The p34 cdc2 -related Cyclin-dependent kinase 11 Interacts with the p47 Subunit of Eukaryotic Initiation Factor 3 during Apoptosis*

Cyclin-dependent kinase 11 (CDK11; also named PITSLRE) is part of the large family of p34 cdc2 -related kinases whose functions appear to be linked with cell cycle progression, tumorigenesis, and apoptotic signaling. However, substrates of CDK11 during apoptosis have not been identified. We used a yeast two-hybrid screening strategy and identified eukaryotic initiation factor 3 p47 protein (eIF3 p47) as an interacting partner of caspase-processed C-terminal kinase domain of CDK11 (CDK11p46). We demonstrate that the eIF3 p47 can interact with CDK11 in vitro and in vivo, and the interaction can be strengthened by stimulation of apoptosis. EIF3 p47 contains a Mov34/JAB domain and appears to interact with CDK11p46 through this motif. We show in vitrothat the caspase-processed CDK11p46 can phosphorylate eIF3 p47 at a specific serine residue (Ser46) and that eIF3 p47 is phosphorylated in vivo during apoptosis. Purified recombinant CDK11p46 inhibited translation of a reporter gene in vitro in a dose-dependent manner. In contrast, a kinase-defective mutant CDK11p46M did not inhibit translation of the reporter gene. Stable expression of CDK11p46 in vivo inhibited the synthesis of a transfected luciferase reporter protein and overall cellular protein synthesis. These data provide insight into the cellular function of CDK11 during apoptosis.

It is becoming increasingly clear that in addition to controlling the cell cycle, cyclin-dependent kinases may have other functions within the cell. The CDK11 (also known as PITSLRE) protein kinases are members of the cyclin-dependent kinase superfamily. Two distinct but closely related human CDK11 genes (Cdc2L 1 and Cdc2L 2) express several CDK11 isoforms (1,2). CDK11 homologues exist in several species including humans, mice, chickens, Caenorhabditis elegans, Drosophila melanogaster, and Xenopus (3)(4)(5). The highly conserved nature of the CDK11 suggests important cellular functions.
Although CDK11 is a member of the cyclin-dependent kinase superfamily, their function within the cell is not totally clear. However, recent studies indicate that the p110 isoform of CDK11 (CDK11 p110 ) may be involved in some aspect of RNA processing or transcription by virtue of the fact that CDK11 co-immunoprecipitate and/or co-purify with multiple transcriptional elongation factors (6,7). Furthermore, CDK11 p110 associates with cyclin L (8). Cyclin L is an RS domain protein that may function in pre-mRNA splicing (8). These observations suggest that CDK11 p110 kinases play some role in the production of translatable RNA transcripts in proliferating cells.
The CDK11 p110 isoforms contain an internal ribosome entry site, which leads to the generation of a CDK11 p58 isoform during the G 2 /M phase of the cell cycle (9). Elevated expression of CDK11 p58 in eukaryotic cells alters normal cytokinesis and can delay cells in late telophase (10). In contrast, reduced expression of CDK11 p58 increases DNA replication and enhances cell growth (11). This p58 isoform of CDK11 can interact with cyclin D3 (12). This suggests that CDK11 p58 /cyclin D3 may play role in mitosis.
In regard to apoptosis, increased expression of CDK11 p58 reduces cell growth due to apoptosis (1). In addition, our group and others have shown that the CDK11 p110 isoform and the CDK11 p58 isoform are cleaved by caspases to generate a smaller 46 -50-kDa protein that contains the catalytic portion of the protein (2,13,14). This smaller CDK11 p46 isoform can be triggered by Fas, tumor necrosis factor A, or staurosporine and phosphorylate histone H1. Caspase inhibitors can modulate the kinase activity of the caspase-processed isoform of CDK11 (2). Collectively, these observations suggest that CDK11 may play a role in apoptotic signaling. However, substrates potentially regulated by CDK11 during apoptosis have not been identified. The present study was performed to address this issue. We identify the p47 subunit of eukaryotic initiation factor 3 (eIF3 1 p47) as a protein associated with the caspaseprocessed isoform of CDK11 (CDK11 p46 ) using a yeast twohybrid screen. The interaction between CDK11 p46 and eIF3 p47 occurs in vitro and in vivo. In addition, CDK11 protein kinase isolated from cells undergoing apoptosis can phosphorylate eIF3 p47 in vitro, and serine phosphorylation of eIF3 p47 occurs in cells during apoptosis. Taken together, these results strongly support our hypothesis that CDK11 may be involved in apoptotic signaling by interacting with eIF3 p47.
Antibodies-GN1 is an affinity-purified rabbit polyclonal antibody raised by injection of rabbit with purified recombinant glutathione S-transferase (GST) fused with the sequence coding for amino acids 341-413 of CDK11 p110 (Rockland, Gilbertsville, PA). CDK11 monoclonal antibody P1C recognizes the last 75 amino acids of CDK11 p110 and was a gift from Drs. Vincent Kidd and Jill Lahti (St. Jude Children's Research Hospital, Memphis, TN) (7). EIF3 p47 antibody is a polyclonal antibody raised by injection of goat with purified recombinant GST-fused full-length eIF3 p47 (Rockland). eIF3 antibody was raised in a goat as described (15). Rabbit polyclonal HA antibody, monoclonal c-Myc antibody and monoclonal anti-phosphoserine antibody were purchased from Sigma.
