RESEARCH ARTICLE

Regulation of global translation during the cell cycleVilte Stonyte, Erik Boye and Beata Grallert*

ABSTRACTIt is generally accepted that global translation varies during the cellcycle and is low during mitosis. However, addressing this issue ischallenging because it involves cell synchronization, which evokesstress responses that, in turn, affect translation rates. Here, we haveused two approaches to measure global translation rates in differentcell-cycle phases. First, synchrony in different cell-cycle phases wasobtained involving the same stress, by using temperature-sensitivemutants. Second, translation and DNA content were measured byflow cytometry in exponentially growing, single cells. We found nomajor variation in global translation rates through the cell cyclein either fission yeast or mammalian cells. We also measuredphosphorylation of eukaryotic initiation factor-2α, an event that isthought to downregulate global translation in mitosis. In contrast withthe prevailing view, eIF2α phosphorylation correlated poorly withdownregulation of global translation and ectopically induced eIF2αphosphorylation inhibited global translation only at high levels.

KEY WORDS: Translation, Cell cycle, EIF2α phosphorylation

INTRODUCTIONIt is one of the basic principles of cell proliferation that there is a linkbetween general cell growth (protein synthesis) and cell-cycleregulation. Such a link is logical and has been hypothesized to exist,but its nature has been elusive. Protein synthesis is one of the mostenergy-demanding cellular processes and is, therefore, carefullyregulated. It is a generally accepted view that global translation isconsiderably reduced during mitosis (reviewed by Sivan and Elroy-Stein, 2008). The reduction is thought to result from alteredphosphorylation state of translation initiation factors. In particular,phosphorylation of the eukaryotic translation initiation factor 2alpha (eIF2α) is induced after a number of different stresses and isthought to be the main reason for repressed translation. Cell-cycle-dependent downregulation of translation in G2/M phase was alsoattributed to increased eIF2α phosphorylation (Datta et al., 1999;Kim et al., 2014; Silva et al., 2015; Tinton et al., 2005).Early translation measurements in synchronizedmammalian cells

revealed a 70% reduction of the global translation rate duringmitosis (Fan and Penman, 1970). More recent studies using differentsynchronization methods suggested that the magnitude of thetranslation reduction depends on the method of synchronization(Coldwell et al., 2013; Shuda et al., 2015b). Also, studies usingbudding yeast indicated that the rate of protein synthesis is constantduring the cell cycle (Elliott and McLaughlin, 1978; Elliott et al.,1979). More recent studies (in mammalian cells) have reported

conflicting results regarding the level of translational reductionduringmitosis (Shuda et al., 2015a; Stumpf et al., 2013; Tanenbaumet al., 2015), and the question of whether and to what extent globaltranslation is downregulated during mitosis remains unanswered.

Measurement of translation in different cell-cycle phases ischallenging because it often involves cell-cycle synchronization,which – in itself – can evoke stress responses that, in turn, will affecttranslation rates. Thus, the exact contribution of the synchronizationmethod versus cell-cycle progression to any observed change intranslation rates or the phosphorylation state of translation initiationfactors is difficult to assess. Here, we use novel approaches tomeasure global translation rates during the cell cycle and whetherthey depend on eIF2α phosphorylation.

RESULTSGlobal translation in synchronized cellsFirst, we utilized temperature-sensitive fission yeast mutants thatarrest at different phases of the cell cycle. We synchronized the cellsby shifting to the restrictive temperature before release into the cellcycle, achieving synchrony at different cell-cycle phases in responseto the same treatment (i.e. temperature shift). Samples for analysisof DNA content, translation rate and eIF2α phosphorylation weretaken every 20 min for 160 or 220 min after release from cell-cyclearrest. DNA content and translation rate were measured in singlecells, by flow cytometry. Translation was assayed by pulse-labelingwith the methionine analogue L-homopropargylglycine (HPG)(Knutsen et al., 2011), which is incorporated into growingpolypeptide chains. It should be noticed that our assay addressesthe regulation of global translation rather than the well-establishedtranslational regulation of individual proteins. To reveal smalldifferences in signal intensity, the samples were barcoded (seeTranslation assays in Materials and Methods) and processedtogether in the very same solution. Barcoding involves stainingeach sample with an amino-reactive fluorescent dye that binds toamine functional groups present primarily on lysine side chains andat the N-terminus of a protein. Since the amount of dye that haslabeled the protein is covalently attached to the cells, excess dyecan be washed away. By staining cell samples with differentconcentrations of the dye, samples are imparted with unique dye-intensity distributions (Krutzik and Nolan, 2006). Barcodedsamples were then pooled and processed together to measuretranslation rates. Using this approach eliminates differences due totechnical issues and samples can be compared with great accuracy.The cdc10-M17mutant was used to synchronize cells in G1, cdc25-22 was used to synchronize cells in G2 and nda3-KM311 was usedto arrest the cells in mitosis. Phosphorylation of eIF2α was assessedby immunoblot analysis.

The rate of translation changed as the cells progressed from theblock and through the cell cycle, apparently consistent with cell-cycle-dependent translation. However, the changes in translationrate followed the same pattern after release from the cell-cycle arrestregardless of when in the cell cycle the cells were synchronized(Fig. 1A–D, Fig. S1). At early time points the translation rate wasReceived 14 May 2018; Accepted 20 July 2018

Department of Radiation Biology, Institute for Cancer Research, Montebello,0379 Oslo University Hospital, Oslo, Norway.

