The Role of DOC-2/DAB2 Protein Phosphorylation in the Inhibition of AP-1 Activity

DOC-2/DAB2, a novel phosphoprotein with signal-transducing capability, inhibits human prostatic cancer cells (Tseng, C.-P., Ely, B. D., Li, Y., Pong, R.-C., and Hsieh, J.-T. (1998) Endocrinology 139, 3542–3553). However, its mechanism of action is not understood completely. This study delineates the functional significance of DOC-2/DAB2 protein phosphorylation and demonstrates that in vivo activation of protein kinase C (PKC) by 12-O-tetradecanoylphorbol-13-acetate (TPA) induces DOC-2/DAB2 phosphorylation, including a serine residue at position 24. Mutation of Ser24 to Ala reduced DOC-2/DAB2 phosphorylation by PKC. Using a synthetic Ser24 peptide (APS24KKEKKKGSEKTD) or recombinant DOC-2/DAB2 as substrates, PKCβII, PKCγ, and PKCδ (but not casein kinase II) directly phosphorylated Ser24 in vitro. This indicates that DOC-2/DAB2 is a PKC-specific substrate. Since expression of wild-type DOC-2/DAB2, but not the S24A mutant, inhibited TPA-induced AP-1 activity in prostatic epithelial cells, phosphorylation of Ser24 appears to play a critical role in modulating TPA-induced AP-1 activity. Taken together, these data suggest that PKC-regulated phosphorylation of DOC-2/DAB2 protein may help its growth inhibitory function.

DOC-2 (deletion in ovarian carcinoma 2), with a molecular mass of 82 kDa, belongs to the Disabled (Dab) gene family. Recently, we demonstrated that DOC-2/DAB2 is up-regulated in degenerated rat ventral prostate induced by androgen deprivation (1). Histologically, elevated levels of DOC-2/DAB2 are associated with an enriched basal cell compartment, a progenitor cell for glandular epithelium. This suggests that DOC-2/ DAB2 may be involved in the homeostasis of prostate regeneration (1). In addition, stable expression of DOC-2/DAB2 in a prostatic cancer cell line significantly reduces its in vitro growth rate concomitant with an increase in G 1 cell fractions. It also decreases anchorage-independent growth on soft agar (1). Other types of cancer cell lines, such as the SKOV3 ovarian cancer cell line (2) and the choriocarcinoma cell lines Jar, JEG, and BeWo (3), demonstrate similar results. Therefore, DOC-2/ DAB2 appears to be a potent negative regulator of carcinoma cell growth.
The primary structure of DOC-2/DAB2 reveals that DOC-2/ DAB2 is a putative signaling molecule with protein-protein interaction and protein phosphorylation as two possible mech-anisms modulating its activity (1). The N-terminal phosphotyrosine-interacting domain shares significant homology with mouse DAB1 (4). Disruption of the mouse Dab1 gene disturbs neuronal layering in the cerebral cortex, hippocampus, and cerebellum (5). The similar phenotypes of mouse Dab1 null mice, Reeler, Scrambler, and Yotari, indicate that mouse DAB1 functions as a signaling molecule regulating cell positioning in the developing brain (6). Indeed, the phosphotyrosine-interacting domain of mouse DAB1 is found to bind SRC proteintyrosine kinase, and therefore, mouse DAB1 may play a key role in key signal transduction pathways involved in the formation of neural networks (4). In addition to the phosphotyrosine-interacting domain, DOC-2/DAB2 also contains a Cterminal proline-rich domain that interacts with the SH3 domain of GRB2 and reduces the binding between GRB2 and SOS. This suggests that DOC-2/DAB2 may modulate growth factor/Ras pathways (7).