Yeast Two-hybrid Assay-Matchmaker GAL4 two-hybrid system 2 (Clontech, Palo Alto, CA) was used to perform yeast two-hybrid screening according to the manufacturer's instructions. The wild type and K451M substituted CDK11 C-terminal portion (CDK11 p46 ) were used as baits. In the K451M altered CDK11 p46 , one lysine residue that is associated with phosphate transfer was substituted with a methionine. This substitution was performed to stabilize the interaction between the kinase and its substrate (16). A total of 5 ϫ 10 6 transformants from a human fetal brain Matchmaker cDNA library (Clontech) were screened in the yeast strain AH109 (Clontech), and 28 colonies were identified as His ϩ Ade ϩ , out of which 13 were positive for ␤-galactosidase. Among the 13 positive clones, 12 were isolated when the K451M mutant CDK11 p46 was used as bait, whereas one clone was isolated using wild type CDK11 p46 as bait. One of these clones (clone 4) was revealed to encode the full-length cDNA to the p47 subunit of eukaryotic translation initiation factor 3 gene by sequence analysis. To confirm the CDK11 p46 /eIF3 p47 interaction in yeast, 0.1 g of clone 4, and wild type bait plasmid were co-transfected into yeast strain AH109 using the lithium acetate transformation method. Growth selection was performed according to the manufacturer's protocol (Clontech).
Construction of CDK11 p46 Vectors-pCMV-Myc-CDK11 p46 containing the CDK11 p46 coding sequence (GenBank TM accession number U04824; nucleotides 1282-2465) with a c-Myc tag at the N terminus was constructed. CDK11 p46 was generated by PCR using p110KS ϩ (a gift from Dr. Vincent Kidd) as template. pGEX-4T-2-CDK11 p46 that produces a GST fusion CDK11 p46 in bacteria was constructed by inserting the same PCR product of CDK11 p46 in frame into the EcoRI site of pGEX-4T-2 vector (Amersham Biosciences). pcDNA3-CDK11 p46 was constructed by cloning the PCR amplified CDK11 p46 with an in frame start codon ATG (nucleotides 1121-2465) into the EcoRI site of a eukaryotic expression vector pcDNA3 (Invitrogen) for stable transfection purposes.
Construction of eIF3 p47 Vectors-pCMV-HA-eIF3 p47, containing the full-length coding sequence to the eIF3 p47 with a HA tag at the N terminus, was constructed. Full-length eIF3 p47 was isolated from the original fetal brain cDNA clone we obtained by yeast two-hybrid assay using EcoRI/XhoI restriction enzymes. The full-length eIF3 p47 was cloned into pGEX-4T-2 to generate GST-eIF3 p47 fusion protein expressed in E. coli. Four DNA fragments encoding truncated eIF3 p47, amino acids 106 -360, 170 -360, 248 -360, and 1-113 were obtained by PCR amplification and cloned into pGEX-4T-2 vector to generate deletion GST-p47 fusion proteins.
Construction of CDK11 p46 Mutants-The K451R mutation was generated by polymerase chain reaction using a QuikChange TM site-directed mutagenesis kit (Stratagene, La Jolla, CA) with CDK11 p46 wild type construct as template. Sequencing analysis was used to confirm the mutation. This mutation has been shown to cause CDK11 to lose its kinase activity (1). Therefore, we created this mutation in the ATP binding domain to inactivate the kinase. The mutagenesis primers were 5Ј-GAAATTGTGGCTCTAAGGCGGCTGAAGATGG-3Ј and 5Ј-CCAT-CTTCAGCCGCCTTAGAGCCACAATTTC-3Ј.
Immunoprecipitation and Western Blotting-Cells were harvested, washed twice with cold PBS, and lysed in lysis buffer (10 mM Hepes, pH 7.2, 142.5 mM KCl, 5 mM MgCl 2 , 1 mM EGTA, 0.2% Nonidet P-40) containing 1 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, and 1% protease inhibitor mixture (Sigma) for 30 min on ice. Following lysis, cells were centrifuged at 13,000 ϫ g for 10 min at 4°C, and the protein content was determined using the bicinchoninic acid assay (Pierce). Total cell lysate (500 g) was precleared with protein Aor G-agarose beads (Oncogene, La Jolla, CA) and goat IgG (Santa Cruz Biotechnology, Santa Cruz, CA) at 4°C for 1 h. Myc-CDK11 p46 fusion protein or eIF3 was then immunoprecipitated using c-Myc monoclonal antibody (Sigma) or goat eIF3 polyclonal antibody (15) and protein A-or G-agarose overnight at 4°C. The immune complex was then washed three times with lysis buffer and subjected to SDS-PAGE. The proteins were transferred to a polyvinylidene difluoride membrane (Bio-Rad), and the blots were probed with different antibodies. A secondary probe with horseradish peroxidase-labeled antibodies (Sigma) was detected by enhanced chemiluminescence (ECL) (Amersham Biosciences).
Immunofluorescence Confocal Microscopy-pCMV-HA-eIF3 p47 transiently transfected A375 cells were grown on coverslips, washed twice with cold PBS, and fixed in 4% paraformaldehyde in PBS for 20 min at room temperature. Cells were rinsed three times with PBS and permeabilized with 100% methanol at Ϫ20°C for 6 min. Cells on coverslips were washed with PBS again and incubated with 5% bovine serum albumin in PBS for 10 min at room temperature, and then the bovine serum albumin was removed. The coverslips were then incubated with CDK11 p110 -specific antibody P1C (1:500 dilution) and rabbit anti-HA antibody (1:50 dilution) for 1 h at room temperature. Coverslips were washed three times with PBS for 5 min each and then incubated with fluorescence-labeled secondary antibodies fluorescein isothiocyanate-conjugated anti-mouse IgG ϩ IgM (1:100 dilution) and Cy5-conjugated anti-rabbit IgG (1:100 dilution) in darkness for 1 h at room temperature. Following incubation, coverslips were washed three times with PBS, mounted with DAKO mounting medium, and stored at 4°C overnight for immunofluorescence confocal microscopy analysis.