*Author for correspondence (beata.grallert@rr-research.no)

B.G., 0000-0001-5917-0555

1

© 2018. Published by The Company of Biologists Ltd | Journal of Cell Science (2018) 131, jcs220327. doi:10.1242/jcs.220327

Journal

ofCe

llScience

low and after release it gradually increased to a rate above thatmeasured before the shift. At late time points translation ratesbecame similar to those measured in exponentially growing cells.There was no correlation between any particular cell-cycle phaseand increase or decrease in translation rates. These results stronglysuggest that global translation is not regulated in a cell-cycle-dependent manner and that the variations observed are caused by thesynchronization.To test the effects of a temperature shift, wild-type fission yeast

cells were subjected to the same changes in temperature asemployed to synchronize the cell-cycle mutants. Interestingly,translation rates followed the same pattern in wild-type cells and thecell-cycle mutants described above (Fig. 1A,C), demonstrating that

the observed changes are due to the temperature shift rather than tothe stage during the cell-cycle at which the particular mutant arrests.Furthermore, in itself, the change in temperature from 25 to 36°Cand back to 25°C induced a transient delay of G2 onset (Fig. S1G),which is probably due to the previously described Rad3ATR-Rad9-dependent mechanism (Janes et al., 2012). Curiously, a shift from30 to 20°C and back to 30°C also induced a cell-cycle delay, but inG1/S (Fig. S1H).

Phosphorylation of eIF2α was high at the early time points in theheat-sensitive mutants, then gradually diminished (Fig. 1E,F),regardless of where in the cell cycle the particular mutant wasarrested. There was no correlation between eIF2α phosphorylationand any particular cell-cycle phase. We used α-tubulin as loading

Fig. 1. Global translation in cells synchronized in the cell cycle. Cells of the indicated strains were grown exponentially at 25°C (A,B,E–G) or 30°C (C,D,H,I),incubated at 36°C or 20°C for one generation time and then shifted back to 25 and 30°C, respectively. Samples were taken at the indicated times after the shift.(A,C) Median intensities of the AF647 (HGP) signal normalized to that of exponentially growing cells. Average of three biological repeats and ± standard errors(±s.e.) are shown. (B,D) Illustrations of cell-cycle progression in the respective mutants. Fig. S1 shows the cell-cycle distributions. (E–I) Quantification ofphosphorylated eIF2α normalized to tubulin in the indicated strains. Average and ±s.e. of three independent experiments are shown. Representative immunoblotsare shown in Fig. S2.

2

RESEARCH ARTICLE Journal of Cell Science (2018) 131, jcs220327. doi:10.1242/jcs.220327

Journal

ofCe

llScience

control, since quantification of the total eIF2α is difficult in fissionyeast due to the lack of a good antibody. To exclude the possibilitythat the ratio of eIF2α to α-tubulin changes during the temperature-shift experiments and/or during the cell cycle, leading us to incorrectconclusions, we tested whether the amount of eIF2α varies ascompared to α-tubulin. There was no change in the eIF2α:α-tubulinratio (Fig. S2A,B), which allowed us to normalize the eIF2α-Psignal to α-tubulin.As a control, and to assess synchrony achieved in the above

experiments, we followed expression of the G1 cyclin Cig2 byimmunoblotting. The previously reported cell-cycle-dependentregulation was obvious in all three strains (Fig. S1), showing thatthe synchrony achieved in the above experiments allows us to detectcell-cycle-dependent changes in protein levels. Furthermore, thetemperature shift resulted in increased levels of phosphorylatedeIF2α also in the wild-type cells (Fig. 1G), confirming that suchtemperature shifts, routinely employed in cell-cycle synchronizationexperiments, invoke a stress response.When cells were shifted from 20 to 30°C, changes in the amount of

phosphorylated eIF2α were much less pronounced, be it in wild-typecells or the cold-sensitive nda3mutant (Fig. 1H,I). Notably, the nda3mutant arrests in metaphase, the very cell-cycle phase during whichlevels of phosphorylated eIF2α are thought to increase and contributeto a downregulation of translation. Furthermore, the biggest change intranslation rate was observed in those cells shifted from 20 to 30°C,both for wild-type cells and the nda3mutant (Fig. 1C), although thistreatment resulted in the smallest change in the amount ofphosphorylated eIF2α (Fig. 1H,I). These results directly contradictthe prevailing view that eIF2α phosphorylation correlates with and isthe reason for downregulation of global translation.To assess the contribution of eIF2α phosphorylation to the observed

changes in translation rates, strains carrying non-phosphorylatable

eIF2α-S52A were used. Cell-cycle synchronization experiments andtranslation measurements were performed as above. Surprisingly,translation rates followed exactly the same pattern in the absence ofeIF2α phosphorylation as in its presence; low immediately after thetemperature shift, then recovering (Fig. 1A,C). Furthermore, in theheat-sensitive mutants translation was much more downregulatedwhen eIF2α could not be phosphorylated (Fig. 1A).

We conclude that the changes in translation rates during the cell-cycle synchronization experiments were not due to cell-cycle-specific regulation of translation, but to the temperature shift itself.Furthermore, phosphorylation of eIF2α is not cell-cycle regulatedand is not required for the downregulation of global translation aftertemperature shift.

