PRAM-1 potentiates arsenic trioxide-induced JNK activation.

The promyelocytic leukemia RARalpha target gene encoding an adaptor molecule-1 (PRAM-1) is involved in a signaling pathway induced by retinoic acid in acute promyelocytic leukemia (APL) cells. To better understand the function of PRAM-1, we have undertaken the identification of its partners through a yeast two-hybrid screen. Here, we show that the proline-rich domain of PRAM-1 interacted with the Src homology 3 (SH3) domain of hematopoietic progenitor kinase 1 (HPK-1)-interacting protein of 55 kDa (HIP-55, also called SH3P7 and Abp1) known to stimulate the activity of HPK-1 and c-Jun N-terminal kinase (JNK). Overexpression of PRAM-1 in the NB4 APL cell line increased arsenic trioxide-induced JNK activation through a caspase 3-like-dependent activity. Dissociation of the SH3 domain from the rest of the HIP-55 protein was observed in the NB4 APL cell line treated with arsenic trioxide due to specific cleavage by caspase 3-like enzymes. The cleavage of HIP-55 correlated with the induction of PRAM-1 mRNA and protein expression. Taken together, our results suggest that the caspase 3-cleaved SH3 domain of HIP-55 is likely involved in PRAM-1-mediated JNK activation upon arsenic trioxide-induced differentiation of NB4 cells.

Acute promyelocytic leukemia (APL) 1 is associated with five reciprocal translocations always involving the retinoic acid receptor ␣ (RAR␣) (1,2). In more than 95% of APL, the specific translocation t (15;17) produces the PML⅐RAR␣ and RAR␣⅐PML fusion proteins (3). The association of histone deacetylases with PML⅐RAR␣ has been described in APL cells (4 -6). PML⅐RAR␣ would therefore recruit a histone deacetylase complex to RA response elements leading to the repression of RA target genes critical to myeloid differentiation. The dissociation of the histone deacetylase complex from the fusion protein is obtained by treating cells with pharmacological concentrations of RA, which could explain the sensitivity of APL cells to RA (4 -6). When treated with RA, these cells can withdraw from the cell cycle and undergo terminal maturation both in vitro (7) and in vivo (8 -10). This suggests a molecular mechanism by which RA-responsive genes critical to myeloid differentiation are repressed in leukemia cells and de novo induced when these cells are treated with RA (7,8). Arsenic trioxide (As 2 O 3 ) has been identified as an alternative therapy in patients with both RAsensitive and RA-resistant APL (9, 10). As 2 O 3 triggers apoptosis and differentiation of APL cells both in vitro and in vivo (10,11). Indeed, PML⅐RAR␣-bound co-repressors are released from DNA upon As 2 O 3 treatment of APL cells, leading to the activation of PML⅐RAR␣ target genes through PML⅐RAR␣ degradation (12).
To identify genes controlling proliferation and/or induced differentiation of both normal and leukemia cells, we have used a differential screening strategy to isolate genes that are activated during induced maturation of APL cells (13). Among four novel genes, PRAM-1 (PML⅐RAR␣ target gene encoding an adaptor molecule-1) is a first example of an adaptor in which expression is inhibited and superinduced when PML⅐RAR␣ is expressed alone and in the presence of RA, respectively (14). PRAM-1 shares structural homologies with SLAP-130 (SLP-76associated protein of 130 kDa)/FYB (Fyn-binding protein) renamed ADAP (adhesion-and degranulation-promoting adaptor protein) (15,16), which is involved in clonal expansion of activated T-cells through transcriptional regulation of the IL2 gene (17,18), mastocyte degranulation (19), as well as integrinmediated adhesion (20,21). PRAM-1 interacts with several signaling intermediates, such as the SLP-76 and SKAP55-HOM adaptors and the Src tyrosine kinase LYN (14).
To decipher PRAM-1 mechanism of action, we have undertaken the identification of its partners through a yeast twohybrid screen. Here we show that the proline-rich domain of PRAM-1 interacts with the SH3 domain of HIP-55. Dissociation of the SH3 domain from the rest of the HIP-55 protein was found in NB4 APL cells treated with As 2 O 3 due to a specific cleavage by caspase 3-like enzymes. We show that PRAM-1 mRNA and protein expressions were induced in As 2 O 3 -treated NB4 cells. Furthermore, overexpression of PRAM-1 in NB4 cells potentiated As 2 O 3 -induced JNK activation.

