| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Department of Dermatology and Skin Science, Vancouver Coastal Health Research Institute, Jack Bell Research Centre, University of British Columbia, Vancouver, British Columbia, Canada;
Department of Biology, California State University–Fresno, Fresno, California, USA; and
ImmuneChem Pharmaceuticals Inc., Burnaby, British Columbia, Canada
1Correspondence: Jack Bell Research Centre, 2660 Oak St., Vancouver, BC, Canada V6H 3Z6. E-mail: gangli{at}interchange.ubc.ca
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: protein half-life • cell proliferation
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Three alternatively spliced transcripts of the ING1 gene have been identified and encode protein variants with predicted sizes of 47, 33, and 24 k Da (2)
. Among these isoforms, p33ING1b is the most abundant and many of the tumor-suppressive functions have been attributed to this variant. p33ING1b has been shown to mediate various cellular functions including senescence, growth arrest, gene regulation, and apoptosis (3)
. p33ING1b also plays an important role in the cellular stress response to ultraviolet radiation (UVR). In melanoma cells UVR induces the expression of this protein in a time- and dose-dependent manner (4
, 5)
. Furthermore, p33ING1b was found to physically associate with GADD45 and PCNA proteins, which are essential for nucleotide excision repair (NER) (5
, 6)
. In fact, overexpression of p33ING1b dramatically enhances NER of UVB-damaged DNA and promotes UVB-induced apoptosis in melanoma cells (5
, 7)
. However, little is known on the regulation of p33ING1b expression associated to DNA damage response.
Posttranslational modification is a major regulatory mechanism through which cellular responses to DNA damaging events are orchestrated. A number of protein kinases, notably the phosphatidylinositol-3 kinase-related kinase (PIKK) family members, including ataxia-telangiectasia mutated (ATM) and ATM and Rad3-related (ATR), are mobilized on genomic injury and play critical roles in the DNA damage response (8
, 9)
. Chemical agents such as the type II topoisomerase inhibitor doxorubicin, ionizing radiation, or UVR are reported to cause DNA damage activating ATM and ATR (8
9
10)
. ATM phosphorylates p53 and the damage checkpoint effector kinases Chk1 and Chk2 on ionizing radiation (IR) and oxidative stress but only phosphorylates Chk2 on UVR (11)
. In contrast, ATR is mainly responsive to bulky DNA adducts such as those induced by UVR (12
, 13)
. UV activation of ATR results in the selective phosphorylation of p53 and Chk1 but not Chk2 (11)
. Chk1 phosphorylates and inhibits the protein phosphatase Cdc25 by cytosol sequestration (14)
, thus inhibiting the cyclin-dependent kinase 1 (Cdk1) required for entry into mitosis (15)
. Therefore, Chk1 plays a key role in inducing a G2 cell cycle arrest, which allows genetic repair prior to mitotic onset. Phosphorylation of p53 in response to DNA damage results in elevated protein levels and increased half-life of this tumor suppressor (16
17
18)
activating a cascade of events toward DNA repair and apoptosis. A current model suggests that p53 becomes phosphorylated within the Mdm2 binding domain on Ser-15, Thr18, and Ser-20 on DNA damage, thus evading proteasome-mediated degradation (16
, 19)
. Since protein phosphorylation is a fundamental mechanism by which intracellular events can be rapidly and transiently regulated, we sought to investigate if p33ING1b is posttranslationally modified on genotoxic stress. We report here the phosphorylation of p33ING1b at serine 126 (Ser-126 on genotoxic stress). Furthermore, we found that the ATM/ATR-responsive checkpoint kinase Chk1 can phosphorylate Ser-126 on DNA-damaging events, while Cdk1 phosphorylates this residue in the absence of stress. Moreover, Ser-126 phosphorylation increases the stability of p33ING1b since abolishment of phosphorylation by replacing Ser-126 with alanine greatly shortened the half-life of this protein. More importantly, this residue proves to be a key regulator of cyclin B1 expression and cell proliferation.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 100 U/ml penicillin, and 100 µg/ml streptomycin supplemented with 10% fetal calf serum at 37°C in a 5% CO2 atmosphere. The kinase inhibitors wortmannin, staurosporine, roscovitine, and kenpaullone; the phosphatase inhibitor okadaic acid; and the inhibitor of the 80S cytosolic ribosome cycloheximide were purchased from Sigma (St. Louis, MO, USA). Transient plasmids, ATM, ATR, Cdc2, and Chk1 siRNAs (ID numbers 118232, 82, 42819, and 108, respectively, Ambion) and ING1b siRNA (sense r[GGUACACGUGUAACAAGAA]dTdT, Qiagen, Valencia, CA, USA) transfections were performed using Effectene (Qiagen) and Silentfect (Bio-Rad, Hercules, CA, USA) reagents, respectively, according to the manufacturer’s protocol. Ultraviolet irradiation was performed by removing media and exposing cells to UVB (290–320 nm) light using a bank of four unfiltered FS40 sunlamps (Westinghouse, Bloomfield, NJ, USA).
