| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,
,1
* Department of Pathology,
Division of Research Immunology/BMT, Childrens Hospital Los Angeles Saban Research Institute, Los Angeles, California, USA;
The Burnham Institute Cancer Center, La Jolla, California, USA;
University of Southern California Keck School of Medicine, Los Angeles, California, USA
1Correspondence: Department of Pathology, MS# 103, Childrens Hospital Los Angeles, University of Southern California Keck School of Medicine, 4650 Sunset Blvd., Los Angeles, CA 90027 USA. E-mail: lingtaow{at}usc.edu
ABSTRACT
Acute promyelocytic leukemia (APL) cells express promyelocytic leukemia/retinoic acid receptor alpha (PML/RAR
) fusion protein, which leads to the blocking of APL cell differentiation. Treatment of APL with all-trans-retinoic acid (ATRA) induces disease remission by in vivo differentiation of APL cells. Differentiation requires cell cycle exit; yet how ATRA couples cell cycle exit to differentiation of APL remains largely unknown. We previously found that ATRA-induced cell differentiation accompanies ubiquitination-proteolysis of ménage à trois 1 (MAT1), an assembly factor and targeting subunit of cyclin-dependent kinase (CDK)-activating kinase (CAK) that regulates G1 exit. We report here that CAK binds to and phosphorylates PML/RAR in actively proliferating APL cells. In response to ATRA, PML/RAR is dissociated from CAK, leading to MAT1 degradation, G1 arrest, and decreased CAK phosphorylation of PML/RAR. CAK phosphorylation of PML/RAR is inhibited when MAT1 levels are reduced. Both MAT1 degradation and PML/RAR hypophosphorylation occur in ATRA-induced G1-arresting cells undergoing differentiation but not in the synchronized G1 cells that do not differentiate. These findings reveal a novel ATRA signaling on APL cell differentiation, in which ATRA coordinates G1 arrest and transition into differentiation by inducing MAT1 degradation and PML/RAR hypophosphorylation through disrupting PML/RAR binding and phosphorylation by CAK.—Wang, J.-g., Barsky, L. W., Davicioni, E., Weinberg, K. I., Triche, T. J., Zhang, X.-k., and Wu, L. Retinoic acid induces leukemia cell G1 arrest and transition into differentiation by inhibiting cyclin-dependent kinase-activating kinase binding and phosphorylation of PML/RAR.
Key Words: MAT1 degradation • cell cycle • proliferation/differentiation transition
APL IS A DISTINCT subtype of acute myeloid leukemia (1)
characterized by a t(15; 17) reciprocal chromosomal translocation that generates a chimeric promyelocytic leukemia/retinoic acid receptor alpha (PML/RAR) (2
3
4)
. PML/RAR blocks myeloid differentiation and suppresses apoptosis (5
6
7)
. Paradoxically, PML/RAR also mediates the differentiation response of acute promyelocytic leukemia (APL) cells to all-trans-retinoic acid (ATRA) stimulation (8
9
10
11)
. Hence, PML/RAR is responsible for both the pathogenesis of APL and its sensitivity to ATRA treatment. ATRA induces leukemic cell differentiation from reduced proliferation on its binding to retinoid receptors, notably PML/RAR (6
, 12)
. APL is the first human malignancy that can be effectively treated with ATRA, yet how binding of ATRA to PML/RAR induces leukemic cell proliferation/differentiation (P/D) transition is poorly understood.
P/D transition requires cell cycle exit, which may result from a dual function of cell cycle regulators that coordinate cell cycle arrest and transition into differentiation. Human cyclin-dependent kinase (CDK)-activating kinase (CAK) is a trimeric enzyme complex consisting of CDK7 (13)
, cyclin H (14)
, and MAT1 (15
,16)
. CAK, through its activation of CDKs (17)
, regulates cell cycle G1 exit, in which cells commonly commit to proliferation (18
19
20
21
22)
or to differentiation (23
, 24)
. CAK is also a part of general transcription factor IIH (TFIIH) (25)
. TFIIH-containing CAK regulates transcription initiation (26
27
28
29)
by phosphorylating the largest subunit of RNA polymerase II (RNA Pol II) (26
, 29)
. Retinoblastoma tumor suppressor protein (pRb), a proliferation repressor (17
, 30
31
32
33)
and a differentiation enhancer (34
, 35)
, is another key substrate for CAK (22
, 24)
. Furthermore, CAK phosphorylates RAR (23
, 36
37
38)
, a key player in mediating myeloid differentiation (6
, 39)
. Decreased CAK phosphorylation of pRb, RAR, and RXR is associated with ATRA-induced P/D transition (23
, 24)
. These studies indicate that CAK plays a key role in the crosstalk between cell cycle, transcription, and differentiation through its phosphorylation regulation of CDKs, RNA Pol II, pRb, and RARs.