Global translation in exponentially growing cellsHaving seen no evidence of cell-cycle-dependent regulation oftranslation in synchronized cells, we set out to measure translationrates in different cell-cycle phases in non-synchronized cells. To thisend, we measured HPG incorporation and DNA content inexponentially growing cells by flow cytometry. Cells in eachcell-cycle phase were gated on two-parametric DNA cytograms(Knutsen et al., 2011) and HPG incorporation per cell wasquantified in each cell-cycle phase. There were no significantdifferences in the rate of translation in the different cell-cycle phases(Fig. 2A,C). It should be noted that this method does not allow us todistinguish cells in mitosis from those in G1. Thus, a hightranslation rate in G1 cells might compensate for a reducedtranslation rate in the mitotic cells so that the relative translation ratefor the mixed M-G1 population appears to be unchanged. However,in such a scenario the distribution of the HPG intensities in theM-G1 population would be broad, but this is not the case(Fig. 2A,C), which argues against this explanation. Another

Fig. 2. Global translation in exponentiallygrowing cells. (A,B) Two-parametric flowcytometry plots of fission yeast cells grown in EMM(A) or isoleucine-minimal medium (B).(C,D) Average of median intensity of the AF647signal normalized to G2 (C) or G1 (D) from at leastthree biological repeats with ±s.e. Gating is shownin Fig. S2.

3

RESEARCH ARTICLE Journal of Cell Science (2018) 131, jcs220327. doi:10.1242/jcs.220327

Journal

ofCe

llScience

concern is that a low number of mitotic cells in the population wouldconceal a low translation rate in mitotic cells. To address this issue,cells of the M-G1 population were sorted onto microscopy slidesand the microtubules were stained. At least 20% of the cells clearlycontained a mitotic spindle (data not shown), demonstrating that thetranslation rates measured in the M-G1 population reliably representthose of mitotic cells. In addition, we analyzed exponentiallygrowing fission yeast cells grown in a medium with isoleucine assole nitrogen source. Under these conditions G1 is longer andcytokinesis occurs in G1 (Carlson et al., 1999), which allows us todistinguish a G1 population containing 1C DNA from mitotic cells.Also under these conditions, translation rates were similar in thedifferent cell-cycle phases (Fig. 2B,D). These results obtained innon-synchronized, exponentially growing cells confirm that globaltranslation does not vary significantly through the cell cycle.Basic cellular processes, such as regulation of translation through

the cell cycle are expected to be conserved in evolution, but theextent of such regulation might vary from organism to organism.Therefore, we investigated whether the level of global translationvaries during the cell cycle in human cells. To this end, wemeasuredtranslation rates in different cell-cycle phases in three differenthuman cell lines. To measure translation, non-synchronized cellswere pulse-labeled with the puromycin analogue O-propargyl-puromycin (OPP) and analyzed by flow cytometry. Cells in G1, Sand G2 were identified on the basis of their DNA content andmitotic cells were identified using the mitotic marker histone 3 (H3)phosphorylated at serine residue 10 (phospho-S10-histone H3).The cell lines investigated were normal epithelial RPE cellsimmortalized by telomerase expression, the osteosarcoma-derivedU2OS cells and cervix carcinoma-derived HeLa cells. There is awide distribution of the intensity of the OPP signal in the G1population, indicating that there are significant differences intranslation rates among G1 cells. This feature is particularly obviousin normal RPE cells but less pronounced in the two cancer cell lines(Fig. 3). The G1 cells with lower translation rates might representcells that have not yet passed the restriction point. There is a gradualincrease in translation from G1 phase through S to G2 in all threecell lines, and a somewhat lower rate in mitotic cells. However, therate of protein synthesis in mitotic cells is higher or similar to that inG1 cells and the extent of reduction from G2 to M ranges from 40%(RPE) to 15% (U2OS). Cells of each cell line were also fixed formicroscopy after pulse labeling with OPP. Mitotic cells wereidentified by tubulin staining. Consistent with the results of flowcytometry, there are no major differences in the intensity of the OPPsignal between mitotic and interphase cells (Fig. 3G).Phosphorylation of eIF2α was investigated in HeLa cells. Non-

synchronized cells were fixed, cell-cycle stagewas analyzed as aboveand cells were collected by fluorescence-activated cell sorting(FACS). Phosphorylation of eIF2α was investigated in the differentpopulations by immunoblotting. Therewere no significant changes inthe levels of eIF2α phosphorylation during the cell cycle (Fig. 3H).The above results strongly suggest that the previously observed,

apparently cell-cycle-dependent, variation in translation rates wasinstead a result of synchronization. In order to directly address this,we synchronized HeLa cells by using Nocodazole and mitoticshake-off, and measured the translation rates. Consistent withprevious studies, translation rates changed dramatically inNocodazole-treatated cells (Fig. 3I,J) and levels of phosphorylatedeIF2α increased upon Nocodazole-induced arrest (Fig. S3).However, in light of our results above, these dramatic changes areunlikely due to cell-cycle regulation but, rather, to the stressresponse following treatment with Nocodazole.

These findings strongly suggest that global translation rates arenot dramatically downregulated in mitotic cells and that earlierstudies overestimated the extent of variation through the cell cycle.

Phosphorylation of eIF2α and global translationSurprisingly poor correlation was observed between the levels ofphosphorylated eIF2α and global translation in the temperature-shift experiments. Previous work demonstrated that eIF2αphosphorylation can attenuate the translation of mRNAs (Hardinget al., 2003; Hinnebusch, 1994). However, this is not the onlyconsequence of eFI2α phosphorylation and the primary role mightbe to upregulate the translation of a specific class of mRNAs(reviewed by Dever, 2002).

