The PIM1 serine/threonine protein kinase mediates growth factor and survival signalling, and cooperates potently with c-MYC during tumorigenesis. PIM1 is overexpressed in many human cancers and is a promising target for drug development. PIM1 levels are regulated mainly through cytokine-induced transcription and protein degradation, but mechanisms regulating its activity and levels remain largely unexplored. Here, we show that PIM1 is modified in vitro and in cultured cells by the Small ubiquitin-like modifier (SUMO) on two independent sites: K169, within a consensus SUMOylation motif (IK169DE171) in the active site of PIM1, and also at a second promiscuous site. Alanine substitution of E171 (within the consensus motif) abolished SUMOylation, significantly increased the half-life of PIM1, and markedly reduced its ubiquitylation. Mechanistically, SUMOylation promoted ubiquitin-mediated degradation of PIM1 via recruitment of the SUMO-targeted ubiquitin ligase, RNF4. Additionally, SUMOylated PIM1 showed enhanced protein kinase activity in vitro. Interestingly, the E171A mutant was active in vitro but displayed altered substrate specificity in cultured cells, consistent with the idea that SUMOylation may govern PIM1 substrate specificity under certain contexts. Taken together, these data demonstrate that the protein kinase activity and levels of PIM1 can be regulated by a covalent post-translational modification.
The proto-oncogene PIM1 (Proviral Integration site for MuLV site-1) encodes a serine/threonine protein kinase that phosphorylates a wide range of substrates involved in the regulation of important cellular processes such as cell cycle progression, protein translation, metabolism and survival1. PIM1 levels are elevated in a significant proportion of haematological malignancies such as lymphomas and solid tumours such as prostate cancer, where its expression correlates with high-tumour grade2, 3. More recently, PIM1 was shown to be upregulated in triple-negative breast cancers where it is thought to engage potently in driving c-MYC-dependent transcription4, 5. PIM kinases have been shown to mediate therapeutic resistance through various mechanisms and in a range of cellular contexts relevant to human cancer development6. PIM inhibitors are therefore considered to be promising targets for cancer therapy and have been a significant focus of drug development.
Unlike most protein kinases, PIM kinases are constitutively active, lack regulatory domains, and do not require any external phosphorylation events for activation. At the protein level, the half-life of PIM1 varies between 5 to 15 minutes in primary cells7 and is found to be increased up to 100 minutes in certain tumour cells such as K562, and BV173 expressing the BCR-ABL fusion protein, potentially through HSP90-mediated protection from ubiquitylation and subsequent proteasomal degradation8, 9. Other authors have also reported that PIM1 can be destabilized by the protein phosphatase PP2A10, 11. Apart from these events, however, there have been no other regulatory post-translational modifications reported for PIM kinases. Furthermore, the pathway(s) by which PIM1 is targeted for ubiquitylation and degradation is incompletely understood and remains a subject of ongoing investigation. A clearer understanding of these events may provide insight into how PIM1 could be efficiently targeted for degradation in cancer therapy.
The Small Ubiquitin-like Modifier (SUMO) is a protein tag that is covalently and reversibly attached to a lysine residue on its protein substrates, in a manner similar to the ubiquitin pathway. SUMOylation regulates important biological processes including gene expression, intracellular transport, chromosome segregation, DNA repair, and protein turnover/stability12,13,14,15.
In the present study we have explored the mechanisms of regulation of PIM1 in greater depth. We show that PIM1 is covalently modified by SUMO1 and SUMO2/3 in cultured cells on at least two lysine residues, one consensus (i.e. within a SUMOylation recognition motif) and one non-consensus. We find that SUMO modification is a key event in the rapid turnover of PIM1, thereby providing mechanistic detail for the normally short-lived nature of this oncogenic protein kinase. We also show that SUMOylation can directly stimulate the protein kinase activity of PIM1 towards established physiological substrates, demonstrating, for the first time, a non-auto-catalytic mechanism of regulation for this protein kinase. These novel findings offer new insights into PIM regulation and reveal potential opportunities for further therapeutic targeting.
PIM1 is modified in vitro and in cultured cells by SUMO
To explore whether PIM1 might be regulated post-translationally, potential sites of post-translational modification present in all three PIM kinases (PIM1, 2 and 3) were investigated using a bioinformatics approach. This analysis revealed that K169 of PIM1, K165 of PIM2 and K172 of PIM3 lie within a classic SUMO consensus motif, ψKxE/D, where ψ (psi) is a hydrophobic residue (usually I/V/L/F/M), K is the modified lysine and x is any amino acid16.
To determine whether PIM1 could be post-translationally modified by SUMO, 35S-labelled PIM1 was incubated in vitro with the SUMO E1 and E2 enzymes SAE1/2 and UBC9 respectively, in the absence and presence of SUMO1 or SUMO2. Analysis of the products by SDS-PAGE and autoradiography showed the appearance of slower migrating PIM1 species with apparent molecular weights (around 55 kDa) consistent with SUMOylated forms of PIM1 (Fig. 1a). PIM1 showed weak SUMO modification in this experiment compared with SP100 (positive control) suggesting that PIM1 may require an E3 SUMO ligase for efficient modification. Higher molecular weight bands were also observed when purified GST-PIM1 was incubated with SUMO1 or SUMO2 in the presence of the SUMO E1 and E2 enzymes in vitro (Fig. 1b).