MAT1 was initially discovered as an assembly factor of CAK (15
, 16
, 40
41
42)
. A 30-kDa MAT1-associated fragment (M30), likely a C-terminal-truncated MAT1 (23)
, is associated with MAT1 in multiple tumor cell lines (23
, 24
, 43)
. Comprehensive studies have demonstrated that MAT1 determines CAK’s substrate specificity (15
, 22
, 28
, 36
, 44
45
46
47)
. MAT1 can shift the substrate preference of cyclin H/CDK7 from CDK2 to RNA Pol II (28
, 44
, 48
, 49)
, tumor suppressor p53 protein (46)
, and octamer transcription factors (47)
. Anti-MAT1 antibody (Ab) inhibits TFIIH-dependent transcription (49)
, and mice lacking MAT1 are defective in RNA Pol II phosphorylation (45)
. Retroviral-MAT1 antisense decreases CAK phosphorylation of pRb (22)
, leading to CAK-dependent G1 arrest (22
, 50)
. Hence, CAK activity in regulating cell-cycle G1 exit and transcription are largely controlled by MAT1.
Recent studies show that ATRA-induced ubiquitination-proteolysis of MAT1 in human leukemia HL60, APL-NB4, and neuroblastoma CHP126 cells (43)
is accompanied by G1 arrest and cell differentiation (23
, 24)
. MAT1 enhances CDK7-cyclin H phosphorylation of RAR (36)
, while ATRA-induced MAT1 ubiquitination decreases CAK phosphorylation of RAR (43)
. Furthermore, decrease of levels of MAT1 protein by adenovirus-MAT1 antisense mimics ATRA-induced pRb hypophosphorylation, RXR hypophosphorylation, G1 arrest, and neurite outgrowth in CHP126 cells (24)
. These studies suggest that ATRA induces CAK-dependent P/D transition by inducing MAT1 degradation and suppressing CAK phosphorylation of retinoid receptors. The effects of ATRA on cell differentiation are mainly mediated by its receptors. However, how ATRA on its receptor binding modulates CAK-dependent G1 arrest and transition into differentiation remains unclear. In this study, we investigate the effect of ATRA on interaction between CAK and PML/RAR in ATRA-induced P/D transition in APL-NB4 cells expressing PML/RAR. Our results demonstrate that PML/RAR binds strongly to CAK in vitro and in vivo in the absence of ATRA, which is associated with PML/RAR hyperphosphorylation. On treatment with ATRA, PML/RAR is dissociated from CAK, leading to MAT1 degradation, G1 arrest, decreased CAK phosphorylation of PML/RAR, and onset of APL cell differentiation.
MATERIALS AND METHODS
Cell culture
Human NB4 cells (provided by Dr. Lanotte) were cultured in RPMI 1640 with 10% FBS and 2 mM L-glutamine. Human breast carcinoma MCF-7 cells were cultured in Dulbecco’s modified Eagle medium (DMEM) plus 10 µg/ml insulin (Sigma, St. Louis, MO), 10% FBS, and 2 mM L-glutamine. Cells were grown in the medium supplemented with 100 U/ml penicillin and 100 µg/ml streptomycin at 37°C in a 5% CO2. ATRA (Sigma) was dissolved in ethanol, and 5 µM of ATRA was used for treatment as described (23
, 51)
.
Cell cycle synchronization and analysis
Nocodazole, thymidine, and aphidicolin were purchased from Sigma. Cell cycle synchronization was performed as described (15
, 52)
. Briefly, NB4 cells were synchronized in prometaphase by treating with 100 ng/ml of nocodazole for 16 h, then released into fresh medium at various intervals. To release NB4 cells from G1/S boundary that was synchronized with thymidine/aphidicolin double-block, cells were first exposed to 2 mM of thymidine for 14 h, and then released into fresh medium for 14 h, followed by a second arrest in the presence of 1 µg/ml of aphidicolin for 14 h. Cells were released from G1/S boundary by feeding with fresh medium for different periods. Cells were harvested, washed, and processed for flow cytometric analysis of the cell cycle profile as described previously (22)
.
