2 mRNAs coding for A 1 and A 2 isoforms of translation factor eEF 1 demonstrate different half-lives while A 1 and A 2 proteins are similarly stable in MCF 7 cells

Eukaryotic translation elongation factor 1A (eEF1A) exists as two 98 % homologous isoforms eEF1A1 and eEF1A2 that are tissue/development specific and differentially linked to apoptosis/cancerogenesis. A2 is overexpressed in a number of tumors while unusual expression of A1 is observed in injured muscles. To approach a possible mechanism underlying induced changes in the relative amounts of the isoforms we examined the intrinsic stability of the proteins and their mRNAs in human cancer cells. Aim. To estimate half-life of the isoforms of eEF1A at mRNA and protein level in human cancer cells. Methods. To measure mRNA stability the transcriptional block technique was applied, with subsequent analysis of the mRNA level by qPCR. To determine the protein decay rate the translation was blocked by cycloheximide and changes in the protein level were detected by Western blot. Results. Calculation of the protein stability revealed half-life of 72 for eEF1A1 and 95 hours for eEF1A2. Half-life of EEF1A1 and EEF1A2 mRNAs were 3 and 60 hours respectively. Conclusions. Despite similar protein stability, the isoforms of eEF1A dramatically differ in the half-lives of their mRNAs, suggesting that the mRNA decay mechanism is one of the main regulators of eEF1A1/A2 amount in MCF7 cancer cells.

Introduction.eEF1A is one of the main components of translational apparatus that provides delivery of aminoacylated tRNA to the A site of ribosome during elongation step of protein biosynthesis [1].eEF1A exists as two isoforms, eEF1A1 and eEF1A2.The reason behind the appearance of two rather than one isoform of eEF1A in higher vertebrates is unknown.During embryonic development eEF1A1 is expressed exclusively and ubiquitously.However, throughout postnatal development in cardiac, muscle and neuronal tissues a switch occurs from the eEF1A1 to eEF1A2 expression.This change is crucial, as the mice with a partial deletion of the EEF1A2 gene die on the 28 th day after birth [2].eEF1A1 and eEF1A2 isoforms are 98 % homologous, so is not surprising that their translation activity is similar [3].However, these proteins are very different in non-canonical functions, for instance, eEF1A1 stimulates apoptosis [4,5] while eEF1A2 demonstrates anti-apoptotic properties [6].Moreover, EEF1A2 is a putative proto-oncogene [7] over expressed in a variety of tumor tissues [8,9].As mentioned before, the expression of A1 and A2 in tissues is mutually exclusive.However there are examples of their induced co-expression under extreme situation like the muscle injury [10] or tumorigenesis [9,11] suggesting special need in appearance of A1 or A2 isoform, respectively, during the background presence of a counterpart protein.Specific non-canonical roles of the isoforms and mechanisms controlling their respective quantities in corresponding tissues are unknown.First step towards elucidating these mechanisms is to examine intrinsic stability of the corresponding mRNAs and proteins.Breast cancer MCF7 cell line is selected for this purpose as these cells demonstrate relatively high expression level of eEF1A2.

Materials and methods.
Cell line, cycloheximide and actinomycin treatment.MCF7 cells were cultured in DMEM («Sigma», USA) growth medium with stabilized L-glutamine, contained 10 % FBS («Sigma») and 1 % Penicillin/Streptomycin («Sigma»).Cells were maintained at 37 °C in a humidified atmosphere containing 5 % CO 2 .To determine mRNAs half-lives, MCF7 cells were treated with 5 µg/ml actinomycin D («Sigma»).Cells were harvested at the time points indicated in the Figures.Amount of EEF1A1/A2 mRNAs were quantified by qPCR.For protein half-life determination, MCF7 cells were incubated with 100 µg/ml cycloheximide («Sigma»).Cells were harvested at the time points indicated in the Figures with consequent Western blot analyses.The Western blot results were quantified by the ChemiDoc system («Bio-Rad», USA) Changes in amount of eEF1A1, eEF1A2 and GAPDH proteins were analysed by Gel Doc software («Bio-Rad»).