Cell Fractionation-A375 cells were washed with PBS and resuspended in hypotonic buffer (10 mM HEPES, pH 7.9, 1.5 mM MgCl 2 , 10 mM KCl, 0.5 mM dithiothreitol) containing 1 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, and 1% protease inhibitor mixture (Sigma) and incubated on ice for 15 min. Cells were then homogenized using the Dounce tissue grinder (Wheaton, Millville, NJ) with 25-30 strokes, and the percentage of nuclei was checked by trypan blue staining. Nuclei were collected at 2000 rpm for 10 min. The supernatant contained the cytoplasm.
Purification of Recombinant Protein from E. coli-GST, GST-CDK11 p46 , GST-CDK11 p46M , and GST-p47 were induced by 0.2 mM isopropyl-1-thio-␤-D-galactopyranoside and expressed in BL21 bacteria for 4 h at 30°C. The recombinant proteins were purified using Bulk GST Purification Module according to the manufacturer's instruction (Amersham Biosciences). Purified proteins were concentrated using Centricon 30 (Amicon Inc.) to an appropriate concentration and stored as aliquots at Ϫ70°C.
GST Pull-down Assay-The assay was performed as described previously (12) with a few modifications. Briefly, GST or GST fusion proteins were expressed in BL21 cells, and equal amounts of bacterial lysates were incubated with 25 l of glutathione-Sepharose beads for 30 min. The beads were then washed three times with PBS and incubated with 5 l of in vitro transcribed and translated [ 35 S]methionine-labeled CDK11 p46 overnight at 4°C. [ 35 S]Methionine-labeled CDK11 p46 was produced using a TNT-coupled reticulocyte lysate system (Promega). The beads were then washed five times with binding buffer and boiled in SDS sample buffer. The bound CDK11 p46 protein was analyzed by autoradiography after resolved by SDS-PAGE.
Kinase Assay-A375 cells were treated with anti-Fas (Upstate Biotechnology) or with both anti-Fas and caspase-3 inhibitor DEVD-FMK (50 M) for 0, 24, 36, and 48 h and harvested, and kinase assays were carried out as described (2). Lysates of A375 cells stably transfected with kinase mutant CDK11 p46M were also used in the kinase assays. 5 g of either GST or GST-eIF3 p47 were used as substrates for kinase assays. Phosphorylation of the substrates was analyzed by 12% SDSpolyacrylamide gel electrophoresis and autoradiography. The gel was exposed overnight at room temperature. Amersham Biosciences Image-Quant software was used to quantitate the relative phosphorylation level of the substrates.
Identification of eIF3 p47 Phosphorylation Site-Caspase-processed CDK11 p110 was immunoprecipitated from anti-Fas-treated A375 cells. Recombinant eIF3 p47 was then phosphorylated by caspase-processed CDK11 p110 as previously described (2) and was separated by SDS-PAGE. The recombinant eIF3 p47 was excised from the gel, digested with pepsin, and analyzed for phosphoamino acids by mass spectrometry as previously described (17,18) by the Arizona Cancer Center Proteomic Core Facility.
In Vitro Translation Inhibition Assay-First, a time course study and a dose response study were performed to determine the linear range of luciferase production using the rabbit reticulocyte lysate system. 50 ng of luciferase mRNA were added into the system and incubated at 30°C for 0 -120 min followed by measurement of luciferase activity every 15 min to determine the luciferase production rate over the time course. The production of luciferase is in a linear range up to 75 min of incubation. 20 -50 ng of luciferase mRNA were incubated with the system at 30°C for 60 min followed by measurement of luciferase activity. The production of luciferase is in a linear range between 20 and 50 ng of luciferase mRNA. Therefore, 40 ng of firefly luciferase mRNA was added to 50 l of a rabbit reticulocyte lysate in vitro translation reaction (Promega, Madison, WI) in the presence of recombinant purified 18 -70 nM of CDK11 p46 , CDK11 p46M , 70 nM GST, or elution buffer. The reaction mixtures were incubated at 30°C for 1 h. Luciferase activity was measured with a luminometer. In addition, newly synthesized 35 S-labeled luciferase proteins were analyzed by subjecting equal amounts of reaction mixtures to 10% SDS-PAGE and autoradiography.

