Pim kinases phosphorylate multiple sites on Bad and promote 14-3-3 binding and dissociation from Bcl-X_L

Pim kinases phosphorylate Bad on three sites in vitro

Bad has been described previously to be phosphorylated on three sites, Ser112, Ser136 and Ser155 in vivo. Peptides from around these sites were used as substrates for purified Pim-1, 2 or 3 in in vitro kinase assays. All three Pim kinases were found to phosphorylate the Ser112 peptide efficiently (Fig 1A–C). In contrast the Ser136 and Ser155 peptides were much less effective substrates for Pim-1 and Pim-3 (Fig 1A and 1C) and very poor substrates for Pim-2 (Fig 1B). As phosphorylation of synthetic peptides does not always give a true reflection of a kinases ability to phosphorylate the corresponding site in a protein, the ability of Pim kinases to phosphorylate recombinant Bad protein was also examined. As expected Bad was phosphorylated by all three Pim kinases, with maximum phosphorylation occurring in the first 5 to 10 min (Fig 1D–F). This activity was due to the Pim kinase, as incubation of the Bad protein in the absence of Pim (Fig D-F), or with kinase dead Pim (data not shown), did not result in any incorporation of ³²P into Bad. To determine the sites phosphorylated by Pim-1, 2 or 3, recombinant Bad was phosphorylated with ³²P labelled ATP and digested with trypsin. Peptides were then resolved by reverse phase HPLC and ms/ms used to identify the phospho peptides eluted (data not shown). The position of the phosphate in each peptide was also confirmed by solid phase sequencing reactions and measuring the cycle at which radioactivity was released. Pim-1 phosphorylated 4 major peptides in Bad, designated P1, P2 P3 and P4 (Fig 2A). Peak P1 was found to correspond to phospho Ser155, P2 to Ser136 while both P3 and P4 corresponded to phospho Ser112 (Table 1). The different elution times of the Ser112 peptides is probably explained by different oxidation states of the methionine residue in the peptide. Analysis of the cpm in each peak showed that Ser112 was the site most highly phosphorylated by Pim-1, with approximately 7 times the level of phosphorylation than occurred at either Ser136 or Ser155. This was consistent with what was found for the rate of phosphorylation of the synthetic peptides for each site (Fig 1A). Pim-2 was found to phosphorylate two major peaks that eluted at an identical position to the P3 and P4 peaks in the Pim-1 phosphorylated Bad (Fig 2B). Ms/ms analysis confirmed that these peptides also corresponded to Bad phosphorylated on Ser112. Small peaks were also seen corresponding to P1 and P2 in the Pim-2 phosphorylated Bad HPLC trace, and these were confirmed to correspond to phospho Ser155 and Ser136. The amount of phosphate incorporated at these sites was 15 times less than that at Ser112 (Table 1).

Pim-3 phosphorylated Bad also gave peaks corresponding to P1, P2, P3 and P4, and again these were confirmed as Ser112, Ser136 and Ser155 by ms/ms and solid phase sequencing (Fig 2C and 2D). Of the three kinases Pim-3 gave the highest phosphorylation of Ser136 and Ser155 compared to Ser112 in Bad (Table 1). Five further smaller phospho peptide peaks (P5-9) were also observed in the HPLC when Pim-3 was used to phosphorylate Bad. These sites were not however significantly phosphorylated by Pim-1 or Pim-2. P5 was observed in the flow through of the column and peptides in this fraction were not identified by mass spectrometry. Prediction from the Bad sequence shows that tryptic digest would give 2 small peptides that contain serine or threonine that could have been phosphorylated. Phosphopeptides from P9 were also not successfully identified by mass spectrometry. P7 was identified as the peptide AATNSHHGGAGAMETR (residues 92 to 107 of Bad) with 1 phospho group. Ms/ms analysis suggested that this was a mixture of phosphorylation on either Thr94 or Ser96. Solid phase sequencing confirmed this result, and also suggested that Ser96 was the predominant site (Fig 2D). P6 and P8 were identified by mass spectrometry as SAGTATQMR (residues 170 to 178 in Bad) phosphorylated at Ser170. The phosphorylation of Ser170 was confirmed by solid phase sequencing for both P6 and P8. Surprisingly, solid phase sequencing also showed that in P8 the 4^th residue in the peptide was also phosphorylated, suggesting that Thr 173 may also be phosphorylated by Pim-3.