Plasmids, in vitro translation, and in vitro CAK assembly
The coding cDNA sequences of CDK7, cyclin H, and MAT1 were cloned into pET protein expression vector with His-tag (Novagen, San Diego, CA), as described previously (13
, 22)
. In vitro transcription-translation of proteins in pET expression vectors was performed using the TNT-coupled reticulocyte lysate system (Promega, Madison, WI), according to the procedures provided by the manufacturer. In vitro CAK assembly was performed as described previously (22)
.
Plasmid construction, glutathione S-transferase fusion protein expression, glutathione S-transferase-pull-down assay, and in vitro protein binding assay
The PML/RAR cDNA from pCMX-PML/RAR construct (provided by Dr. Evans) was cloned into the pGEX-2T expression vector (Amersham Pharmacia, Piscataway, NJ) using our established techniques (13
, 22)
. Glutathione S-transferase (GST)-PML/RAR induction and purification were performed as described previously (22)
. Interaction between GS-bound GST-PML/RAR and each of in vitro translated CAK subunits was determined by using GST pull-down assay (22)
. Interaction of GST-PML/RAR with either recombinant CAK complex or cellular CAK complex was determined by using immunoprecipitation-dependent protein-binding assay. CAK complexes were precipitated from in vitro-assembling reaction mixtures (see Plasmids, in vitro translation, and in vitro CAK assembly above) or from NB4 cells using anti-CDK7 Ab immunoabsorbed onto protein A-agarose beads. The CAK complexes retained on the beads were then incubated with GST-PML/RAR proteins in protein binding buffer (PBB) at 4°C for 1 h as described (22)
. After washing the binding mixtures six times with PBB, proteins retained on the beads were determined by Western blot analysis.
In vivo phosphorylation analysis and in vitro kinase assay
In vivo CAK phosphorylation of CAK-bound PML/RAR and pRb was determined by metabolic labeling in NB4 cells with [32P]-orthophosphate as described (23)
. In vitro kinase assays were performed in the presence of [
32P]-ATP as described (22)
. To assess CAK activity, CAK complexes immunoprecipitated from ATRA-treated or synchronized cells were incubated with either 500 ng of GST-PML/RAR or 100 ng of histone H1. The reaction mixtures were resolved by SDS-PAGE, electrotransferred onto polyvinylidene difluoride (PVDF) membrane, and autoradiographed.
Immunological methods
Western blot analysis and immunoprecipitation were performed as described (22)
. Enhanced chemiluminescence (ECL) (Applied Biosystems, Foster City, CA) was used for detection of immunoreactive proteins in Western blot analysis. Antibodies used for immunoprecipitation and Western blot analysis, including rabbit polyclonal antibodies against MAT1 (FL-309), CDK7 (C-19), RAR (C-20), pRb (C-15), cyclin H (H-279), goat polyclonal antibodies against MAT1 (M-20), CD11b (M-19), mouse monoclonal antibody (mAb) against CDK7 (C-4), and PML (PG-M3) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA).
Cell proliferation and morphology analysis
Cell replication in proliferating, G1 arresting, and differentiating cells was determined by cell counting as described before (50)
. Nuclear segmentation of NB4 cells treated with or without ATRA was determined with Wright-Giemsa (Sigma) stain as described (23)
.
