Mechanisms of rapid mRNA degradation

Rapid mRNA degradation is a means to limit protein translation in time or space. It may have evolved in order to contain the level of proteins that cause cell transformation when over-expressed. There exists evidence of prolonged mRNA half-life in tumor cells. For this reason we study mechanisms of rapid mRNA decay. The mRNA of about 5 percent of all genes is rapidly degraded in the cytoplasm. This is best documented for mRNA of certain cytokines and growth factors, like IL‑2, IL‑3, TNF‑α, G‑CSF GM‑CSF or VEGF, as well as growth-inducing transcription factors, like c‑Myc and c‑Fos. They all decay with half-lives of 30 to 60 min. For some of them it is known that AU‑rich sequences in the non-coding 3’‑untranslated regions are necessary and sufficient to induce rapid degradation. For others, rapid degradation may require more than one 3’‑element. c‑myc and c‑fos mRNAs contain additional “coding region determinants” that destabilize the RNA, in concert with 3’‑elements.

We have constructed retroviral vectors for the tet-off system which permit to turn off rapidly transcription of green fluorescent protein (GFP) cDNA reporter constructs without affecting general cellular transcription. With this method mRNA decay can be accurately measured by RT‑PCR. We graft cis-acting elements from 3’‑untranslated or coding regions of unstable mRNAs into stable GFP mRNA and study their effect. In parallel we have constructed retroviral vectors for RNA interference and for inducible expression of myc-tagged trans-acting proteins. Thus, we can strongly reduce or increase the level of candidate proteins and study their importance in mRNA stability of a given construct. We further use myc-tagged proteins combined with immunoprecipitation to analyze RNA‑protein and protein‑protein interactions in cells. With these tools we have mainly analyzed the rapid degradation of IL‑6 and c‑myc mRNA. The goal of these investigations is to understand each step from the recognition of a specific element by a trans-acting protein to the actual RNA decay. Beyond, it will be of interest to look in cancer cells with high RNA stability whether we can pinpoint changes induced by cell transformation.

In human IL‑6 mRNA we have identified two instability elements in the 5’‑half of the 3’‑untranslated region. The first one corresponds to a sequence resembling an AU-rich element (ARE). The second one, 80 nucleotides further 5’, comprises a sequence predicted to form a stem-loop structure. Neither element alone was sufficient to confer strong instability suggesting that they might cooperate. IL‑6 mRNA was specifically co-immunoprecipitated from cells expressing myc‑tagged AUF1 p37 and p42 isoforms, indicating their RNA binding. This interaction required the ARE. Suppression of endogenous AUF1 by RNA interference stabilized IL‑6 mRNA indicating a functional role for AUF1 in IL‑6 mRNA degradation. However, overexpression of AUF1 p37 and p42 also stabilized IL‑6 mRNA and this required the ARE. It suggests that excess AUF1 may either titrate out an AUF1‑interacting degradation factor or it may compete with other decay-promoting proteins for binding to the ARE. Alternatively, excess AUF1 might modify protein-protein interactions on the mRNA and thereby hinder its own degradation promoting function. Similar observations have been made for other ARE‑containing mRNAs.

For mouse c‑myc mRNA we found that both the coding region and 3’‑untranslated region conferred instability to a GFP reporter mRNA. However, only the coding region was sufficient to induce a 45 min half-life, whereas the 3’‑untranslated region showed a half-life of 120 min. In agreement with previous studies on human c‑myc mRNA, destabilization by c‑myc coding sequences required mRNA translation. With stop codons at different positions, instability increased the further translation was allowed to advance. For full degradation, translation had to proceed up to the first half of exon 3. It was not possible to attribute instability to a single location in the mRNA. Shorter reporter constructs rather indicated multiple elements in exons 2 and 3 that may either cooperate or act sequentially to destabilize c‑myc mRNA. Translation of the c‑myc coding sequence triggers rapid deadenylation which may precede the decay of the RNA body.

Construction of a conditional ferritin H knockout mouse

Iron is an essential metal for life and at the same time a hazard since, in its free form, it catalyzes the formation of hydroxyl radicals, which are thought to be a natural cause of mutations in DNA. Therefore, free iron must be delicately controlled to avoid deprivation or excess. The protein complex of ferritin stores excess free iron and is thought to play a central role in the protection against radical formation. To test this hypothesis, we have constructed a mouse strain with a conditional ferritin H gene knock-out.

In our new mouse strain we use the Cre‑Lox strategy to delete the ferritin H gene and take advantage of crosses with mice harboring an inducible Cre. We find that deletion of ferritin H in the germ-line provokes lethality of embryos. However, when ferritin H is deleted to 95% in the liver, spleen and bone marrow of mice at 10 weeks of age, they survive without visible disadvantage. We are presently analyzing various parameters of changes in iron physiology and the immune system after ferritin H deletion. One of the striking findings is an induction of hepcidin mRNA in liver and a repression of iron transporter mRNAs in intestine. This is compatible with a physiological feedback mechanism by which the body iron level is controlled (Fig. 1). The hypothesized signal transduction pathways will guide us in future experiments. To this end we plan to cross our mice with mouse strains defective in HFE or TfR2.

In order to test the effects of the loss of ferritin H in tissue culture cells we have derived cell lines from primary embryonal fibroblasts. We notably find a 30‑fold increased toxicity of iron salts in cells that lack the ferritin H gene. It indicates that ferritin H is necessary to protect cells from free iron. By adding compounds that prevent the formation of reactive oxygen species, ferritin H knockout cells survive better. This is compatible with the idea that iron toxicity acts through the formation of reactive oxygen species. We hope to gain further insight into the question whether the loss of ferritin H contributes directly to DNA mutagenesis in cell cultures and in vivo. When human patients with hemochromatosis are inappropriately treated they show a strongly increased rate of liver cancer, which is thought to arise from the continuous insult by excess iron. The new mouse strain should help us to clarify this issue.

Figure 1:   Model of a regulatory loop which controls intestinal iron absorption.
Hepcidin is a peptide hormone secreted by the liver and normally induced by high body iron levels. It is known that high hepcidin levels repress intestinal iron transport proteins required for nutritional iron absorption. This feedback mechanism is perturbed in hemochromatosis patients, which accumulate too much iron. The deletion of ferritin H in the liver of adult mice induces hepcidin mRNA and decreases intestinal iron transporter mRNAs, compatible with the idea that stored iron is released when ferritin disappears. It is presently unknown what signals control the hepcidin gene and whether hepcidin exerts its effects through an intestinal receptor. Mutations in HFE and transferrin receptor 2 (TfR2) provoke hemochromatosis, for reasons that are not understood, but might be due to defects in signaling.

Keywords

mRNA degradation, IL‑6 in cancer, ferritin knockout mice