RT-PCR and Quantitative Real Time One-step RT-PCR-Total
RNAs were extracted from A375-pcDNA3 and A375-CDK11 p46 cells transfected with pGL3 control vector and A375 cells untransfected with pGL3 in triplicate using RNeasy Mini Kit (Qiagen, Valencia, CA). Reverse transcription was performed with 2 g of total RNA using the Omniscript RT kit (Qiagen). 2 l of the cDNA mix was added to 20 l of the PCR containing 2.5 units of Taq polymerase (TaKaRa, Otsu, Shiga, Japan), 1ϫ buffer, and 1 pmol of primers. For histone, PCR was carried out at 95°C for 5 min followed by 30 cycles at 95°C for 30 s, 55°C for 30 s, and 72°C for 45 s with a final extension at 72°C for 5 min using the primers Histone-F (5Ј-CCACTGAACTTCTGATTCGC-3Ј) and Histone-R (5Ј-GCGTGCTAGCTGGATGTCTT-3Ј). For luciferase, PCR was carried out at 95°C for 5 min followed by 30 cycles at 95°C for 1 min, 65°C for 1 min, and 72°C for 1 min with a final extension at 72°C for 10 min using the primers LucF (5Ј-AATTGCTTCTGGTGGCGCTC-3Ј) and LucR (5Ј-GGGGTGTTGGAGCAAGATGG-3Ј). The amplified fragments were analyzed by 1% agarose gel electrophoresis and stained with ethidium bromide. Real time RT-PCR was performed using SYBR® Green PCR Core Reagents (PerkinElmer Life Sciences) and Omniscript RT kit (Qiagen) and amplified in a PerkinElmer Life Sciences prism 5700 sequence detection system according to the manufacturer's instructions. Dissociation curves were performed to determine the specificity of the amplicon. Each sample was completed in triplicate, and no template control was included. Threshold cycle (Ct) during the exponential phase of amplification was determined by real time monitoring of fluorescent emission. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a control gene. Primers for GAPDH were purchased from PerkinElmer Life Sciences. Primers for luciferase are 5Ј-CGTCTTAATGTATAGATTTGAAGAAG-3Ј and 5Ј-TGGCGAAGAA-GGAGAATAG-3Ј. Luciferase mRNA levels were demonstrated as the absolute number of copies normalized against GAPDH mRNA. Difference in amplification was indicated as 1/(Ct luciferase Ϫ Ct GAPDH) (19,20).
Overall Cellular Protein Synthesis Inhibition Assay-A375 or A375-CDK11 p46 cells were plated at 2 ϫ 10 4 cells/well in 96-well plates in RPMI 1640 medium without leucine (U.S. Biological, Swampscott, MA) in triplicate and incubated at 37°C for the indicated time. Cells were either treated with ethanol or treated with 10 ng/ml of staurosporine for the indicated time. Apoptosis and Cell Survival Assays-Cells were stained with 7-aminoactinomycin D (7-AAD) and analyzed for apoptosis using a FACStar flow cytometer (Becton Dickinson, San Jose, CA) as described (2). Cell morphology analysis was performed by cytospin preparation of cells followed by methanol fixation for 6 min. Slides were air dried and stained for 15 min in Giemsa stain. Slides were then rinsed in deionized water and air-dried followed by light microscopy analysis at ϫ15 magnification. Cell viability was measured with a 3-(4,5-dimethylthiazol-2yl)-2,5-diphenyltetrazolium bromide assay as described (2).

RESULTS
The Kinase Domain of the CDK11 Protein Kinase (CDK11 p46 ) Interacts with the Mov34/JAB_MPN Domain of eIF3 p47-The yeast two-hybrid method was used to screen a human fetal brain cDNA library using the caspase processed C-terminal portion containing the kinase domain of CDK11 p110 (CDK11 p46 ) as bait, and 13 CDK11 p46 -interacting cDNA clones were isolated. DNA sequencing and data base searching revealed that the nucleotide sequence of clone 4 encoded the full-length human p47 subunit of eukaryotic initiation factor 3 (eIF3 p47). The specificity of the interaction was confirmed by co-transformation of the bait and clone 4 again into yeast AH109, and various positive controls and negative controls were included in the nutritional deficiency selection. Characterization of the other 12 clones is in progress in our laboratory.
The interaction of CDK11 p46 with eIF3 p47 was further tested in vitro using a GST pull-down assay. The in vitro translated 35 S-labeled CDK11 p46 could directly and specifically bind to GST-p47 eIF3 fusion protein (Fig. 1B). EIF3 p47 is a Mov34 family member. The function of the Mov34/JAB_MPN domain in eIF3 p47 has not been established. We speculated that Mov34/JAB_MPN related domain in eIF3 p47 might be involved in protein-protein interactions. To determine the re- gion in eIF3 p47 interacting with CDK11 p46 , we deleted specific sequences within eIF3, made fusion proteins, and performed GST pull-down assays (Fig. 1A). In the GST pull-down assay, in vitro translated 35 S-labeled CDK11 p46 protein could be pulled down by GST-p47 amino acids 106 -360 and GST-p47 amino acids 170 -360 but not by GST-p47 amino acids 248 -360 or GST-p47 amino acids 1-113 (Fig. 1, A and B). This suggests that the region within eIF3 p47 that directly interacts with CDK11 p46 is between amino acids 113 and 248, which contains the Mov34/JAB_MPN domain. Therefore, the interaction of the caspase-processed isoform of CDK11 (CDK11 p46 ) with eIF3 p47 may be mediated through the Mov34/JAB_MPN domain in vitro, and the major region of interaction appears to be the C-proximal half of the Mov34 domain.
