(2010) 29, 5490–5499 & 2010 Macmillan Publishers Limited All rights reserved 0950-9232/10 www.nature.com/onc ORIGINAL ARTICLE PIM1 is destabilized by ribosomal stress causing inhibition of progression

V Iadevaia1,4, S Caldarola1,4, L Biondini1, A Gismondi1, S Karlsson2, I Dianzani3 and F Loreni1

1Department of Biology, University Tor Vergata, Roma, Italy; 2Department of Molecular Medicine and Therapy, Lund Stem Cell Center, Lund University Hospital, Lund, Sweden and 3Department of Medical Sciences, University of Eastern Piedmont, Novara, Italy

PIM1 is a constitutively active serine/threonine kinase and cellular hypoxia (Wang et al., 2001; Bachmann and regulated by , growth factors and hormones. It Moroy, 2005; Shah et al., 2008). Oncogenic properties has been implicated in the control of cell cycle progression have been observed in animal models where PIM1 and and its overexpression has been associated enhances the development of lymphoma and leukemia with various kinds of lymphoid and hematopoietic (Selten et al., 1985; van Lohuizen et al., 1989). In these malignancies. The activity of PIM1 is dependent on the experimental systems, PIM1 cooperates with in phosphorylation of several targets involved in transcrip- lymphomagenesis, possibly by suppressing MYC- tion, cell cycle and apoptosis. We have recently observed induced apoptosis. More recently, it has been shown that PIM1 interacts with ribosomal protein (RP)S19 and that PIM1 colocalizes with MYC at sites of active cosediments with ribosomes. Defects in ribosome synthesis ; thus, contributing to the expression of (ribosomal stress) have been shown to activate a p53- part of the MYC-regulated (Zippo et al., 2007). dependent growth arrest response. To investigate if PIM1 Animals deficient in PIM1 are viable and show only a could have a role in the response to ribosomal stress, we minor defect in hematopoiesis (Laird et al., 1993). This induced ribosome synthesis alterations in TF-1 and K562 suggests that PIM1 contributes to hematopoietic cell erythroid cell lines. We found that RP deficiency, induced growth and survival, although its role is probably by RNA interference or treatment with inhibitor of secondary to other signaling pathways. According to nucleolar functions, causes a drastic destabilization of this, combined deficiency of PIM1 and AKT1 caused a PIM1. The lower level of PIM1 induces an increase in the severe impairment in hematopoietic cell growth, survival cell cycle inhibitor p27Kip1 and blocks cell proliferation and proliferation (Hammerman et al., 2005). Several even in the absence of p53. Notably, restoring PIM1 level regulators of cell cycle progression and apoptosis have by transfection causes a recovery of cell growth. Our data been identified as PIM1 targets, implicating it in the indicate that PIM1 may act as a sensor for ribosomal control of cell cycle and cell survival (Wang et al., 2001; stress independently of or in concert with the known Bachmann and Moroy, 2005). Among them, the cyclin- p53-dependent mechanisms. dependent kinase (CDK) inhibitor p27Kip1, shown to be Oncogene (2010) 29, 5490–5499; doi:10.1038/onc.2010.279; phosphorylated and downregulated by PIM1 (Morishita published online 19 July 2010 et al., 2008), has been proposed as a specific controller of hematopoietic cell proliferation (Steinman, 2002; Soeiro Keywords: ribosome synthesis; cell cycle arrest; erythro- et al., 2006). Deregulation of the expression of PIM1 leukemia cells; PIM1; RPS19; p27Kip1 kinase has been associated with the development of human malignancies, including lymphomas, leukemias, , squamous cell carcinoma, gastric carcinoma and colorectal carcinoma (reviewed by Shah Introduction et al., 2008). PIM1 kinase has been recently identified as an PIM1 oncogene is highly expressed not only in interactor of ribosomal protein (RP)S19 (Chiocchetti hematopoietic cells, but also in other tissues, such as et al., 2005). The interaction, found in a two-hybrid epithelial cells, prostate and hippocampus (Bachmann screen, was confirmed by co-immunoprecipitation ex- and Moroy, 2005). It is regulated by a variety of growth periments. Moreover, PIM1 was shown to cosediment factors and cytokines at the transcriptional, post- with ribosomal particles in sucrose gradient fractiona- transcriptional, translational and post-translational tion analysis. To investigate the physiological function levels in response to stress conditions, such as ischemias of the PIM1–ribosome interaction, we decided to test if alteration of ribosome synthesis could affect PIM1 Correspondence: Dr F Loreni, Department of Biology, University Tor expression. Ribosome synthesis is a complex process Vergata, Via Ricerca Scientifica, Roma 00133, Italy. taking place sequentially in the nucleolus, in the E-mail: [email protected] 4These authors contributed equally to this work. nucleoplasm and in the cytoplasm. It involves four Received 9 January 2010; revised 9 May 2010; accepted 9 June 2010; ribosomal RNA molecules, about 80 RPs and nearly published online 19 July 2010 200 non-ribosomal factors that are required for the Ribosomal stress affects PIM1 level V Iadevaia et al 5491 synthesis, maturation and export of the two ribosomal subunits (Fatica and Tollervey, 2002). A number of reports suggest that perturbations of ribosome biogen- esis, owing to a variety of causes (ribosomal stress), can activate a specific checkpoint and block cell prolifera- tion mostly through a p53-dependent mechanism (Pestov et al., 2001; Rubbi and Milner, 2003; Anderson et al., 2007; Panic et al., 2007; Danilova et al., 2008; McGowan et al., 2008; Fumagalli et al., 2009). This occurs, for example, in the case of conditional deletion of RPS6 (Volarevic et al., 2000; Sulic et al., 2005) or in response to drugs that disrupt nucleolar structures (Rubbi and Milner, 2003). Alteration of ribosome synthesis has also been implicated in some pathologies that are associated to mutations in ribosomal compo- nents, for instance, Diamond–Blackfan anemia (DBA), or in molecules involved in ribosome maturation (reviewed by Liu and Ellis, 2006 and Caldarola et al., 2009). To address if PIM1 could be involved in the response to ribosomal stress, we used TF-1 and K562 erythroid- cultured cells. We found that RP deficiency and other kinds of ribosomal stress cause a decrease of PIM1 level. Figure 1 PIM1 association with the ribosome. (a) Cytoplasmic Kip1 extracts from TF-1 and K562 cells were separated by ultra- As a consequence, the PIM1 target, p27 , increases; centrifugation in a pellet (P), including ribosomes and ribosomal thus, blocking cell cycle progression. subunits and a supernatant (S), containing free cytoplasmic proteins. The two fractions were analyzed by western blot with primary antibodies against PIM1 (PIM), RPS19 (S19) and b-actin Results (ACT). Loading ratio between P and S was 10:1. (b) An aliquot of K562 cells was treated with the crosslinking-inducing agent (dithiobis(succinimidyl propionate (DSP)). Cytoplasmic extracts Interaction of PIM1 with the ribosome of treated and control cells were separated as in (a), in the presence We have shown that PIM1 interacts with the ribosome of high-salt concentration. Blots were decorated as in (a), including in human embryonic kidney cells (Chiocchetti et al., antibodies against eukaryotic initiation factor 4E. Loading ratio 2005). To verify that this occurs also in hematopoietic between P and S was 3:1. (c) Cytoplasmic extract from TF-1 cells was separated by ultracentrifugation on a linear sucrose gradient. cells, cytoplasmic extracts from both TF-1 and K562 Fractions were collected while monitoring absorbance at 260 nm. cells were separated by ultracentrifugation into two The pellet, containing polysomes, was pooled to the first fraction. fractions: (1) P, pellet, which includes polysomes and Fraction aliquots were precipitated by trichloroacetic acid and ribosomal subunits; and (2) S, super, which includes free analyzed by western blot with primary antibodies against PIM1 cytosolic proteins. Western blot analysis of the two (PIM), RPS19 (S19), RPL7a (L7a), b-tubulin (TUB) and AKT (AKT). The upper panel shows the absorbance profile with the fractions (Figure 1a) showed that most of the cytoplas- position of ribosomal subunits and monomer 80S indicated mic PIM1 is observed in the polysomal pellet in both cell by arrows. Lanes corresponding to fractions 9, 10, 11 contain lines. Importantly, treatment of K562 cells with the one-half, one-third, one-sixth of the other fractions, respectively. crosslink-inducing agent dithiobis(succinimidyl propio- Quantifications of the percentage of PIM1 and the ratio between PIM1 and RPS19 in the different fractions is reported on the right nate) (DSP) can stabilize PIM1–ribosome interaction. In of the panel as a column plot. fact, DSP treatment of the cells prevents dissociation of PIM1 induced by high salt concentration in cell extract (Figure 1b), indicating that the PIM1–ribosome inter- TF-1C and K562C, both infected with a lentiviral vector action occurs in the cell under physiological conditions. inducible for the expression of small interfering RNA Further separation of TF-1 cytoplasmic extracts on specific for RPS19 mRNA (Miyake et al., 2005). The sucrose gradients showed that PIM1 is associated both two parental cell lines differ in p53 expression: TF-1 with polysomes and with free 40S ribosomal subunit expresses a wild-type p53 (Urashima et al., 1998), (Figure 1c). Quantification of the percentage of PIM1 in whereas K562 does not show detectable levels of p53. the different fractions shows that the vast majority is As a control for p53 activity in TF-1C, we treated the associated with active polysomes and 40S ribosomal cells with the genotoxic agent bleomycin, observing p53 subunit (Figure 1c). Interestingly, the ratio between increase by western blot and p21Waf1 activation by real- PIM1 and RPS19 appears higher in polysomes (fraction 1) time reverse transcription–polymerase chain reaction compared with free 40S (fraction 6). This indicates that (RT–PCR) and western blot (Supplementary Figures PIM1 may interact preferentially with active ribosomes. S1A and S1B). Doxycycline (dox) treatment of TF-1C and K562C cells causes a decrease of RPS19, clearly Effect of RPS19 deficiency visible on western blot after 4 days of induction To study the effect of alteration of ribosome synthesis (Figure 2a). Consistent with the data from other on PIM1 expression, we used two modified cell lines, laboratories (Idol et al., 2007; Badhai et al., 2009), we

