Transcriptional adaptation in the context of genetic loss-of-function experiments

El-Brolosy MA, Kontarakis Z, Rossi A, Kuenne C, Günther S, Fukuda N, Kikhi K, Boezio GLM, Takacs CM, Lai SL, Fukuda R, Gerri C, Giraldez AJ, Stainier DYR, entitled "Genetic compensation triggered by mutant mRNA degradation" in Nature. 2019 Apr;568(7751):193-197.

Modern biomedical scientific method is generally based on analyses when gene function is suppressed. However, whilst many researchers have reservations about RNAi and morpholinos because of questionable specificity due to potential off-target effects, we seldom ask if our knockout mouse or zebrafish mutants are accurately portraying what happens when that single gene is disrupted. Sometimes we can be surprised when, following targeted deletion of the transcriptional start site, a second transcriptional start site results in almost normal gene function. The importance of proving non-sense mediated decay (NSMD) to confirm loss of gene function is also well known. Over the last few years the Stainier group at the Max Planck Institute in Bad Nauheim, Germany has been asking why some mutants do not produce the expected phenotype, while others do. This has culminated in the description of a novel mechanism - transcriptional adaptation - with potentially wide-ranging consequences. Transcriptional adaptation appears to be linked to NSMD, such that mutants that have a premature termination codon or a deleted last exon and show a decrease or absence of mutant mRNA, and thus active NSMD, show the transcriptional adaptation. In contrast, when the expression level of the mutant mRNA is maintained (i.e. there is no NSMD), transcriptional adaptation is not observed. Transcriptional adaptation not only involves compensation by immediate gene family members but also more broadly, and unexpectedly, affects transcription of many other genes, changing the overall landscape of gene expression.

Stainier’s team has shown that transcriptional adaptation occurs across a range of species including zebrafish, mouse and human. The upregulated bystander genes appear to share sequence similarity with the target gene and increased expression levels are due to increased transcription, rather than to increased stability of their existing mRNA. El-Brosoly et al explain this by showing that RNA decay factors, guided by RNA fragments produced during the decay, interact with histone modifiers and chromatin remodelers. Whilst the underlying molecular mechanism needs further analysis, this work challenges our scientific perspective. Fundamentally, we now must ask whether the observed mutant phenotype in gene knockout studies is the result of a single gene inactivation or due to a more general and wide ranging alteration to the transcriptional landscape.

With respect to designing future studies, the Stainier group suggest transcriptional adaptation can be avoided in several ways. One way is by preventing the production of the entire transcript through destruction of the promoter or initiation sequences, thus preventing production of any mRNA. However, production of these promoter-less mutants is not always possible, for example because of other genes present in trans. Alternatively, one could produce subtle mutations which abolish the function of the gene without affecting its expression level, in order to understand, in a more elegant (though complex) way, how a gene functions. Rather challenging, modern genome editing techniques should allow the production of these kinds of alterations to the genome.

Perhaps the most important take home message from this study is that we should  take a moment to review our human and mouse studies. Is the phenotype as expected, or is it weakened or enhanced by wide ranging transcriptional compensation? Genomic analyses have identified many variants which have not undergone functional testing. It is possible that these proposed disease-causing mutations result in no phenotype because of transcriptional adaptation. In contrast, an apparently innocuous mutation could lead to a severe phenotype, because it does not lead to NSMD and thus there is no transcriptional adaptation to compensate for loss of gene function. Either way, developmental biology is more important than ever to understand how genes cause inherited disease and malformation.