To directly address the importance of eIF2α phosphorylation onglobal translation rates, we expressed eIF2α kinase 2 (EIF2AK2,hereafter referred to as PKR), one of the four human eIF2α kinases,in fission yeast and measured phosphorylated eIF2α and globaltranslation rates. PKR expression was controlled by the regulatablenmt1 promoter, which is induced upon thiamine removal from themedium (Basi et al., 1993; Maundrell, 1993). We used two differentversions of the promoter, resulting in two different expression levelsof PKR. Cells were grown exponentially with the promoterrepressed before PKR expression was induced, and globaltranslation rates as well as eIF2α phosphorylation were measuredduring the first 24 h (in six generations) after induction. PKRexpression was detected at 13 h after induction and levels ofphosphorylated eIF2α reached their maximum at 16–19 h (Fig. 4A,B, Fig. S4). The extent of eIF2α phosphorylation induced by PKRdriven by the weaker promoter was comparable to that induced bymilder stresses (Fig. 4C, Fig. S4) but, surprisingly, we did not seeany significant decrease in global translation rates in this case. Therate of translation remained similar to that before induction of PKRexpression. (Fig. 4D). However, in cells that expressed PKR via thefull-strength nmt promoter, translation was strongly reduced and,consistently, these cells did not form colonies when the promoterwas derepressed (not shown). These results are consistent withprevious findings, suggesting that extreme and lasting eIF2αphosphorylation can inhibit global translation and is lethal (Deveret al., 1993; Zhan et al., 2002). We conclude that the extent of eIF2αphosphorylation is crucial for its effect on downregulation of globaltranslation. Very high levels of phosphorylated eIF2α blocktranslation, but intermediate levels seem to have little influence onglobal translation.

DISCUSSIONGlobal translation rate changes little during the cell cycleMany recent studies dispute the generally accepted view that globaltranslation varies in a cell-cycle-dependent manner and is lowduring mitosis. Our results suggest that the discrepancies arise fromexperimental challenges. Studies of cell-cycle-related events ofteninvolve synchronization of cell cultures. In our work here, weemployed temperature-sensitive yeast mutants. It should be noticedthat studies on heat stress generally employ higher temperatures(>40°C) and the temperatures we used are close to those common inthe natural environment of fission yeast cells. However, here weshow that even the temperature shifts routinely used to synchronizethe temperature-sensitive S. pombe mutants invoke a cellular stressresponse by themselves and influence global translation rates,supporting the idea that previously reported cell-cycle-dependentchanges in translation rates are caused by the method ofsynchronization. Using the same stress to synchronize cells indifferent cell-cycle phases allowed us to separate the effects of

4

RESEARCH ARTICLE Journal of Cell Science (2018) 131, jcs220327. doi:10.1242/jcs.220327

Journal

ofCe

llScience

Fig. 3. Global translation through the cell cycle in human cells. (A–C) Two-parametric flow cytometry plots of the indicated cell lines. Yellow lines represent themean intensity of AF647 (OPP) for each cell-cycle phase. (D–F) Bar graphs representing mean AF647 (OPP) intensity ±s.d. (G) Microscopic images of RPE, HeLaandU2OS cells labeled with OPP. Tubulin and DNA staining are shown for the identification of mitotic cells (arrowheads). (H) Quantification of phosphorylated eIF2αnormalized to eIF2α in the indicated cell-cycle phases. Exponentially growing HeLa cells were fixed, and stained for H3-P and DNA content to identify cells in eachcell-cycle phase. Then 50,000 cells from each phasewere sorted tomeasure the level of phosphorylated eIF2α. Average and ±s.e. of three independent experimentsare shown. Representative immunoblots are shown in Fig. S3. (I) Two-parametric flow cytometry plots of asynchronously growing andNocodazole-arrested cells andcells 4 h after release from the Nocodazole block. (J) Bar graphs representing mean AF647 (OPP) intensity ±s.d. after Nocodazole block and release. Levels ofphosphorylated eIF2α versus eIF2α phosphorylation are shown in Fig. S3.

5

RESEARCH ARTICLE Journal of Cell Science (2018) 131, jcs220327. doi:10.1242/jcs.220327

Journal

ofCe

llScience

cell-cycle progression from temperature shift on global translationrates. It is possible that, in our experiments, modest cell-cycle-dependent variations in global translation rates are concealed byimperfect synchrony. However, the synchrony achieved in theblock-and-release experiments (Fig. S1) should have allowed us toobserve the dramatic changes described previously. Furthermore,using flow cytometry to measure translation in exponentiallygrowing cells allowed us to investigate global translation rates indifferent cell-cycle phases in non-stressed cells.One caveat of analyzing the cell cycle of fission yeast by flow

cytometry is that mitotic cells can only be identified after separation ofthe daughter nuclei, but cells in the early phases of mitosis cannot bedistinguished from cells in G2. Thus, a reduction of global translationrates in metaphase would not be detected when using asynchronouslygrowing cells and flow cytometry alone, although it would have beendetected in the block-and-release experiments. Collectively, thesedata demonstrate that global translation is not significantly differentbetween any of the cell-cycle phases in fission yeast cells.

In the human cell lines we also saw only small changes inthe translation rate, consistent with recent studies reporting onlyminor variations. Mitotic cells were identified on the basis ofphosphorylated histone H3, a mitotic marker that is present both inmetaphase and anaphase. Notably, our approach did not involve anysynchronization method, exposure to chemicals or changes in thecellular environment, which makes our results less subject to artifactsand technical problems. Furthermore, when we synchronized thecells, we also observed the previously reported variations, confirmingthe notion that the changes in translation are due to thesynchronization-induced stress rather than cell-cycle progression.