Results and Discussion.As eEF1A2 is a putative proto-oncogene and such proteins usually are quickly degraded [13] we expected the life time of eEF1A2 to be significantly less than eEF1A1.There were several datasheets on the eEF1A1 and eEF1A2 half-lives obtained from global proteomic studies examining a vast majority of cellular proteins at once [14][15][16].Unfortunately, these data did not give a definite answer, even in the case the same cell line was used in different studies (Table ).To clarify the issue we measured half-lives of eEF1A1 and eEF1A2 directly, using cycloheximide-treated MCF-7 cells and two kinds of antibodies, the first recognizing both A1 and A2 isoforms and the second being exclusively specific for A2 [17].As the amount of A1 in MCF7 cells is much larger (by orders) than A2 we consider the former antibodies response reflects mostly the A1 amount.Cycloheximide is an antibiotic which specifically blocks elongation step of protein biosynthesis [18].That is why it is often used for protein stability measurements [19].
Surprisingly, the eEF1A1 and eEF1A2 isoforms demonstrated similar decay rates (Fig. 1, A). Subsequent calculations revealed that the both isoforms belong to the class of long lived proteins with estimated halflives of 72 for eEF1A1 and 95 hours for eEF1A2 (Fig. 1, B). GAPDH which is a known long-lived protein was planned to be used as a loading control.However, as the half-life of GADPH (61 h) was found to be comparable with that for eEF1A1 and eEF1A2 (Fig. 1, A) we had to use strictly similar amount of cells for preparing the extract to load per line.Thus, both eEF1A1 and eEF1A2 isoforms possess comparable half-lives in the breast cancer MCF-7 cells.We suggest that protein stabilization does not seem to contribute to cancer-related change in the intracellular eEF1A1/eEF1A2 ratio and the alterations in relative amount of eEF1A1 and eEF1A2 in cancer cells are governed by the level and translation efficacy of their mRNAs.

Protein name
Half-lives of proteins, h HeLa [14] HeLa [16] NIH 3T3 [15] C2C12 [14] eEF1A1 We examined the parameter of mRNA stability by comparing the decay rates of the EEF1A1 and EEF1A2 mRNAs.mRNAs coding for the isoforms are highly homologous in open reading frames (ORF), however, they are rather different in the 3' and 5' untranslated regions (UTRs), which may provide a background for specific posttranscriptional control.The presence of such structural elements as CPE (cytoplasmic polyadenylation element), ARE (AU-rich element) etc. in the EEF1A1 rather than in the EEF1A2 mRNA suggests lower stability of the former.Varieties of methods were developed to quantify the decay rates of mRNA in different cell lines.The most exploited is «transcriptional shut off» method to block transcription by RNA polymerase inhibitors, with subsequent measurement of the residual amounts of mRNAs of the interest by Northern Blot or qPCR.
Therefore, MCF7 cells were treated with actinomycin D to block RNA polymerase II activity.Cells were collected after 2, 4, 8, 28, 48 and 80 h of incubation with consequent analysis of the EEF1A1/1A2 mRNA level by qPCR.Indeed, mRNA of eEF1A2 was found to be considerably more stable than eEF1A1 (Fig. 2).Approximate half-life times were 60 h for eEF1A2 and 3 h for eEF1A1 mRNAs.In the literature, there is the only indication of half-life measurement for the eEF1A1 mRNA [15] where Schwanhausser et al. observed that half-life of the eEF1A1 mRNA is 17,5 h.However, the authors measured mRNA decay rates in non-cancerous mouse fibroblasts (NIH3T3), while our experiments we-re carried out in MCF7 human cancer cells.Consequently such divergence in the results can be partly explained by the difference in 3' UTRs of human and mouse EEF1A1 mRNAs [20].3' UTR of the human EEF1A1 mRNA is much longer and contains several additional instability elements.Besides, rate of the EEF1A1 mRNA decay in cancer and non-cancerous cells may be different.