Expression of recombinant p33ING1b and site-directed mutagenesis
Recombinant GST-tagged p33ING1b was expressed in E. Coli BL-21 DE3 from the pGEX-ING1b vector and isolated as described (21)
. The pCIneo-ING1b-FLAG vector (21)
was used for transfection of MMRU cells and was further modified by gene splicing by overlap extension (22)
using the following primers: fw: 5'-GGCAACGCAGGCAAGGC-3' rv: 5'-CCTTGCCTGCGTTGCCC-3' and fw: 5'-GGCAACGAAGGCAAGGC-3' rv: 5'-CCTTGCCTTCGTTGCCC-3' to create S126A and S126E mutants, respectively.
Immunoprecipitation and Western blot analysis
Nuclear pellets were obtained as described previously (23)
, and the nuclear proteins were solubilized in RIPA buffer (20 mM Tris, pH 7.2; 150 mM NaCl; 1% Triton X-100; 1% sodium deoxycholate; 0.1% SDS) plus protease and phosphatase inhibitors. Immunoprecipitation was performed as described (24)
by incubating 500 µg of nuclear extracts with 2 µg/ml of either mouse anti-Cdk1 or Chk1 antibodies (Santa Cruz Biotechnology, Santa Cruz, CA, USA) or anti-FLAG M2 antibodies (Sigma). Alkaline calf intestinal phosphatase (CIP) (New England Biolabs, Beverly, MA, USA) treatment of immunoprecipitated p33ING1b was done on-beads using NEBbuffer-3 and 10 U CIP per immunoprecipitation. Reaction was performed at 37°C for 2 h and stopped on addition of sample buffer.
Whole cell proteins were extracted as described previously (25)
. Concentration of proteins was determined using the DC Protein Assay (Bio-Rad), and Western blot analysis was performed as described previously (26)
. Low bisacrylamide gels (acrylamide:bisacrylamide 118.5:1 ratio instead of the usual 37.5:1) were used to enhance the gel mobility retardation resulting from phosphorylation. The following antibodies were used for Western blot: mouse anti-FLAG M2, ATM, ING1b (monoclonal), Chk1, Cdk1, p53, ß-actin, and rabbit anti-Lamin B1 antibodies (Santa Cruz). The relative optical density of the band was estimated using NIH ImageJ 1.31v software. All measures were made with exposures within the linear range of the film (Hyperfilm ECL, Amersham Biosciences, Piscataway, NJ, USA). Intensity of the signal of interest was corrected for the different amounts of cellular protein loaded on the gel by detecting ß-actin as the input control.
Immunofluorescence
MMRU cells were grown on coverslips at a density of 2 x 105 cells/well in a 6-well plate. Twenty-four hours following transfection, cells were simultaneously fixed and permeabilized as described (27)
. The permeabilized cells were incubated with mouse anti-FLAG diluted 1:500 in PBS at room temperature for 1 h, rinsed 3 times with PBS, and incubated for 1 h at room temperature with CY3-conjugated goat anti-mouse antiserum (Jackson ImmunoResearch, West Grove, PA, USA) diluted 1:500 in PBS. Cells were counterstained with 650 ng/ml HOECHST 33258 diluted in PBS to visualize DNA. Slides were visualized under a Zeiss Axioplan 2 microscope.
Half-life estimation of wild-type and S126A mutant p33ING1b by pulse-chase
MMRU cells were transfected with pCIneo-ING1b-FLAG or pCIneo-S126A-FLAG. Twenty-four hours after transfection, the cells were pulsed with media containing [35S]methionine/cysteine mixture (0.2 mCi/ml, Perkin-Elmer) for 48 h. Cells were chased with cold methionine-containing media and harvested at various time points. Nuclear proteins were extracted from these cells and subjected to immunoprecipitation using p33ING1b specific antibodies. Half-life of wild-type and S126A mutant p33ING1b was estimated as described previously (28)
.
Mass spectrometry
Protein bands were manually excised from SDS-PAGE gel and digested with AspN using an Abimed Digest Pro robot, and the resultant peptides were dried. A newly described methodology (29)
was used to reduce the ionization of the positively charged ions, therefore enabling easier detection of negative (e.g., phosphorylated) ions. The mass spectra were acquired on an Applied Biosystems Voyager DE-Pro in both positive and negative ion reflector detector mode. The spectra were calibrated using the autolytic trypsin peaks at m/z 842.5099 and 2211.0968. The phosphorylated peptides were also subjected to on-target phosphatase treatment. Approximately 1 µl of 0.05 U/µl CIP was spotted onto each sample on the MALDI target. The target was placed in a humidified chamber at 37°C for 2 h. The CIP was blotted and allowed to dry. The samples were then run on the mass spectrometer, and a neutral loss of H3PO4 (–98 Da) was expected from the phosphorylated samples (30)
.