RESULTS
Binding of PML/RAR with CAK
PML/RAR, retaining RARS77 residue known to be phosphorylated by CAK (36)
, modulates NB4 cell differentiation on ATRA binding (8
9
10
11)
. To study the involvement of CAK activity in ATRA-induced NB4 cell P/D transition, we determined whether it interacted with PML/RAR in vitro and in NB4 cells. In GST-pull-down assay, in vitro translated recombinant CDK7, MAT1, or cyclin H was incubated with glutathione-sepharose (GS)-bound GST-PML/RAR. After washing the binding mixtures, CAK subunits retained on GS-bound PML/RAR were determined by immunoblotting. Our results showed that GST-PML/RAR had higher binding affinity to His-CDK7 compared to His-MAT1 and His-cyclin H (Fig. 1
A, lanes 2 vs. 3 and 4). In coimmunoprecipitation assay, in vitro assembled recombinant CAK complexes or cellular CAK complexes were immunoprecipitated by anti-CDK7 Ab immunoabsorbed onto protein A-agarose. The CAK complexes retained on the beads were then incubated with GST-PML/RAR proteins. After washing the binding mixtures, the resulting precipitates were immunoblotted with anti-RAR, anti-MAT1, anti-CDK7, or anticyclin H antibodies. We found that GST-PML/RAR interacted with recombinant CAK because it was coimmunoprecipitated together with His-CDK7, His-MAT1, or His-cyclin H (Fig. 1A
, lane 1). Cellular CAK interacted with both GST-PML/RAR and endogenous PML/RAR, as both GST-PML/RAR and cellular PML/RAR were coimmunoprecipitated by anti-CDK7 Ab (Fig. 2
B, lane 1). This indicates that PML/RAR also binds to CAK in vivo. To evaluate that CAK-PML/RAR interaction in vivo, the putative CAK-PML/RAR complexes were coprecipitated from NB4 cells using anti-PML, anti-MAT1, or anti-CDK7 Ab. The resulting precipitates were then immunoblotted with anti-RAR, anti-CDK7, anticyclin H, and anti-MAT1 antibodies. Our results showed that PML/RAR was precipitated together with MAT1, CDK7, and cyclin H by either anti-PML, anti-MAT1, or anti-CDK7 Ab (Fig. 1, C and D
). pRb was also precipitated together with MAT1 by anti-CDK7 Ab from NB4 cells (Fig. 1E
). Together, these results demonstrate that CAK interacts with PML/RAR and pRb in vivo.
|
|
CAK-PML/RAR dissociation precedes MAT1 degradation and reduction of CAK abundance in ATRA-induced P/D transition.
PML/RAR binds to CAK in the absence of ATRA (Fig. 1, C and D
). We therefore investigated whether ATRA modulates PML/RAR’s binding to CAK. CAK-PML/RAR complexes were precipitated with anti-CDK7 Ab from NB4 cells with different periods of ATRA treatment. The resulting precipitates were then immunoblotted with anti-RAR and anti-MAT1 antibodies. CAK complex exists as a 1:1:1 M ratio of cyclin H, CDK7, and MAT1. Thus, the levels of MAT1 in the precipitates resulting from anti-CDK7 precipitation represent the net abundance of CAK, whereas the levels of PML/RAR in the precipitates represent the levels of PML/RAR bound to CAK. The levels of PML/RAR bound to CAK were reduced within 12 h of ATRA treatment, whereas there was no change in CAK abundance, as evidenced by constant levels of MAT1 (Fig. 2A
). Thereafter, the levels of PML/RAR bound to CAK remained at constant low levels, whereas CAK abundance represented by MAT1 levels was decreased after 48 h of ATRA treatment (Fig 2B
). We next examined whether these temporal changes in CAK-PML/RAR dissociation and MAT1 degradation correlated to the development of ATRA-induced NB4 cell P/D transition. We treated NB4 cells with either 1 or 5 µM of ATRA for up to 6 days. By assessing cell proliferation with cell counting, we found that NB4 cell proliferation was inhibited during the first 3 days of treatment with either 1 µM or 5 µM ATRA, followed by constant low levels of proliferation for up to 6 days (Fig. 2C
). pRb expression was also inhibited after 72 h of ATRA treatment, which was associated with decreased hyperphosphorylated form of pRb (Fig. 2D
). After 3 to 4 days of ATRA treatment, cells then underwent differentiation, as evidenced by halted proliferation (Fig. 2C
), nuclear segmentation (Fig. 2E
), and CD11b expression (Fig. 2F
). Taken together, these results show that ATRA dissociates PML/RAR from CAK, leading to MAT1 degradation and decreased CAK abundance. Reduction of MAT1 is then followed by proliferation inhibition, decreased hyperphosphorylated form of pRb, nuclear segmentation, and CD11b induction. Because PML/RAR mediates the effect of ATRA on cell differentiation while MAT1 assembles CAK, ATRA-induced PML/RAR dissociation from CAK may be involved in decreasing CAK abundance by inducing MAT1 degradation in NB4 cells (43)
, leading to CAK-dependent NB4 cell P/D transition.