It is known that transcription of the EEF1A1 mRNA occurs very powerfully, the promoter region of eef1a gene is the second most efficient after cytomegalovirus one [21].Moreover the EEF1A1 mRNA belongs to the TOP class of mRNAs [22] and consequently A1 can be rapidly synthesized in response to the growth stimuli.It should be mentioned that the EEF1A2 mRNA also contains a non-canonical TOP element [23], however its functionality is still questionable.We discover now that to prevent over saturation of a cell with rather stable protein eEF1A1 a novel mechanism was elaborated which includes rapid degradation of EEF1A1 mRNA soon after its synthesis.Thus, repression of transcription of the eEF1A1 mRNA appears to be though not very fast but reliable way to decrease the amount of eEF1A1 protein with subsequent inhibition of translation elongation.The proto-oncogenic eEF1A2 protein may be much less controlled by means of down-regulation of transcription or translation.Interestingly, eEF1A2 is exclusively expressed in such very important cells as myocytes and neurons [3,24].Virtual absence of the possibility to quickly alter its amount may provide additional reliability and stability of the protein synthesis in these terminally dif- ferentiated cells.On the other hand, over expression of eEF1A2 could be very harmful for cells and lead to oncogenesis [8].To control the amount of eEF1A2, a cell has elaborated another mechanism, exactly down-regulation of eEF1A2 by microRNA-663 and 744 recently described by us [25].The organism seems to keep the eEF1A2 expression under a tight control, limiting its appearance in terminally differentiated cells.On the contrary, eEF1A1 is normally found everywhere in the organism except muscles and neurons.The eEF1A1 mRNA, opposite to A2, has short, TOP class 5' UTR and long 3' UTR which opens wide «place d'arme» to the transfactors [20,26].Consequently, this mRNA has much shorter half-life, that also contributes to the regulatory process.
For instance, the neurons respond to electric stimuli via certain parts of the dendrites rather than via whole soma [27].In that case, local protein biosynthesis and/or cytoskeleton reorganization should be activated to afford neuronal plasticity [28].We assume that the eEF1A2 mRNA being stable may provide homeostasis and stability of the soma, while the EEF1A1 mRNA could be responsible for the fast structural changes via cytoskeleton reorganization and local protein biosynthesis in distal compartments of neuron.It is already known that upon LTP (long term potentiation) formation in such distal compartments as dendrites, mTOR-induced translation of the eEF1A1 mRNA is observed [29][30][31].Another example is muscles, where the isoform eEF1A2 is the only one, while the induction of EEF1A1 expression is observed during regeneration processes after muscle injury [10].After healing is finished the eEF1A1 mRNA can degrade quickly because of low half-life, permitting to restore the normal functioning.Interestingly, muscles healing process largely recapitulates embryonic myogenesis where the only eEF1A1 isoform is present [32].
Besides, high level of homology and similar catalytic function during protein biosynthesis, the isoforms of eEF1A are very different in their numerous non-canonical functions.
Combining our and literature data we put forward a novel concept stating that the existence of the two highly homologous isoforms of human eEF1A can be explained by the necessity of cells to control selectively expression of the eEF1A1/A2 isoforms at different levels and by different mechanisms depending on their canonical and non-canonical functions.
eEF1A1 is constitutively expressed in majority of the organs in the organism providing a basic level of protein synthesis responsive to different kinds of regulation.Appearance of eEF1A2 in the same tissues seems not to respond to the same regulatory factors, therefore, may be very harmful for cells.
Conclusions.The data show that the isoforms of eEF1A demonstrate very different half-lives of mRNA but similar stability of corresponding proteins.We assume that the mRNA decay mechanism is one of the main regulators of eEF1A1 expression, at least, in the human breast cancer MCF7 cells.