Generation of phospho-specific Ser-126 antibody
Anti-p-Ser-126 antisera was raised in New Zealand rabbits using the synthetic phosphorylated peptide ELGDTAGNpSGKAGADRP by a previously described method (21)
. To isolate the specific anti-p-Ser-126 antibodies, 10 mg of the phosphorylated synthetic peptide was conjugated with Sepharose CL beads (Sigma) at pH 9.5 following manufacturer’s instructions. The antibody was affinity purified by running 
80 ml of rabbit serum through the affinity column as described previously (21)
. Anti-p-Ser-126 antibody was then separated from antibodies that would bind nonphosphorylated p33ING1b by incubating it with glutathione agarose beads-bound recombinant GST-p33ING1b at 4°C overnight with gentle rocking and the nonbound fraction kept for use.
In vitro kinase assays
Chk1, Cdk1, and ATM were immunoprecipitated as aforementioned. Rinsed beads were resuspended in 20 µl kinase buffer (20 mM MOPS [pH 7.2], 25 mM ß-glycerol phosphate, 1 mM dithiothreitol, 1 mM CaCl2). Kinase assays were performed by mixing 0.5 mM of substrate peptide (full-length p33ING1b or the LGDTAGNSGKAGADRPK peptide corresponding to amino acids 119–135 of p33ING1b) diluted in kinase buffer to the above mentioned immunoprecipitates. The reaction was started by the addition of 5 µCi of [
-32P]ATP (Perkin-Elmer, Wellesley, MA, USA), 50 µM ATP in kinase buffer in the presence of phosphatase inhibitors. Reactions were incubated at 30°C for 30, 90, or 120 min and stopped by transferring a 25 µl aliquot onto the center of phosphocellulose P81 paper. P81 squares were washed with 0.75% phosphoric acid three times and once with acetone prior to scintillation counting. Radioactive labeling of full-length p33ING1b was performed in the same manner but using 5 µg of MMRU nuclear cell extracts as the kinase source and incubating the reactions for 30 min. Reactions were stopped on the addition of loading buffer and resolved by SDS-PAGE prior to autoradiography. For in vitro kinase assays using GST-ING1b as substrate, 200 ng of recombinant full-length human Chk1 (SignalChem, Richmond, British Columbia, Canada) or 2 U of Cdk1/cyclin B1 complex (EMD Biosciences, San Diego, CA, USA) were incubated in kinase buffer with 500 ng of isolated GST-ING1b substrate. The reaction was started by the addition of 100 µM ATP in kinase buffer.
Cell proliferation assays
MMRU cells (1 x 105) were seeded on 6-well plates and incubated in DMEM media until a 20% of confluence was reached. The cells were then transiently transfected with empty vector pCIneo, pCIneo-ING1b-FLAG, pCIneo-S126A-FLAG, or pCIneo-S126E-FLAG or with scramble control or ING1b siRNAs. The cells were fixed with 3.7% p-formaldehyde at different time points, and the cell proliferation was quantitated by sulforhodamine B (SRB) staining (31)
.
Cell synchronization
To synchronize MMRU cells at G1 phase, 1 x 106 cells were seeded onto 100 mm dishes and incubated in 5% FBS-containing DMEM media until a 50% of confluence was reached. The cells were then serum-starved for 48 h, followed by incubation in 6 µM aphidicolin in 10% FBS-containing DMEM media for 14 h. The cells were then released from G1 by rinsing them three times in 10% FBS-containing DMEM and incubating them in this media for different time points. The efficiency of synchronization and the cell progression after release was assessed by flow cytometry. For flow cytometry, the cells were pelleted by centrifugation at 200 g for 3 min and resuspended in 1 ml of hypotonic fluorochrome buffer (0.1% Triton X-100, 0.1% sodium citrate, 25 µg/ml RNase A, 50 µg/ml PI). After incubation at 4°C for 30 min, the samples were analyzed by a Coulter EPICS XL-MCL flow cytometer (Beckman Coulter, Fullerton, CA, USA).
Statistical analysis
Student’s t test was used to perform statistical analyses. A P value < 0.05 was considered significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 33)
. Since p33ING1b appears as barely detectable doublet in untreated cells (Fig. 1
A), it is likely that some posttranslational modification event occurs early on after translation. Therefore, at least two populations of this protein coexist in the cells under resting conditions. Increasing doses of the DNA-damaging agents etoposide and doxorubicin resulted in elevated expression of both bands. While the band representing unmodified p33ING1b increased modestly, the higher molecular weight variant was robustly accumulated 24 h after exposure. Importantly, only the phosphorylated form is detected after treatment with the higher dose of doxorubicin (1 µg/ml). These results are suggestive of a stress-induced expression and posttranslational modification of p33ING1b.