MATERIALS AND METHODS
Cell Lines, Culture, and Differentiation-NB4 cells (34) were cultured in RPMI 1640 medium (Invitrogen) with 10% fetal bovine serum (PAA Laboratories), 2 mM glutamine and 1% penicillin-streptomycin (Invitrogen). The COS-7 cell line was grown in Petri dishes in Dulbecco's modified Eagle's medium supplemented with 5% fetal bovine serum and antibiotics. Exponentially grown NB4 cells were seeded at 2 ϫ 10 5 cells/ml 16 h prior to As 2 O 3 (ICN) treatment. Cell differentiation was determined by: (i) the percentage of nitro blue tetrazolium (Sigma)positive cells and (ii) morphological changes on cytospin slides stained with May-Grü nwald-Giemsa (VWR Scientific). For the caspase inhibition, cells were preincubated for 30 min with Z-DEVD-fmk (Calbiochem) before As 2 O 3 treatment.
Plasmid Constructs-The PRAM-1 coding sequence was subcloned into the pGBKT7 (Clontech) and into a pSG5 (35)-derived vector to direct the expression of PRAM-1. The PRAM-1⌬-(183-613) mutant was generated by the internal PvuII fragment deletion. The full-length PRAM-1 cDNA was subcloned into the pBabe retroviral vector (36). The human HIP-55 open reading frame was obtained by PCR amplification using bone marrow cDNA and subcloned into: (i) a pSG5-derived vector to direct the expression of HIP-55 fused to the FLAG epitope at its amino terminus (pSG5FN-HIP-55) and (ii) pGST, a derivative of pGEX-3X, in-frame with the glutathione S-transferase (GST) sequence. Constructs encoding the SH3 domain of HIP-55-(362-430) were amplified by PCR. The hemagglutinin-tagged HPK-1 (pcDNA3-HPK-1:HA) expression vector was a gift from Dr. F. Kiefer (23). The pGST-c-Jun-(1-223) was a gift from Dr. B. Chatton. Construction of the pSG5FN-HIP-55D361A mutated vector was achieved using the QuikChange sitedirected mutagenesis kit (Stratagene). For this, we used a mutated oligonucleotide sequence as indicated in boldface: D361A, 5Ј-CTCT-GAGCACATGCACCACCACATTC-3Ј. All constructs were verified by sequencing.
Yeast Two-hybrid Screen-To identify PRAM-1 cellular partners, a human bone marrow library was screened using the Matchmaker Two-Hybrid System-3 protocol. Briefly, a Gal4-DBD (DNA binding domain) fusion of PRAM-1 (pGBKT7-PRAM-1) was employed as a bait in the screening carried out by mating AH109-MATa pre-transformed with pGBKT7-PRAM-1 to Y187-MAT␣ library pre-transformed cells. The AH109 transformation was carried out by standard Li-acetate protocol; the library was maintained in the Y187 strain according to manufacturer's recommendations. Recombinant pGBKT7-PRAM-1 clones were tested for protein expression by Western blotting (positive), autonomous lacZ reporter activation by the bait alone (negative), cell proliferation effects through the bait (negative), mating efficiency with Y187-MAT␣, and were constantly maintained under selection of the auxotrophy marker Trp Ϫ . Once AH109-MATa-pGBKT7-PRAM-1 were confirmed through the above protocol, mating of a sufficient large pool of these cells to the pretransformed Y187-MAT␣ library-carrying strain (titer, 5 ϫ 10 7 colony-forming units/ml) was conducted according to the manufacturer's recommendations. Mating efficiency was determined in this case at 77%, with a total of 1 ϫ 10 8 independent clones, thus covering the library (3.5 ϫ 10 6 independent inserts) ϳ29 times. Mated cells were spread onto selective media (SD, Clontech) for Trp (bait), Leu (prey), His (reporter 1), Ala (reporter 2), and lacZ (reporter 3) (SD-HALTX). After plating, positive clones were collected from days 8 to 20, and restreaked on SD-HALTX, SD-LTX, and SD-LX plates for verification of growth phenotype. Clones confirmed through this secondary screening were amplified in SD-L media, selecting only for the prey library insert. Plasmid DNA was finally isolated for sequencing according to standard procedures.
