from kdm5JmjC* flies. Based on striking similarities between kdm5A512P and kdm5JmjC*, we extended our transcriptional analyses to include three additional ID mutants: kdm5A224T which affects demethylase activity, and also kdm5 F765L and kdm5L854F that lie in an inter-domain region and in the C5HC2 domain, respectively. Comparison of all four ID genotypes uncovered a significant number of genes that were similarly dysregulated in all mutant strains. Consistent with our previous analyses of kdm5A512P these were also enriched for ribosomal function genes. These studies also revealed that the four ID mutants examined, regardless of their demethylase activity status, have significant overlap with the kdm5JmjC* mutant. We propose that the mutation in the kdm5JmjC* strain that disrupts the structure of the Jumonji C domain may also affect the adjacent C5HC2 motif suggesting that this motif of unknown function is involved in KDM5's and neuronal functions. In conclusion, we have generated a Drosophila model to examine the molecular, cellular and behavioral phenotypes caused by mutations in KDM5. Our study has revealed an exciting link between KDM5 and the activation of genes required for ribosome structure and function. Defects in learning and memory are a hallmark of ID. We have recapitulated this ID defect in Drosophila. By investigating the transcriptional and cellular deficits, we also uncovered translation as a key biological process that may contribute to ID in patients with KDM5 mutations. It is therefore conceivable that restoring the reduced translation may ameliorate the cognitive dysfunction observed with these mutations. This possibility warrants a more in-depth study of the translation link as a significant underlying cause of ID caused by KDM5 mutations.