Physiological levels of phosphorylated eIF2α do notsignificantly repress global translationUnder stressful conditions cells reduce the rate of global translationto conserve resources (Holcik and Sonenberg, 2005). At the sametime, synthesis of proteins necessary to survive the stress ismaintained or even increased. Many different forms of stress result

Fig. 4. Phosphorylation of eIF2α and global translation.Cells carrying the indicated plasmids were grownexponentially with the promoter repressed and one samplewas taken to measure translation. The promoter wasinduced for the indicated times. (A,B) Quantification ofphosphorylated eIF2α (eIF2α-P) phosphorylationnormalized to α-tubulin at the indicated time points whenPKR is expressed from the two different promoters. Noticethe different scales on the y-axes. Representativeimmunoblots are shown in Fig. S4. (C) Quantification ofphosphorylated eIF2α normalized to tubulin after theindicated stresses. Average and ±s.e. of three independentexperiments are shown. Representative immunoblots areshown in Fig. S4. (D) Median intensities of the AF647 (HGP)signal normalized to that of exponentially growing cells(promoter repressed). Average of three biological repeatsand ±s.e. are shown.

6

RESEARCH ARTICLE Journal of Cell Science (2018) 131, jcs220327. doi:10.1242/jcs.220327

Journal

ofCe

llScience

in phosphorylation of eIF2α in eukaryotic cells (Clemens, 2001;Sonenberg and Hinnebusch, 2009), which it is thought to berequired for both responses – downregulation of global translationand upregulation of translation of certain mRNAs. In addition,phosphorylation of eIF2α is also implicated in the cell-cycle-dependent regulation of translation. Here, we find that increasedlevels of phosphorylated eIF2α do not correlate with any particularcell-cycle phase but, rather, with the stress involved insynchronization, be it temperature shift or exposure toNocodazole. We conclude that eIF2α phosphorylation is notregulated in a cell-cycle-dependent manner.There is compelling evidence that eIF2α phosphorylation can

attenuate the translation of mRNAs (Harding et al., 2003; Hinnebusch,1994). The regulation of eIF2α phosphorylation is relevant for anumber of diseases, such as neurodegenerative disorders, cancer andautoimmune diseases (Fullwood et al., 2012; Koromilas, 2015;Marchal et al., 2014; Ohno, 2014; Ravindran et al., 2016; Way andPopko, 2016). In all these fields, increased levels of phosphorylatedeIF2α has commonly been taken to be a readout of reduced globaltranslation. However, the two parameters have rarely been measured inthe same experiment. Furthermore, at least under some stress situationsother initiation factors can substitute for eIF2, as recently shown foreIF5B under hypoxia (Ho et al., 2018). Our results also suggest poorcorrelation between eIF2α phosphorylation and repressed globaltranslation. First, eIF2α phosphorylation is clearly not required forthe temperature-shift-induced downregulation of translation (Fig. 1),consistent with previous findings after UVC irradiation, oxidativestress and ER stress (Hamanaka et al., 2005; Knutsen et al., 2015;Shenton et al., 2006). Second, in the absence of eIF2α phosphorylationtranslation is repressed not less but rather more dramatically aftertemperature shift (Fig. 1). Third, ectopically induced eIF2αphosphorylation did not noticeably downregulate global translationin unstressed fission yeast cells, unless it was induced to high levels(Fig. 4).We suggest that the impact of phosphorylated eIF2α on globaltranslation has been overestimated in the literature and that eIF2αphosphorylation cannot be used as a marker of downregulatedtranslation. Our results demonstrate that the amount of phosphorylatedeIF2α is crucial to determine whether it impacts on global translationand it has only a minor effect on the global translation at levelsobserved after mild stresses, mimicked by the expression of PKR fromthe weaker promoter in this study. In contrast, for the typical studies ofstress responses involving the eIF2α kinases, often extreme conditionsare employed, resulting in a severe block to global translation initiation.Under physiological conditions cells probably rarely experience thesemassive insults. Collectively, and as also suggested previously (Dever,2002), these results imply that the main physiological role of eIF2αphosphorylation is not the downregulation of global translation but,most likely, the translation of certain mRNAs.

MATERIALS AND METHODSCells and cell handlingAll fission yeast strains used in this study are derivatives of S. pombeL972 h− wild-type strain (Leupold, 1950). Strains used in this study arelisted in Table S1.

Cells were maintained and cultured as previously described (Moreno1991). The cells were grown in liquid Edinburgh minimal medium (EMM)with appropriate supplements at 25°C (or at 30°C for nda3-KM311cells) to acell concentration of 2–4×106/ml. The cells were synchronized in G1 or G2phase by incubating cdc10-M17 or cdc25-22 cells, respectively, at 36°C for4 h (or 5 h for cdc10-M17 eIF2alphaS52A strain) before release into the cellcycle at 25°C; in M phase by incubating nda3-KM311 cells at 20°C for 4 hbefore release into the cell cycle at 30°C; and in early G2 by centrifugalelutriation as previously described (Hagan et al., 2016).

To obtain a population of mononuclear G1 cells, cultures weremaintained at 30°C in minimal medium where NH4Cl was replaced with20 mM L-isoleucine (Carlson et al., 1999).

Cultures of S. pombe transformants (together with a wild-type controlculture) were grown to a cell concentration of 8×106/ml (OD595=0.4) inminimal medium where NH4Cl was replaced with 3.75 g/l L-glutamic acid,monosodium salt (Pombe Minimal Glutamate medium, PMG). To inducehuman PKR expression, cells cultured in PMG containing 5 µg/ml thiamine(Sigma-Aldrich) were harvested by centrifuging for 3 min at 1800 g, washedthree times with PMG without thiamine, and resuspended in PMG lackingthiamine for the induction of nmt1 and nmt41 promoters.