To determine whether SUMOylation of PIM1 can occur within a cellular context, SUMOylation assays were performed in COS7 cells transfected with plasmids expressing MYC-tagged PIM1 together with plasmids expressing 6His-tagged SUMO1, 2 or 3 (Fig. 1c). While PIM1 was not modified in the absence of ectopically expressed SUMO proteins, SUMO conjugates of PIM1 were detected when either SUMO1, 2, or 3 were co-expressed. Notably, higher molecular weight PIM1-SUMO conjugates run as a doublet around 55 kDa (also observed in Fig. 1a) suggesting that SUMO modification may occur on at least two independent sites. Affinity purification of 6His-tagged PIM1 alone under denaturing conditions in H1299 cells was sufficient to observe SUMOylation by endogenous SUMO2 cells, indicating that (a) it is a substrate for endogenous SUMOylation enzymes, and (b) SUMOylation of PIM1 is not a cell line-specific phenomenon (Fig. 1d). Furthermore, expression of the WT SENP1, but not an inactive SENP1 mutant, led to disappearance of the higher molecular weight PIM1 bands, confirming that the doublet band observed was indeed SUMO-modified PIM1 (Fig. 1e).
To assess whether PIM1 is endogenously SUMOylated, extracts from K562 cells (which express relatively high levels of PIM1) were analyzed by western blotting following treatment, or mock treatment, with the SUMO protease, SENP1 in vitro. As shown in Fig. 1f (long exposure), the levels of higher molecular weight bands corresponding to SUMO-modified forms were reduced in the presence of SENP1 with no detectable changes in the levels of the unmodified kinase (short exposure, Fig. 1f). This observation supports the idea that PIM1 SUMOylation occurs endogenously in untransfected cells.
PIM1 is SUMOylated at Lysine169 and at a second promiscuous site
To determine whether K169 (within the consensus SUMO-motif: Fig. 2a) could act as a SUMO-acceptor, an arginine substitution mutant was generated using site-directed mutagenesis. SUMOylation assays were carried out using WT PIM1 and the K169R PIM1 mutant, co-expressed with 6His-SUMO2 (Fig. 2b). As observed previously, SUMOylated PIM1 (WT) ran as a doublet of approximate molecular weight 55 kDa. Interestingly, the K169R mutant was still capable of undergoing SUMOylation, but only a single species at 55 kDa molecular weight was observed. This suggests that K169 is a major site of SUMOylation, but that another non-consensus site can be SUMOylated independently of K169.
Consensus SUMOylation sites are also characterized by the presence of an invariant glutamic acid at the +2 position relative to the acceptor lysine residue. Strikingly, alanine substitution of the consensus glutamic acid residue within the postulated SUMOylation motif in PIM1 (E171A) abolished SUMOylation of the protein, suggesting that this residue is critical for overall SUMOylation of PIM1, irrespective of the target residue(s). Curiously, a conservative E171Q substitution only partially eliminated PIM1 SUMOylation similar to the K169R mutant. E181 was also mutated to alanine as it formed an inverted SUMO motif (D/ExKΨ) with lysine 183 (K183). However, SUMOylation of this mutant was similar to that of wild type (WT) PIM1. Arginine substitution of each of the remaining 11 lysines in PIM1 did not reveal any other specific site of SUMO modification, including the K67M mutant, which is catalytically inactive17, indicating that SUMOylation of PIM1 occurs independently of its protein kinase activity. Overall, these analyses identify K169 as a specific site of SUMO modification in PIM1, and suggest that there is a second (non-consensus) site(s) of modification, which appears to be promiscuous. The SUMOylation status of these mutants, together with their functional properties (see below) is shown in Table 1.
PIAS1 and PIAS3 can act as E3 SUMO ligases for PIM1
Members of the Protein Inactivator of Activated STATs (PIAS) family, which negatively regulate the JAK-STAT pathway, were the first E3 SUMO ligases to be identified18, 19. Interestingly, PIM1 itself is a downstream target of the JAK-STAT pathway, which led us to hypothesize that PIAS proteins could function as E3 SUMO ligases for PIM1.
To test this idea, SUMOylation of PIM1 was examined in COS7 cells transfected with plasmids expressing WT PIM1, 6His-SUMO2, and either PIAS1, PIASy (also called PIAS4) or PIAS3. As above, SUMOylation of PIM1 (with SUMO2) by endogenous modifying enzymes could be observed (Fig. 3a, lanes 2 and 7). Under basal conditions, only PIAS3 greatly enhanced PIM1 SUMOylation suggesting that it can act as an E3 SUMO ligase for PIM1 (Fig. 3a, lanes 4 and 9). Treatment of the cells with the proteasome inhibitor, MG132, led to an increase in the levels of PIM1 SUMOylation even in the absence of an E3 SUMO ligase suggesting that SUMOylated PIM1 is targeted for degradation by the proteasome. Strikingly, PIAS1 could also stimulate PIM1 SUMOylation under these conditions (Fig. 3a, lanes 7 and 8), although to a lesser extent than PIAS3. In contrast, no such stimulation was seen when PIASy was used (Fig. 3a, lanes 7 and 10). Further analysis by co-immunoprecipitation demonstrated that PIM1 and PIAS3 associate, consistent with their interaction within the cell (Fig. 3b). Co-immunoprecipitation of PIM1 with PIAS1 was also observed but only when the cells had been treated with MG132 (Fig. 3c). No association could be detected between PIM1 and PIASy (Fig. 3d).