One g of T7 promoter-fused cDNA was incubated according to the manufacturer's recommendations (Promega) in the presence of [ 35 S]methionine for 2 h at 30°C. In vitro binding assays were performed using ϳ1 g of GST fusion protein and 8 l of the translation products of a coupled in vitro transcription/translation of cDNAs for 30 min at room temperature. After several washes to remove unbound material in washing buffer (50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 10 mM MgCl 2 , 0.3 mM dithiothreitol, 5% glycerol, 0.1% Nonidet P-40), the proteins were recovered in Laemmli buffer, separated by SDS-polyacrylamide gel electrophoresis and analyzed by Coomassie staining or autoradiography of dried gels.
Protein Cleavage Assay-Recombinant HIP-55 proteins were incubated in 50 l of caspase reaction buffer (10 mM HEPES, 100 mM NaCl, 10 mM dithiothreitol, 1 mM EDTA, 0.1% CHAPS, pH 7.2) with 100 ng of recombinant caspase 3 (Calbiochem) at 37°C. The reaction was terminated by adding Laemmli buffer and a 2-min boiling. The samples were resolved by SDS-polyacrylamide gel electrophoresis and analyzed by Western blot.
Northern Blots-Total RNA extraction and hybridization were as described previously (14,37). The HIP-55 probe corresponded to the HIP-55 coding sequence.
Transient Transfections and Protein Extracts-COS-7 cells were transfected using calcium phosphate co-precipitation (38) of the appropriate amount of DNA vectors adjusted to 14 g/9-cm Petri dish with pBluescript carrier DNA. The medium was changed after 16 h, and 36 h after transfection, the cells were washed one time in phosphate-buffered saline, collected, and resuspended in whole cell extract buffer (50 mM Tris-HCl, pH 7.9, 0.15 M NaCl, 1 mM EDTA, 0.1% Nonidet P-40, 10% glycerol, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 1 mM Na 3 VO 4 , and 0.1% protease inhibitor mixture) for use with mammalian cell extracts. After two freeze-thaw cycles in liquid nitrogen, the resulting cell lysates were cleared by a 30-min 10,000 ϫ g centrifugation at 4°C.
In Vitro Kinase Assays-Briefly, 2 g of GST⅐c-Jun fusion protein immobilized on glutathione-Sepharose beads were incubated with 5 g of cell extracts. Kinase reactions were performed for 30 min at 30°C in the presence of 15 Ci of [␥-32 P]ATP. The beads were washed six times with kinase buffer (20 mM MOPS, pH 7.6, 10 mM MgCl 2 , 2 mM EGTA, 1 mM dithiothreitol, 1 mM Na 3 VO 4 , 0.1% Triton X-100), resuspended in Laemmli buffer and boiled for 2 min.
Production of Retroviral Vectors and Retroviral Gene Transfer-Production of retroviral vectors and retroviral gene transfer were conducted as described previously (39). NB4 cells were infected with pBabe-PRAM-1 or with control pBabe-eGFP viruses and selected with puromycin (0.5 g/ml). Several independent pools of cells were expanded for further analysis.