Pim kinases are constitutively active in HEK-293 cells and phosphorylate Bad on Ser112, 136 and 155

To examine the regulation of Pim-1, 2 and 3 in cells, FLAG tagged Pim kinases were expressed in HEK-293 cells by transient transfection. Endogenous levels of Pim-1, 2 or 3 could not be detected in HEK-293 cells with the antibodies available. After serum starvation the cells well treated with various stimuli and the activity of the Pim kinases determined after immunoprecipitation with an anti-FLAG antibody. All three of the Pim kinases were found to be active in serum starved cells, and their activity was not further increased by stimulation of the cells with mitogenic signals (PMA or IGF), cellular stress (anisomycin, sorbitol, UV-C or hydrogen peroxide) or AICar, an activator of AMPK (Fig 3A).

GST-Bad expressed by transient transfection of serum starved HEK-293 cells, in the absence of the co-expression of any kinase, was found to be phosphorylated on Ser136, bur not Ser112 or Ser155, as judged by immunoblotting using antibodies specific for the various phosphorylation sites. Consistent with previous studies [33], stimulation of the cells with PMA resulted in a modest increase in Ser112 phosphorylation whilst forskolin resulted in increases in Ser155 phosphorylation and, to a lesser extent, Ser136 (Fig 3B). Co-expression of Pim-1 and Bad resulted in increased phosphorylation of Ser136 and phosphorylation of Ser112 and Ser155, when compared to Bad expressed in the absence of Pim-1. Similar results were also obtained for co-expression of Pim-2 or Pim-3, although Pim-2 had a reduced effect on Ser136 phosphorylation in Bad compared to co-expression with Pim-1 and Pim-3 (Fig 3B). The phosphorylation of Bad by Pim kinases was not further increased by PMA stimulation, although Ser155 phosphorylation was further increased by Forskolin stimulation. The phosphorylation of Bad was dependent on the kinase activity of Pim kinases, as co-expression of kinase dead versions of Pim-1, 2 or 3 and Bad did not result in the phosphorylation of Bad (Fig 3C).

The ability of Pim-1 to phosphorylate Bad in which one of the phosphorylation sites had been mutated was also determined by co-expression. As expected the mutation of Ser112 to alanine prevented phosphorylation of this site by Pim-1, but did not significantly affect the phosphorylation of either Ser136 or Ser155 (Fig 4A). Similarly, mutation of Ser155 prevented phosphorylation of this site by Pim-1 but did not significantly affect phosphorylation of Ser112. Surprisingly mutation of Ser155 to alanine did cause a reduction in the level of Ser136 phosphorylation by Pim-1 (Fig 4A) and a similar reduction in Ser136 phosphorylation was also seen in Ser112/Ser155Ala mutants (Fig 4B). Mutation of Ser136 to alanine not only blocked phosphorylation of Ser136 by Pim-1, but also greatly reduced Ser112 and Ser155 phosphorylation (Fig 4A). Mutation of Ser136 to an acidic residue to mimic phosphorylation was able to restore phosphorylation of Ser112 and Ser155 (Fig 4A).

Similar effects of the mutants were also seen in the absence of co-expressed Pim-1, when cells were stimulated with agents to induce the phosphorylation of specific sites within Bad. PMA promotes Bad Ser112 phosphorylation, probably via RSK, however this was prevented by Ser136 mutation. Forskolin stimulated Ser155 phosphorylation via PKA, and this was also prevented by mutation of Ser136 [see Additional file 1].