Oncogene Ribosomal stress affects PIM1 level V Iadevaia et al 5492 observed that RPS19 deficiency in K562C cells caused a Cell counting and fluorescence-activated cell sorting reduction of other RPs from the small subunit, but not (FACS) analysis showed that the induction of RPS19 from the large one (Supplementary Figure S2A). This deficiency causes an evident reduction of cell prolifera- alteration is also visible as a decrease of the 40S peak on tion and an accumulation of cells in the G0/G1 phase of the absorbance profile (polysomal profile) of cytoplas- the cell cycle (Figures 2b and c). It is important to note mic extract of dox-induced K562C cells (Supplementary that alteration of cell cycle and proliferation can also be Figure S2B). observed in the p53-null K562C cells, suggesting that a

Figure 2 Effects of RPS19 deficiency. (a) TF-1C and K562C cells were treated for the indicated number of days with doxycycline (dox). Total protein extracts were separated on SDS–PAGE and transferred onto nitrocellulose membrane. Blots were decorated with primary antibodies against RPS19 (S19) and b-actin (ACT). (b) Aliquots of cells treated as in (a) were counted in triplicate and the results are reported in a linear plot with s.e.m. barely visible. (c) TF-1C and K562C cells untreated or treated for 4 days with dox were analyzed by fluorescence-activated cell sorting (FACS) in triplicate. The percentage of cells in the different cell cycle phases G0/G1 (G1), S (S) and G2/M (G2) is reported in a column plot±s.e.m. (d) Total extracts from TF-1C and K562C treated as in (c) were analyzed by western blot with primary antibody against RPS19 (S19), PIM1 (PIM), p27Kip1 (p27), p53 (p53) and glyceraldehyde 3-phosphate dehydrogenase (GAPDH). An example of western blot is shown on the left. Quantification of at least three blots from two independent experiments is reported on the right as a column plot of the mean±s.e.m. of the values normalized for GAPDH and considering untreated sample as 1. (e) K562C cells were treated for 4 or 6 days (4d, 6d) with dox. Total protein extracts were separated on SDS–PAGE and transferred onto nitrocellulose membrane. Blots were decorated with primary antibodies against RPS19 (S19), PIM1 (PIM), total p27Kip1 (p27), threonine157-phosphorylated p27 (pp27) and b-actin (ACT). Quantification of the ratio between phospho-p27Kip1 and p27Kip1 is reported on the right as a column plot considering untreated sample as 1.