Human U2OS and HeLa cells were cultivated in DMEM (Dulbecco’sModified Eagle’s Medium) (Invitrogen) supplemented with 10% fetalbovine serum (FBS) (Gibco) and 1% penicillin/streptomycin (P/S) (Gibco).Human retinal pigment epithelium (RPE) cells immortalized with hTERTwere cultivated in DMEM/F12 Glutamax supplement (Invitrogen)supplemented with 10% FBS, 1% P/S and 0.01 mg/ml hygromycin B(Sigma). Cells were tested for contamination once every two months.

Translation assaysTo label newly synthesized proteins, 50 µM of L-homopropargylglycine(HPG, Thermo Fisher Scientific) was added to 1 ml samples of the main yeastculture taken out 10 min before the indicated time points. To stop translation,0.1 mg/ml of cycloheximide (CHX) was added after 10 min. Cells were fixedin ice-cold methanol or 70% ethanol, washed in 0.5 ml TBS and barcodedusing up to five different concentrations (450, 124.8, 31.2, 6.24 and 0.78 ng/ml) of Pacific Blue dye (PB; Thermo Fisher Scientific P10163) for 30 min inthe dark at room temperature. Stained samples were then washed three timesin 0.5 ml TBS and pooled. The samples were permeabilized with 0.5 ml 1%Triton X-100 in TBS and blocked with 1% BSA in TBS. To detect HPG,Alexa Fluor 647 was linked to the incorporated HPG in a ‘click’ reaction(Liang and Astruc, 2011) using the Click-iT cell reaction buffer kit (ThermoFisher Scientific C10269) following the manufacturer’s protocol to ligate theHPG alkyne with a fluorescent azide. Incorporation was quantified by usingflow cytometry (LSR II flow cytometer, BDBiosciences). SYTOXGreen dye(Thermo Fisher Scientific) was used to stain the DNA. Cell doublets wereexcluded from the analysis as described previously (Knutsen et al., 2011).Samples without HPG were used as negative controls. O-propargyl-puromycin (OPP), (Thermo Fisher Scientific) was added to 4 µM for20 min, the cells were then trypsinized and fixed in 70% ethanol. To detectincorporated OPP, the fixed cells were washed once in PBS with 1% FBS.OPP was ligated with Alexa Fluor 647 in a ‘click’ reaction following themanufacturer’s instructions. The samples were incubated for 5 min indetergent buffer [0.1% Igepal CA-630, 6.5 mMNa2HPO4, 1.5 mMKH2PO4,2.7 mM KCl, 137 mM NaCl, 0.5 mM EDTA (pH 7.5)] containing 4% non-fat milk to block non-specific binding. The cells were incubated for 1 h withprimary antibody against phospho-S10-histone H3 (1:500, Millipore 06-570)in detergent buffer containing 2% non-fat milk, washed once in PBS with 1%FBS and incubated for 30 min with Alexa Fluor 488-linked secondaryantibody (1:500, Thermo Fisher Scientific A-11034) in detergent buffer. Allincubations were carried out in the dark at room temperature. The cells werewashed once in PBS with 1% FBS and stained with 1.5 µg/ml of Hoechst33258 (Sigma) in PBS. The samples were analyzed using flow cytometry(LSR II flow cytometer, BD Bioscience, San Jose, CA) and data wereanalyzed using the FlowJo software (https://www.flowjo.com).

Fluorescence-activated cell sortingExponentially growing cells were fixed with 70% ethanol and stainedfor phospho-S10-histone H3 as described above and detected usingAlexa Fluor 647-coupled secondary antibody. DNA was stained with8 µg/ml propidium iodide. 50,000 cells from each cell-cycle phase wereharvested using a FACS Aria II cell sorter.

UVC irradiationFission yeast cells were irradiated with 254 nm UV light (UVC) in asuspension in EMM (or PMG) medium under continuous stirring to ensureequal irradiation dose (Nilssen et al., 2003). The incident dosewas measuredwith a radiometer (UV Products). A surface dose of 1100 J/m2 (at a dose rate

7

RESEARCH ARTICLE Journal of Cell Science (2018) 131, jcs220327. doi:10.1242/jcs.220327

Journal

ofCe

llScience

of ∼250 J/m2/min) induces a checkpoint response but results in >90% cellsurvival. Samples for protein analysis were taken immediately afterirradiation.

H2O2 treatmentCells grown in PMG medium were treated with H2O2 at the indicatedconcentrations for 15 min before samples were taken.

Leucine starvationAn auxotroph strain was grown in PMG medium supplemented withleucine. The cells were washed with PMG medium three times andincubated in medium not containing leucine for the indicated times.

ImmunoblottingTotal protein extracts of yeast cells were obtained using a low-salt buffer(25 mM MOPS pH 7.1, 60 mM β-glycerophosphate, 15 mM p-nitrophenylphosphate, 15 mM MgCl2, 15 mM EGTA pH 8.0, 1 mM DTT, 0.1 mMNa3VO4, 1% Triton X-100) supplemented with protease inhibitors (Roche).Cell debris was removed by centrifugation at 14,000 g for 15 min at 4°C.The extracts were mixed with 4× LDS sample buffer (Thermo FisherScientific) and 50 mM DTT. Human cells were lysed in Laemmli samplebuffer.

Cell extracts were separated on polyacrylamide gels, transferred ontoPVDF membranes and probed with antibodies against phosphorylatedeIF2α (1:750, CST 3398), eIF2α (1:1000, Santa Cruz sc-11386) PKR(1:3000, Abcam 32052), α-tubulin (1:30,000, Sigma-Aldrich T5168) andγ-tubulin (1:30,000, Sigma T6557). Signal intensities were quantified usingImageJ software.

AcknowledgementsWe thank L. Lindbergsengen and M. O. Haugli for excellent technical assistance.The funders had no role in study design, data collection and interpretation, or thedecision to submit the work for publication.