Pim promotes 14-3-3 binding to Bad via phosphorylation of Ser136 and dissociation from Bcl-X_L

Phosphorylation of a protein can create binding sites for 14-3-3 proteins, and Ser112 and 136 in Bad lie in a potential consensus 14-3-3 motif [38]. Ser155 does not lie in a consensus 14-3-3 binding motif, however in vitro phosphorylation of Bad by PKA (which predominantly phosphorylates Ser155) does promote 14-3-3 interaction [33]. To test the ability of Bad to interact with 14-3-3, GST-Bad was co-expressed in HEK-293 cells with Pim-1, and 14-3-3 overlay assays were used to determine if Bad could bind 14-3-3 proteins. In the overlay assay, recombinant 14-3-3 protein was used to detect the presence of proteins on a Western blot capable of binding 14-3-3. GST-Bad isolated by glutathione-Sepharose pull downs from the cell lysates was tested for its ability to bind to 14-3-3. In the absence of Pim-1 co-expression, GST-Bad was not significantly phosphorylated on Ser112, 136 or 155 and did not bind 14-3-3 in the overlay assay (Fig 4A and 4C). Co-expression of Pim-1 however resulted in phosphorylation of all three sites and promoted interaction of the GST-Bad with 14-3-3 in the overlay assay. Mutation of either Ser112 or Ser155 to alanine resulted in a modest reduction in the ability of the Pim-1 phosphorylated GST-Bad to interact with 14-3-3 proteins in the overlay assay (Fig 4C). As expected, mutation of Ser136 to alanine, which also blocks the phosphorylation of Ser112 and Ser155, completely blocked the ability of GST-Bad to interact with 14-3-3 proteins. Mutation of Ser136 to an acidic residue also blocked the ability of Bad to bind to 14-3-3 proteins in the overlay assay, even though Pim-1 was able to phosphorylate both Ser112 and Ser155 in the Ser136Asp or Ser136Glu Bad mutants (Fig 4C). These data suggest that Ser136 is the major 14-3-3 binding site. Consistent with this, a double mutant of both Ser112 and Ser155 to alanine, which was phosphorylated on Ser136 (Fig 4B), was still able to interact with 14-3-3 proteins in the overlay assay (Fig 4D). Similar to Pim-1, when Pim-2 or Pim-3 were co-expressed with GST-Bad, they were also able to promote 14-3-3 binding in the overlay assay (Fig 4E). To test if Pim kinases promoted the interaction of Bad and 14-3-3 proteins in cells, Pim kinases and GST-Bad were co-expressed in HEK-293 cells and the GST-Bad isolated by glutathione-Sepharose pull downs and blotted for the presence of endogenous 14-3-3 proteins using a pan 14-3-3 isoform antibody (Fig 4F). In the absence of Pim kinase co-expression, 14-3-3 proteins were not detected in the GST-Bad pull downs, however expression of Pim-1, 2 or 3 was found to promote the binding of 14-3-3 to GST-Bad in cells (Fig 4F). The ability of Pim kinases to promote the interaction of Bad and 14-3-3 was dependent on their kinase activity, as co-expression of kinase dead Pim-1, 2 or 3 did not promote 14-3-3 binding as judged by either co-precipitation or 14-3-3 overlay assays (Fig 4E and 4F).

Phosphorylation of Bad has been shown to prevent an interaction with Bcl-X_L [39]. The ability of Pim kinases to prevent the interaction of GST-Bad with endogenous Bcl-X_L was tested in pull down experiments. HeLa cells were used for these experiments as these cells express detectable levels of endogenous Bcl-X_L.