DOC-2/DAB2 contains several consensus protein kinase C (PKC), 1 casein kinase II (CKII), and cAMP-dependent protein kinase phosphorylation sites (1), implying that protein phosphorylation may modulate DOC-2/DAB2 activity. Colony-stimulating factor-1 induces DOC-2/DAB2 protein phosphorylation in a mouse macrophage cell line (8). We also found that the phosphorylation of DOC-2/DAB2 is modulated by receptor protein-tyrosine kinase pathways, such as the epidermal growth factor, 2 and the PKC activation pathway, such as stimulation by 12-O-tetradecanoylphorbol-13-acetate (TPA) as described in this study. PKC is the receptor of TPA and is a member of the serine/threonine kinase family composed of at least 12 isoforms (9). Structural and functional studies indicate that PKC isoforms are likely to have distinct and distinguishable functions (10 -15).
One of the properties ascribed to activation of PKC by TPA is the ability to alter gene expression. Among genes transcriptionally induced by TPA are ornithine decarboxylase, collagenase, and stromelysin. The promoter regions of several TPA-inducible genes share a conserved TPA-responsive element recognized by the transcription factor AP-1 (16). This study investigated the impact of DOC-2/DAB2 protein phosphorylation on the PKC-mediated signal transduction cascade. We have mapped the TPA-induced DOC-2/DAB2 protein phosphorylation site to Ser 24 , which appears to modulate the DOC-2/DAB2 inhibition of AP-1 transcription activity. Results indicate that phosphorylation of Ser 24 is mediated by PKC␤II, PKC␥, and PKC␦, but not CKII. This suggests that the PKC phosphorylation of Ser 24 in DOC-2/DAB2 may be an underlying mechanisms for its tumor-suppressive function.
Cell Cultures-COS, NbE, and C4-2 cells were maintained in T medium supplemented with 5% fetal bovine serum as described previously (1).
Construction of Plasmids-The SalI-NotI fragments of pCI-neo-p82 and pCI-neo-p59 (1) were subcloned into pET-21b(ϩ); and the XbaI-NotI fragments of pET-21b(ϩ)-p82 and pET-21b(ϩ)-p59, containing p82 and p59 cDNA, respectively, were subcloned back into the pCI-neo vector to make the T7-p82 and T7-p59 expression plasmids. The ⌬B mutant was created by deleting a 1.2-kilobase Bsu36I fragment from pCI-T7-p82. The ⌬N mutant was generated by four-fragment ligation with appropriate restriction enzymes, resulting in the deletion of the first 636 base pairs of p82. Site-directed mutagenesis with polymerase chain reaction was used to create ⌬B-S24A, ⌬B-S32A, ⌬B-S24A,S32A, and ⌬B-S241A,S249A. The BsaBI fragments of both p82 and p59 were subcloned into BsaBI-digested ⌬B-S24A to create a single amino acid mutant of both p82 and p59.
In Vitro Protein Kinase Assay-The immunocomplex PKC assay was performed as described (21) in the presence or absence of the activators. The CKII protein kinase assay was performed as described by the manufacturer (Promega, Madison, WI) with equal molar concentrations of the indicated substrates.
Immunoprecipitation Assay-The recombinant GRB2-GST fusion protein was induced by isopropyl-␤-D-thiogalactopyranoside for 4 -6 h. The bacteria were spun down, and lysed in phosphate-buffered saline supplemented with 1% Triton X-100 and 1 mM phenylmethylsulfonyl fluoride. The lysate was sonicated, and then all insoluble material was spun down. The GRB2-GST protein was immobilized on glutathione-Sepharose (Amersham Pharmacia Biotech) according to the manufacturer's instruction. The supernatant was incubated with glutathione-Sepharose for 30 min, washed three times, and resuspended as a 50% slurry in phosphate-buffered saline supplemented with 1% Triton X-100.
Wild-type DOC-2/DAB2 protein and its mutant were expressed in COS cells 24 h after DNA transfection. Cells were collected in 0.5 ml of phosphate-buffered saline supplemented with 1% Triton X-100 and a mixture of protease inhibitors. After a low speed centrifugation, 0.4 ml of the supernatant was incubated overnight at 4°C with 60 l of GRB2-GSTglutathione-Sepharose or GST-glutathione-Sepharose alone. After centrifugation, the pellet was washed twice and dissolved in the sample buffer. After electrophoresis, the gel was transferred to a filter and probed with antiserum against the T7 tag.