SUMOylation increases the protein kinase activity of PIM1 in vitro
The location of K169 in the active site of PIM1 suggests that a SUMO moiety could either inhibit kinase activity by sterically blocking access to the substrate, or promote kinase activity by regulating kinase-substrate interactions. To explore this idea, in vitro protein kinase assays were carried out using purified 6His-PIM1 that had been SUMOylated with GST-tagged SUMO2 in vitro and subsequently affinity-purified for analysis (Supplementary Fig. 1). For an accurate assessment of relative protein kinase activity, equal amounts of SUMOylated PIM1 were compared with PIM1 that had been de-SUMOylated in vitro using the catalytic domain (aa 415–643) of the SUMO protease SENP120. The data (Fig. 4a) indicate that SUMOylated PIM1 was significantly more active in phosphorylating Histone H3.3 at Ser10 when compared with unmodified PIM1 at all time points. Western blotting for total Histone H3.3 confirmed equal amount of substrate in these reactions. Similar results were obtained when BAD was used as a substrate (Supplementary Fig. 2).
To rule out that SENP1 affected PIM1 protein kinase activity, unmodified PIM1 was first incubated with increasing concentrations of SENP1 catalytic domain followed by kinase assay. However, no obvious difference was noted in the level of phosphorylation of Histone H3.3 substrate in the absence or presence of the SENP1 fragment (Fig. 4b) confirming that SUMOylated PIM1 is more active towards Histone H3.3 than unmodified PIM1.
Key residues required for SUMOylation selectively impact the activity and substrate specificity of PIM1
Prior to investigating the intracellular effects of substituting key residues in PIM1 that are required for SUMOylation (K169R, E171A), mutant proteins were expressed in bacteria as 6His-PIM1 proteins, affinity-purified, and tested to determine whether they retained protein kinase activity. PIM1 can autophosphorylate on serine, threonine and tyrosine21, 22. SDS-PAGE of the purified proteins revealed a distinguishable mobility shift of WT PIM1 to an apparent higher molecular weight (characteristic of a phosphorylation event) as compared with the K67M inactive mutant, which cannot autophosphorylate17 (Fig. 5a, lower panel). The mobility of the K169R mutant was identical to the K67M mutant, suggesting impairment of its autophosphorylation function. In contrast, the E171A mutant, like WT PIM1, remained active. Treatment of the purified proteins with lambda phosphatase confirmed that the difference in mobility was due to autophosphorylation of PIM1 (Fig. 5b). Western analysis using a pan-phosphotyrosine antibody (Fig. 5a, upper panel) also indicated that only WT PIM1 and the E171A mutant could autophosphorylate.
To determine whether the K169R and E171A mutants were catalytically active on established PIM1 substrates, in vitro protein kinase assays were performed. Western blotting analysis confirmed that that WT, K169R and E171A PIM1 proteins, but not the K67M mutant, were active against Histone H3.3 (pSer1023) (Fig. 5c), although less active in the case of the K169R mutant. Similar results were seen when BAD (pSer11224) was used as substrate. Both E171A and K169R could phosphorylate c-MYC (pSer6225) but at a lower level as compared with WT PIM1. These data suggested that the E171A mutant largely retained activity but that generally the K169R mutant was catalytically impaired.
To explore loss of SUMOylation under cellular conditions, the phosphorylation status of PIM1 substrates was examined in U2OS-FRT cells inducibly expressing YFP alone or YFP-PIM1 (WT or mutant) (Supplementary Fig. 3a,b). The data (Fig. 5d) suggest that WT PIM1, but not the K67M mutant, is active towards Histone H3 at Ser10, c-MYC at Ser62, BAD at Ser112 and ERK1/2 at Thr202/Tyr204 (Note that activation of ERK1/2 is an indirect effect of PIM1 activation in cells26). No apparent activity was noted towards p70S6K or 4E-BP1. The K169R mutant showed little or no activity towards BAD, ERK1/2 and c-MYC, thereby reflecting its in vitro activity towards two of these substrates (BAD and c-MYC), and its autophosphorylation, suggesting that K169 may indeed entail an important catalytic function. The E171A mutant showed impaired activity towards BAD and ERK1/2, consistent with a role in activation by SUMOylation, but did not show any reduced activity towards c-MYC, possibly reflecting other cellular events affecting its substrate selectivity. Both mutants, like WT PIM1, showed a slight stimulation of Histone H3.3 Ser10 phosphorylation. Curiously, WT PIM1 and the two SUMO mutants reduced the levels of phospho (active)-AKT reflecting a compensatory adjustment since these two protein kinase share many substrates27. We have not been able to observe any obvious phenotypes resulting from PIM1 expression in these cells (data not shown). Previous studies have suggested that overexpression of PIM1 may be superfluous in cells already expressing all PIM isoforms9. Taken together, these findings suggest that SUMOylation may regulate the phosphorylation of selected substrates by PIM1 although, as stated above, it is difficult to rule out possible catalytic consequences of the mutations.
PIM1 that cannot be SUMOylated is significantly more stable than WT PIM1
Next, we examined the potential impact of SUMOylation on PIM1 protein stability by cycloheximide-chase analysis in H1299 cells transiently transfected with PIM1 (WT or mutant) plasmids. The half lives of the WT, K169R and K67M PIM1 proteins were approximately 120, 90, and 75 min respectively (Fig. 6a and b). However, the E171A mutant appeared to be much more stable with a half-life of greater than 4 h.