As expected, GST-Bad, which was not significantly phosphorylated in the absence of kinase co-expression, was able to interact with Bcl-X_L in HeLa cells. Co-expression of Pim-1, 2 or 3 resulted in phosphorylation of GST-Bad and reduced the interaction of GST-Bad with Bcl-X_L. This was dependent on the kinase activity of Pim, as co-expression of kinase dead Pim mutants did not have this effect (Fig 5A). Similar to the 14-3-3 interaction, Ser136 appeared to be the critical residue involved in this process. Mutation of Ser112, Ser155 (Fig 5B), or both Ser112 and Ser155 to alanine (Fig 5C) in Bad did not affect the ability of Pim-1 to prevent the interaction of GST-Bad and Bcl-X_L. Mutation of Ser136 in Bad to alanine (which blocked phosphorylation of all three sites) or an acidic residue (which did not affect Ser112 or Ser155 phosphorylation) blocked the ability of Pim-1 to promote the dissociation of GST-Bad and Bcl-X_L (Fig 5B).

Plasmids

The coding region for an amino-terminal extended version of human Pim-1 (NCBI Acc. NM_002648) predicted to be synthesised by alternative translation initiation at an upstream CUG codon [46] was amplified from IMAGE consortium EST clone 4591723 using primer 5'-gcgaattcgccaccatggactacaaggacgacgatgacaagctg ccgcacgagccccacgag-3' incorporating a 5' FLAG tag sequence and primer 5'- gcgaattcctatttgctgggccccggcgac-3'. The PCR product was ligated into pCR2.1-TOPO vector (Invitrogen), sequenced, and sub-cloned as an Eco R1-Eco R1 insert into the pCMV5 expression vector [47]. The same coding region was cloned into expression vector pEBG2T [48] without the 5' FLAG tag sequence using appropriate primers as a Spe 1-Spe 1 insert. The coding region for an amino-terminal extended version of human Pim-2 (NCBI Acc. NM_006875.2) predicted to be synthesised by alternative translation initiation at an upstream CUG codon was amplified from IMAGE consortium EST clone 3913552 using primer 5'- gcgaattcgccaccatggactacaaggacgacgatgacaagctg gcgcgcgcggcgaatctcaac-3' incorporating a 5' FLAG tag sequence and primer 5'- gcggatccttagggtagcaaggaccaggccaaag-3'. The PCR product was ligated into pCR2.1-TOPO vector, sequenced, and sub-cloned as an Eco R1-Bam H1 insert into the pCMV5 expression vector. The same coding region was cloned into expression vector pEBG2T without the 5' FLAG tag sequence using appropriate primers as a Bam H1-Not 1 insert. Nucleotides 125–1455 (stop codon of CDS) of murine Pim-3 (NCBI Acc. NM_145478) mRNA were amplified from IMAGE consortium EST clone 4924687 using primer 5'-gcgaattcgagcgtccggcttccccag-3' and primer 5'-gcgaattctcacttgtcatcgtcgtccttgtagtccaagctctcactgctggaagtg-3' incorporating a FLAG tag sequence at the 3' end. The PCR product was ligated into pCR2.1-TOPO vector, sequenced, and sub-cloned as an Eco R1-Eco R1 insert into expression vector pCMV5. The same region of Pim-3 mRNA was cloned (lacking the Pim-3 stop codon) with a C-terminal GST sequence (generated by overlap PCR) into pCMV5 as a Kpn1/HindIII insert. The Pim-3 sequence was also cloned with an N-terminal His Tag in to a pFastBAC vector for expression in Baculovirus expression. Kinase dead versions of the Pim-1, Pim-2 and Pim-3 expression constructs were generated by mutating the 'DFG' motif to 'AFG' in each case using the QuikChange Site-Directed Mutagenesis Kit (Stratagene). A cDNA encoding full-length murine Bad was amplified from the pGEX Bad construct described in [33] using primer 5'- gcggatccatgggaaccccaaagcagccctc-3' and primer 5'- gtgcggccgctcactgggagggggtggagcc-3'. The PCR product was ligated into pCR2.1-TOPO vector, sequenced, and sub-cloned as a BamH 1-Not 1 insert into expression vector pEBG6P-1 (pEBG2T containing a PreScission Protease site). S112A, S155A, S136A, S136D and S136E mutants were generated using the QuikChange Site-Directed Mutagenesis Kit (Stratagene).