RESULTS
TPA Induces DOC-2/DAB2 Protein Phosphorylation at Serine 24 -DOC-2/DAB2 appears to be a phosphoprotein (8). However, the nature of the phosphorylation and the upstream protein kinase that mediates DOC-2/DAB2 phosphorylation have not yet been elucidated. To study DOC-2/DAB2 protein phosphorylation, NbE cells were treated with the PKC activator TPA (100 ng/ml). Both DOC-2/DAB2 (i.e. p82) and its splicing variant, p59, showed a mobility shift on SDS-polyacrylamide gel electrophoresis (PAGE) within 5 min, and this was sustained for at least 30 min after TPA treatment (Fig. 1A). Trans-fection of the T7-tagged p82 or p59 expression plasmid into COS or NbE cells (data not shown) followed by TPA treatment also resulted in a similar mobility shift of exogenously expressed DOC-2/DAB2. This finding suggests that a protein modification occurred in DOC-2/DAB2 after TPA treatment.
To determine whether the mobility shift of protein on SDS-PAGE is caused by changes in protein phosphorylation, we labeled T7-p59-transfected cells with [ 32 P]orthophosphate, and cell extracts were immunoprecipitated with the anti-T7 tag antibody 30 min after TPA treatment. A basal level of p59 phosphorylation occurred in control cells, whereas TPA treatment resulted in an increase in p59 phosphorylation (Fig. 1B, left panel). The majority of the phosphorylation occurred at the serine residue as determined by phosphoamino acid analysis (Fig. 1B, right panel). A series of DOC-2/DAB2 deletion mutants was constructed to map the phosphorylation site (Fig. 1C,  upper panel). Although the basal level of protein phosphorylation was still detected in the N-terminal deletion mutant (⌬N), TPA-induced protein phosphorylation was abolished (Fig. 1C, lower panel). In contrast, TPA induced the phosphorylation of a p82 mutant (⌬B) with a deletion of the majority of the C terminus (Fig. 1C, lower panel). To avoid possible transfection artifacts, we demonstrated similar levels of protein expression by each transfection as determined by Western blotting (Fig.  1C, lower panel). The higher extra band seen in ⌬B experiments by Western blotting is immunoglobulin protein, which has a similar molecular mass compared with ⌬B. These results suggest that the N terminus of DOC-2/DAB2 may harbor key phosphoserine residue(s) responsive to TPA.
We have further used ⌬B to map the TPA-induced protein phosphorylation site with CNBr phosphopeptide mapping ( Fig.  2A). TPA treatment enhanced the phosphorylation of at least two additional peptides, with estimated molecular masses of 13.6/13.2 and 7.3 kDa, respectively, compared with the ethanol control (Fig. 2B). The protein sequence of DOC-2/DAB2 revealed four consensus PKC/CKII phosphorylation sites (i.e. SKKE) at serines 24, 32, 241, and 249. The double mutation of Ser 241 and Ser 249 (⌬B-S241A,S249A) did not affect TPA-induced ⌬B phosphorylation (Fig. 2B). However, the double mutation of Ser 24 and Ser 32 (⌬B-S24A,S32A) completely abolished the protein phosphorylation of ⌬B by TPA, indicating that phosphorylation of either Ser 24 or Ser 32 is induced by TPA (Fig.  2B). Therefore, a single mutation of Ser 24 or Ser 32 was used to precisely map the TPA-induced protein phosphorylation site in DOC-2/DAB2. As shown in Fig. 2C, TPA-induced DOC-2/DAB2 protein phosphorylation was detected only in ⌬B and ⌬B-S32A, but not ⌬B-S24A. In p82-S24A and p59-S24A, the 7.3-kDa phosphopeptide induced by TPA was also eliminated (Fig. 2D). Thus, Ser 24 is one of the TPA-induced protein phosphorylation sites in DOC-2/DAB2.