To further demonstrate the interaction in human cells, an expression vector with an HA epitope-tagged full-length eIF3 p47 was constructed and transfected into A375 cells. Immunofluorescence confocal microscopy using an anti-HA antibody revealed that p47 localized both in nucleoplasm and cytoplasm, especially the region around nucleus (Fig. 2A). CDK11 p110 localizes predominantly to the nucleoplasm as revealed by P1C antibody. No fluorescence was seen with the HA antibody in untransfected cells (data not shown). Merging of the two images showed a yellow color in the nucleoplasm ( Fig. 2A), indicating that a portion of eIF3 p47 co-localizes in the nucleus with CDK11. To confirm this observation, we performed cell fractionation and Western blot analysis. Again, we observed both a cytoplasmic and nuclear pool of eIF3 p47, whereas CDK11 was seen in the nucleus (Fig. 2B). Thus, CDK11 and a portion of eIF3 p47 can both be found in the nucleus. In addition, reciprocal immunoprecipitation and Western analysis with Myc-tagged CDK11 p46 and HA-tagged eIF3 p47 from transfected HEK293 cells showed that CDK11 p46 and eIF3 p47 could associate in cells (Fig. 2C). These results together with FIG. 2. The co-localization of CDK11 and eIF3 p47 in human cells. A, co-localization of CDK11 p110 and eIF3 p47 in the nucleus. A375 cells were transfected with pCMV-HA-p47, and 24 h post-transfection, a confocal immunofluorescence microscopy assay was performed using P1C antibody and anti-HA antibody. The subcellular location of CDK11 p110 (green) and eIF3 p47 (red) and their co-localization (yellow) are shown (magnification, ϫ100). B, localization of CDK11 p110 and eIF3-p47 by cell fractionation and Western blot. 20 g of fractionated A375 cell lysates were used for each lane and CDK11 p110 -specific antibody GN1 or anti-eIF3 antibody was used for immunoblotting. C, the interaction between transfected CDK11 p46 and eIF3 p47 in human cells. HEK293 cells were cotransfected with pCMV-Myc-CDK11 p46 and pCMV-HA-p47. 48 h post-transfection, cells were harvested, and cell extracts were made. Immunoprecipitation was performed with either mouse IgG or anti-Myc antibody followed by Western blotting with anti-Myc or anti-HA antibody.
FIG. 3. The interaction between endogenous CDK11 p46 and eIF3 p47 during apoptosis. A375 cells were treated with 0.5 g/ml anti-Fas antibody to stimulate apoptosis for 0, 24, 36, and 48 h, and then the cells were lysed and used for immunoprecipitation. CDK11 p46 overexpressed A375 cells were used in lanes 5 and 11. In lanes 1-4, immunoprecipitation was performed with goat anti-GST IgG, whereas in lanes 5-9, immunoprecipitation was performed with goat anti-eIF3 antibody followed by immunoblotting with CDK11 p110 -specific antibody, GN1. 20 g of cell lysates were used in lanes 10 and 11 for Western blot.
Endogenous eIF3 p47 Interacts with CDK11 during Apoptosis-Next we tested whether p47 is a potential substrate of CDK11 during apoptosis. During apoptosis, the p110 isoform of CDK11 is cleaved by caspase-3 into a p46 -50 C-terminal fragment containing the kinase domain and a p60 N-terminal fragment (2). We saw a weak interaction between the endogenous CDK11 p110 isoform and eIF3 in untreated A375 cells (Fig. 3,  lane 6). Interestingly, when apoptosis is induced by anti-Fas, the caspase-processed C-terminal fragment of CDK11 p110 interacted strongly with eIF3. The interaction was strongest at 36 h after stimulation of apoptosis (Fig. 3, lanes 7-9). The specificity of the interaction was confirmed by including nonspecific goat IgG as negative controls (Fig. 3, lanes 1-4), and the correct sizes of the expected protein bands are shown in lanes 5, 10, and 11 (Fig. 3). In CDK11 p46 stable transfected cells, the interaction between eIF3 and CDK11 p46 (p46-p50) was strong (Fig. 3, lane 5), which means that the overexpressed CDK11 p46 associates with eIF3, presumably through its p47 subunit. In order to prove that the cells did undergo apoptosis after anti-Fas or staurosporine treatment during the time course, the percentage of apoptotic cells was measured by flow cytometry (Fig. 4A), and the cell viability was analyzed by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay (Fig. 4B). In addition, apoptosis was confirmed by light microscopy analysis (data not shown). The results indicated that the cells underwent apoptosis gradually after stimulation and the percentage of apoptotic cells approached 100% by 72 h after treatment (Fig. 4A). These observations are consistent with the previous findings by our group (2). CDK11 p46 Phosphorylates eIF3 p47 during Apoptosis in Vitro-Since CDK11 p46 contains the kinase domain of CDK11 and it has been shown by our group and others that the kinase activity of CDK11 is stimulated during apoptosis (1, 2), we hypothesized that CDK11 p46 may phosphorylate eIF3 p47 during apoptosis. CDK11 was immunoprecipitated from anti-Fastreated A375 cell lysates, and GST control or GST recombinant p47 protein were used as substrates in kinase assays. Immunoprecipitated CDK11 from A375 cells undergoing apoptosis was capable of phosphorylating recombinant eIF3 p47 protein (Fig. 5A). In contrast, CDK11 p46 did not phosphorylate GST alone. Furthermore, the addition of a caspase-3 inhibitor (DEVD-FMK) diminished the degree of eIF3 p47 phosphorylation (Fig. 5B, lane 3), and A375 cells stably transfected with kinase mutant CDK11 p46M displayed suppressed phosphorylation of eIF3 p47 (Fig. 5B, lane 4). These results indicate that the caspase-processed isoform of CDK11 p110 isolated from cells undergoing apoptosis phosphorylates eIF3 p47 in vitro. The results with the specific caspase-3 inhibitor and the mutant form of CDK11 p46 kinase indicate that the phosphorylation of eIF3 p47 is due to the CDK11 p46 , not to some other kinase that copurifies in the immunoprecipitate.