To rule out that these effects were due to differences in transfection efficiencies or expression levels, cycloheximide-chase analysis was carried out using single-copy isogenic HeLa-FRT-derived lines that permitted inducible expression of WT and mutant PIM1 proteins as YFP-fusion proteins (Supplementary Fig. 3c,d). In this system the differences in turnover were qualitatively similar to those observed following transient expression, with the E171A mutant showing a significant increase in stability (Fig. 6c and d). These data suggest that SUMOylation is required for efficient turnover of PIM1, although we cannot rule out the possibility that E171 may play a more direct role in engaging the protein degradation machinery. The rapid turnover of the K169R mutant, which can still be partially SUMOylated, may be associated with loss of auto-catalytic activity, which has been previously linked to reduced PIM1 stability10, 11.
SUMOylation downregulates PIM1 levels by promoting RNF4-dependent ubiquitylation and proteasomal degradation
To further investigate the role of SUMOylation in regulating PIM1 turnover, we silenced UBC9 expression (the sole SUMO E2 enzyme) in H1299 and HeLa cells. In both cases, decreased levels of UBC9 were accompanied by increased levels of endogenous PIM1 (Fig. 7a), thereby supporting a role for SUMO modification in promoting PIM1 turnover. However, we cannot rule out that this is an indirect effect as changes in UBC9 levels also affect global SUMOylation in cells. When 6His-tagged PIM1 was expressed in H1299 cells and subsequently isolated on Ni2+-agarose beads, a series of high molecular weight bands was observed, consistent with multiple attachments of SUMO and/or ubiquitin. These bands disappeared following silencing of UBC9 expression (Fig. 7b), consistent with the idea that UBC9 promotes the appearance of poly-SUMO- or SUMO-ubiquitin hybrid chains on PIM1.
SUMOylation can promote ubiquitin-mediated degradation of some substrates by recruiting the SUMO-targeted E3 ubiquitin ligase, RNF428. To test this idea, PIM1 was transiently expressed in COS7 cells, together with 6His-SUMO2 and UBC9 to induce polySUMOylation, and in the presence or absence of WT or mutant RNF4 proteins. SUMOylation assay (Fig. 7c) confirmed that expression of SUMO2 and UBC9 with PIM1 led to the formation of polySUMOylated PIM1. Co-expression of RNF4 caused a decrease in the levels of polySUMOylated PIM1. This effect was most likely due to selective ubiquitin-mediated degradation of polySUMOylated PIM1, as mutant forms of RNF4, where either the RING domain (mRING) or the SUMO interaction motifs mSIM) were mutated did not lead to disappearance of SUMOylated PIM1. Additionally, silencing of endogenous RNF4 caused a substantial increase in endogenous PIM1 levels (Fig. 7d). These observations support that idea that SUMOylation of PIM1 acts as a signal for RNF4-mediated ubiquitylation, leading to proteasomal degradation.
We also performed ubiquitylation assays in H1299 cells co-transfected with PIM1 and 6His-tagged ubiquitin plasmids. Western blot analysis of the captured proteins showed that the WT, K169R and K67M PIM1 proteins were ubiquitylated to similar extents (Fig. 7e). However, the ubiquitylation of the E171A mutant was significantly reduced relative to WT PIM1. Since E171A can only be SUMOylated very weakly (Fig. 2b) these data support that idea that SUMOylation can promote PIM1 ubiquitylation. To confirm that RNF4 is an E3 ubiquitin ligase for PIM1, we performed ubiquitylation assays in parental H1299 cells and in H1299 cells in which RNF4 had been knocked out by CRISPR. The data indicate that the ubiquitylation of PIM1 was significantly reduced in the RNF4 knock-out cells as compared with the parental line (Fig. 7f). Ubiquitylation of PIM1 was also examined in the absence and presence of co-expressed PIAS3. As shown in Fig. 7g, PIAS3 also stimulated the ubiquitylation of PIM1. Taken together, these experiments support a model in which SUMOylation of PIM1 can lead to the subsequent ubiquitylation and degradation of PIM1.
In the present study we have established that SUMOylation regulates two independent functions of PIM1. Firstly, we have shown that SUMOylation mediates RNF4-dependent ubiquitylation and proteasomal turnover of PIM1. We cannot rule out at this stage, however, that other mechanisms may also operate to mediate PIM1 turnover. Secondly, our data indicate that SUMOylation of purified PIM1 in vitro directly stimulates its protein kinase activity towards two established PIM1 substrates: Histone H3 (Ser10) and BAD (Ser112). It is challenging to confirm unequivocally that this regulation occurs within a cellular context as it is not possible to mimic PIM1 SUMOylation in cultured cells. Moreover, under steady state conditions only a small proportion of molecules are SUMOylated at any given time14, 29. Nevertheless, an alanine substitution mutant of E171 (a key amino acid in PIM1 required for SUMOylation to occur) is fully active in vitro yet shows a significant reduction in BAD-Ser112 phosphorylation in cells. This is consistent with loss of the ability of this mutant to undergo SUMOylation. These two novel observations establish for the first time, to our knowledge, that PIM1 can be regulated through non-autocatalytic post-translational modification events. Further experimentation did not reveal any detectable changes arising from loss of SUMOylation in subcellular localization (Supplementary Fig. 4), association with partner/substrates proteins such as c-MYC (Supplementary Fig. 5), or the ability to stimulate c-MYC-dependent transcription (Supplementary Fig. 6). However, it is yet possible that there are other PIM1-associated functions that may be affected by this modification.