Competing interestsThe authors declare no competing or financial interests.

Author contributionsConceptualization: B.G.; Methodology: V.S., B.G.; Validation: V.S.; Formal analysis:V.S.; Investigation: V.S., B.G.; Writing - original draft: B.G.; Writing - review & editing:V.S., E.B., B.G.; Supervision: B.G.; Funding acquisition: E.B., B.G.

FundingWe are grateful to Kreftforeningen (the Norwegian Cancer Society), Helse Sør-ØstRHF (the Norwegian South-Eastern Health Authority) and RadiumhospitaletsLegater for funding.

Data availability

Supplementary informationSupplementary information available online athttp://jcs.biologists.org/lookup/doi/10.1242/jcs.220327.supplemental

ReferencesBasi, G., Schmid, E. and Maundrell, K. (1993). TATA box mutations in theSchizosaccharomyces pombe nmt1 promoter affect transcription efficiency butnot the transcription start point or thiamine repressibility. Gene 123, 131-136.

Carlson, C. R., Grallert, B., Stokke, T. and Boye, E. (1999). Regulation of the startof DNA replication in Schizosaccharomyces pombe. J. Cell Sci. 112, 939-946.

Clemens, M. J. (2001). Initiation factor eIF2 alpha phosphorylation in stressresponses and apoptosis. Prog. Mol. Subcell. Biol. 27, 57-89.

Coldwell, M. J., Cowan, J. L., Vlasak, M., Mead, A., Willett, M., Perry, L. S. andMorley, S. J. (2013). Phosphorylation of eIF4GII and 4E-BP1 in response tonocodazole treatment: a reappraisal of translation initiation during mitosis. CellCycle 12, 3615-3628.

Datta, B., Datta, R., Mukherjee, S. and Zhang, Z. (1999). Increasedphosphorylation of eukaryotic initiation factor 2α at the G2/M boundary inhuman osteosarcoma cells correlates with deglycosylation of p67 and adecreased rate of protein synthesis. Exp. Cell Res. 250, 223-230.

Dever, T. E. (2002). Gene-specific regulation by general translation factors.Cell108,545-556.

Dever, T. E., Chen, J. J., Barber, G. N., Cigan, A. M., Feng, L., Donahue, T. F.,London, I. M., Katze, M. G. and Hinnebusch, A. G. (1993). Mammalianeukaryotic initiation factor 2 alpha kinases functionally substitute for GCN2 proteinkinase in the GCN4 translational control mechanism of yeast. Proc. Natl. Acad.Sci. USA 90, 4616-4620.

Elliott, S. G. and McLaughlin, C. S. (1978). Rate of macromolecular synthesisthrough the cell cycle of the yeast Saccharomyces cerevisiae. Proc. Natl. Acad.Sci. USA 75, 4384-4388.

Elliott, S. G., Warner, J. R. and McLaughlin, C. S. (1979). Synthesis of ribosomalproteins during the cell cycle of the yeast Saccharomyces cerevisiae. J. Bacteriol.137, 1048-1050.

Fan, H. andPenman, S. (1970). Regulation of protein synthesis inmammalian cells.II. Inhibition of protein synthesis at the level of initiation during mitosis. J. Mol. Biol.50, 655-670.

Fullwood, M. J., Zhou, W. and Shenolikar, S. (2012). Targeting phosphorylation ofeukaryotic initiation factor-2alpha to treat human disease. Prog. Mol. Biol. Transl.Sci. 106, 75-106.

Hagan, I. M., Grallert, A. and Simanis, V. (2016). Synchronizing progression ofschizosaccharomyces pombe cells from G2 through repeated rounds of mitosisand S phase with cdc25-22 arrest release. Cold Spring Harb. Protoc. 2016.

Hamanaka, R. B., Bennett, B. S., Cullinan, S. B. and Diehl, J. A. (2005). PERKand GCN2 contribute to eIF2α phosphorylation and cell cycle arrest afteractivation of the unfolded protein response pathway.Mol. Biol. Cell16, 5493-5501.

Harding, H. P., Zhang, Y., Zeng, H., Novoa, I., Lu, P. D., Calfon, M., Sadri, N., Yun,C., Popko, B., Paules, R. et al. (2003). An integrated stress response regulatesamino acid metabolism and resistance to oxidative stress. Mol. Cell 11, 619-633.

Hinnebusch, A. G. (1994). The eIF-2 alpha kinases: regulators of protein synthesisin starvation and stress. Semin. Cell Biol. 5, 417-426.

Ho, J. J. D., Balukoff, N. C., Cervantes, G., Malcolm, P. D., Krieger, J. R. and Lee,S. (2018). Oxygen-sensitive remodeling of central carbon metabolism by archaiceIF5B. Cell Rep. 22, 17-26.

Holcik, M. and Sonenberg, N. (2005). Translational control in stress and apoptosis.Nat. Rev. Mol. Cell Biol. 6, 318-327.

Janes, S., Schmidt, U., Ashour Garrido, K., Ney, N., Concilio, S., Zekri, M. andCaspari, T. (2012). Heat induction of a novel Rad9 variant from a cryptictranslation initiation site reduces mitotic commitment. J. Cell Sci. 125, 4487-4497.

Kim, Y., Lee, J. H., Park, J.-E., Cho, J., Yi, H. and Kim, V. N. (2014). PKR isactivated by cellular dsRNAs during mitosis and acts as amitotic regulator.GenesDev. 28, 1310-1322.

Knutsen, J. H. J., Rein, I. D., Rothe, C., Stokke, T., Grallert, B. and Boye, E.(2011). Cell-cycle analysis of fission yeast cells by flow cytometry. PLoS ONE 6,e17175.