Identification of the Phosphorylation Site in eIF3 p47 and the in Vivo Phosphorylation Status of p47 during Apoptosis-To determine the phosphorylation site in the eIF3 p47 protein phosphorylated by CDK11, recombinant eIF3 p47 was incubated with caspase-processed CDK11 that immunoprecipitated FIG. 5. EIF3 p47 is phosphorylated by the caspase-processed CDK11 p110 isoform during apoptosis. A, A375 cells were treated with 0.5 g/ml anti-Fas antibody or 10 ng/ml staurosporine to stimulate apoptosis for 0, 24, 36, 48 h, and then the cells were harvested and lysed. 300 g of each cell lysate was used for immunoprecipitation using CDK11 p110specific antibody GN1 and protein A-agarose. The precipitate was then incubated with 5 g of the substrate GST or GSTfused eIF3 p47 (GST-p47) in the presence of [␥-32 P]ATP in kinase buffer followed by SDS-PAGE and autoradiography or phosphorimaging to quantitate the incorporated 32 P. The figure shows a representative result from at least three independent experiments. B, A375 cells were treated with 0.5 g/ml anti-Fas antibody and/or 50 M caspase-3 inhibitor (DEVD-FMK) for 36 h. In lane 4, A375 cells stably transfected with kinase mutant CDK11 p46M were used. Cells were then lysed for the kinase assay as described for A. 5 g of GST-p47 were used as substrate.
from anti-Fas-treated cell lysate in kinase buffer. After the kinase reaction, recombinant eIF3 p47 protein was separated by SDS-PAGE. EIF3 p47 was digested with pepsin in gel and subjected to liquid chromatography/MS/MS analysis. The spectrum for the p47 phosphopeptide (residues Ala-41 to Pro-49), [M-H 3 PO 4 ] ϭ 784 m/z is shown in Fig. 6A. Identification of a phosphorylated Ser 46 is seen by analysis of b ions and y ions. A gain of phosphate at the y4 ion, 485 m/z, derived from fragmentation between Ser-45 and Ser-46 was observed. The reason that y ions are not completely shown is that multiple internal cleavages instead of sequential cleavages from the carboxyl to amino terminus occurred. Once we found that Ser-46 within FIG. 6. Determination of the phosphorylation site in eIF3 p47 by the caspase-processed CDK11 p110 isoform and characterization of serine phosphorylation during apoptosis in vivo. A, MS identification of the eIF3-p47 in vitro phosphorylation site. CDK11 p110 was immunoprecipitated from anti-Fas-treated A375 cell lysate using CDK11 p110 -specific antibody GN1 and incubated with recombinant GST-p47. Then an in-gel pepsin digest was carried out. MS spectra of the identified serinic-phosphopeptide showed that Ser-46 is the site of phosphorylation. Annotated ions are shown as nominal m/z values and resulted from fragmentation of an amide bond with charge retention on the carboxyl (y ions) or amino terminus (b ions). B, serine phosphorylation of eIF3-p47 during apoptosis in vivo. A375 cells were treated with staurosporine for 0, 24, 36, 48, and 72 h. Cell lysates were immunoprecipitated with eIF3 antibody and then immunobloted with phosphoserine antibody or eIF3 p47-specific antibody. eIF3 p47 was phosphorylated by CDK11 in vitro, we investigated the serine phosphorylation status of eIF3 p47 in cells undergoing apoptosis in vivo. A375 cells were treated with staurosporine to trigger apoptosis, and cell lysates were immunoprecipitated with eIF3-specific antibody followed by Western analysis with phosphoserine-specific antibody or eIF3 p47-specific antibody. In A375 cells stimulated to undergo apoptosis, serine phosphorylation of eIF3 p47 was seen, and the maximum phosphorylation level was seen at 36 h after treatment (Fig. 6B). These results are consistent with our in vitro kinase assay and show that serine phosphorylation of eIF3 p47 occurs during apoptosis in vivo. CDK11 p46 , but not the Kinase Mutant CDK11 p46M , Inhibits in Vitro Translation-We further investigated the functional consequences of the association of caspase-processed CDK11 p110 with eIF3 p47. We determined the effect of CDK11 p46 on protein synthesis in vitro. A rabbit reticulocyte in vitro translation system was used for this purpose. In this system, a given mRNA is translated into protein. For our studies, we choose to use the luciferase gene product for two reasons. First, we could monitor the luciferase activity as a measurement of protein production, and second, incorporation of radioactive [ 35 S]methionine into the synthesized protein can act as a different measurement of protein production. We established that our assay conditions were within the linear range. This was achieved by measuring luciferase activity every 15 min throughout the incubation and adding a defined amount of luciferase mRNA (40 ng) into the system. We found that the rate of luciferase production is within the linear range between 0 and 75 min of incubation time (Fig. 7A) (R 2 ϭ 0.9906, p Ͻ 0.01). Next, we investigated whether the dose response of luciferase mRNA to translation in vitro is linear. We found that adding luciferase mRNA between 20 and 50 ng is within the linear range (Fig. 7A) (R 2 ϭ 0.9791, p Ͻ 0.01). Therefore, we chose 40 ng of luciferase mRNA and incubated for 60 min in subsequent experiments. Luciferase mRNA was translated in the presence or absence of added recombinant CDK11 p46 or CDK11 p46M protein. Both luciferase activity and incorporated [ 35 S]methionine were measured. CDK11 p46 significantly