Cell culture

HEK-293 cells were maintained in DMEM containing 10% FBS (Sigma), 2 mM L-glutamine, 50 units/ml penicillin G and 50 μg/ml streptomycin (Invitrogen). Cells were transfected using PEI as described [49]. Prior to stimulation and analysis cells were starved in DMEM lacking serum for eighteen hours at 37°C and 5% CO₂. Cells were stimulated with IGF (100 ng/ml 5 min), sorbitol (0.5 M for 30 min), AICar (2 mM for 1 hour), UV-C (200 Jm²/10 sec followed by 1 hour incubation at 37°C), hydrogen peroxide (1 mM for 1 hour), PMA (400 ng/ml for 15 min), anisomycin (10 μg/ml for 1 hour) or Forskolin (20 μM for 30 min). Cells were lysed in 50 mM Tris-HCl pH 7.5, 1 mM EGTA, 1 mM EDTA, 1 mM sodium orthovanadate, 50 mM sodium fluoride, 1 mM sodium pyrophosphate, 0.27 M sucrose, 1% (v/v) Triton X-100, 0.1% (v/v) 2-mercaptoethanol. Lysates were clarified at 13,000 rpm for 5 min at 4°C and the supernatants quick frozen in liquid nitrogen and stored at -80°C until used.

Peptide kinase assay

For in vitro kinase assays 1 unit of activity was defined as the amount of kinase that catalysed the phosphorylation of 1 nmol of substrate in 1 min into the peptide RSRMSSYPDRG at a 30 μM peptide concentration. To determine the ability of Pim kinases to phosphorylate peptides derived from Bad, each peptide substrate was used at the concentrations indicated and incubated with 10 mU of kinase in a 50 μl volume for 10 min at 30°C in 10 mM MgCl₂ – 0.1 mM [γ³²P]-ATP, 50 mM Tris-HCl, pH 7.5, 10 mM EGTA, 0.1% 2-mercaptoethanol and 2.5 μM PKI. Reactions were terminated by spotting on to p81 paper, followed by immersion in 50 mM ortho-phosphoric acid. All papers were then washed three times in 50 mM ortho-phosphoric acid to remove unincorporated ATP, once in acetone and then dried and the ³²P incorporation was determined by Cerenkov counting. The peptides used were RSRMSSYPDRG (corresponding to Ser112), RGSRSAPPNL (corresponding to Ser136) and RELRRMSDEFEGS (corresponding to Ser155).

To determine the activity of Pim kinases expressed in Hek-293 cells, FLAG-Pim kinase was immunoprecipitated from 50 μg of soluble cell lysate using an anti FLAG antibody (Sigma) and assayed by standard protocols [50]. Briefly, immunoprecipitates were washed twice in 0.5 M NaCl, 50 mM Tris-HCl pH 7.5, 0.1 mM EGTA and 0.1% (v/v) 2-mercaptoethanol and once in 50 mM Tris-HCl pH 7.5, 0.1 mM EGTA and 0.1% (v/v) 2-mercaptoethanol. Precipitates were then resuspended in 35 μl of reaction buffer (containing Tris/HCl pH7.5, EGTA, PKI and RSRMSSYPDRG substrate peptide,), and the reaction was started by the addition of 10 μl 50 mM MgAc, 0.5 mM [γ³²P]-ATP and incubated at 30°C for 15 min. Final concentrations of reagents in the assay were 50 mM Tris/HCl pH7.5, 0.1 mM EGTA 0.1% 2-mercaptoethanol 2.5 μM PKI (an inhibitor of PKA), 30 μM substrate peptide peptide, 10 mM MgAc and 0.1 mM [γ³²P]-ATP. Reactions were stopped by transfer onto P81 paper and washing in 75 mM orthophosphoric acid.