PRAM-1 Interacted with the SH3 Domain of HIP-55-
In an attempt to identify binding partners for PRAM-1, a yeast twohybrid screen was used with full-length PRAM-1 as a bait, together with a cDNA library from human bone marrow cells. Four independent clones were recovered: two corresponding to the full-length SKAP-55HOM coding sequence (40,41) and two to part of HIP-55 (26,27). The region covered by the two identical HIP-55 clones corresponded to amino acids 65-430 containing two tyrosine-based motifs, a caspase 3 recognition site, and a C-terminal SH3 domain (Fig. 1A) (26,27,33). The yeast AH109 strain was then transformed with SKAP-55HOM, HIP-55, or the pACT2 empty plasmid together with the PRAM-1 bait plasmid and grown onto selective media. As shown in Fig. 1B, SKAP-55HOM and HIP-55 cDNAs were found to specifically interact with PRAM-1 in yeast. A specific interaction between PRAM-1 and SKAP-55HOM was also found in myeloid cells (14). We examined therefore the in vitro binding of HIP-55 to PRAM-1. Affinity-purified GST and GST-HIP-55 proteins expressed in E. coli were incubated with in vitro translated PRAM-1 and HPK-1 proteins. As shown in Fig. 1C, PRAM-1 and HPK-1 did associate with both GST-HIP-55 and GST-HIP-55-(362-430) containing the SH3 domain but not with the GST alone. Deletion of type I and II SH3 recognition motifs within the PRAM-1 protein abolished PRAM-1 binding to the HIP-55 SH3 domain (Fig. 1D). Altogether, our results indicate that the SH3 domain of HIP-55 can associate with PRAM-1 through the proline-rich domain of PRAM-1.
Because cleavage of HIP-55 by caspase 3 during apoptosis dissociated its SH3 domain from its actin-binding domain (33), we determined whether HIP-55 could be cleaved in response to As 2 O 3 in NB4 cells. Western blot analysis using an anti-HIP-55 antibody revealed the decrease of full-length HIP-55 and the appearance of a cleavage product of ϳ46 kDa in NB4 cells treated with increasing amounts of As 2 O 3 ( Fig. 2A). The cleavage of HIP-55 was suppressed by the Z-DEVD-fmk caspase inhibitor, suggesting that it was mediated through a caspase 3-like enzyme(s) (Fig. 2B). In vitro cleavage of recombinant HIP-55 by caspase 3 also resulted in a 46-kDa product (Fig.  2C). Mutation of the aspartic acid at position 361 to an alanine residue in HIP-55 protein abrogated its cleavage by caspase 3 (Fig. 2C), indicating that the EHID 361 motif of HIP-55 is the caspase 3-specific cleavage site. Altogether, these results strongly suggest that caspase 3 is likely to mediate cleavage of HIP-55 in NB4 cells treated with As 2 O 3 .
Overexpression of PRAM-1 in NB4 Cells Potentiated As 2 O 3induced JNK Activation-Increased PRAM-1 mRNA and protein expression occurred 6 and 12 h, respectively, after exposure of NB4 cells to As 2 O 3 (Fig. 3A), which induced apoptosis with partial granulocytic differentiation (data not shown). Dose-response studies with As 2 O 3 -treated cells showed that 24 h of treatment with 2 M As 2 O 3 induced peak levels of the PRAM-1 protein, which correlated with the cleavage of HIP-55 (Fig. 3B). Furthermore, this also correlated with JNK kinase activation as revealed by an in vitro kinase assay (Fig. 4A). To determine whether PRAM-1 was involved in the activation of JNK by As 2 O 3 , NB4 cells were stably transfected with a vector encoding PRAM-1 or eGFP as a control and treated with 2 M As 2 O 3 (Fig. 4B). Whereas the different pools of cells transfected with either vector showed a cleavage of HIP-55 in response to As 2 O 3 , overexpression of PRAM-1 resulted in a higher activation of the JNK kinase. The fact that HIP-55 is a modulator of HPK-1 (26,29) suggested that PRAM-1 was involved in As 2 O 3induced activation of the JNK kinase in NB4 cells, likely through its interaction with the SH3 domain of HIP-55 released upon caspase 3 cleavage. To test this hypothesis, NB4 cells were pretreated with the Z-DEVD-fmk caspase inhibitor before As 2 O 3 treatment. As shown in Fig. 5, the caspase inhibitor, which suppressed the cleavage of pro-caspase 3, also suppressed the cleavage of HIP-55 and abolished PRAM-1-mediated activation of JNK (Fig. 5A). This had no effect on As 2 O 3induced apoptosis as assessed through cell viability (Fig. 5B) and PARP cleavage (Fig. 5C). These results indicated that the stimulation of As 2 O 3 -induced JNK activity by PRAM-1 was dependent upon caspase 3 activation. Altogether, our results lend further support to the idea that the SH3 domain of HIP-55 resulting from the cleavage of HIP-55 by caspase 3 is involved in PRAM-1-mediated JNK activation upon As 2 O 3 -induced differentiation of NB4 cells. DISCUSSION Retinoic acid and arsenic trioxide induce clinical remission of patients with APL with t(15; 17) translocation. Indeed, PML⅐RAR␣-bound co-repressors are released from DNA upon both RA and As 2 O 3 treatment of APL cells leading to the activation of genes repressed by PML⅐RAR␣ (12). This suggested that genes induced in common by RA and As 2 O 3 are likely to be involved in induced myeloid differentiation (42). We found that PRAM-1 is repressed by PML⅐RAR␣ and induced by RA (14) and As 2 O 3 , suggesting that PRAM-1 is likely to be one of these genes.