The idea that SUMOylation increases the protein kinase activity of PIM1, yet also increases its degradation, may appear initially to be counter-intuitive. One plausible model to explain this, however, would be that, similar to the “ubiquitin clock” model proposed for certain transcription factors such as MYC30, the activity of PIM1 may be governed by the length of time it takes to transition from the SUMOylated and active form to a SUMOylated and ubiquitylated form that is destroyed by the proteasome. A model for the regulation of PIM1 by SUMOylation, consistent with this idea, is shown in Fig. 8. Additionally, it is known that other activated protein kinases can be degraded by the ubiquitin proteasome system or through the lysosomal pathway in cells. For example, a study performed with Protein kinase C suggested that the active conformation is necessary to elicit its degradation31, 32. Moreover, activation of the receptor tyrosine kinase, epidermal growth factor receptor (EGFR) by its ligand, epidermal growth factor (EGF) was shown to not only activate its kinase activity, but also trigger its degradation33, 34. Therefore, although PIM1 is constitutively active, the idea that SUMOylation can further enhance its activity, yet also result in its degradation, is consistent with these examples.
We were able to detect SUMOylation of endogenous PIM1 in cultured cells upon proteasome inhibition (Fig. 1) but have been unable to observe any changes in this modification in response to agents/conditions known to affect SUMOylation including DNA damage (induced by etoposide treatment) or proteotoxic stress (induced by thapsigargin or tunicamycin). Changes in the SUMOylation status of PIM1 may occur only within a selected or localized pool of PIM1, or in response to a particular cytokine or growth factor. Moreover, SUMOylation of PIM1 may lead to rapid protein degradation making it difficult to distinguish between SUMOylated and ubiquitylated PIM1 by western blotting of the total cellular PIM1 pool. Alternatively, SUMOylation-mediated turnover may serve as a normal house-keeping function. Nevertheless, alteration or impairment of the mechanisms we have identified may be a factor in human cancer development and could therefore hold opportunities for enhancing drug targeting of PIM1 kinase activity in cancer treatment.
Our data add to a growing body of evidence that SUMOylation of kinases may regulate at least a subset of these important enzymes. For example, SUMOylation of pancreatic Glucokinase35, AKT36,37,38, AMPKβ239 and Protein kinase R40 stimulates their abilities to phosphorylate their respective substrates. Conversely, SUMOylation inhibits other kinases such as MEK41 and AMPKα142. These different outcomes suggest that there is no unifying mechanism by which SUMOylation regulates kinase activity. Additionally, studying the effect of SUMOylation is particularly challenging in cases such as PIM1 when the modified lysine is present in the kinase domain, because it becomes difficult to distinguish between the functions of the lysine residue associated with the intrinsic activity of the kinase and SUMOylation of the substrate. In this respect, our data from the analysis of mutating E171, which is not an essential catalytic component but is critical for SUMOylation, provide a more reliable indication of the likely outcome of blocking SUMOylation. In spite of these challenges, future investigation of the role played by this regulatory mechanism in influencing the oncogenic function of PIM1 may provide a deeper understanding of the relationship of PIM1 to cancer development together with potential ideas for exploitation.
Materials and Methods
Cell culture and reagents
COS7, H1299, DU145, HeLa-FRT and U2OS-FRT were grown in DMEM medium and K562 cells were cultured in RPMI-1640 medium, both supplemented with 10% FBS (Biosera) and 2 mM L-glutamine (Thermo Fisher Scientific). Parental H1299 cells and H1299 cells in which RNF4 had been knocked out using CRISPR were grown in RPMI-1640 medium supplemented with 10% FBS (Biosera) and 2 mM L-glutamine (Thermo Fisher Scientific). These cells were a kind gift from Prof Ron Hay (University of Dundee). A description of the method used to generate this line is described elsewhere43.
Antibodies and Western blotting
SDS-PAGE and western blotting was carried out using standard methods. The following antibodies were used: PIM1 (12H8, Santa Cruz), PIM1 (A300–313A, Bethyl Laboratories), Actin (ab6276, Abcam), GAPDH (G8795, Sigma), SUMO2 (51–9100, Zymed), His-tag (27-4710-01, GE Healthcare), HA-tag (12CA5, Sigma), Flag-tag (F1804, Sigma), MYC-tag (9E10, Hybridoma supernatant), GFP-tag (sc-8334, Santa Cruz), GST-tag (sc-459, Santa Cruz), total ERK1/2 (ER16, Transduction lab), and phospho S10 Histone H3 (06–570, Millipore). The following antibodies were purchased from Cell Signaling Technology: phospho S62 c-MYC (13748), total c-MYC (5605), total Histone H3 (4499), phospho S112 Bad (5284), total Bad (9239), phospho T37/46 4E-BP1 (2855), total 4E-BP1 (9644), phospho T389 p70S6K (108D2), total p70S6K (49D7), phospho S473 AKT (D9E), total AKT (40D4), phospho T202/Y204 ERK1/2 (9106) and phospho tyrosine-100 (9411). Sheep polyclonal UBC9 and chicken polyclonal RNF4 were from Ron Hay, University of Dundee. Secondary antibodies were purchased from Biorad and Thermo Fisher Scientific.