inhibited the synthesis of luciferase (Fig. 7B). The specificity of this observation was verified by adding GST, another purified Escherichia coli expressed protein, to the reaction, and it had no effect on translation compared to the buffer (Fig. 7B, top panel). This observation was further confirmed by performing a doseresponse assay with recombinant CDK11 p46 protein. A concentration-dependent inhibitory effect of CDK11 p46 protein on translation was observed (Fig. 7B). Moreover, the phosphorylation of eIF3 p47 by CDK11 p46 appears to be required for the inhibition of translation, since the kinase-defective mutant CDK11 p46M did not inhibit the synthesis of luciferase (Fig. 7B). CDK11 p46 , but Not Kinase Mutant CDK11 p46M , Inhibits Translation in Vivo-To examine the ability of CDK11 p46 to inhibit translation in human cells, a luciferase reporter plasmid was transiently transfected into A375 cells stably transfected with either pcDNA3, CDK11 p46 , or kinase-defective mutant CDK11 p46M . Luciferase activity was then measured 24 h after transfection. Luciferase synthesis was significantly inhibited in A375 cells that overexpress CDK11 p46 compared with the vector-transfected cells (Fig. 8A, left panel). In contrast, luciferase activity was slightly elevated in the CDK11 kinase-FIG. 7. Inhibition of translation by CDK11 p46 but not kinase mutant CDK11 p46M in vitro. A, the rate of luciferase production and the dose response to luciferase mRNA were determined. 50 ng of luciferase mRNA were translated in vitro for 15-120 min, and luciferase activity was then measured with a luminometer. Twenty to fifty nanograms of luciferase mRNA were translated in vitro, and luciferase activity was measured. The incubation time or ng of luciferase mRNA was plotted against luciferase activity. B, CDK11 p46 , but not kinase mutant CDK11 p46M , inhibits translation in a dosedependent manner in vitro. Luciferase mRNA (40 ng) was translated in an in vitro translation system in the presence of 18, 35, or 70 nM recombinant purified CDK11 p46 or CDK11 p46M protein or 70 nM purified recombinant GST. Luciferase protein synthesis was analyzed by gel electrophoresis followed by autoradiography. Luciferase activity was measured with a luminometer. The averages of results from three experiments are shown. *, p Ͻ 0.05. defective mutant CDK11 p46M -transfected cells relative to vector-transfected cells (Fig. 8A, right panel). To rule out the effect of CDK11 p46 on transcription of the luciferase gene, RT-PCR was performed on three different transfected clones from either vector alone-or luciferase reporter gene-transfected cells. The histone gene served as a control (246 bp). The luciferase mRNA of vector-or CDK11 p46 -transfected cells was the same (Fig. 8B, 454 bp), indicating that CDK11 p46 inhibited the translation of the luciferase reporter gene and not transcription. To further quantify the luciferase mRNA, real time PCR was performed from pcDNA3 vector-or CDK11 p46 -transfected cells. The Ct of luciferase was normalized against GAPDH. No difference was observed between vector-and CDK11 p46 -transfected A375 cells (Fig. 8C). Untransfected A375 cells and no template control served as negative controls (Fig. 8C). These results suggest that the caspase-processed CDK11 p110 isoform inhibits protein synthesis in vivo at the translational level. CDK11 p46 Suppresses Overall Protein Synthesis-To further investigate the impact of CDK11 p46 on overall cellular protein synthesis, we established two stably transfected cell lines. One was stably transfected with pcDNA3 vector, and the other was stably transfected with CDK11 p46 . Overall protein synthesis, as measured by pulse labeling with [ 3 H]leucine in vivo, was significantly lower in CDK11 p46 -overexpressing cells compared with vector-transfected cells (Fig. 9A). Therefore, CDK11 p46 can also inhibit the overall rate of cellular protein synthesis.
Since it was reported that the rate of protein synthesis is rapidly down-regulated in mammalian cells following the induction of apoptosis (21)(22)(23)(24), we also measured protein synthesis in A375 cells after staurosporine treatment. Consistent with these findings, the overall cellular protein synthesis decreased during apoptosis (Fig. 9B).
Overexpressed CDK11 p46 Causes Higher Apoptosis Rate-We next addressed whether overexpression of CDK11 p46 increases apoptosis. Vector control or CDK11 p46 -overexpressing A375 cells were stained with 7-AAD and subjected to flow cytometry. Results showed increased apoptosis in cells stably transfected with CDK11 p46 compared with cells transfected with vector only (Fig. 10A). This observation was confirmed by cell morphological analysis using light microscopy and Giemsa staining, which showed that CDK11 p46 -overexpressing cells had more apoptotic cells than vector only-transfected cells (Fig.  10B). These results demonstrate that CDK11 p46 promotes apoptosis. DISCUSSION CDK11 appears to be downstream of caspase-3 in apoptotic signaling (2). Upon apoptotic stimulation, CDK11 is cleaved by caspases and activated (1,13,14). However, the potential substrates of CDK11 that may be involved in cell death are not known. In the present study, we identify that eIF3 p47 interacts with the caspase-processed CDK11 isoform using a yeast two-hybrid screening strategy.