PRAM-1 is a novel adaptor molecule likely to be involved in an RA-signaling pathway, and we have reported the association between PRAM-1, LYN, SLP-76, and SKAP55-HOM in RAtreated NB4 cells (14). Here, we have reported that the prolinerich domain of PRAM-1 interacted with the SH3 domain of the HIP-55 adaptor. HIP-55 was originally identified as an HPK-1interacting protein (26). HPK-1 was shown to activate the JNK kinase pathway (22) and to function as a negative regulator for T-cell receptor-mediated AP-1 activation (25). Upon stimulation of the T-cell receptor, HIP-55 was shown to activate HPK-1 or other upstream kinases, which, in turn, activate the JNK kinase (29). SLP-76, SKAP55-HOM, and HIP-55 are essential components of T-cell receptor-signaling cascades regulating gene transcription, T-cell receptor and integrin clustering, endocytosis, and actin reorganization. The fact that PRAM-1 interacts with the SH3 domain of HIP-55 reinforces the view that PRAM-1 and ADAP may occupy a similar functional niche. Furthermore, in NB4 cells, overexpression of PRAM-1 potentiated As 2 O 3 -induced JNK activation.
Cleavage of HIP-55 during apoptosis induced by Fas liga- tion in Jurkat cells is mediated through a caspase 3-like enzyme (33). This resulted in the dissociation of the SH3 domain of HIP-55 from its actin-binding domain. Indeed, activation of caspase 3 was observed after As 2 O 3 treatment of NB4 cells as described previously (43). Our data showed that HIP-55 was cleaved by a caspase 3-like activity in As 2 O 3treated NB4 cells. The resulting SH3-containing fragment may interact with a number of molecules and affect their biological functions. Indeed, the SH3 domain of HIP-55 was shown to bind dynamin (31), HPK-1 (26), Fyn, and Cbl (30). Moreover, our data strongly suggest that the fragment released upon caspase 3-like cleavage interacted with PRAM-1 and HPK-1. It should also be noticed that HPK-1 cleavage by a caspase 3-like enzyme separates the N-terminal kinase domain from the C-terminal regulatory domain and enhances HPK-1 kinase activity (44). Furthermore, As 2 O 3 -mediated G 2 /M arrest and apoptosis of promonocytic U937 cells is preceded by caspase 3 activation (45). Altogether, these data emphasize the importance of proteolytic cleavage products in the regulation of cell fate and signal transduction pathways.
Caspase 3 was also shown to be involved in the phorbol ester-induced differentiation of U937 cells in the absence of cell death (46). We showed that, in NB4 cells treated with As 2 O 3 , the Z-DEVD-fmk caspase inhibitor suppressed cleavage of HIP-55 as well as PRAM-1-mediated activation of JNK without affecting cell death. Considering that in our present observation, PRAM-1 interacted with the SH3 domain of HIP-55 and potentiated As 2 O 3 -induced JNK activation together with the well established fact that HIP-55 mediated JNK activation upon T-cell receptor stimulation (26,29), it is tempting to speculate that the SH3 domain of HIP-55 resulting from the caspase 3 cleavage of HIP-55 is involved in PRAM-1-mediated JNK activation. This will have to be further investigated.
The fact that PRAM-1 is an adaptor molecule involved in both RA-and As 2 O 3 -signaling pathways suggests that PRAM-1 may be central to molecular complexes important for maturation of APL cells. The formation of such a complex may reactivate a normal signaling pathway that might be interrupted in leukemia cells.