ATRA selectively induces MAT1 degradation in G1 arresting cells that commit to differentiation
To determine whether MAT1 degradation specifically occurred in ATRA-induced G1 arresting cells toward transition into differentiation, NB4 cells were arrested in the G1 phase with different periods of ATRA treatment. In parallel, NB4 cells were also synchronized at the G1/S boundary using thymidine/aphidicolin double block or in prometaphase using nocodazole treatment. The synchronized cells were then released from cell cycle block at different time points crossing a full cell cycle. We determined a cell cycle profile in these different cultures using flow cytometric analysis. Over 80% of cells were arrested in G1 phase after 72 h of ATRA treatment (Fig. 3
A). In parallel, 
70% of cells were blocked in G1 phase with thymidine/aphidicolin treatment (Fig. 3B
) or progressed into G1 phase between 3 to 9 h after released from nocodazole arrest (Fig. 3C
). In duplicate cultures, the total protein levels of MAT1 were determined using Western blot analysis. Although MAT1 levels were slightly reduced after 9 h of ATRA treatment, significant reduction of MAT1 proteins occurred after 48 h of ATRA treatment (Fig. 3D
). This significant MAT1 reduction at the total protein levels was accompanied by decreased CAK abundance in the precipitates resulting from anti-CDK7 precipitation after 48 h of ATRA treatment (2A vs. 2B), demonstrating that decreased CAK assembly occurs when MAT1 levels are markedly reduced. Such MAT1 reduction and decreased CAK abundance led to G1 arrest after 48–72 h of ATRA treatment (Fig. 3A
). Previous studies showed that ATRA induces PML/RAR degradation (53)
. We also found that PML/RAR levels were decreased after 12 h of ATRA treatment and then remained at constant low levels (Fig. 3E
). In contrast, the levels of MAT1 and PML/RAR remained unchanged in cells that were synchronized in G1/S boundary with thymidine/aphidicolin (Fig. 3F
) or released into G1 phase from nocodazole block (Fig. 3G
), which were similar to the steady levels of cyclin H and CDK7 (Fig. 3F
and 3G
). These results demonstrate that MAT1 is selectively degraded in ATRA-induced G1 arresting cells undergoing transition into differentiation but not in the synchronized G1 cells that do not differentiate. Furthermore, although PML/RAR was rapidly decreased after 12 h of ATRA treatment (Fig. 3E
), ATRA-induced MAT1 degradation was delayed, occurring after 48 h of ATRA treatment (Fig. 3D
). Hence, the rapid reduction of PML/RAR by ATRA may be also involved in modulating MAT1 degradation in NB4 cells.
|
Decreased CAK phosphorylation of PML/RAR occurs only when MAT1 levels are reduced
MAT1 levels were decreased after 48 h of ATRA treatment at both total protein levels and within the CAK complexed-precipitates (Fig. 2B
and 3D
), leading to decreased CAK abundance (Fig. 2B
). We therefore examined whether ATRA-induced MAT1 degradation inhibited CAK phosphorylation of PML/RAR. CAK bound-PML/RAR was precipitated by anti-CDK7 Ab from NB4 cells treated with ATRA. Using GST-PML/RAR as an additional substrate in an in vitro kinase assay, CAK activity in the precipitates was assessed by its phosphorylation of both GST-PML/RAR and CAK-bound PML/RAR simultaneously in the presence of [32P]-ATP. We found (Fig. 4
) that the phosphorylation of CAK-bound PML/RAR was inhibited to undetectable level after 48 h of ATRA treatment. In parallel, GST-PML/RAR phosphorylation was progressively diminished after 48 h of ATRA treatment, and it then became undetectable after 72 h of ATRA treatment. Although loss of phosphorylation in PML/RAR occurred earlier than in GST-PML/RAR, significant reduction in phosphorylation of both PML/RAR and GST-PML/RAR occurred after 48 h of ATRA treatment. Such decreased phosphorylation of both GST-PML/RAR and CAK-bound PML/RAR corresponded with decreased CAK activity, as shown by autohypophosphorylated [32P]-CDK7 after 48 h of ATRA treatment. By immunoblotting the autoradiography blot with anti-RAR, anti-CDK7, and anti-MAT1 antibodies, we found that, in the presence of a consistent amount of GST-PML/RAR, decreased CAK phosphorylation of GST-PML/RAR occurred when MAT1 levels were decreased after 48 h of ATRA treatment. Consistently, although CAK-bound PML/RAR phosphorylation was inhibited to an undetectable level when MAT1 levels were reduced after 48 h of ATRA treatment, the levels of CAK-bound PML/RAR proteins in the precipitates remained at constant low levels for up to 168 h after 12 h of ATRA treatment. CDK7 Ab distinguished both forms of the autohyperphosphorylated CDK7 (P-K7) and the autohypophosphorylated CDK7 (K7). The change from P-K7 to K7 accompanied both MAT1 reduction and a switch from PML/RAR (or GST-PML/RAR) hyper- to hypophosphorylation. Hence, these results demonstrate that decreased CAK phosphorylation of either GST-PML/RAR or CAK-bound PML/RAR results from ATRA-induced MAT1 reduction.
