Controlling Gene Expression Noise by Preventing Interchromosomal Interactions


Coordinate control of gene expression noise and interchromosomal interactions in a MAP kinase pathway


Emma McCullagh Anupama Seshan Hana El-Samad Hiten D. Madhani


Published in: Nature Cell Biology
Volume: 12, Issue: 10, Pages: 954-962
Published: 10//2010
DOI: 10.1038/ncb2097

Altmetric Statistics:

Our Discussion:

In this paper, the authors examine intrinsic and extrinsic noise in the yeast pheromone response MAP kinase pathway. The transcription factor Ste12 is regulated by this kinase pathway, and there are two partially redundant repressors, Dig1 and Dig2, that interact with Ste12 and inhibit transcriptional activation at Ste12-dependent promoters. The authors confirmed that single gene deletions of DIG1 or DIG2 had little effect on repression of Ste12 transcription and only the dig1∆ dig2∆ double mutant showed a drastic loss of repression. However, the authors were surprised to find that basal levels of Ste12-dependent transcription in the dig1∆ mutant was much noisier than wt, dig2∆, or (not surprisingly) dig1∆ dig2∆ cells. Loss of Dig1 increased both intrinsic and extrinsic noise, the latter measured using two reporters driven by identical promoters.

How is Dig1 reducing noise? After ruling out a feedback mechanism, the authors postulate a role for Dig1 in preventing promoter-promoter interactions resulting from Ste12 homo- and heterodimers, which would result in highly coordinated transcription at multiple Ste12-dependent promoters. This coordinated transcription exacerbates the effects of extrinsic noise due to Ste12 fluctuations. Lo and behold, when the authors examined Ste12 localization in the dig1∆ cells, it formed nuclear puncta instead of being dispersed within the nucleoplasm. The authors then performed ChIP on Dig1 and show that it remains on the promoters at all times, even during pheromone signalling. In addition, wt cells never show Ste12 puncta, even during pheromone signalling. Furthermore ChIP-chip experiments show that multiple Ste12-regulated genes interact in the dig1∆ mutant. These results suggest that, in addition to acting as a repressor, a unique function of Dig1 is to constitutively prevent promiscuous Ste12 interactions leading to promoter-promoter interactions, and coordinated transcription resulting in increased extrinsic noise.

Although this isn’t a criticism of the work per se, I think the author’s rationale for these experiments is a case of brilliance by hindsight. I seriously doubt that the authors (or anyone else) would have guessed that promoter tethering and colocalization was responsible for increased extrinsic noise. Instead, I think the authors were simply trying to determine if there was elevated noise in Ste12 levels in the dig1∆ mutant by examining Ste12-GFP and noticed the puncta. This also brings up some serious criticisms of the paper.

1) The authors do not measure the amount of noise in Ste12 levels in wt vs dig1∆ cells. They fail to rule out a simple explanation for differences in extrinsic noise.

2) The relationship between Ste12 puncta and noise is correlative. This is fine if all the observations are consistent, but Ste12 puncta occur in dig1∆ dig2∆ cells even though the transcription noise is very low. One could argue that reduced intrinsic noise from higher levels of expression is responsible. However, the authors should have used the double reporter system in this mutant to determine if extrinsic noise was still higher than normal.

3) A third criticism is that the authors do not show coordinated transcription of the genes they say are associated in the Ste12 puncta. This could be easily accomplished by examining wt and dig1∆ single cells containing multiple fluorescent proteins (CFP, GFP and mCherry for example) driven by different Ste12-dependent promoters. This data would have significantly strengthened the authors’ contention that the inter chromosomal interactions result in increased extrinsic noise resulting from coordinated transcription.


Leave a Reply Text

Your email address will not be published. Required fields are marked *