Oncogene Ribosomal stress affects PIM1 level V Iadevaia et al 5493 p53-independent mechanism is activated in these cells. efficiency. Northern analysis of K562C cells did not Western analysis of both TF-1C and K562C cells show any difference in PIM1 mRNA level during RPS19 showed that, during induction, in parallel to RPS19 knockdown (Figure 3a); similarly, polysomal associa- decrease, there is a comparable reduction of PIM1 level, tion of PIM1 messenger did not decrease during dox whereas the level of the CDK inhibitor p27Kip1 increases treatment (Figure 3b), ruling out regulation at the (Figure 2d). In addition, in TF-1 cells there is an level of transcription, mRNA stability and translation. increase in the level of p53. To address the relationship We also analyzed p27Kip1 mRNA level by real-time between PIM1 decrease and p27Kip1 increase, we RT–PCR, observing no significant alteration in RPS19- analyzed the phosphorylation status of threonine157 deficient K562C and TF-1C cells (Figure 3c). To address that is one of the PIM1 target sites on p27Kip1 (Morishita the possible variation of protein stability, we treated et al., 2008) during induction of K562C cells. As shown K562C cells with the translation inhibitor cycloheximide in Figure 2e, PIM1 reduction correlates with the and monitored the decrease of PIM1 levels in control decrease of threonine157 phosphorylation. As phos- and induced cells. The decay of PIM1 level is very rapid phorylation on this site (together with threonine198) has (Figure 3d). However, the induction of RPS19 inter- been shown to inhibit p27Kip1 activity and induce its ference causes a further destabilization of the protein, degradation (Larrea et al., 2009), our results indicate which after 15 min is almost undetectable. To further that the decline of PIM1 level could cause stabilization show that the decrease of PIM1 level, observed in and activation of p27Kip1. RPS19-deficient cells, was caused by post-translational destabilization, we treated both control and induced K562C and TF-1C cells with an inhibitor of the Alteration of PIM1 turnover proteasome (MG132), which is the main mediator of To investigate the change of PIM1 and p27Kip1 levels PIM1 degradation (Shay et al., 2005). Consistent with observed in TF-1C and K562C cells following dox the previous experiment, inhibition of the proteasome in treatment, we analyzed mRNA levels and translation RPS19-deficient cells causes the recovery of PIM1 to

Figure 3 Analysis of PIM1 and p27Kip1 expression. (a) K562C cells were treated for 4 days with dox. Total RNA, extracted from control and induced cells, was analyzed by northern blot with probes specific for PIM1, RPS19 and glyceraldehyde 3-phosphate dehydrogenase (GAPDH). (b) Cytoplasmic extracts from K562C cells untreated (Àdox) or treated for 4 days with doxycycline ( þ dox) were separated on sucrose gradients, and RNA extracted from each fraction was analyzed on northern blots with a PIM1 probe. Fractions corresponding to polysomes are indicated. (c) Total RNA, extracted from TF-1C and K562C cells untreated or treated for 4 days with dox, was analyzed by real-time RT–PCR with primers specific for p27Kip1 and GAPDH. The results (three RT–PCR from three RNA preparations for TF-1 and three RT–PCR from two RNA preparations for K562 cells) are reported as a column plot of the mean±s.e.m. of p27Kip1 mRNA normalized for GAPDH mRNA, considering uninduced cells as 1. (d) K562C cells cultured for 4 days without or with dox were treated with 50 mg/ml of cycloheximide for the indicated time (min). Total protein extracts were analyzed by western blot with the indicated primary antibodies. Quantification of three blots from the same experiment is reported in the lower part of the panel as a linear plot of the mean±s.e.m. of the values normalized for b-actin, considering untreated sample as 100. (e) K562C cells cultured as in (d) were treated for 2 or 4 h with MG132. Total protein extracts analyzed by western blot with primary antibody against PIM1 (PIM), p27Kip1 (p27) and b-actin (ACT). An example of western blot is shown on the left. Quantification of at least three blots from two experiments is reported on the right as a column plot of the mean±s.e.m. of the values normalized for b-actin and considering untreated sample as 1.

Oncogene Ribosomal stress affects PIM1 level V Iadevaia et al 5494 approximately the level of control (uninduced) cells (Figure 3e and Supplementary Figure 3A). This suggests that RPS19 deficiency stimulates the proteasome activity on PIM1. In the same experiment, we have analyzed the level of p27Kip1. In this case, we observed that protea- some inhibition increases p27Kip1 level in control cells, but not in induced cells. Therefore, the increase observed during RPS19 deficiency is owing to a lower proteasome activity on p27Kip1, possibly as a conse- quence of PIM1 reduction (Figure 3e). The conclusion is that RPS19 reduction causes an increase in proteasome- mediated degradation of PIM1 and an inhibition of p27Kip1 degradation.