Plasmids and transfection
Plasmids expressing 6His-SUMO1, 6His-SUMO2, 6His-SUMO3, HA-PIAS1, HA-PIASy, Flag-PIAS3 have been described previously44. Plasmids expressing WT MYC-PIM1, 6His-ubiquitin and GST-PIM1 were generated previously45. Rat WT RNF4, RNF4 SIM mutant (mSIM) and RNF4 RING mutant (mRING) and human UBC9 in pcDNA3 were from Ron Hay, University of Dundee. Plasmids expressing Flag-tagged WT SENP1 and mutant SENP1 (C602S) were provided by Laureano de la Vega, University of Dundee. Flag-tagged c-MYC plasmid was from Vicky Cowling, University of Dundee. The Myc-responsive 4X E-box promoter/pGL3 vector was a kind gift from Prof. Dr. Thorsten Berg, University of Leipzig.
To express YFP-tagged PIM1, cDNA encoding human PIM1 was amplified from MGC clone pOTB7-PIM1, and ligated into the EcoRI site of pEYFP-C1 vector (Clontech). YFP-PIM1 or YFP alone was PCR amplified from pEYFP-C1/PIM1 plasmid and ligated into the KpnI site of pcDNA5/FRT/TO vector (Thermo Fisher Scientific) for generation of tet-inducible cell lines. A 6His-PIM1 mammalian expression plasmid was generated by cloning PCR amplified PIM1 cDNA into the EcoRI site of pcDNA3 vector. The vector, pHAT2, used for the expression of 6His-PIM1 in bacterial cells was provided by Dr. Marko Hyvonen (University of Cambridge). The PIM1 cDNA was cloned into the EcoRI site of pHAT2. For retrovirus production, the PIM1 cDNA was cloned into the EcoRI site of pBABE-puro vector.
Plasmids were transfected into cells using Lipofectamine 2000 (Thermo Fisher Scientific) using DNA (μg) to Lipofectamine 2000 (μl) ratio of 1:3 according to manufacturer’s instructions.
PIM1 mutants were generated using the QuikChange Lightning Site-directed Mutagenesis kit (Agilent Technologies) as per manufacturer’s instructions. The primer sequences have been described in Table 2. The coding region of all plasmids was sequenced to confirm the presence of the desired mutation.
These were carried out using Lipofectamine RNAiMAX reagent (Thermo Fisher Scientific) according to manufacturer’s instructions. A pool of 4 siRNAs (ON-TARGETplus SMARTpool, Dharmacon) was used to silence the expression of endogenous UBC9, RNF4 and PIM1 with Non-targeting siRNA pool as the negative control.
In vivo SUMOylation and Ubiquitylation assay
SUMOylated or ubiquitylated proteins were pulled down from the cell lysate using Ni2+-NTA agarose beads (Qiagen) as described in the published protocol28.
These were carried out as described previously46.
Recombinant protein expression and purification
6His-tagged or GST-tagged PIM1 proteins were expressed in Rosetta 2 (DE3) pLys cells (Novagen). Protein induction was achieved by IPTG addition (100 μM final concentration) and further incubation for 3–4 hours at 30 °C. The cell pellet was lysed in B-PER bacterial protein extraction reagent (Thermo Fisher Scientific) and affinity purified using Ni2+-NTA agarose beads (Qiagen) or GST-sepharose (GE Healthcare) using standard protocol. Proteins were stored in buffer containing 50 mM Tris pH 7.5, 500 mM NaCl and 1 mM DTT after elution and dialysis. Purified recombinant SUMO1 and 2, SAE1/2, UBC9, IR1 + M fragment of RanBP2 and SENP1 catalytic domain were a gift from Ron Hay, University of Dundee.
In vitro protein kinase assay
PIM1 kinase assays were carried out using recombinant 6His-PIM1 (WT or mutant) with Histone H3.3 (M2507, NEB UK), c-MYC (MRC-PPU Reagents, University of Dundee) or BAD (SRP5164, Sigma) as substrate. Reactions were carried out in a total volume of 20 μl containing 50 mM Tris-HCl pH 7.5, 10 mM MgCl2, 0.2 mM ATP, 1 μg substrate and approximately 0.5–1 μg PIM1 kinase. Reactions were incubated at 30 °C for 30 min, and terminated by addition of 2X SDS sample buffer (with DTT). Phosphorylated proteins were detected by western blotting using phospho-specific antibodies.
Small scale in vitro SUMOylation reactions
Purification of SUMOylated PIM1 kinase from in vitro reaction
In vitro SUMOylation of 6His-PIM1 was carried out overnight at 37 °C in 100 μl reaction containing 50 mM Tris pH 7.5, 3 mM ATP, 5 mM MgCl2, 5 mM DTT, 20 μg 6His-PIM1, 50 μg GST-SUMO2, 5 μg UBC9, 0.9 μg SAE1/2 and 1.5 μg IR1 + M fragment of RanBP2. The reaction was diluted 5-fold with 50 mM Tris pH 7.5 and SUMOylated proteins were captured by adding 400 μl (50% slurry) glutathione sepharose beads (GE Healthcare) overnight at 4 °C. The beads were washed three times in wash buffer (50 mM Tris pH 7.5 and 150 mM NaCl) and eluted by adding 400 μl elution buffer (20 mM reduced glutathione, 50 mM Tris pH 8, 150 mM NaCl and 1 mM DTT) for 2 hours at 4 °C.