Knutsen, J. H. J., Rødland, G. E., Bøe, C. A., Håland, T. W., Sunnerhagen, P.,Grallert, B. and Boye, E. (2015). Stress-induced inhibition of translationindependently of eIF2α phosphorylation. J. Cell Sci. 128, 4420-4427.

Koromilas, A. E. (2015). Roles of the translation initiation factor eIF2alpha serine 51phosphorylation in cancer formation and treatment. Biochim. Biophys. Acta 1849,871-880.

Krutzik, P. O. and Nolan, G. P. (2006). Fluorescent cell barcoding in flow cytometryallows high-throughput drug screening and signaling profiling. Nat. Methods 3,361-368.

Leupold, U. (1950). Die Vererbung von Homothallie und Heterothallie beiSchizosaccharomyces pombe. C. R. Trav. Lab. Carlsberg Ser. Physiol. 24,381-480.

Liang, L. and Astruc, D. (2011). The copper(I)-catalyzed alkyne-azidecycloaddition (CuAAC) “click” reaction and its applications. An overview. Coord.Chem. Rev. 255, 2933-2945.

Marchal, J. A., Lopez, G. J., Peran, M., Comino, A., Delgado, J. R., Garcıa-Garcıa, J. A., Conde, V., Aranda, F. M., Rivas, C., Esteban, M. et al. (2014). Theimpact of PKR activation: from neurodegeneration to cancer. FASEB J. 28,1965-1974.

Maundrell, K. (1993). Thiamine-repressible expression vectors pREP and pRIP forfission yeast. Gene 123, 127-130.

Moreno, S., Klar, A. and Nurse, P. (1991). Molecular genetic analysis of fissionyeast Schizosaccharomyces pombe. Methods Enzymol. 194, 795-823.

Nilssen, E. A., Synnes, M., Kleckner, N., Grallert, B. and Boye, E. (2003). Intra-G1 arrest in response to UV irradiation in fission yeast. Proc. Natl Acad. Sci. USA100, 10758-10763.

Ohno, M. (2014). Roles of eIF2alpha kinases in the pathogenesis of Alzheimer’sdisease. Front. Mol. Neurosci. 7, 22.

Ravindran, R., Loebbermann, J., Nakaya, H. I., Khan, N., Ma, H., Gama, L.,Machiah, D. K., Lawson, B., Hakimpour, P., Wang, Y.-C. et al. (2016). Theamino acid sensor GCN2 controls gut inflammation by inhibiting inflammasomeactivation. Nature 531, 523-527.

Shenton, D., Smirnova, J. B., Selley, J. N., Carroll, K., Hubbard, S. J., Pavitt,G. D., Ashe, M. P. and Grant, C. M. (2006). Global translational responses tooxidative stress impact upon multiple levels of protein synthesis. J. Biol. Chem.281, 29011-29021.

Shuda, M., Chang, Y. and Moore, P. S. (2015a). Mitotic 4E-BP1hyperphosphorylation and cap-dependent translation. Cell Cycle 14, 3005-3006.

8

RESEARCH ARTICLE Journal of Cell Science (2018) 131, jcs220327. doi:10.1242/jcs.220327

Journal

ofCe

llScience

Shuda, M., Velasquez, C., Cheng, E., Cordek, D. G., Kwun, H. J., Chang, Y.andMoore, P. S. (2015b).CDK1substitutes formTORkinase to activatemitotic cap-dependent protein translation. Proc Natl. Acad. Sci. USA 112, 5875-5882.

Silva, R. C., Dautel, M., Di Genova, B. M., Amberg, D. C., Castilho, B. A. andSattlegger, E. (2015). The Gcn2 regulator Yih1 interacts with the cyclindependent kinase Cdc28 and promotes cell cycle progression through G2/M inbudding yeast. PLoS ONE 10, e0131070.

Sivan, G. and Elroy-Stein, O. (2008). Regulation of mRNA Translation duringcellular division. Cell Cycle 7, 741-744.

Sonenberg, N. and Hinnebusch, A. G. (2009). Regulation of translation initiation ineukaryotes: mechanisms and biological targets. Cell 136, 731-745.

Stumpf, C. R., Moreno, M. V., Olshen, A. B., Taylor, B. S. andRuggero, D. (2013).The translational landscape of the mammalian cell cycle. Mol. Cell 52, 574-582.

Tanenbaum, M. E., Stern-Ginossar, N., Weissman, J. S. and Vale, R. D. (2015).Regulation of mRNA translation during mitosis. eLife 4, e07957.

Tinton, S. A., Schepens, B., Bruynooghe, Y., Beyaert, R. and Cornelis, S.(2005). Regulation of the cell-cycle-dependent internal ribosome entry site of thePITSLRE protein kinase: roles of Unr (upstream of N-ras) protein andphosphorylated translation initiation factor eIF-2α. Biochem. J. 385, 155-163.

Way, S.W. and Popko, B. (2016). Harnessing the integrated stress response for thetreatment of multiple sclerosis. Lancet Neurol. 15, 434-443.

Zhan, K., Vattem, K. M., Bauer, B. N., Dever, T. E., Chen, J.-J. and Wek, R. C.(2002). Phosphorylation of eukaryotic initiation factor 2 by heme-regulatedinhibitor kinase-related protein kinases in Schizosaccharomyces pombeis important for fesistance to environmental stresses. Mol. Cell. Biol. 22,7134-7146.

9

RESEARCH ARTICLE Journal of Cell Science (2018) 131, jcs220327. doi:10.1242/jcs.220327

Journal

ofCe

llScience