Other conditions of ribosomal stress Next, we asked if the deficiency of other RPs would cause the same effect. To answer this question, we transfected TF-1 cells with siRNA specific for RPS6 and RPL7a. The analysis of protein levels by western blot showed that, in both cases, the decrease of the RP causes a reduction of PIM1 level (Figure 4a). This result suggests that the diminution of PIM1 is not dependent on the deficiency of a specific RP, but is rather owing to the decrease of active ribosomes caused by RP insufficiency. As a further example of alteration of ribosome synthesis, we used inhibition of RNA poly- merase I activity. It has been shown that treatments that Figure 4 PIM1 level in other conditions of ribosomal stress. alter nucleolar function, for instance, inhibition of (a) TF-1C cells were transfected with unrelated small interfering ribosomal RNA synthesis, induce a p53-dependent RNA (siRNA) (siC) or siRNA specific for RPS6 (siS6) and RPL7a (siL7a). Total protein extracts were analyzed by western blot with stress response (Rubbi and Milner, 2003; Mayer and the indicated primary antibodies. An example of western blot is Grummt, 2005). We decided to test if treatment with shown in the upper part. Quantification of four blots from two agents that interfere with ribosomal RNA production experiments is reported in the lower part as a column plot of the would affect PIM1 level similarly to RP deficiency. mean±s.e.m. of the PIM1 values normalized for b-actin and K562 cells were treated for 16 h with actinomycin D, considering control sample as 1. (b) K562 cells were treated for 15 h with 5 mM camptothecin (CPT), 2 nM actinomycin D (ACTD), camptothecin and cisplatinum, all known to affect 20 mM cisplatinum (CISP), 5 ng/ml nocodazole (NOCO), 0.3 mM ribosomal RNA synthesis at different steps. As a hydroxyurea (HYDR) or mock treated (Cont). Total protein control, we treated the cells with the cell cycle inhibitors extracts were analyzed by western blot with the indicated primary nocodazole and hydroxyurea, which do not affect antibodies. An example of western blot is shown in the upper part. Quantification of at least three blots from at least two independent nucleolar function. Western analysis showed that treat- experiments is reported in the lower part as a column plot of the ment with all three inhibitors of nucleolar function mean±s.e.m. of the PIM1 values normalized for glyceraldehyde causes a decrease of PIM1 level, whereas nocodazole 3-phosphate dehydrogenase (GAPDH) and considering control and hydroxyurea do not affect it (Figure 4b). Actino- sample as 1. (c) K562C cells, cultured for 16 h without or with 2 nM mycin D effect on PIM1 level was also observed in actinomycin D, were treated or not with MG132 for 2 h. Total protein extracts were analyzed by western blot with the indicated TF-1C cells even after 3 h of treatment (Supplementary primary antibodies. Quantification of two independent experiments Figure S3B). is reported in the lower part of the panel as a column plot of the To investigate if the decrease of PIM1 level induced means normalized for b-actin, considering untreated sample as 1. by RNA polymerase I inhibitors was caused by (d) Total protein extracts of lymphoblastoid cell lines derived from healthy donors (C1, C2) or from DBA patients (P1, P2), were proteasome-mediated degradation, we used the protea- analyzed by western blot with the indicated primary antibodies. An some inhibitor MG132. K562C cells were incubated for example of western blot is shown in the upper part. Quantification 16 h with actinomycin D and then treated for 2 h with of 3 (P1) or 2 (P2) blots is reported in the lower part as a column MG132. Western analysis (Figure 4c) showed that plot of the mean±s.e.m. (only P1) of the PIM1 values normalized proteasome inhibition induces a recovery of PIM1 for b-actin or GAPDH, and considering control sample as 1. levels. This indicates that actinomycin D, similarly to RPS19 deficiency, induces PIM1 destabilization. Analo- gous effect of MG132 on PIM1 level was observed in TF-1C cells (Supplementary Figure S3C). responsible for the disease has not yet been identified. Further correlation between defective ribosome However, on the basis of the available data on the synthesis and reduction of PIM1 level was obtained by disease, we can assume that these cells have a ribosomal analyzing lymphoblastoid cell lines derived from two defect. Western analysis showed that extracts from DBA DBA patients. In both cases, the genetic alteration cell lines had a lower PIM1 level compared with controls