PKC Is a DOC-2/DAB2 Protein Kinase-Since S 24 KK appears to fit the consensus phosphorylation site for either PKC or CKII, we tested whether PKC␤II, PKC␥, PKC␦, and CKII could use the synthetic Ser 24 peptide (APS 24 KKEKKKG-SEKTD) of DOC-2/DAB2 as an in vitro substrate. In the presence of PKC activators, all PKC isoforms tested demonstrated comparable PKC activity toward the Ser 24 peptide (Fig. 3A). In contrast, CKII failed to phosphorylate the Ser 24 peptide, whereas myelin basic protein was phosphorylated significantly (Fig. 3B). The phosphorylation of Ser 24 by PKC was abolished when an Ala 24 peptide (APA 24 KKEKKKGSEKTD) of DOC-2/ DAB2 was used (Fig. 3A), indicating that the Ser 24 peptide is a specific substrate for PKC.
To determine whether PKC directly phosphorylates Ser 24 of DOC-2/DAB2 in vitro, the recombinant proteins His-T7-⌬B and His-T7-⌬B-S24A were generated (Fig. 3C). As shown in Fig.  3D, His-T7-⌬B was phosphorylated by PKC␤II, PKC␥, and PKC␦, whereas His-T7-⌬B-S24A phosphorylation was substantially reduced in vitro. The CNBr phosphopeptide map revealed that the 7.3-kDa phosphopeptide was the key substrate for PKC␤II, PKC␥, or PKC␦ (Fig. 3E) and was consistent with the in vivo data shown in Fig. 3C. Clearly, protein phosphorylation of DOC-2/DAB2 can be controlled by PKC activity.
Phosphorylation of Ser 24 Is Essential for the Inhibitory Effect of DOC-2/DAB2 on TPA-induced AP-1 Activity-To further delineate the possible impact of Ser 24 phosphorylation on DOC-2/DAB2 function, we examined the effect of DOC-2/DAB2 on the PKC-mediated signaling pathway. Since TPA enhances the binding of the AP-1 transcription factor to a specific cis-element of many growth-related genes and increases gene transcription, we cotransfected the indicated DOC-2/DAB2 expression plasmids with a luciferase reporter gene construct (i.e. Ϫ73/ ϩ63-Col-luc) containing a TPA-responsive element derived from the collagenase gene promoter into a human prostate cancer cell line (C4-2). We found that p82 elicited a dose-dependent inhibition of TPA-induced AP-1 activity (Fig. 4A). Similarly, expression of p59 and ⌬B, but not ⌬N, in C4-2 cells demonstrated the inhibitory effect on TPA-induced AP-1 activity, indicating that the N terminus of DOC-2/DAB2 is critical in regulating AP-1 activity (Fig. 4B). We also observed a significant decrease in the AP-1 activity of the p59-transfected C4-2 sublines (1) with TPA compared with that of wild-type or control plasmid-transfected C4-2 cells (Table I). Such a decrease suggests that DOC-2/DAB2 protein can function as a potent negative regulator to modulate TPA-induced AP-1 activity.
To further determine the significance of Ser 24 phosphorylation of DOC-2/DAB2, we cotransfected a series of mutants with the reporter gene construct into C4-2 cells. As shown in Fig. 4C, expression of ⌬B-S24A, p82-S24A, and p59-S24A showed greater luciferase activity than p82. Thus, Ser 24 in DOC-2/ DAB2 is critical for maintaining its inhibitory effect on TPAinduced AP-1 activity. Moreover, ⌬B-S24A and p82-S24A neutralized the inhibitory activity of ⌬B and p82 on AP-1 activity in a dose-dependent manner (Fig. 5). In p59-expressing C4-2 cells (Table I), increased expression of p59-S24A also decreased the inhibitory effect of p59 on AP-1 activity. These results clearly indicate that Ser 24 phosphorylation of DOC-2/DAB2 plays a critical role in modulating AP-1 inhibitory activity.