|
To determine whether the higher molecular weight variant of p33ING1b, seen in the low-bisacrylamide SDS-PAGE, is due to phosphorylation of this protein, we induced its accumulation by incubating the cells in the presence of doxorubicin and the cell-permeable phosphatase inhibitor okadaic acid for 24 h. This potent inhibitor of protein phosphatase 1A and 2A caused the disappearance of the lower band (presumably from unmodified protein) (Fig. 1B
), indicating that the upper band is likely a result of protein phosphorylation. Furthermore, the addition of the serine-kinase inhibitor staurosporine dramatically reduced the appearance of the higher molecular weight band (presumably phosphorylated proteins) 24 h after exposure (Fig. 1B
). Wortmannin, an inhibitor of phosphoinositide kinase-3 (PI-3K) and PI-3K related kinases, also reduced the upper band albeit to a lesser extent. Notably, both kinase inhibitors reduced the expression of total p33ING1b, suggesting that its stability is dependent on the phosphorylation event. Taken together, these results strongly suggest that p33ING1b is a phosphoprotein.
To confirm that p33ING1b is a phosphoprotein, transiently expressed FLAG-tagged p33ING1b was immunoprecipitated from cell extracts of doxorubicin-treated melanoma MMRU cells with a mouse anti-FLAG antibody. Half of the immunoprecipitate was incubated in an alkaline phosphatase reaction, while the other half was incubated in a mock reaction. Western blot analysis using antibodies directed against phosphorylated serine residues indicate that at least some serine residue(s) within the p33ING1b protein can be phosphorylated (Fig. 1C
). This was further supported by the loss of phospho-serine signal, but not of the overall p33ING1b signal, following a phosphatase treatment. In vitro radiolabeling of GST-tagged p33ING1b further confirmed that p33ING1b can be phosphorylated using cell extracts obtained from doxorubicin-treated melanoma MMRU cells as kinase source (Fig. 1D
).
Ser-126 is phosphorylated on genotoxic stress
Ser-126 was first identified as a potential phosphorylable residue by mass spectrometric analysis. The analysis was performed on recombinant GST-ING1b exposed to either untreated or doxorubicin-treated MMRU cell extracts as kinase source. Phosphorylation of Ser-126 was detected in recombinant GST-ING1b exposed to doxorubicin-treated MMRU cell extracts, CIP treatment confirmed the phosphorylation of this residue (Table 1
). To confirm that Ser-126 is a bona fide phosphorylation site, a rabbit polyclonal antibody was raised and affinity purified against a synthetic peptide corresponding to phosphorylated Ser-126 and surrounding residues (amino acids 118–134) as described in Material and Methods.
|
To assess the specificity of the antibody, a Western blot analysis was done comparing bacterially expressed recombinant GST-ING1b left untreated or phosphorylated using MMRU cell extracts. The polyclonal anti-p-S126 antibody did not recognize bacterially produced recombinant GST-ING1b unless it was phosphorylated in vitro using doxorubicin-treated MMRU cell extracts (Fig. 2
A). In addition, the antibody did not recognize GST-ING1b that underwent a phosphorylation-dephosphorylation cycle (Fig. 2A
), confirming its specificity for the phosphorylated residue.
|
To further test that the specificity of this antibody is for the phosphorylated-Ser-126 residue, we first mutated p33ING1b by site-directed mutagenesis at Ser-126 to alanine (to prevent phosphorylation of this residue) or to glutamic acid (as phosphomimetic), then the isolated antibody (anti-p-Ser-126) was used to detect phosphorylation of p33ING1b through Western blot analysis of cellular extracts obtained from MMRU cells transiently expressing wild-type or nonphosphorylable S126A or phosphomimetic S126E mutant p33ING1b (Figs. 2B, C
). The isolated antibody bound preferentially, with a 7.5-fold affinity, to the phosphomimetic S126E p33ING1b of cellular extracts obtained from resting cells compared to vector control, (Fig. 2B
). Importantly, anti-p-Ser-126 bound preferentially to both wild-type and phosphomimetic p33ING1b (4.5- and 7.4-fold, respectively, compared to the vector control), but not the S126A mutant, after UV irradiation (Fig. 2C
). This finding confirmed both the specificity of this antibody for the phosphorylated Ser-126 residue on p33ING1b and the induction of phosphorylation of this residue on UV irradiation of MMRU cells.
To determine the kinetics of Ser-126 phosphorylation of endogenous p33ING1b on genotoxic events, we exposed MMRU cells to 20 mJ/cm2 UVB or 1 µg/ml doxorubicin and monitored phosphorylation levels at various time points using the anti-p-Ser-126 antibody. A more than 2-fold increase in endogenous p33ING1b Ser-126 phosphorylation was detected as quickly as 15 min following UVR and a 1.5-fold increase after 30 min after doxorubicin treatment (Figs. 2D, E
). However, a basal level of Ser-126 phosphorylation is detected in the untreated asynchronous cells, suggesting that phosphorylation of this residue can also occur in the absence of stress.
