Harvard University COVID-19 updates

Department News



Erin O’Shea (l) and Brian Zid

The central dogma of molecular biology proposes that information flows from DNA to RNA through the process of transcription and from RNA to protein through translation. In eukaryotes, these two processes are thought of as disconnected: nuclear factors control transcription, and a different set of factors control translation in the cytoplasm. During times of stress, cells exhibit large transcriptional changes including the upregulation of many genes important for survival. Concurrent with these large transcriptional changes, cells dramatically decrease overall protein synthesis, thereby dampening the stress response at the level of protein expression. Brian Zid and Erin O’Shea wondered if there may be preferential translation of transcripts under stress. To find out, they studied glucose starvation in yeast, a condition in which translation is rapidly repressed while large transcriptional changes are taking place. 

Using ribosome profiling, a technique that measures the number of ribosomes on mRNAs using next-generation sequencing, the authors found that a subset of transcriptionally upregulated mRNAs were preferentially translated during glucose starvation. A conserved phenomenon upon stress is the formation of mRNA-protein aggregates, called stress granules. Using fluorescent microscopy, they found that mRNAs that were preferentially translated remained diffuse throughout the cytoplasm, while transcriptionally upregulated but poorly translated mRNAs were found to aggregate into stress granules.

Surprisingly, the information specifying differential localization and protein production of these two classes of mRNAs is not found directly in the mRNA sequence, but instead is encoded in the promoter sequence driving mRNA production. The authors found that promoter responsiveness to the transcription factor heat shock factor (Hsf1) specifies diffuse cytoplasmic localization and higher protein production upon glucose starvation, whereas promoter elements upstream of poorly translated mRNAs direct these mRNAs to stress granules under glucose starvation.

This alters the current paradigm that transcription and translation are disconnected in eukaryotes; instead, the authors find that these spatially distinct processes are coupled during nutrient limitation. A linkage between transcriptional regulation and cytoplasmic localization may be a general adaptation during times of stress, enabling the cell to coordinately regulate the production of entire classes of proteins. Under non-stress conditions, upregulation of a class of transcripts by a transcription factor would produce similar amounts of protein from each of the mRNAs, as translation would proceed at a generally high rate. Under stressful conditions when overall translation is reduced, selective translation may be required to produce proteins needed for adaptation to the new condition. This suggests that the translation of gene sets can be coordinated using a single nuclear factor, without the need to modulate the sequence of each mRNA.

This work, which will appear in the August 7th issue of Nature, offers a new direction for investigating whether promoter-dependent mRNA localization regulates protein expression in a variety of cell states. The O’Shea lab is currently working to identify factors that may be co-transcriptionally loaded onto mRNAs to determine an mRNA’s fate.

Read more in Nature or download PDF.

This research was supported by the American Cancer Society and the New England Division Funding A Cure initiative (B.M.Z.) and the Howard Hughes Medical Institute (E.K.O.)

View Erin O’Shea’s Faculty Profile