N-terminal DOC-2/DAB2 Does Not Interact with GRB2 Protein-Previously, DOC-2/DAB2 was demonstrated to interact with GRB2 (7), which may be a potential underlying mechanism for its AP-1 inhibitory function. Therefore, we tested whether ⌬B or ⌬N interacts with GRB2. As expected, in the presence of GRB2, p82, p59, and ⌬N could be precipitated (Fig.   6, A and B). In contrast, in the presence of ethanol (Fig. 6A) or TPA (Fig. 6B), ⌬B failed to be precipitated by GRB2 protein. This indicates that the N-terminal DOC-2/DAB2 protein does not interact with GRB2 (Fig. 6, A and B). This interaction appears to be specific because GST-glutathione-Sepharose alone did not precipitate any DOC-2/DAB2 protein (Fig. 6C). Taken together, these data suggest that the mechanism for the AP-1 inhibitory effect of DOC-2/DAB2 may be mediated by GRB2-independent pathway(s).

DISCUSSION
Although DOC-2/DAB2 is believed to be a candidate tumor suppressor in several cancer cell lines (1-3), the mechanism of its inhibition of cell growth is unknown. In this work, we describe the modulation of DOC-2/DAB2 protein phosphorylation by the PKC activator TPA and the critical role of the phosphorylation in determining its function. We demonstrate that DOC-2/DAB2 serves as a negative regulator in the PKCmediated signaling axis and may inhibit cell growth through inhibiting AP-1 activity that has been associated with cell proliferation and tumorigenicity (23). In addition to our studies, Xu et al. (7) have shown that the C terminus of DOC-2/ DAB2 prevents GRB2 from binding to SOS. Although, the functional impact of this interaction is unclear, it suggests that DOC-2/DAB2 may interfere with the signal transduction pathway activated by receptor protein-tyrosine kinase.
Several mitogenic stimuli, including colony-stimulating factor-1, TPA, phosphatidylcholine-specific phospholipase C, and sphingomyelinase, are known to induce phosphorylation of DOC-2/DAB2. However, unlike the DAB1 protein, which contains a phosphotyrosine moiety, our data and others (8) demonstrate that the major protein phosphorylation site of DOC-2/DAB2 is the serine residue. We mapped a key PKCmodulated DOC-2/DAB2 protein phosphorylation site to Ser 24 , a conserved residue of DOC-2/DAB2 among human, mouse, and rat species (1). The impact of Ser 24 phosphorylation on the function of DOC-2/DAB2 is very significant because mutation of Ser 24 to Ala abolishes the inhibitory effect of DOC-2/DAB2 on TPA-induced AP-1 activity (Fig. 4). Interestingly., we observed that the Ser 24 mutants of DOC-2/DAB2 can compete with their wild-type proteins in a dose-dependent manner ( Fig.  5 and Table I), indicating that Ser 24 in DOC-2/DAB2 is a critical amino acid motif modulating its activity. Furthermore, TPA-induced AP-1 activity is substantially inhibited in the p59-expressing C4-2 sublines (i.e. p59-18 and p59-23) (Table I). This may be one of the underlying mechanism(s) contributing to the slower growth of both p59-18 and p59-23 in vitro (1). Similarly, by transfecting the AP-1 reporter construct into NbE cells with detectable endogenous DOC-2/DAB2 levels (1), we found that TPA failed to increase AP-1 activity in this cell line (data not shown). Taken together, these data indicate that DOC-2/DAB2 represents a novel factor in the modulation of the PKC-elicited signaling pathway.