Ser-126 phosphorylation increases the half-life of p33ING1b but does not alter subcellular localization
Altered expression and subcellular localization as well as gene mutation of the p33ING1b tumor suppressor have been found in various cancer types (3)
. Since Ser-126 is 16 residues from the nuclear localization sequence, we attempted to study whether phosphorylation of p33ING1b at Ser-126 would affect the localization of the protein. We then transfected wild-type or S126A mutant FLAG-tagged p33ING1b into MMRU melanoma cells prior to immunofluorescent staining to examine the protein subcellular localization. Both wild-type p33ING1b and the S126A mutant showed nuclear signals, suggesting that phosphorylation of Ser-126 does not modulate the cellular localization of the protein (Supplemental Fig. S1).
Since incubation of MMRU cells with the kinase inhibitors staurosporine and wortmannin reduced both the level of posttranslationally modified p33ING1b and the overall protein level (Fig. 1B
), we studied the role of Ser-126 phosphorylation on the protein stability of p33ING1b. The turnover rate of p33ING1b was assessed in MMRU cells by overexpressing FLAG-tagged protein and harvesting cycloheximide-treated cells at fixed time intervals (Fig. 3
A). Analysis of the optical density of the p33ING1b bands detected by Western blot established a half-life of 16.8 h for wild-type p33ING1b compared with 5.7 h for its S126A mutant counterpart in MMRU cells (Fig. 3B
). Half-life estimation was confirmed through pulse-chase assays (Supplemental Fig. S2). These results suggest that phosphorylation at Ser-126 enhances protein stability of p33ING1b inducing accumulation of this protein as shown by band-shift assays of endogenous p33ING1b (Fig. 1B
).
|
A previous study has shown that p33ING1b can stabilize p53 by disrupting the interaction between p53 and MDM2 (34)
. However, this is likely independent of p33ING1b phosphorylation on Ser-126 since serine to alanine mutation had little effect on p53 protein stability. As previously established, the short-lived p53 protein underwent rapid degradation on cycloheximide treatment. However, the p53 turnover rate was not influenced by the expression of mutant p33ING1b compared to the wild-type counterpart. In both cases the half-life of p53 was of 50 min (Fig. 3A
).
Ser-126 is phosphorylated by both Cdk1 and Chk1
Our initial attempt to identify the kinase responsible for serine 126 phosphorylation was focused on the protein kinase C (PKC) based on two observations: 1) Staurosporine, a potent PKC inhibitor caused drastic dephosphorylation of p33ING1b (Fig. 1B
); and 2) The peptide sequence N-S-G-K, where ‘S’ corresponds to amino acid 126 of p33ING1b matches the consensus PKC substrate sequence (X-S/T-X-R/K). However, a p33ING1b peptide containing serine 126 could not be phosphorylated in vitro by PKC, while the PKC control peptide QKRPSQRSKYL easily incorporated radiolabeled-phosphate groups (data not shown). To predict candidate kinases, which can phosphorylate Ser-126 of p33ING1b, the primary sequence of this protein was analyzed using NetPhosK 1.0 prediction algorithm (35)
(Table 2
). Kinases scoring above 0.4 (Cdk1, CaM-II and GSK3) were chosen to be screened. Since Ser-126 is phosphorylated in response to genotoxic stress (Figs. 1
and 2)
and the fact that NetPhosK 1.0 prediction is limited to 17 protein kinases not containing the DNA-damage responsive ATR and Chk1 kinases, we, therefore, decided to add the ATM/ATR-downstream and Cdk1-upstream Chk1 kinase in the screening. Therefore, the peptide ELGDTAGNSGKAGADRP corresponding to p33ING1b amino acids 118–134 was assayed as substrate on the commercially available KinaseProfilerTM screening service (Upstate Biotechnology, Lake Placid, NY, USA). KinaseProfilerTM was assessed for rabbit CaMKII and human Cdk1-cyclin B, Chk1, and GSK3
(ATM and ATR were unavailable). Unexpectedly, both Chk1 and Cdk1 could phosphorylate Ser-126 in vitro (Supplemental Fig. S3).
|
To confirm that Chk1 and/or Cdk1 can phosphorylate p33ING1b at Ser-126 in MMRU cells, we immunoprecipitated Chk1 and Cdk1 proteins from untreated, doxorubicin- or UV-treated cells and used them for kinase assays. Our data indicated that Cdk1 immunoprecipitated from resting cells could phosphorylate Ser-126, while doxorubicin and UV irradiation inhibited Cdk1 kinase activity to phosphorylate Ser-126 of p33ING1b (Fig. 4
A). In contrast, Chk1 easily phosphorylated Ser-126 when immunoprecipitated from UV- or doxorubicin-treated cells but weakly when immunoprecipitated from resting cells (Fig. 4B
). The effect of genotoxic agents on Cdk1 kinase activity is likely due to its expression since little Cdk1 is immunoprecipitated from UV- or doxorubicin-treated cells as previously shown (36)
. While Chk1 expression does not change dramatically (Fig. 4C
), it has been shown that genotoxic stress induces its activity through ATR-dependent phosphorylation (37)
. Interestingly, both Cdk1 and Chk1 could phosphorylate not only the Ser-126 peptide but also the recombinant GST-ING1b protein (Fig. 4D
). Recombinant Cdk1 phosphorylated GST-ING1b more avidly than recombinant Chk1, which may be attributed to absence of genotoxic stress. Accordingly, Cdk1 phosphorylated its control substrate histone H1 more avidly than Chk1 phosphorylating p53.