Oncogene Ribosomal stress affects PIM1 level V Iadevaia et al 5495 (Figure 4d), confirming the link between ribosome Following our previous finding of an interaction functionality and the kinase amount. between PIM1 and the ribosome, we addressed the possibility of a role of this kinase in response to the ribosomal alteration. PIM1 is able to phosphorylate PIM1 overexpression attenuates ribosomal stress several targets involved in cell cycle progression and At this point, we hypothesized that the growth arrest apoptosis, and it has been implicated in signaling induced by ribosomal stress, observed in our experi- pathways relevant for cell survival (Bachmann and mental systems, could be owing to the destabilization of Moroy, 2005). We observed that RP deficiency induced PIM1. This would affect cell growth and proliferation by RNAi or treatment with inhibitors of nucleolar Kip1 through the stabilization of the PIM1 target p27 .To functions cause destabilization of PIM1. The increase in verify our hypothesis and establish a causal relationship degradation is mediated by the proteasome and could be between PIM1 decrease and the alterations observed in owing to reduced availability of ribosomes for binding RPS19-deficient cells, we carried out a rescue experi- with a consequent rapid degradation of free (unbound) ment by transiently transfecting PIM1-expressing plas- PIM1 (Figure 6c). In fact, we observed a decrease of 40S mids into TF-1C and K562C cells after induction of subunit and possibly of active ribosomes during the RPS19 interference. FACS analysis and growth curves induction of RPS19 deficiency. However, it could be showed that transient expression of PIM1 causes a that ribosomal stress activates molecules involved in recovery of cell cycle progression and proliferation both PIM1 degradation, such as PP2A (Losman et al., 2003). in TF-1C and in K562 cells (Figures 5a and b). In In any case, we think that PIM1 decrease triggers at addition, western analysis showed that restoring the least some of the effects of ribosomal stress. In fact, both PIM1 level during RPS19 interference causes an evident in TF-1C and in K562C cells, the lower level of PIM1 Kip1 reduction of p27 levels in TF-1C cells (Figure 5c). correlates with (1) an increase in the CDK inhibitor Kip1 The effect of PIM1 transfection on p27 level is less p27Kip1 that is hypophosphorylated on threonine157 (at marked in K562C cells. This suggests that, in this cell least in K562C cells), (2) proliferation decline and (3) type, the cell cycle effects are not mediated exclusively inhibition of cell cycle progression. Importantly, restor- Kip1 by p27 . Interestingly, the level of p53, present only in ing PIM1 level by transfection, during the induction of TF-1C cells, is also affected by PIM1 transfection. RPS19 deficiency, significantly attenuates all these To further verify our working model, we decided effects: p27Kip1 resumes a low level, and cell proliferation Kip1 to knock down the expression of p27 by RNA and cell cycle progression are recovered. Moreover, the interference (RNAi) in RPS19-deficient TF-1C and knockdown of p27Kip1 in RPS19-deficient cells causes, K562C cells to obtain a recovery of the cell cycle at least partially, a recovery of cell cycle progression. progression. For this purpose, TF-1C and K562C It should be noted that the recovery of cell growth cells were induced with dox for 4 or 6 days. Part of occurs without affecting RPS19 level. This indicates that the cells was transfected with an unrelated siRNA (siC) the proliferation block was not owing to intrinsic Kip1 or with a p27 -specific siRNA either at the beginning insufficiency of the translation machinery, but to a or after 2 days of induction. Western analysis of cell specific checkpoint mechanism. The working model we Kip1 extracts indicated that the transfection of p27 -specific would like to propose is that PIM1 can act as a sensor siRNA caused a clear reduction of protein level for ribosomal stress. In fact, using RPS19 as a docking (Figure 6a). FACS analysis of both TF-1C and K562C site, PIM1 interacts with active ribosomes and is cells (Figure 6b) showed that the increase in the stabilized by the binding. A decrease of available percentage of G0/G1 cells caused by RPS19 deficiency ribosomes, owing to deficiency of an RP or ribosomal Kip1 is attenuated by the knockdown of p27 , confirming RNA, would cause degradation of unbound PIM1. In the hypothesis that PIM1 affects cell cycle through this addition, the well-documented function of PIM1 as a CDK inhibitor. regulator of cell cycle and cell growth fits well with the role of transducing the stress signal to various key cellular components. Among the PIM1 targets possibly Discussion involved in the response to ribosomal stress, we tested p27Kip1 because: (1) it is rapidly destabilized by PIM1 It is now well documented that ribosomal stress, defined phosphorylation (Morishita et al., 2008) and (2) it is a as an alteration of ribosome synthesis, can induce the key controller of cell cycle progression of hematopoietic activation of the tumor suppressor p53 blocking cell cells (Steinman, 2002). We found that p27Kip1 level proliferation and activating apoptosis (Pestov et al., increases in response to alteration of ribosome synthesis 2001; Sulic et al., 2005; Anderson et al., 2007; Panic in a PIM1-dependent way, both in the absence (K562C) et al., 2007; Danilova et al., 2008; McGowan et al., 2008; and in the presence (TF-1C) of p53. Furthermore, the Fumagalli et al., 2009). Here, we report the analysis of decrease of p27Kip1, caused by RNAi, partially restores the effect of RPS19 deficiency on the metabolism of two cell cycle. However, the role of p27Kip1 in mediating the erythroid cell lines. Our data suggest the existence of effect of RPS19 deficiency appears less evident in K562C p53-independent response to ribosomal stress. In fact, cells. Therefore, it could be that in these cells, and we found that both in the presence (TF-1) and in the possibly in TF-1C cells as well, other PIM1 targets, such absence (K562) of p53, the insufficiency of RPS19 as p21Waf, CDC25A, CDC25C, and so on, might have a causes cell cycle arrest and a block in cell proliferation. role in the response to ribosomal stress.

Oncogene Ribosomal stress affects PIM1 level V Iadevaia et al 5496

Figure 5 PIM1 transfection in RPS19-deficient cells. (a) TF-1C and K562C cells were treated for the indicated number of days with doxycycline (dox). At day 3 of dox treatment, part of the cells was transfected with PIM1- or RPS19-expressing plasmids. At the indicated time points, aliquots of cells were counted in triplicate and the results are reported as a linear plot of the mean±s.e.m. (b) TF-1C and K562C cells untreated with dox (Àdox), treated for 5 days with dox ( þ dox) or treated for 5 days with dox and transfected with PIM1 at day 3 ( þ dox þ PIM) were analyzed by fluorescence-activated cell sorting (FACS) in triplicate. The percentage of cells in the different cell cycle phases, G0/G1 (G1), S (S) and G2/M (G2), is reported in a column plot of the mean±s.e.m. (c) Total extracts, from TF-1C and K562C cells treated as in (b), were analyzed by western blot with the indicated primary antibodies. An example of western blot is shown on the left. Quantification (at least three blots from two independent experiments) is reported on the right as a column plot of the mean±s.e.m. of the values normalized for glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and considering untreated sample as 1.

It is important to underline that this is the first involved in the development of human malignancies. It characterization of a stable interaction of a kinase with could be that the mechanism we have described is an the mammalian ribosome. Moreover, PIM1 has been integral part of the complex relationship between