DeSUMOylation coupled kinase assay
Elutions from in vitro SUMOylation reactions were divided into two equal parts: one with 20 nM active SENP1 catalytic domain and the other without SENP1. First, the reactions were incubated at 30 °C for 1 hour for deSUMOylation to occur, following which 5 μg Histone H3.3 (substrate) and 1x kinase buffer (50 mM Tris pH 7.5, 10 mM MgCl2, 0.2 mM ATP) was added. The reactions were further incubated at 30 °C for the kinase assay reaction, and aliquots were collected at 0, 15, 30 and 45 min for analysis of histone H3.3 phosphorylation by western blotting using a phospho-specific antibody.
Lambda phosphatase treatment
Dephosphorylation of PIM1 was carried out using lambda protein phosphatase kit (NEB, UK) according to manufacturer’s instructions.
Generation of tetracycline (tet) inducible cell lines
HeLa-FRT cells were a kind gift from Prof. Stephen Taylor (University of Manchester) and U2OS-FRT were a kind gift from Prof. Kevin Hiom (University of Dundee). Tet-inducible cell lines were generated according to Flp-In T-REx instruction manual (Thermo Fisher Scientific).
Blanco-Aparicio, C. & Carnero, A. Pim kinases in cancer: Diagnostic, prognostic and treatment opportunities. Biochem. Pharmacol. 85, 629–643 (2013).
Brault, L. et al. PIM serine/threonine kinases in the pathogenesis and therapy of hematologic malignancies and solid cancers. Haematologica 95, 1004–1015 (2010).
Nawijn, M. C., Alendar, A. & Berns, A. For better or for worse: the role of Pim oncogenes in tumorigenesis. Nat. Rev. Cancer 11, 23–34 (2011).
Braso-Maristany, F. et al. PIM1 kinase regulates cell death, tumour growth and chemotherapy resistance revealing a novel target in triple-negative breast cancer. Breast Cancer Res. Tr. 159, 179–180 (2016).
Horiuchi, D. et al. PIM1 kinase inhibition as a targeted therapy against triple-negative breast tumors with elevated MYC expression. Nat. Med 22, 1321–1329 (2016).
Warfel, N. A. & Kraft, A. S. PIM kinase (and Akt) biology and signaling in tumors. Pharmacol. Ther. 151, 41–49 (2015).
Amson, R. et al. The human protooncogene product p33pim is expressed during fetal hematopoiesis and in diverse leukemias. Proc. Natl. Acad. Sci. USA 86, 8857–8861 (1989).
Mizuno, K. et al. Regulation of Pim-1 by Hsp90. Biochem. Biophys. Res. Comm. 281, 663–669 (2001).
Shay, K. P., Wang, Z., Xing, P. X., McKenzie, I. F. & Magnuson, N. S. Pim-1 kinase stability is regulated by heat shock proteins and the ubiquitin-proteasome pathway. Mol. Cancer Res. 3, 170–181 (2005).
Losman, J. A., Chen, X. P., Vuong, B. Q., Fay, S. & Rothman, P. B. Protein phosphatase 2A regulates the stability of Pim protein kinases. J. Biol. Chem. 278, 4800–4805 (2003).
Ma, J., Arnold, H. K., Lilly, M. B., Sears, R. C. & Kraft, A. S. Negative regulation of Pim-1 protein kinase levels by the B56beta subunit of PP2A. Oncogene 26, 5145–5153 (2007).
Denuc, A. & Marfany, G. SUMO and ubiquitin paths converge. Biochem. Soc. Trans. 38, 34–39 (2010).
Geoffroy, M. C. & Hay, R. T. An additional role for SUMO in ubiquitin-mediated proteolysis. Nat. Rev. Mol. Cell. Biol. 10, 564–568 (2009).
Gill, G. SUMO and ubiquitin in the nucleus: different functions, similar mechanisms? Genes Dev. 18, 2046–2059 (2004).
Hay, R. T. Decoding the SUMO signal. Biochem. Soc. Trans. 41, 463–473 (2013).
Rodriguez, M. S., Dargemont, C. & Hay, R. T. SUMO-1 conjugation in vivo requires both a consensus modification motif and nuclear targeting. J. Biol. Chem 276, 12654–12659 (2001).
Padma, R. & Nagarajan, L. The human PIM-1 gene product is a protein serine kinase. Cancer Res. 51, 2486–2489 (1991).
Chung, C. D. et al. Specific inhibition of Stat3 signal transduction by PIAS3. Science 278, 1803–1805 (1997).
Liu, B. et al. Inhibition of Stat1-mediated gene activation by PIAS1. Proc. Natl. Acad. Sci. USA 95, 10626–10631 (1998).
Shen, L. N., Dong, C., Liu, H., Naismith, J. H. & Hay, R. T. The structure of SENP1-SUMO-2 complex suggests a structural basis for discrimination between SUMO paralogues during processing. Biochem. J. 397, 279–288 (2006).
Palaty, C. K. et al. Identification of the autophosphorylation sites of the Xenopus laevis Pim-1 proto-oncogene-encoded protein kinase. J. Biol. Chem. 272, 10514–10521 (1997).
Telerman, A., Amson, R., Zakut-Houri, R. & Givol, D. Identification of the human pim-1 gene product as a 33-kilodalton cytoplasmic protein with tyrosine kinase activity. Mol. Cell. Biol. 8, 1498–1503 (1988).