In vitro Bad phosphorylation

Human Pim-1 and Pim-2 were cloned with an N-terminal GST-tag and expressed in HEK-293 cells. Murine Pim-3 was cloned with an N-terminal His-tag and expressed in a Baculovirus expression system using Sf 9 insect cells. Murine Bad was cloned with an N-terminal GST-tag and expressed in bacteria. All recombinant proteins were purified using standard affinity chromatography techniques on glutathione or Ni-NTA sepharose as appropiate. Purified recombinant Pim-kinases (10 mU) were incubated with 2 μg of recombinant Bad over the indicated time period in 50 mM Tris-HCl, pH 7.5, 10 mM EGTA, 0.1% 2-mercaptoethanol, 2.5 μM PKI, 10 mM MgCl₂ and 0.1 mM [γ³²P]-ATP. Reactions were stopped by the addition of SDS loading buffer and samples analysed on a 4–12% Novex gel (Invitrogen). Protein was detected by Colloidal blue staining and the appropriate band was excised and the amount of ³²P incorporation was determined by Cerencov counting.

Phosphorylation site identification

Samples were run on a 4–12% polyacrylamide gel, and the substrate band was excised and the amount of ³²P incorporated determined by Cerenkov counting. For phospho site mapping, the protein was reduced with dithiothreitol, alkylated with iodoacetamide and digested with trypsin. The resulting peptides were applied to a Vydac 218TP5215 C18 column equilibrated in 0.1% trifluoroacetic acid and the column developed with a linear gradient of acetonitrile/0.1% trifluoroacetic acid at a flow rate of 0.2 ml/min while collecting 0.1 ml fractions. ³²P radioactivity was recorded with an on-line monitor. Phosphorylation site mapping was performed essentially as described previously [51]. Identification of ³²P labelled peptides was performed by MALDI-TOF and MALDI-TOF-TOF mass spectrometry on an Applied Biosystems 4700 using a matrix of 10 mg/ml alpha-cyanocinnamic acid in 50% acetonitrile/0.1% trifluoroacetic acid/10 mM ammonium phosphate. Sites of phosphorylation within the peptides were determined by a combination of MALDI-TOF-TOF mass spectrometry and solid phase Edman sequencing. Solid phase sequencing was performed on an Applied Biosystems Procise 494C after coupling the peptide covalently to a Sequelon-arylamine membrane and measuring by Cerenkov counting the ³²P radioactivity released after each cycle.

GST-Bad pull down

Lysates were incubated with glutathione-sepharose for 2 hours at 4°C. Immunoprecipitates were pelleted by centrifugation for 1 min at 13,000 g and washed twice in 50 mM Tris-HCl, pH 7.5, 500 mM NaCl, 0.1 mM EGTA and 0.1 % 2-mercaptoethanol and once in 50 mM Tris-HCl, pH 7.5, 0.1 mM EGTA and 0.1 % 2-mercaptoethanol. Protein complexes were eluted in SDS sample buffer.

Immunoblotting

Soluble cell extract was run on 4–12% Novex gels (Invitrogen) and transferred onto nitrocellulose membranes and blotted using standard techniques. Antibodies that recognise p-Bad (Ser112), p-Bad (Ser 136), pBad (Ser155) and Bcl-X_L were from Cell Signalling Technology. An antibody that detected total Bad protein was raised in-house. The Flag antibody was from Sigma and HRP conjugated secondary antibodies were from Pierce (Cheshire, UK) and blots were developed using the enhanced ECL reagent (Amersham, Buckinghamshire, UK).

14-3-3 overlay assays

14-3-3 overlay assays were performed as described [52]. Briefly, proteins from cell extracts or GST pull downs were separated on 4–12% Novex gels (Invitrogen) and transferred to nitrocellulose membrane. Membranes were then treated as for immunoblots. Digoxygenin (DIG)-labelled 14-3-3 proteins were incubated with the membranes in place of the primary antibody, followed by incubation with a secondary anti-DIG horseradish peroxidase antibody. Blots were developed using the enhanced ECL reagent (Amersham, Buckinghamshire, UK).