|
While doxorubicin caused an increase in endogenous Ser-126 phosphorylation, the serine-kinase inhibitor staurosporine and Cdk inhibitor kenpaullone caused a decrease in phosphorylation levels (Fig. 4E
). These results suggest that in untreated cells Cdk1 keeps a basal level of phosphorylation on Ser-126 residue of p33ING1b, while on genotoxic stress this residue is phosphorylated by Chk1 as a downstream effector of the ATM/ATR cascade. To further confirm that Ser-126 phosphorylation is important for protein stability, we treated MMRU cells with the Cdk inhibitor roscovitine and analyzed the half-life of p33ING1b protein. Consistent with the finding that mutation of Ser-126 to alanine greatly shortens the half-life of p33ING1b protein (Fig. 3)
, the Cdk inhibitor roscovitine also drastically reduced the half-life of wild-type p33ING1b to 6.5 h (Fig. 4F
).
To determine the role of the ATM/ATR cascade on the phosphorylation of Ser-126 and the stability of endogenous p33ING1b, we assessed the effect of validated ATM or ATR RNAi on these parameters. Short interfering RNA against both ATM and ATR inhibited p33ING1b phosphorylation at Ser-126 and p33ING1b expression in resting as well as UVB irradiated cells (Figs. 5
A, B). The role of ATR seems to be particularly important in UV irradiated cells since ATR RNAi treatment decreases Ser-126 phosphorylation by 80%, meanwhile ATM RNAi treatment decreases this phosphorylation by only 47%.
|
To further confirm the role of Cdk1 and Chk1, we assessed the effect of validated Cdk1 or Chk1 RNAi on the phosphorylation of Ser-126 and the stability of endogenous p33ING1b. The knockdown of either Cdk1 or Chk1 inhibited both p33ING1b phosphorylation at Ser-126 and p33ING1b expression in resting as well as UVB-irradiated cells (Figs. 5C, D
). Noteworthy, Chk1 effect seems to be particularly important in UV-irradiated cells since its knockdown decreases Ser-126 phosphorylation by 88%, and total p33ING1b level by 93%, while knocking down Cdk1 decreases Ser-126 phosphorylation by only 41% and p33ING1b expression by 36% in UV-irradiated cells. Taken together, these results suggest that p33ING1b is phosphorylated at Ser-126 by both Cdk1 and Chk1. In particular, Chk1 is a more efficient kinase for Ser-126 phosphorylation after UV irradiation.
p33ING1b phosphorylation at Ser-126 is cell-cycle restricted
The observation that p33ING1b could be phosphorylated at Ser-126 by Cdk1 under nonstress conditions (Fig. 4A, D
) suggested that this phosphorylation may be cell-cycle restricted. To assess this possibility, we synchronized MMRU cells at G1 phase, released them for different time points. We detected Ser-126 phosphorylation and the expression of cyclin D1 (G1/S phases marker) and cyclin B1 (G2/M phases marker) in nuclear extracts of these cells. The cell progression was also confirmed by flow cytometry (data not shown). As shown in Fig. 6
A, Ser-126 phosphorylation is scarcely detectable in late S-phase (6 h after releasing them from G1), increasing through G2/M phase and reaching its maximum 15 h after releasing them from G1. These results suggest that p33ING1b phosphorylation at Ser-126 residue occurs at G2/M phases in non UVB-irradiated cells. However, UVB-irradiation increased the level of p33ING1b phosphorylation at Ser-126 (Fig. 6B
), which reached its maximum at late S-phase (only 6 h after being released from G1) and then gradually decrease. UVB-induced rapid increase of p33ING1b phosphorylation at Ser-126 after cell synchronization (Fig. 6B
) is consistent with the result obtained from cells without synchronization (Fig. 2D
).
|
p33ING1b phosphorylation at Ser-126 controls the expression of cyclin B1 and cell proliferation
Overexpression of p33ING1b is known to down-regulate cyclin B1 (38)
. To assess whether Ser-126 plays a role in controlling the expression of this protein, we expressed wild-type p33ING1b, S126A, or S126E mutants in MMRU cells. As shown in Fig. 7
A, only the phosphomimetic S126E down-regulated cyclin B1 in non UVB-irradiated cells. On UVB-irradiation, ectopic expression of wild-type p33ING1b and phosphomimetic S126E could down-regulate cyclin B1 in MMRU cells (Fig. 7B
). However, the S126A mutant of p33ING1b could not decrease cyclin B1 protein levels. To further confirm whether p33ING1b down-regulates cyclin B1 expression at a physiological condition, we knocked down the endogenous ING1 expression by siRNA. As expected, cyclin B1 expression increases in cells treated with ING1b siRNA that had reduced levels of p33ING1b (Fig. 7C
).