The eIF3 p47 is a subunit of the mammalian eIF3 multipro- FIG. 8. Inhibition of the synthesis of luciferase reporter gene by CDK11 p46 but not kinase mutant CDK11 p46M in vivo. A, CDK11 p46 , not CDK11 p46M , inhibits translation of the luciferase reporter gene in vivo. Triplicate pcDNA3-, pcDNA3-CDK11 p46 -, or pcDNA3-CDK11 p46M -transfected cells were transfected with pGL3-SV40, which contains a luciferase reporter gene. 24 h post-transfection, cells were lysed, and luciferase activity was measured by a luminometer. *, p Ͻ 0.05; ****, p Ͻ 0.0001. B, CDK11 p46 did not inhibit transcription of the reporter gene. pcDNA3-or pcDNA3-CDK11 p46 -transfected cells were transfected with pGL3-SV40, which contains a luciferase reporter gene. After 24 h, cells were harvested, and total RNA was isolated for RT-PCR analysis. The expected PCR products for histone and luciferase are 246 and 454 bp, respectively. Three different clones of each transfected cells were tested. C, quantitative real time PCR also indicates that there is no difference in luciferase mRNA level between vector-and CDK11 p46transfected cells. The same RNA samples were analyzed for luciferase transcripts by one-step real time RT-PCR. Luciferase mRNA levels were normalized against GAPDH and illustrated as mean Ϯ S.D. NTC, no template control; LucϪ/Lucϩ, luciferase reporter gene untransfected/ transfected cell lines. tein complex that binds to the 40 S ribosome and promotes the binding of methionyl-tRNA and mRNA (21,25). The eIF3 p47 protein appears to be a Mov34 family member (26,27). Mov34 family members are involved in regulation of proteasome, translational initiation, and transcription (26). In the present study, we can show that the carboxyl-terminal domain of CDK11 p110 directly interacts with eIF3 p47 in GST pull-down assays. Furthermore, deletion of DNA sequences of the Mov34/ JAB_MPN domain in eIF3 p47 eliminates the interaction between these two proteins. Thus, the Mov34/JAB_MPN domain within eIF3 p47 may be important for the protein-protein interaction with the carboxyl-terminal domain of CDK11 p110 .
We also demonstrate by confocal microscopy that both eIF3 p47 and CDK11 p46 co-localize in the nucleocytoplasm. We also show for the first time that there is both a cytoplasmic and nuclear pool of eIF3 p47. This phenomenon is not uncommon. Several other proteins, such as cyclin B1 (28), RanBPM (29), and eukaryotic initiation factors 4E (eIF4E) (30,31), eIF4G (32), and eIF5A (33) can also be found in the nucleus as well as the cytoplasm. Since eIF3 p47 is a Mov34 family member, it is tempting to speculate that it may have a nuclear function in mRNA metabolism such as splicing or transport. We are currently pursuing these possibilities.
In this study, we show that endogenous CDK11 p46 and eIF3 p47 protein can interact and that stimulation of apoptosis enhances this interaction. We also show that the kinase activity of CDK11 is activated during apoptosis, that the kinase obtained can phosphorylate eIF3 p47 apparently at serine 46, and that the phosphorylation of eIF3 p47 by CDK11 can be blocked by caspase-3 inhibitor (DEVD-FMK). In addition, mu-tation of the phosphate transfer site in CDK11 p46 abrogates the phosphorylation of eIF3 p47 protein. Finally, we show that serine phosphorylation of eIF3 p47 occurs in cells undergoing apoptosis, with maximal phosphorylation occurring at 36 h both in vitro and in vivo. Although not definitive, these observations strongly suggest that eIF3 p47 may be a substrate of the caspase-processed CDK11 p110 isoform during apoptosis. We noticed that the protein level of eIF3 p47 also increased during apoptosis with maximal expression at 36 h after stimulation. The biological mechanism of this phenomenon is under investigation.
When cells are committed to apoptosis, a remarkable inhibition of the rate of overall protein synthesis is observed in a variety of cell types (24,34,35). This down-regulation of protein synthesis may either protect cells against noxious agents and ensure the conservation of resources needed for survival or activate apoptosis (36). This inhibition of protein synthesis occurs at the level of polypeptide chain initiation and is accompanied by the phosphorylation of the ␣ subunit of eIF2 and the caspase-dependent cleavage of eIF4GI, eIF4GII, eIF4B, p35 subunit of eIF3, minor proportions of the ␣ subunit of eIF2, and the eIF4E-binding protein 4E-BP1 (21,37). The rate of protein synthesis is regulated by the phosphorylation status of several eIFs. eIF4E phosphorylation was reported to strongly decrease following anti-CD3 and anti-CD4 stimulation in immature DP thymocytes, which resulted in rapid decreases in protein synthesis and apoptosis (38). Phosphorylation of eIF2␣ by doublestranded RNA-dependent serine/threonine protein kinase was also correlated with down-regulation of translation initiation and apoptosis (21,36,39,40). The data from our study suggest that phosphorylation of eIF3 p47 may also contribute to the alteration of protein synthesis during apoptosis. The biological consequences of the phosphorylation of eIF3 p47 during apoptosis are under investigation. In A375 cells stably expressing either the caspase-processed carboxyl-terminal domain of CDK11 p110 or a phosphate transfer site mutant CDK11 p46M , we show that luciferase activity and the production of luciferase protein are compromised in CDK11 p46 but not in CDK11 p46M cells. In addition, overexpressed CDK11 p46 from stably transfected cells are capable of interacting with eIF3 in vivo. These data suggest that the function of the caspase-processed CDK11 p110 isoform may be to inhibit translation during apoptosis. However, whether or not this inhibition of protein translation occurs in an eIF3 p47-dependent or -independent manner remains to be clarified.
In summary, the CDK11 protein kinase may be involved in apoptotic signaling. This study demonstrates that the caspaseprocessed CDK11 p110 isoform can interact with eIF3 p47, which is a Mov34 family member. The correlation of inhibition of protein synthesis and eIF3 p47 phosphorylation by CDK11 p46 suggests, but does not prove, a causal relationship between the two phenomena. Further analysis of this interaction might result in a clearer understanding of the regulation and function of CDK11 and eIF3 p47, thus providing new insights into apoptotic signaling and protein synthesis.