We still do not know how DOC-2/DAB2 inhibits AP-1 activity in the nucleus, although Xu et al. (7) demonstrated that DOC-2/DAB2 can bind to GRB2, which may account for the inhibition of the GRB2-mediated signaling pathway. However, our results indicate that N-terminal DOC-2/DAB2 (i.e. ⌬B), with or without phosphorylation, does not interact with GRB2 (Fig. 6). Therefore, this AP-1 inhibitory effect must be mediated by GRB2-independent pathway(s). Alternatively, DOC-2/DAB2 may inhibit the function of signal molecules in PKC-elicited pathway(s). However, we observed that Ser 24 and Ser 24 /Ser 32 mutants appear to affect the phosphorylation status of other serine residues as indicated by CNBr phosphopeptide mapping (Fig. 2), suggesting a critical role for Ser 24  DAB2. Moreover, analysis of the amino acid sequence surrounding Ser 24 revealed a potential nuclear localization signal (S 24 KKEKKKG). The sequence KKEK has been shown to associate with the actin-interacting capability of villin (23), suggesting that the subcellular localization of DOC-2/DAB2 may be affected by protein phosphorylation. Interestingly, we observed that TPA induces accumulation of DOC-2/DAB2 in the particulate fraction (i.e. membrane fraction) of prostatic epi-thelial cells. 2 The accumulated DOC-2/DAB2 protein in the particulate fraction appears to be phosphorylated as determined by the mobility shift on SDS-PAGE. This observation is in accord with reports suggesting that phosphorylation within the sequence immediately N-terminal to the minimal nuclear localization signal modulates the transport kinetics of the respective protein (24,25). The effect of Ser 24 phosphorylation on DOC-2/DAB2 protein localization, which may explain how DOC-2/DAB2 regulates AP-1 activity, warrants further investigation.
Another important issue is the mechanism controlling the phosphorylation of Ser 24 . Among the protein kinases we tested (PKC␤II, PKC␥, PKC␦, and CKII), PKC preferentially phosphorylates DOC-2/DAB2 both in vivo and in vitro. PKC has been implicated in many aspects of cellular functions, including cell proliferation and differentiation, T-cell activation, and gene activation (10 -15). Although functional redundancies between PKC isoforms have been shown, other reports suggest that individual PKCs may have distinct functions (10 -15, 26 - The plasmid control (pCI-neo; 0.24 g) was used in these experiments. b The plasmid control (pCI-neo; 2.4 g) and p59-S24A (0.24 g) were used in this experiment. c The data were calculated from three independent transfections; the numbers represent the mean Ϯ S.D. condition. An aliquot of recombinant GRB2-GST-glutathione-Sepharose (ϳ200 g; A and B) or GST-glutathione-Sepharose (C) was incubated with 400 l of cell lysates. After overnight incubation and washing, bound proteins were subjected to SDS-PAGE analyses and blotted with either the anti-T7 tag antibody for DOC-2/GRB2 (upper panels) or the anti-GST antibody for GRB2 (middle panels). An aliquot of each cell lysate (ϳ30 g) was also subjected to SDS-PAGE analysis and blotted with the anti-T7 tag antibody (lower panels) as an internal control. IP, immunoprecipitation; WB, Western blot. 28) and substrate specificity (13, 29 -30). Similarly, as shown in Fig. 3E, we observed a slightly different phosphorylation profile for PKC␦. The largest phosphopeptide is not phosphorylated by PKC␦. This suggests the distinct activity of PKC isoforms toward the phosphorylation of ⌬B. In addition, there was an increased expression of both DOC-2/DAB2 and PKC␦ mRNAs in the enriched basal epithelial cells of degenerated prostate in castrated rat (1), 2 suggesting the involvement of PKC␦ in the control of DOC-2/DAB2 activity in prostate gland homeostasis. Kinetic studies with purified recombinant PKC will be required to further elucidate whether Ser 24 of DOC-2/ DAB2 is a preferential substrate for any PKC isoform.
In summary, our results indicate that DOC-2/DAB2 is phosphorylated at Ser 24 by PKC, which plays a critical role in controlling AP-1 activity. The regulation of DOC-2/DAB2 protein phosphorylation may in turn regulate its inhibitory function in cell proliferation and tumorigenicity. It may define a novel mechanism for the transduction of negative growth signals from the membrane to the nucleus.