|
To assess if Ser-126 phosphorylation would also affect the ability of p33ING1b to regulate cell proliferation, control vector, wild-type p33ING1b and S126A or S126E mutants were transfected into MMRU cells, which were subsequently analyzed for cell proliferation at various time points. In non UVB-irradiated cells (Fig. 7D
), only the phosphomimetic S126E mutant significantly inhibited cell proliferation with respect to vector transfected cells (P<0.01 at 72 h and 96 h). On UVB-irradiation, the expression of wild-type p33ING1b and phosphomimetic S126E mutant significantly inhibited cell proliferation with respect to vector transfected cells (P <0.01 at 72 h and 96 h for wild-type p33ING1b; P <0.01 at 72 h and P <0.001 at 96 h for S126E), while the S126A mutant of p33ING1b was unable to suppress MMRU cell proliferation (Fig. 7E
). To test the role of endogenous p33ING1 in cell proliferation, we treated the cells with ING1b siRNA (Fig. 7F
). Our data indicated that knockdown of p33ING1b significantly increased cell proliferation (P <0.05 at 24 and 48 h; P <0.01 at 36 h). Taken together, these results suggest that p33ING1b phosphorylation at Ser-126 residue can affect the expression of cyclin B1 and control cell proliferation in melanoma cells.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Yet kinetic studies demonstrate that NER factors are mobilized seconds after cellular exposure to UV (39)
. Since only limited information is available on regulatory mechanisms of ING proteins, we investigated whether the p33ING1b protein would be subjected to posttranslational modifications on genotoxic stress.
In this study, we found that phosphorylation at Ser-126 residue occurs, accompanying with a striking net accumulation of p33ING1b on treatment with genotoxic agents (Fig. 1)
. This increase in p33ING1b expression is consistent with its role in cellular stress response to UVR we have previously observed (4
, 5)
. Recently, phosphorylation of p33ING1b at Ser-199 has been described as a regulatory mechanism of the subcellular localization of this protein (40)
. Since nuclear staining was observed for both wild-type p33ING1b and S126A mutant (Supplemental Fig. S1), it appears that Ser-126 phosphorylation does not regulate p33ING1b subcellular localization. Indeed the identification through mass spectrometry of Ser-126 as the phosphorylable residue induced by DNA damage was of special interest since we have previously found a mutation (Ser to Asn) on this residue in a Clarks level III melanoma biopsy (41)
. Noteworthy, preventing phosphorylation of p33ING1b at Ser-126 through site-directed mutagenesis to alanine results in a marked 3-fold decrease in the half-life of the protein (Fig. 3)
confirming that the decrease in p33ING1b protein level after treatment with kinase inhibitors observed on low-bisacrylamide gels is largely due to reduced protein stability (Fig. 1B
). A balance seems to exist between a pool of nonphosphorylated and phosphorylated p33ING1b, regardless of whether it is endogenous or ectopic under normal conditions (Figs. 1
and 2)
. However, on genotoxic stress, p33ING1b is quickly phosphorylated to prolong its turnover rate as our phospho-specific antibody detected a two-fold increase in p33ING1b phosphorylation of Ser-126 15 min after UV irradiation (Fig. 2D
). The delay in Ser-126 phosphorylation on doxorubicin treatment is likely due to the lag in which the drug incorporates into the cell nucleus and affects cellular functions (Fig. 2E
). The initial finding that both Chk1 and Cdk1 can phosphorylate p33ING1b presented a paradox. On one hand, Chk1 is a direct PIKK target, thus leading to the phosphorylation of p33ING1b but also to the inhibition of Cdc25, the phosphatase that activates Cdk1-cyclin B, required for mitotic entry. However, this same Cdk1 protein, which promotes cell division, also phosphorylates p33ING1b. Closer investigation revealed that while Cdk1 mainly phosphorylates Ser-126 under resting conditions, it is not efficient to do so when cells are exposed to genotoxic stress (Fig. 4A
). This is likely due to Chk1 activation on DNA damage and subsequent inhibition of Cdk1 (14
, 15)
. In the absence of genotoxic stress, inactive Chk1 does not efficiently phosphorylate p33ING1b. However, genotoxic stress triggers the ATM/ATR signaling pathway, hence activated Chk1 can thus phosphorylate p33ING1b (Fig. 4B
). The observation that the Cdk1 inhibitor roscovitine greatly reduced the half-life of wild-type p33ING1b in the absence of genotoxic stress (Fig. 4E
) confirms that phosphorylation of this residue is important for the stability of p33ING1b. Our results, therefore, suggest that p33ING1b is a downstream component in the ATM/ATR signaling pathways in response to genotoxic events. This hypothesis is strongly supported by our results obtained using validated siRNAs targeting ATM, ATR, Cdk1, and Chk1 (Fig. 5)
. Knocking down these kinases resulted in an overall decrease of the phosphorylation level of Ser-126 residue on endogenous p33ING1b, which in turn decreases its stability. Furthermore, after UV irradiation the inhibition of Ser-126 phosphorylation and reduction of p33ING1b expression was more dramatic when using ATR or Chk1 targeted siRNAs.
