Most human genetic disorders arise from mutations that disrupt coding sequences or conserved local or distant (although this is much rarer) regulatory regions, including promoters and mRNA processing signals. Tufarelli and colleagues have now extended this list by showing that a genetic disorder — α-thalassemia, in this case — can be brought about by transcription of antisense RNA that leads to methylation and gene silencing.

The authors had previously faced a dilemma — they identified a patient with α-thalassemia but no corresponding mutation in the coding or regulatory regions of the adult α2-globin gene ( HBA2 ). This patient did, however, carry a deletion in the α-globin cluster. Despite the positively acting cis-elements being intact, HBA2 was silenced and its CpG island was heavily methylated resulting in α-thalassemia.

It turned out that the deletion positions a truncated copy of a widely expressed component of the splicing machinery ( LUC7L ) next to HBA2. LUC7L is transcribed from the opposite strand to HBA2, and because antisense RNA has been implicated in maternal imprinting and X inactivation the authors used RT-PCR to look for LUC7L transcripts. The results showed that in the deleted globin cluster, transcripts that initiate from the LUC7L promoter fail to terminate (the termination site having been removed by the deletion) and extend into the CpG-island region of HBA 2.

To investigate the role of this antisense transcript on the silencing and methylation of HBA2, Tufarelli et al. made a mouse model to mimic the effects of this deletion. They found that, as in their patient, the expression of Hba-a2 was abolished in the presence of the antisense transcripts. There was also a strong correlation between the methylation of the Hba2 CpG island and the presence of the antisense transcript.

The authors also established a model of this deletion in embryonic stem (ES) cells. Working in tissue culture allowed them to determine that methylation was CpG-island-specific and that removing the minimal promoter of HBA2 silenced the gene but did not induce methylation — further indicating that antisense transcription was necessary for the methylation and subsequent silencing.

Finally, the authors showed that other antisense RNAs that are complementary to the HBA2 promoter also resulted in methylation and silencing, which indicated that the mechanism is not specific to the aberrant LUC7L transcript.

So, although the authors do not propose an exact mechanism by which antisense RNA brings about methylation and silencing, they point out that given their example and what we know about X inactivation and imprinting, regulation of gene expression by antisense RNA might be a general natural phenomenon. No doubt Tufarelli et al. will use their mouse model and the ES system that they have set up to look into the mechanism of this process more closely.