Oncogene Ribosomal stress affects PIM1 level V Iadevaia et al 5497

Figure 6 p27Kip1 inhibition by RNA interference in RPS19-deficient cells. (a) TF-1C and K562C cells untreated with dox (Àdox), treated for 4 days with dox ( þ dox), transfected with unrelated siRNA and treated for 4 days with dox ( þ dox þ siC), and transfected with p27Kip1 siRNA and treated for 4 days with dox ( þ dox þ sip27) were analyzed by western blot with the indicated primary antibodies. An example of a blot is shown in the upper part of the panel. Quantification of the signals is reported as a column plot in the lower part of the panel. (b) TF-1C and K562C cells treated as in (a) were analyzed by fluorescence-activated cell sorting (FACS). The percentage of cells in G0/G1 phase is reported as a column plot of the mean of two independent experiments. (c) Model of the relationship between ribosomal stress and PIM1. In normal growth condition, PIM1 associates with the ribosome and phosphorylates p27Kip1 destabilizing it; during ribosomal stress, the decrease of ribosomes renders PIM1 available for degradation by the proteasome. As a consequence, p27Kip1 is stabilized and the cell cycle is arrested. ribosome and cancer that has been addressed by several were maintained in suspension in RPMI 1640 medium studies (reviewed by Ruggero and Pandolfi, 2003). supplemented with 10% fetal bovine serum, 50 mg/ml penicillin In any case, our data suggest an entirely new perspective and 50 mg/ml streptomycin. in the analysis of the PIM1-dependent oncogenesis. TF-1C and K562C cells, expressing inducible siRNA targeting RPS19 mRNA, were prepared in Karlsson’s Finally, the relationship between PIM1 and the p53- 6 dependent response to ribosomal stress remains to be laboratory (Miyake et al., 2005). Cells (5 Â 10 ) were tran- siently transfected using 15 mg of plasmid DNA and 30 mlof elucidated and will be an important subject for future JetPei transfection reagent (Polyplus transfection), or with investigation. 100–200 nM siRNA and 10 ml of Interferin transfection reagent (Polyplus transfection) according to the manufacturer’s pro- tocol. After 24 h, the cells were harvested and analyzed by Materials and methods western blot, or in the case of RNAi, after 48 h they were re- transfected and then harvested and analyzed by western blot. and transient transfection Plasmids encoding full-length human PIM1 and RPS19 have Human erythroleukemic TF-1 and K562 cells were maintained been described previously (Chiocchetti et al., 2005). The RNAi in RPMI media supplemented with 10% fetal bovine serum, target sequences were as follows: sense, 50-GTAAGAAAGC and antibiotics (100 IU/ml penicillin and 100 mg/ml strepto- CCTTAAATA-30 for human RPS6, sense, 50-CACCACCTT mycin). TF-1 medium was also supplemented with 5 ng/ml GGTGGAGAACAA-30 for human RPL7a and sense, 50-GCA granulocyte–macrophage colony-stimulating factor. Human GGAAUAAGGAAGCGACCUGCAA-30 for human p27Kip1. lymphoblastoid cells lines were established by Epstein–Barr As a control, siRNA targeting the unrelated protein ZNF9 virus immortalization of patients’ blood lymphocytes. Cells (sense, 50-GACAAGUGAAGUCAACUGU-30) was used.