Having the duality of Ser-126 phosphorylation in mind, it is not surprising that the reduction of phosphorylated p33ING1b caused by the PIKK inhibitor wortmannin was limited compared to staurosporine in resting cells (Fig. 1B
). In the lack of genotoxic stress, Chk1 would not be activated by ATM/ATR and wortmannin would have a limited effect. However, staurosporine is recognized as a highly potent albeit nonselective inhibitor of PKC with reported IC50 values as low as 3 nM in cultured cells (42)
. However, it has been recently reported to inhibit both Chk1 and Cdk1 within a comparable range (43)
, thus explaining the drastic reduction of phosphorylated p33ING1b as well as overall reduction of protein levels that we observed on treatment with staurosporine (Fig. 1B
). The fact that Ser-126 motif (TAGNSGKA) does not comply with either Chk1 or Cdk1 consensus substrates (R-X-X-S/T and S/T*-P-x-K/R, respectively) does not discard p33ING1b from being phosphorylated by these kinases. In fact, a wide variety of substrates has been described for both Chk1 (44)
and Cdk1 (45)
. Of particular interest is the observation that Ser-20 on p53, which does not comply with Chk1 consensus target, can be phosphorylated in vitro by this kinase (44)
. It has also been described that the presence of a full consensus site is not sufficient for Cdk phosphorylation (45)
and that substrate recognition by Cdks is influenced by the associated cyclin subunit (46)
. Therefore, the ability of a kinase to phosphorylate a target must be evaluated not only on primary sequence but also on physiological conditions, subcellular localization of the protein, and protein-protein interactions. Although it remains to be tested, it is also interesting to note that p33ING1b contain a potential cyclin-dependent kinase binding motif (a.a. 76–79). This motif may act as a substrate recognition site that directly interacts with cyclin to promote phosphorylation by the Cdk-cyclin complex (47)
and may help Cdk1-cyclin B1 phosphorylate p33ING1b. As shown in this work, both Cdk1 and Chk1 have the ability to phosphorylate p33ING1b at Ser-126 in vitro (Fig. 4D
). Whether these kinases can phosphorylate other residues on p33ING1b or other kinases could phosphorylate Ser-126 remains to be explored. Our results showing that phosphorylation at Ser-126 is cell-cycle restricted, being maximal during G2/M phase, when the activity of Cdk1/cyclin B1 complex is maximal (Fig. 6A
), reinforces the idea that this complex controls the phosphorylation of p33ING1b in MMRU cells. However, UVB-irradiation induces p33ING1b phosphorylation at Ser-126 (Figs. 2D
and 6B
), which reach its maximum during S phase.
A past study demonstrated that CCNB1, which expresses cyclin B1, is transcriptionally down-regulated by p33ING1b (38)
. This observation suggests that p33ING1b may play a key role in cell cycle regulation because cyclin B1 is essential for the entry into mitosis (48)
and cell proliferation (49)
. We observed that in nonirradiated cells overexpression of phosphomimetic S126E reduces the levels of cellular cyclin B1 (Fig. 7A
), while under UVB stress overexpression of both wild-type p33ING1b and phosphomimetic S126E but not the unphosporylable S126A mutant can reduce the levels of this cyclin (Fig. 7B
). Therefore, we propose the existence of a potential negative feedback loop where on mitotic onset Cdk1-cyclin B1 phosphorylates Ser-126, leading to p33ING1b accumulation. This in turn contributes to the down-regulation of CCNB1 and therefore inactivation of Cdk1 by the end of mitosis and into subsequent cell cycle progression (Fig. 8
). On DNA damaging events Chk1 could also contribute to the down-regulation of cyclin B1 and indirectly reduce Cdk1 activity.
|
The overall effect of p33ING1b on cyclin B1 and, therefore, mitosis and cell proliferation is supported by our observation that on UVB-irradiation wild-type p33ING1b and the phosphomimetic S126E, but not the unphosphorylable S126A mutant, can decrease cell proliferation (Fig. 7E
). This finding is consistent to previous reports describing that overexpression of ectopic p33ING1b enhances the G2/M DNA damage checkpoint induced by adriamycin and decreases cell proliferation (50)
. Thus, phosphorylation of p33ING1b at Ser-126 seems to play an important role controlling key functions in cell progression. Even though we have only begun to understand the regulation of cell proliferation mediated by p33ING1b phosphorylation at Ser-126, future studies aimed at achieving a quantitative determination of the relationships between the phosphorylation of this residue and the promoter activity of cyclin B1 and/or the stability of this protein or its interaction with Cdk1 should provide insights into the molecular mechanism by which p33ING1b regulates cell proliferation.
|
ACKNOWLEDGMENTS |
|---|
Received for publication January 4, 2007. Accepted for publication May 17, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