|
Decreased CAK phosphorylation of PML/RAR accompanies progressive degradation of MAT1 in vivo
PML/RAR phosphorylation was abolished after 72 h of ATRA treatment (Fig. 4)
, which was associated with the development of cell differentiation (Fig. 2, C-F
). We therefore examined the relationship between MAT1 reduction and decreased CAK phosphorylation of PML/RAR in differentiating NB4 cells after 72 h of ATRA treatment. NB4 cells following different periods of ATRA exposure were in vivo labeled with [32P] orthophosphate. CAK and PML/RAR were then coimmunoprecipitated using anti-CDK7 or anti-MAT1 Ab. Autoradiography analyses of these complexed-precipitates revealed that CAK hyperphosphorylated PML/RAR in proliferating cells in the absence of ATRA. Densitometry analysis (data not shown) showed that CAK phosphorylation of PML/RAR was inhibited in differentiating cells after 72 h of ATRA treatment, with over 90% of inhibition after 120 h of ATRA treatment compared to proliferating cells without ATRA treatment (Fig. 5
A, B). CAK phosphorylation of pRb was similarly inhibited (Fig. 5C
). Inhibition of PML/RAR and pRb phosphorylation by ATRA was associated with decreased CAK activity, as evidenced by decreased [32P]-CDK7 autophosphorylation (Fig. 5A, B
). To determine the levels of MAT1, CAK abundance, and CAK-bound PML/RAR in the precipitates, the autoradiography blots were probed with anti-RAR, anti-CDK7, or anti-MAT1 Ab. We found that, in the presence of a consistent low level of PML/RAR in differentiating cells, the development of CDK7 autohypophosphorylation and decreased PML/RAR phosphorylation corresponded with progressively decreased MAT1 levels (Fig. 5, A
and 5B
, lanes 2 vs. 3). Progressively decreased MAT1 levels were associated with progressively reduced CAK abundance. This was shown by either decreased MAT1 levels in the complexed-precipitates resulting from anti-CDK7 precipitation (Fig. 5A
) or decreased CDK7 levels in the complexed-precipitates resulting from anti-MAT1 precipitation (Fig. 5B
). Together, these results demonstrate that the development of MAT1 degradation on ATRA:PML/RAR binding leads to progressively decreasing CAK phosphorylation of PML/RAR and pRb in differentiating cells. Because CAK-PML/RAR association prevents MAT1 degradation and sustains CAK hyperphosphorylation of PML/RAR, CAK-PML/RAR dissociation may induce MAT1 degradation to decrease CAK phosphorylation of PML/RAR (Figs. 2A
, 2B
, 5A
, and 5B)
.
|
ATRA selectively inhibits CAK activity in G1 arresting cells committed to differentiation
ATRA decreased CAK phosphorylation of PML/RAR (Fig. 4)
. We therefore examined whether CAK activity in phosphorylation of PML/RAR was selectively decreased in ATRA-induced G1 arresting cells committed to differentiation. We either arrested NB4 cells in G1 phase with ATRA treatment or released NB4 cells into G1 stage from nocodazole block. CAK complexes precipitated with MAT1 Ab from these different G1 cells were assessed for their activity using histone H1 as a substrate in an in vitro kinase assay. Autoradiography showed that CAK activity was inhibited in ATRA-induced G1 arresting cells undergoing differentiation, as evidenced by decreased CAK phosphorylation of histone H1 (Fig. 6
A). Inhibition of CAK activity was also confirmed as shown by autohypophosphorylation of [32P]-CDK7. In contrast, phosphorylation of histone H1 by CAK remained unchanged in the synchronized G1 cells released from prometaphase block imposed by nocodazole (Fig. 6B
). To examine MAT1 levels and CAK abundance in the precipitates, these autoradiography blots were probed with anti-MAT1 and anti-CDK7 antibodies. We found that MAT1 was decreased in ATRA-treated cells (Fig. 6A
) but remained at constant high levels in the synchronized G1 cells (Fig. 6B
). Consistently, CAK abundance represented by CDK7 levels in the immunoprecipitates resulting from anti-MAT1 precipitation was decreased with ATRA treatment (Fig. 6A
) but remained at constant high levels in the synchronized G1 cells (Fig. 6B
). These results demonstrate that CAK activity is selectively inhibited in ATRA-induced G1 arresting cells undergoing differentiation but not in the synchronized G1 cells that do not differentiate. Previous studies demonstrated that PML/RAR mediates the effect of ATRA in transactivating differentiation-responsive genes (6
, 12)
. Because decreased CAK phosphorylation of PML/RAR after 72 h of ATRA treatment (Figs. 4
and 5)
accompanied cell differentiation (Fig. 2, C-F
), PML/RAR hypophosphorylation resulting from decreased CAK activity may be involved in modulating ATRA-dependent transcriptional control of cell differentiation.