Oncogene Ribosomal stress affects PIM1 level V Iadevaia et al 5498 RNA analysis (Minneapolis, MN, USA), kindly provided by Giuseppe For polysomal RNA analysis, cells (1–2 Â 106) that had been Viglietto, Catanzaro, Italy), rabbit polyclonal anti-eIF4E (Cell washed once with phosphate-buffered saline buffer (150 mM Signaling) and rabbit polyclonal anti-b-tubulin (Santa Cruz). NaCl, 2.7 mM KCl, 8 mM NaH2PO4 and 1.4 mM K2PO4) were Primary antibodies were revealed using horseradish peroxi- lysed with 300 ml of lysis buffer (10 mM NaCl, 10 mM MgCl2, dase-conjugated goat anti-rabbit or anti-mouse Ab (Jackson 10 mM Tris–HCl (pH 7.5), 1% Triton X-100, 1% sodium Immunoresearch, Suffolk, UK) and the ECL chemilumines- deoxycholate, 36 U/ml RNase inhibitor (Promega, Milan, cence detection system (Pierce, Rockford, IL, USA). Quanti- Italy), 1 mM dithiothreitol) and transferred into a microcen- fication analyses were performed by LAS3000 Image System trifuge tube. After 5 min of incubation in ice with occasional (Fuji, Milan, Italy) and ImageQuant software (GE Healthcare, vortexing, the lysate was centrifuged for 8 min at 10 000 r.p.m. Milan, Italy). at 4 1C. The supernatant was frozen in liquid nitrogen and stored at À70 1C to be analyzed later, or immediately layered in a 15–50% (w/v) sucrose gradient containing 30 mM Tris–HCl Ribosome isolation and crosslinking assay TF-1 or K652 cells were pelleted and resuspended in (pH 7.5), 100 mM NaCl and 10 mM MgCl2, and centrifuged in a Beckman SW41 rotor for 110 min at 37 000 r.p.m. lysis buffer (10 mM NaCl, 10 mM MgCl2,10mM Tris–HCl Fractions were collected while monitoring the optical density (pH 7.5), 0.5% NP-40, aprotinin 1 mg/ml, leupeptin 1 mg/ml, at 254 nm and total RNA was extracted from each fraction by pepstatinA 1 mg/ml, phenylmethylsulfonyl fluoride 100 mg/ml). the proteinase K method. For northern analysis, RNA was After incubation in ice for 1 min, the extract was centrifuged fractionated on formaldehyde–agarose gels and transferred for 1 min in a microcentrifuge at a maximum speed at 4 1C and onto GeneScreen Plus membrane (Perkin-Elmer Life Sciences, the supernatant (cytoplasmic extract) was frozen in liquid Milan, Italy). Northern blotting was performed essentially nitrogen. To isolate ribosomes and ribosomal subunits, the as recommended by the manufacturer. Radioactive probes cytoplasmic extracts were spun at 100 000 g for 2 h on 15% were prepared by the random primer technique using sucrose cushion. After centrifugation, the pellet (which DNA fragments isolated from plasmids containing PCR- includes polysomes and ribosomal subunits) was resuspended amplified cDNA sequences. Primers for amplification in Loading Buffer (63 mM Tris–HCl (pH 6.8), 5% glycerol, 1% were designed according to sequences present in the NCBI SDS, 2.5% bromophenol-blue), whereas the supernatant was Data Bank. precipitated with 10% trichloroacetic acid and resuspended in For real-time RT–PCR, total RNA was extracted from cells Loading Buffer. To separate polysomes and ribosomal using Eurogold Trifast (Euroclone, Milan, Italy), according to subunits, cytoplasmic extracts were loaded onto a 10–30% the manufacturer’s protocol. RNA (2.5 mg) was treated with linear sucrose gradient containing 30 mM Tris–HCl (pH 7.5), DNAse I (Promega), then reverse transcribed into single- 100 mM NaCl and 10 mM MgCl2. Gradients were centrifuged in strand cDNA, using Moloney murine leukemia virus reverse a Beckman SW 41 rotor for 5 h at 37 000 r.p.m. and then transcriptase (Promega) and random primers (Invitrogen, collected in 11 fractions while monitoring the absorbance at Milan, Italy). cDNA was diluted at a concentration of 80 ng/ 260 nm. The pellet and trichloroacetic acid-precipitated frac- ml in nuclease-free water and stored in aliquots at À80 1C until tions were washed with acetone, dried, resuspended in Loading used. Real-time PCR was performed with the LightCycler Buffer and analyzed by western blot. For crosslinking (Roche Diagnostics, Milan, Italy) using SYBR green detection experiments, DSP was prepared as a stock solution of 50 mg/ (Kapa SYBR Fast qPCR kit; Kapa Biosystems, Woburn, MA, ml in dimethyl sulfoxide and added to the cell culture medium USA), the cDNA as the template and the primer mix of to a final concentration of 1 mg/ml (2.5 mM). Cells were then interest as follows: glyceraldehyde 3-phosphate dehydrogenase incubated at 37 1C, and after 10 min, the DSP was quenched by (GAPDH) NM_002046.3: sense primer, 50-ACCAGGGCTG adding Tris–HCl (pH 8.0) to a final concentration of 100 mM. CTTTTAACTCTGGT-30; antisense primer, 50-GCAAATTT After 10 min incubation at room temperature, the cells were CCATGGCACCGTCAAGG-30; p27Kip1 NM_004064: sense lysed and the cytoplasmic extracts were incubated with high- primer, 50-ACGTGAGAGTGTCTAACGG-30; antisense primer, salt buffer (0.5 M KCl) for 30 min in ice. The extracts were then 50-AGTGCTTCTCCAAGTCCC-30; p21Waf1 NM_000389: separated into a ribosomal pellet and ‘free’ cytosolic proteins sense primer, 50-CGGCAGACCAGCATGACAGATTT-30; as described above. antisense primer, 50-TCAAAGGCCCGCTCTACATCTT-30. The amount of mRNA transcripts encoding these genes was ÀDDCt Proliferation assay and flow cytometry analysis determined using the expression 2 , considering the thresh- K562 and TF-1 cells (uninduced or induced with 2 mg/ml dox) old cycle (Ct) of the sample relative to the internal reference were seeded in triplicate. Cell aliquots were harvested and GAPDH (DCt) and to untreated cells (DDCt). counted with a Nebauer chamber after Trypan blue staining. Cultures were diluted to maintain approximately constant cell Western blotting concentration (5–10 Â 105 cells per ml). Proteins were separated on 12% sodium dodecyl sulfate– For flow cytometry, cells were washed in phosphate- polyacrylamide gel, and transferred onto nitrocellulose Pro- buffered saline, fixed for 30 min at 4 1C with methanol:acetone tran membrane (Schleicher and Schuell, Milan, Italy), and (4:1), treated for 20 min with RNAse A (100 mg/ml) and stained incubated with the following primary antibodies and antisera: with propidium iodide (1 mg/ml) for 20 min at room tempera- mouse monoclonal antibody specific for RPS19 (Orru et al., ture. Then, cells were analyzed by a FACSCalibur instrument 2007), mouse monoclonal anti-GAPDH (Chemicon, Milan, (Becton Dickinson, Milan, Italy) and the percentage of cells in Italy), rabbit polyclonal anti-RPS6 (Cell Signaling, Boston, the different cell cycle phases was evaluated with CellQuest MA, USA), rabbit polyclonal anti-b-actin (Sigma, Milan, software. Italy), mouse monoclonal anti-PIM1 (Santa Cruz, Heidelberg, Germany), mouse monoclonal anti-p53 (Santa Cruz), rabbit polyclonal anti-L7a (kindly provided by Giulia Russo, Naples, Italy), rabbit polyclonal anti-AKT (Santa Cruz), Conflict of interest mouse monoclonal anti-p27Kip1 (BD, Milan, Italy), rabbit polyclonal anti-phospho-p27Kip1 (threonine157; R&D Systems The authors declare no conflict of interest.

Oncogene Ribosomal stress affects PIM1 level V Iadevaia et al 5499 Acknowledgements gratefully acknowledged. This work was also supported by grants from Diamond Blackfan Anemia Foundation Inc. and We thank Marcello Giorgi for expert technical assistance. The the Italian Ministry for University and Research (MIUR) financial support of Telethon, Italy (Grant No. GGP07242) is FIRB and PRIN.