Zippo, A., De Robertis, A., Serafini, R. & Oliviero, S. PIM1-dependent phosphorylation of histone H3 at serine 10 is required for MYC-dependent transcriptional activation and oncogenic transformation. Nat. Cell Biol. 9, 932–944 (2007).
Aho, T. L. et al. Pim-1 kinase promotes inactivation of the pro-apoptotic Bad protein by phosphorylating it on the Ser112 gatekeeper site. FEBS Lett. 571, 43–49 (2004).
Zhang, Y., Wang, Z., Li, X. & Magnuson, N. S. Pim kinase-dependent inhibition of c-Myc degradation. Oncogene 27, 4809–4819 (2008).
Wang, J. et al. Pim1 kinase is required to maintain tumorigenicity in MYC-expressing prostate cancer cells. Oncogene 31, 1794–1803 (2012).
Muraski, J. A. et al. Pim-1 regulates cardiomyocyte survival downstream of Akt. Nat. Med. 13, 1467–1475 (2007).
Tatham, M. H., Rodriguez, M. S., Xirodimas, D. P. & Hay, R. T. Detection of protein SUMOylation in vivo. Nat. Protoc. 4, 1363–1371 (2009).
Johnson, E. S. SUMO. Ann. Rev. Biochem 73, 355–382 (2004).
Geng, F., Wenzel, S. & Tansey, W. P. Ubiquitin and proteasomes in transcription. Annu. Rev. Biochem. 81, 177–201 (2012).
Lu, Z. et al. Activation of protein kinase C triggers its ubiquitination and degradation. Mol. Cell. Biol. 18, 839–845 (1998).
Kang, B. S., French, O. G., Sando, J. J. & Hahn, C. S. Activation-dependent degradation of protein kinase C eta. Oncogene 19, 4263–4272 (2000).
Galcheva-Gargova, Z., Theroux, S. J. & Davis, R. J. The epidermal growth factor receptor is covalently linked to ubiquitin. Oncogene 11, 2649–2655 (1995).
Huang, F., Kirkpatrick, D., Jiang, X., Gygi, S. & Sorkin, A. Differential regulation of EGF receptor internalization and degradation by multiubiquitination within the kinase domain. Mol. Cell 21, 737–748 (2006).
Aukrust, I. et al. SUMOylation of Pancreatic Glucokinase Regulates Its Cellular Stability and Activity. J. Biol. Chem. 288, 5951–5962 (2013).
de la Cruz-Herrera, C. F. et al. SUMOylation regulates AKT1 activity. Oncogene 34, 1442–1450 (2015).
Li, R. et al. Akt SUMOylation Regulates Cell Proliferation and Tumorigenesis. Cancer Res. 73, 5742–5753 (2013).
Lin, C. H., Liu, S. Y. & Lee, E. H. SUMO modification of Akt regulates global SUMOylation and substrate SUMOylation specificity through Akt phosphorylation of Ubc9 and SUMO1. Oncogene 35, 595–607 (2016).
Rubio, T., Vernia, S. & Sanz, P. Sumoylation of AMPKbeta2 subunit enhances AMP-activated protein kinase activity. Mol. Biol. Cell 24, 1801–1811, S1801–1804 (2013).
de la Cruz-Herrera, C. F. et al. Activation of the double-stranded RNA-dependent protein kinase PKR by small ubiquitin-like modifier (SUMO). J. Biol. Chem. 289, 26357–26367 (2014).
Kubota, Y., O’Grady, P., Saito, H. & Takekawa, M. Oncogenic Ras abrogates MEK SUMOylation that suppresses the ERK pathway and cell transformation. Nat. Cell Biol. 13, 282–291 (2011).
Yan, Y. et al. SUMOylation of AMPKalpha1 by PIAS4 specifically regulates mTORC1 signalling. Nat. Commun. 6, 8979 (2015).
Maure, J. F., Moser, S. C., Jaffray, E. G. A. F. A. & Hay, R. T. Loss of ubiquitin E2 Ube2w rescues hypersensitivity of Rnf4 mutant cells to DNA damage. Sci. Rep 6, 26178 (2016).
Jacobs, A. M. et al. SUMO modification of the DEAD box protein p68 modulates its transcriptional activity and promotes its interaction with HDAC1. Oncogene 26, 5866–5876 (2007).
Hogan, C. et al. Elevated levels of oncogenic protein kinase Pim-1 induce the p53 pathway in cultured cells and correlate with increased MDM2 in Mantle Cell Lymphoma. J. Biol. Chem. 283, 18012–18023 (2008).
Marcar, L. et al. MAGE-A Cancer/Testis Antigens Inhibit MDM2 Ubiquitylation Function and Promote Increased Levels of MDM4. PLoS One 10, e0127713, doi:10.1371/journal.pone.0127713 (2015).
We are grateful to Ellis Jaffray and Prof. Ron Hay for specialized reagents and invaluable assistance. This work was supported by Medical Research Scotland (grant number: PHD-601-2011) and part-sponsorship from CXR Biosciences.
The authors declare that they have no competing interests.
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Electronic supplementary material
About this article
Cite this article
Iyer, R.S., Chatham, L., Sleigh, R. et al. A functional SUMO-motif in the active site of PIM1 promotes its degradation via RNF4, and stimulates protein kinase activity. Sci Rep 7, 3598 (2017). https://doi.org/10.1038/s41598-017-03775-w