|
DISCUSSION
Retinoid receptors mediate the effect of ATRA on cell differentiation. We investigated the involvement of PML/RAR in mediating ATRA-induced G1 arrest and transition into differentiation. Our studies suggest a dual mechanism of RA action in regulating P/D transition, in which ATRA modulates CAK-dependent G1 arrest by preventing CAK-PML/RAR association and CAK phosphorylation of PML/RAR, which might be subsequently involved in modulating ATRA-dependent transcriptional control of cell differentiation.
ATRA-induced MAT1 degradation couples reduced proliferation to cell differentiation by inducing CAK-dependent G1 arrest and decreasing CAK phosphorylation of PML/RAR.
It has been long established that pRb hypophosphorylation suppresses proliferation to sustain differentiation, while retinoids function as ligands of retinoid receptors that transactivate differentiation-responsive genes (54)
. How is this reduced proliferation linked to transcriptional control of cell differentiation? Decreased CAK phosphorylation of pRb and retinoid receptors undergoing ATRA-induced MAT1 ubiquitination (43)
accompanies G1 arrest and transition into differentiation (23
, 24)
. Thus, ATRA may couple G1 arrest to transcriptional control of cell differentiation by inducing MAT1 degradation to decrease CAK phosphorylation of pRb and retinoid receptors. Knowing that MAT1 reduction induces CAK-dependent pRb hypophosphorylation and G1 arrest (22
, 50)
, we examined the relationship between MAT1-dependent CAK activity and PML/RAR phosphorylation in ATRA-induced NB4 cell P/D transition. Our results demonstrate that CAK phosphorylation of either GST-PML/RAR or CAK-bound PML/RAR is inhibited only when MAT1 levels are decreased after 48 h of ATRA treatment (Fig. 4)
. Similarly, cell-based in vivo phosphorylation analyses show that decreased CAK phosphorylation of PML/RAR occurs only when MAT1 levels are decreased (Fig. 5)
. The induction of MAT1 degradation (Figs. 2B
and 3D
) corresponds with the occurrence of G1 arrest and decreases CAK activity (Figs. 3
, 4
5
6)
. These results present evidence that, on CAK-PML/RAR dissociation in the presence of ATRA, MAT1 degradation simultaneously induces CAK-dependent G1 arrest and decreases CAK phosphorylation of PML/RAR. Because PML/RAR hypophosphorylation resulting from decreased CAK activity (Figs. 4
5
6)
accompanies cell differentiation (Fig. 2, C-F
), PML/RAR hypophosphorylation may modulate the effect of ATRA to transactivate differentiation-responsive genes in CAK-dependent G1 arresting cells (Fig. 7
).