References

Anderson SJ, Lauritsen JP, Hartman MG, Foushee AM, Lefebvre JM, Morishita D, Katayama R, Sekimizu K, Tsuruo T, Fujita N. (2008). Shinton SA et al. (2007). Ablation of ribosomal protein L22 Pim promote cell cycle progression by phosphorylating and selectively impairs alphabeta T cell development by activation of a down-regulating p27Kip1 at the transcriptional and posttranscrip- p53-dependent checkpoint. Immunity 26: 759–772. tional levels. Cancer Res 68: 5076–5085. Bachmann M, Moroy T. (2005). The serine/threonine kinase Pim-1. Int Orru S, Aspesi A, Armiraglio M, Caterino M, Loreni F, Ruoppolo M J Biochem Cell Biol 37: 726–730. et al. (2007). Analysis of the ribosomal protein S19 interactome. Mol Badhai J, Frojmark AS, Razzaghian HR, Davey E, Schuster J, Cell Proteom 6: 382–393. Dahl N. (2009). Posttranscriptional down-regulation of small Panic L, Montagne J, Cokaric M, Volarevic S. (2007). S6-haploinsuf- ribosomal subunit proteins correlates with reduction of 18S rRNA ficiency activates the p53 tumor suppressor. Cell Cycle 6: 20–24. in RPS19 deficiency. FEBS Lett 583: 2049–2053. Pestov DG, Strezoska Z, Lau LF. (2001). Evidence of p53-dependent Caldarola S, De Stefano MC, Amaldi F, Loreni F. (2009). Synthesis cross-talk between ribosome biogenesis and the cell cycle: effects of and function of ribosomal proteins—fading models and new nucleolar protein Bop1 on G(1)/S transition. Mol Cell Biol 21: perspectives. FEBS J 276: 3199–3210. 4246–4255. Chiocchetti A, Gibello L, Carando A, Aspesi A, Secco P, Garelli E Rubbi CP, Milner J. (2003). Disruption of the nucleolus mediates et al. (2005). Interactions between RPS19, mutated in Diamond– stabilization of p53 in response to DNA damage and other stresses. Blackfan anemia, and the PIM-1 oncoprotein. Haematologica 90: EMBO J 22: 6068–6077. 1453–1462. Ruggero D, Pandolfi PP. (2003). Does the ribosome translate cancer? Danilova N, Sakamoto KM, Lin S. (2008). Ribosomal protein S19 Nat Rev Cancer 3: 179–192. deficiency in zebrafish leads to developmental abnormalities and Selten G, Cuypers HT, Berns A. (1985). Proviral activation of the defective erythropoiesis through activation of p53 . putative oncogene Pim-1 in MuLV induced T-cell lymphomas. Blood 112: 5228–5237. EMBO J 4: 1793–1798. Fatica A, Tollervey D. (2002). Making ribosomes. Curr Opin Cell Biol Shah N, Pang B, Yeoh KG, Thorn S, Chen CS, Lilly MB et al. (2008). 14: 313–318. Potential roles for the PIM1 kinase in human cancer—a molecular Fumagalli S, Di Cara A, Neb-Gulati A, Natt F, Schwemberger S, and therapeutic appraisal. Eur J Cancer 44: 2144–2151. Hall J et al. (2009). Absence of nucleolar disruption after impairment Shay KP, Wang Z, Xing PX, McKenzie IF, Magnuson NS. (2005). of 40S ribosome biogenesis reveals an rpL11-translation-dependent Pim-1 kinase stability is regulated by heat shock proteins and the mechanism of p53 induction. Nat Cell Biol 11: 501–508. ubiquitin–proteasome pathway. Mol Cancer Res 3: 170–181. Hammerman PS, Fox CJ, Birnbaum MJ, Thompson CB. (2005). Pim Soeiro I, Mohamedali A, Romanska HM, Lea NC, Child ES, and Akt are independent regulators of hematopoietic cell Glassford J et al. (2006). p27Kip1 and p130 cooperate to regulate growth and survival. Blood 105: 4477–4483. hematopoietic cell proliferation in vivo. Mol Cell Biol 26: 6170–6184. Idol RA, Robledo S, Du HY, Crimmins DL, Wilson DB, Ladenson JH Steinman RA. (2002). Cell cycle regulators and hematopoiesis. et al. (2007). Cells depleted for RPS19, a protein associated with Oncogene 21: 3403–3413. Diamond–Blackfan anemia, show defects in 18S ribosomal RNA Sulic S, Panic L, Barkic M, Mercep M, Uzelac M, Volarevic S. (2005). synthesis and small ribosomal subunit production. Blood Cells Mol Inactivation of S6 ribosomal protein gene in T lymphocytes Dis 39: 35–43. activates a p53-dependent checkpoint response. Genes Dev 19: Laird PW, van der Lugt NM, Clarke A, Domen J, Linders K, McWhir J 3070–3082. et al. (1993). In vivo analysis of Pim-1 deficiency. Nucleic Acids Res Urashima M, Teoh G, Chauhan D, Ogata A, Shirahama S, Kaihara C 21: 4750–4755. et al. (1998). MDM2 protein overexpression inhibits apoptosis of Larrea MD, Wander SA, Slingerland JM. (2009). p27 as Jekyll TF-1 granulocyte-macrophage colony-stimulating factor-dependent and Hyde: regulation of cell cycle and cell motility. Cell Cycle 8: acute myeloblastic leukemia cells. Blood 92: 959–967. 3455–3461. van Lohuizen M, Verbeek S, Krimpenfort P, Domen J, Saris C, Liu JM, Ellis SR. (2006). Ribosomes and marrow failure: coincidental Radaszkiewicz T et al. (1989). Predisposition to lymphomagenesis in association or molecular paradigm? Blood 107: 4583–4588. pim-1 transgenic mice: cooperation with c-myc and N-myc in Losman JA, Chen XP, Vuong BQ, Fay S, Rothman PB. (2003). murine leukemia virus-induced tumors. Cell 56: 673–682. Protein phosphatase 2A regulates the stability of Pim protein Volarevic S, Stewart MJ, Ledermann B, Zilberman F, Terracciano L, kinases. J Biol Chem 278: 4800–4805. Montini E et al. (2000). Proliferation, but not growth, blocked Mayer C, Grummt I. (2005). Cellular stress and nucleolar function. by conditional deletion of 40S ribosomal protein S6. Science 288: Cell Cycle 4: 1036–1038. 2045–2047. McGowan KA, Li JZ, Park CY, Beaudry V, Tabor HK, Sabnis AJ Wang Z, Bhattacharya N, Weaver M, Petersen K, Meyer M, Gapter L et al. (2008). Ribosomal mutations cause p53-mediated dark skin et al. (2001). Pim-1: a serine/threonine kinase with a role in cell and pleiotropic effects. Nat Genet 40: 963–970. survival, proliferation, differentiation and tumorigenesis. J Vet Sci Miyake K, Flygare J, Kiefer T, Utsugisawa T, Richter J, 2: 167–179. Ma Z et al. (2005). Development of cellular models for ribosomal Zippo A, De Robertis A, Serafini R, Oliviero S. (2007). PIM1- protein S19 (RPS19)-deficient Diamond–Blackfan anemia dependent phosphorylation of histone H3 at serine 10 is required for using inducible expression of siRNA against RPS19. Mol Ther 11: MYC-dependent transcriptional activation and oncogenic transfor- 627–637. mation. Nat Cell Biol 9: 932–944.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Oncogene