Subtlety is underrated: Physiological UPR activation in meiosis

Discovery of various stress factors in cell biology in the 1980s–1990s—with its emphasis on Hammer time—has mirrored the golden age of hip-hop: wallop cells with a large dose of something noxious, find sensitive mutants, organize them by epistasis, and make the resulting pathway synonymous with its less-than-subtle stimulus. With some help from nominative determinism, the first stress response to be molecularly defined, namely, the heat shock response, conjured sensationalist images in the public mind: “This is your cell. This is heat shock. This is your cell without the heat shock response. Any questions?” Despite this attempt to deter contrarianism, some researchers began to wonder: can cells use stress response pathways to adapt to physiological fluctuations and respond to environmental stimuli, not just to cope with disaster?

Yeast gametogenesis offers a tractable system for studying a physiological UPR stress that accompanies a cell differentiation program. In response to starvation, diploid budding yeast cells undergo meiosis followed by sporulation to form hardy haploid gametes (spores). A previous genomic analysis of translation during this developmental process observed that Hac1, the transcription factor for the UPR, is synthesized in its active form in two discrete waves: once before cells enter meiosis and again in mid-meiosis (1) (Fig. 1 A). Both waves occur without the addition of any stress other than the media change that activates cell signaling to induce sporulation, suggesting UPR activation is a physiological response.

The authors in reference (2) have now determined that the first wave of UPR activation has an adaptive function in sporulation while the second wave may not. To demonstrate this, the authors first showed that hac1Δ cells arrested before the first meiotic division and largely failed to form spores. They then used a Tet-inducible promoter to generate only a pulse of active Hac1 during premeiotic growth but deactivated it before transferring cells to sporulation medium. Despite the absence of the second wave of Hac1 activation in these engineered cells, they still sporulated with wild-type efficiency and timing. Thus, UPR activation in the first wave is sufficient for cells to go through meiosis, so the second wave is therefore dispensable for gamete production.

Despite convincing evidence for Hac1’s essential role in meiosis, the authors in reference (2) observed only a mutated transcriptomic UPR signature. Of the 406 UPR target genes defined in yeast in response to drug treatments that severely disrupt ER protein folding, only a few were induced in a Hac1-dependent manner in premeiotic cells (3). KAR2, encoding the UPR-defining Hsp70 chaperone (yeast BiP), was the only promoter to which Hac1 bound. When the authors overexpressed active Hac1 in mid-meiotic cells, the full canonical program came back on, so the chromatin is permissive to UPR activation. The exquisite sensitivity of KAR2 as a UPR gene target parallels a previous finding in mitotic cells. Here, KAR2 was among the small set of genes that were induced even at very low levels of Hac1 expression by engineered titration (4). This set of experiments demonstrates that transient physiological UPR activation during the meiotic program induces a qualitatively different transcriptional state than acute pharmacological ER stress.

Although the UPR is not strongly induced, premeiotic cells do undergo substantial transcriptional reprogramming compared with mitotic cells. This most obvious change to the transcriptome is the repression of ribosomal protein and biogenesis genes, ultimately resulting in reduced protein synthesis. The decrease in ribosome biogenesis and translation turns out to be more important than activation of the UPR for sporulation: slowing translation by a variety of means bypassed the requirement for Hac1 in meiosis. Low doses of cycloheximide, lactimidomycin, or rapamycin, genetic perturbations to ribosomal proteins, or reducing protein translocation into the ER lumen via mutations or by artificially clogging the translocon (5) all rescued sporulation in hac1Δ cells (Fig. 1, B–D). The implication here is that Hac1 somehow inhibits protein synthesis and/or ribosome biogenesis (see below).

When attempting to sporulate, there are cell biological consequences for cells lacking HAC1. Premeiotic hac1Δ cells accumulate puncta containing reticulons, organelle morphogenetic proteins that tubulate the ER (6). The foci colocalize with general ER markers and do not appear to be protein aggregates. Interestingly, this microscopic phenomenon is not fully penetrant, but it predicts meiotic outcome. Only cells containing foci fail to complete meiosis, while the few hac1Δ cells lacking foci (presumably those with stochastically low levels of protein synthesis) sporulate nearly normally. Furthermore, every manipulation that rescues meiosis also prevents focus formation. It is unclear what these structures represent, but their reticulon specificity suggests underlying aberrant membrane geometry. Similar foci have recently been observed following acute ER stress in mitotic cells (7), suggesting this may be a generic failure mode during unsubtle ER stress—i.e., only when the stress surpasses the capacity of the UPR or the UPR is genetically compromised.

How does Hac1 repress ribosome biogenesis/translation? The authors find a promising lead by presenting evidence for direct gene regulation of DOT6, which encodes a repressor of RiBi genes. However, the challenge in interpreting these transcriptomic data is that ribosome gene expression inextricably tracks growth rate (8), so it is not trivial to separate cause from consequence in starvation media, particularly when comparing fold changes between genotypes whose baselines differ. Moreover, global mRNA composition is systematically constrained by growth rate, so that much of any transcriptome (not just the ribosomal genes) is reporting on how fast cells are dividing rather than on any regulatory input (9, Preprint). Hac1 protein and/or HAC1 mRNA may have several contributions to growth rate, affecting gene expression at multiple levels.

Given these new results, it appears that budding yeast have at least two mechanisms of translational repression under ER stress, the Hac1-dependent one described here, and delayed Ras/PKA signaling that operates independently of the UPR but converges on Dot6 (4). This yeast redundancy mirrors the situation in metazoans but with a convergent twist. In the latter, one ER stress receptor (PERK) signals inhibition of translation initiation, while another (Ire1, the same receptor that also activates the metazoan version of Hac1) causes ER-associated mRNA decay.

The broader lesson may be that interrogating stress responses by unleashing chemical or genetic catastrophe on the cells tells you what the pathways can do but not necessarily what the cell uses them for. The specific ecological history of each organism selects adaptive functions and retains, refines, and redeploys parts as fitness demands. In ecology and evolution, we often find subtlety where chance met necessity.

This work was supported in part by National Institutes of Health/National Institute of General Medical Sciences grants RM1 GM153533 (to D. Pincus) and R35 GM127136 (to V. Denic) and the National Science Foundation grant QLCI QuBBE OMA-2121044 (to D. Pincus).

Author contributions: David Pincus: conceptualization, funding acquisition, and writing—original draft, review, and editing. Vladimir Denic: writing—review and editing.