|
ATRA-induced CAK-PML/RAR dissociation promotes MAT1 degradation that leads to suppress CAK phosphorylation of PML/RAR
In the absence of ATRA, PML/RAR forms a complex with CAK in actively proliferating NB4 cells. Binding of PML/RAR with ATRA rapidly dissociates PML/RAR from CAK (Fig. 2, A
vs. B), leading to subsequent MAT1 degradation and decreased CAK phosphorylation of PML/RAR (Figs. 2B
, 3D
, and 4
5
6
). Decreased CAK phosphorylation of RAR can be overcome by inhibiting MAT1 ubiquitination in HL60 cells with MG 132 treatment (43)
. Our results therefore suggest that ligand binding dissociates PML/RAR from CAK to induce MAT1 degradation, leading to inhibit CAK phosphorylation of PML/RAR. Indeed, binding of ATRA to retinoid receptors induces ubiquitination-proteolysis of MAT1 in NB4, HL60, and CHP126 cells (43)
. Furthermore, in HL60R cells harboring a mutant RAR with deleted AF-2 activation domain core (AF-2ADc), ATRA failed to induce CAK-RAR dissociation, MAT1 degradation, and RAR hypophosphorylation (23)
. This observation suggests that AF-2ADc plays an important role in mediating the effect of ATRA on CAK-PML/RAR dissociation and MAT1 degradation. AF-2ADc transduces RA signaling (55)
. Comparison of the crystal structures of unliganded and liganded RARs (56
, 57)
reveals that ligand-binding induces a conformational change in the AF-2ADc through helix 12 (54
, 58)
, resulting in the release of corepressors (58)
and generating a new surface for binding coactivators (59
60
61)
. Through this conformation change, RARs also interact with regulatory complex of the 26S proteasome (62
, 63)
. In this regard, liganded PML/RAR once released from CAK may either induce a gene product or activate an existing molecule to promote MAT1 ubiquitination in NB4 cells. This adaptor molecule may target MAT1 to its preexisting E3 ligase. On the other hand, MAT1 is a Ring-figure protein. A number of recent studies have shown that several Ring-finger proteins act as self-E3 ligases to promote self-degradation (64
65
66)
. It is possible that in a proliferating state of NB4 cells, MAT1 itself does not possess an E3 ligase activity. Liganded PML/RAR may mediate binding of an activator to MAT1 and induce MAT1 E3 ligase activity for MAT1 self-degradation. In either of the above possibilities, CAK-PML/RAR dissociation may induce the expression of a cofactor that promotes MAT1 degradation to suppress CAK phosphorylation of PML/RAR.
PML/RAR hypophosphorylation couples CAK-dependent G1 arrest to ATRA-dependent transcriptional control of cell differentiation
CAK phosphorylation of PML/RAR is inhibited when MAT1 levels are reduced after 48 h of ATRA treatment (Fig. 4)
in G1 arresting cells committed to differentiation (Figs. 2
and 3)
. What is the role of PML/RAR hypophosphorylation in ATRA-induced P/D transition? According to the current model of gene regulation by retinoids (54
, 60)
, in a context of chromatin where the nucleosomes do not impede the binding of RAR/RXR heterodimers to their DNA recognition sequences, unliganded and DNA-bound retinoid receptors repress transcription by recruiting corepressors (6
, 54
, 59
, 67)
. Thus, to activate gene expression, retinoid receptors will have to contend with the repressive chromatin structures in order to dissociate repressors and to allow the recruitment of coactivators (54
, 59
, 67)
. PML/RAR represses transcription to a greater extent than wild-type (WT) RAR in the absence of ATRA (6
, 68
, 69)
, whereas PML/RAR mediates the effect of ATRA in inducing APL cell differentiation (6
, 12)
. Our data show that decreased CAK phosphorylation of PML/RAR after 72 h of ATRA treatment (Figs. 4
5
6)
accompanies cell differentiation (Fig. 2, C-F
). Hence, ATRA-induced PML/RAR hypophosphorylation might mediate cell differentiation by releasing transcription-repression of ATRA-responsive genes. It is likely that transcription of ATRA-target genes is inhibited by PML/RAR hyperphosphorylation in the absence of ATRA. ATRA-induced conformation change in PML/RAR resulting from decreased CAK phosphorylation of PML/RAR might facilitate its dissociation from corepressors and recruitment of coactivators in order to transactivate ATRA-target genes.
The effect of ATRA:PML/RAR binding on cell differentiation is generally considered to act in a transcription-dependent mechanism. Our studies here reveal that ATRA may induce MAT1 degradation by dissociating PML/RAR from CAK to coordinate G1 arrest and transition into differentiation. Hence, the complex formation of PML/RAR with CAK and its regulation by ligand represent a novel mechanism of ATRA signaling in mediating cell differentiation. Using a loss of both MAT1 function and RAR phosphorylation in cultured leukemia cells transduced with lentiviral-MAT1 shRNA or lentiviral-phosphorylation-defective RARS77A, the roles of MAT1 reduction and PML/RAR hypophosphorylation in coupling ATRA-induced myeloid leukemia cell P/D transition are currently being evaluated as a logical extension to this work in our laboratory. We anticipate that the detailed molecular mechanisms of ATRA action in posttranslational modification of MAT1 reduction and PML/RAR hypophosphorylation will present new ideas in the design of improved intervention strategies against leukemia.
ACKNOWLEDGMENTS
We thank Dr. M. Lanotte for providing NB4 cell line and Dr. R. M. Evans for providing pCMX-PML/RAR construct. This work was supported by Grants from the National Institutes of Health to L.W, X-k.Z, and K.I.W.
Received for publication April 10, 2006. Accepted for publication May 15, 2006.
REFERENCES
