Topoisomerase enzymes facilitate gene transcription by resolving DNA tangles. Malfunction of these enzymes seems to compromise the expression of very long genes, potentially mediating neurodevelopmental disorders. See Article p.58
Topoisomerases are a family of enzymes that catalyse the unwinding and unknotting of DNA sequences. By introducing transient 'nicks', these enzymes can relieve the topological pile-up of DNA that is caused by processes such as replication and transcription. On page 58 of this issue, King et al.1 provide evidence that topoisomerases are required for the proper expression of extremely long genes in neurons. This insight has implications for our understanding of the fundamentals of both transcription and neurodevelopmental disorders*.
Genomic imprinting is an evolutionarily conserved mammalian phenomenon in which expression of a gene occurs preferentially from one parental chromosome. In human neurons, for example, the UBE3A gene is expressed only from the maternal chromosome, and deletion or mutation of this allele (gene copy) causes a severe neurodevelopmental disorder called Angelman syndrome2.
In a previous study3, King and colleagues screened for small molecules that, when applied to neurons in culture, activated the normally silenced paternal allele of UBE3A. Surprisingly, they found that topoisomerase inhibitors activated this allele. This ability of the enzymes to correct gene 'dosage' by derepressing silent alleles seemingly held much promise for the treatment of disorders that involve imprinted genes or that map to the X chromosome in females. But the mechanism underlying this activation remained unclear.
The earlier report also showed that topoisomerase inhibitors reduced the expression of UBE3A-ATS, a very long transcript expressed in neurons. UBE3A-ATS overlaps with paternal UBE3A on the opposite strand and is associated with the silencing of this allele (Fig. 1a). Repression of UBE3A-ATS by topoisomerase inhibitors, therefore, suggested that topoisomerases are involved in maintaining the expression of extremely long genes.
King et al. now report that, indeed, treatment of mouse and human neurons with topoisomerase inhibitors results in the widespread silencing of very long genes (those longer than 67 kilobases). This repression depends on the dose of the inhibitor and is highly correlated with increased gene length. Sustained repression of topoisomerase expression using short hairpin RNA (shRNA) sequences also resulted in reduced expression of long genes, excluding the possibility of off-target effects of the inhibitors.
To investigate the mechanism of topoisomerase action, King and colleagues mapped genome-wide binding sites of RNA polymerase II (Pol II) — the enzyme that catalyses DNA transcription — before and after treatment with topoisomerase inhibitors. The authors noted a significant enrichment of Pol II in promoter regions after treatment and a corresponding paucity of Pol II in the body of long genes (Fig. 1b). For short genes, however, Pol II density across the gene body was slightly increased. These results suggest that topoisomerases are specifically involved in the elongation step of transcription during the expression of long genes.
King et al. also found that topoisomerase inhibitors decreased the expression of an impressive proportion (27%) of long genes that are candidates for an association with autism spectrum disorders (ASDs). The authors show that the inhibitors significantly downregulate the collective expression of such ASD candidate genes, further supporting the link between topoisomerase mutations and reduced expression of long ASD genes. Consistently, recent work4,5 has uncovered rare de novo mutations in topoisomerases from patients with ASD.
Notably, these findings suggest a possible role for topoisomerases in other genetic disorders in which the causal gene is exceptionally long. It is plausible that mutations that reduce the expression of these enzymes could allow the appropriate expression of all but a few very long genes. For instance, CFTR, the gene mutated in cystic fibrosis, spans more than 200 kilobases. And DMD, the causal gene in many forms of muscular dystrophy, spans a staggering 2.2 megabases. It would be surprising if topoisomerases did not contribute to these disorders in certain rare cases at least.
Nevertheless, the work also shows that topoisomerase inhibitors are not a panacea for disorders that would benefit from the activation of a normally silent allele. As these inhibitors are likely to have a widespread effect on the expression of all long genes, even appropriate targets such as UBE3A are unlikely to be activated without nonspecific effects. Furthermore, the delicate interplay between gene dosage and traits associated with such disorders makes a broad regulator of transcription such as topoisomerase a less attractive target for drug design.
Despite these concerns, King and colleagues' paper presents an intriguing and fundamentally novel role for topoisomerases in gene regulation. Although these enzymes were known to be required for the expression of longer transcripts in yeast6, the present work cements their importance in mammals. Moreover, the studies show a requirement for both topoisomerase I and II enzymes in the expression of long genes, whereas only topoisomerase II is required in yeast. Such observations reveal an increased significance for topoisomerases in the regulation of the expanded and much more complex mammalian genome.
By excluding other possibilities, this work also strongly indicates that topoisomerases modify DNA topology during the expression of long transcripts. In prokaryotes (bacteria and archaea), transcription causes dynamic DNA supercoiling, which acts as both a negative and a positive regulator of transcription7. Transcription-associated supercoiling also occurs in mammals8, but its functional relevance is unclear. King and co-authors' findings point towards the possible importance of overcoming supercoiling during the transcription of long genes. Thus, modifying the action or recruitment of topoisomerases should provide a way of exploring how the expression of long mammalian genes is modulated.
*This article and the paper under discussion1 were published online on 28 August 2013.
About this article
Neuroscience & Biobehavioral Reviews (2014)