Coordinated Activation of the Nuclear Ubiquitin Ligase Cul3-SPOP by the Generation of Phosphatidylinositol 5-Phosphate*

Phosphoinositide signaling pathways regulate numerous processes in eukaryotic cells, including migration, proliferation, and survival. The regulatory lipid phosphatidylinositol 4,5-bisphosphate is synthesized by two distinct classes of phosphatidylinositol phosphate kinases (PIPKs), the type I and II PIPKs. Although numerous physiological functions have been identified for type I PIPKs, little is known about the functions and regulation of type II PIPK. Using a yeast two-hybrid screen, we identified an interaction between the type IIβ PIPK isoform (PIPKIIβ) and SPOP (speckle-type POZ domain protein), a nuclear speckle-associated protein that recruits substrates to Cul3-based ubiquitin ligases. PIPKIIβ and SPOP interact and co-localize at nuclear speckles in mammalian cells, and SPOP mediates the ubiquitylation of PIPKIIβ by Cul3-based ubiquitin ligases. Additionally, stimulation of the p38 MAPK pathway enhances the ubiquitin ligase activity of Cul3-SPOP toward multiple substrate proteins. Finally, a kinase-dead PIPKIIβ mutant enhanced ubiquitylation of Cul3-SPOP substrates. The kinase-dead PIPKIIβ mutant increases the cellular content of its substrate lipid phosphatidylinositol 5-phosphate (PI5P), suggesting that PI5P may stimulate Cul3-SPOP activity through a p38-dependent signaling pathway. Expression of phosphatidylinositol-4,5-bisphosphate 4-phosphatases that generate PI5P dramatically stimulated Cul3-SPOP activity and was blocked by the p38 inhibitor SB203580. Taken together, these data define a novel mechanism whereby the phosphoinositide PI5P leads to stimulation of Cul3-SPOP ubiquitin ligase activity and also implicate PIPKIIβ as a key regulator of this signaling pathway through its association with the Cul3-SPOP complex.

Phosphoinositide signaling pathways modulate a diverse array of cellular processes in eukaryotes. Modification of the inositol ring by lipid kinases and phosphatases produces distinct phosphatidylinositol phosphate (PIP) 2 isomers. These phosphatidylinositol phosphate isomers in turn selectively modulate the activities of effector proteins. In the cytosol, phosphoinositides regulate numerous processes, including actin polymerization, focal adhesion dynamics, ion channel activity, growth factor receptor signaling, and vesicle trafficking (1)(2)(3)(4). In the nucleus, an autonomous phosphoinositide cycle regulates processes, including differentiation, proliferation, cell cycle progression, and apoptosis (5).
The ubiquitylation and degradation of proteins is essential both for constitutive protein turnover as well as for the modulation of signaling pathways in response to extracellular stimuli. Ubiquitylation involves a three-step process in which ubiquitin is primed by an activating enzyme (E1), transferred to a conjugating enzyme (E2), and finally attached to a designated protein substrate by a ubiquitin ligase (E3). Specificity is mediated primarily by the multiprotein E3 ubiquitin ligases, which recruit unique substrates through a number of modular specificity fac-tors. One of the best known E3 ubiquitin ligases in eukaryotes is the SCF (Skp1/Cul1/F-box) complex, a modular ubiquitin ligase assembled on the cullin protein Cul1. In addition to Cul1, six other cullins (Cul2, -3, -4A, -4B, 5, and -7) have been identified in humans. Cul3-based ubiquitin ligases are an emerging member of this family (13)(14)(15)(16). Substrate specificity of Cul3based ubiquitin ligases is dictated by BTB (Broad complex/ Tramtrack/bric-a-brac) domain-containing proteins that bind directly to Cul3 through their BTB domain and bind substrates through a second protein-protein interaction domain (13,14,16,17). Orthologs of Cul3 and BTB proteins have been identified in eukaryotes ranging from Caenorhabditis elegans to humans, and several substrates of Cul3-based ligases have been identified (18 -20).
A central theme in the regulation of phosphoinositide signaling pathways is the interaction of enzymes such as PIPKs with upstream regulators and downstream effectors at discrete subcellular sites. To better understand the function and regulation of nuclear phosphoinositide signaling pathways, we sought to identify proteins that interact with PIPKII␤. Yeast two-hybrid screening identified an interaction between PIPKII␤ and speckle-type POZ domain protein (SPOP), a nuclear speckle-associated BTB domain protein, and substrate adaptor for Cul3based ubiquitin ligases (13,14,16,17,19). We demonstrate that PIPKII␤ and SPOP interact in vitro and in vivo and co-localize at nuclear speckles in HeLa cells. We also demonstrate that Cul3-SPOP mediates the ubiquitylation of PIPKII␤ in vivo, and that the ubiquitylation of multiple Cul3-SPOP substrates is potently stimulated by the MKK6-p38 MAPK pathway. Finally, we demonstrate that PI5P, the product of the PI-4,5-P 2 4-phosphatases and the lipid substrate of PIPKII␤, stimulates Cul3-SPOP activity, and this was blocked by the p38 inhibitor SB203580. Taken together, our data support a novel signaling pathway in which PI5P and PIPKII␤ regulate Cul3-SPOP ubiquitin ligase activity through p38 MAPK.

EXPERIMENTAL PROCEDURES
Cell Culture and Antibodies-HEK293 and HeLa cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum at 37°C under a humidified atmosphere with 5% CO 2 . MG132 (Boston Biochem, Boston), SB203580, and SP600125 (EMD Bioscience, San Diego) were dissolved in Me 2 SO and added to the media where indicated. ␣-GST and ␣-T7 antibodies were purchased from EMD Bioscience. ␣-Myc and ␣-hemagglutinin antibodies were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). ␣-FLAG antibody was purchased from Sigma.
cDNA Synthesis and Cloning-Total cellular RNA for cDNA synthesis was prepared from HEK293 cells using the RNeasy mini kit (Qiagen) following the manufacturer's instructions. cDNAs for SPOP, Cul3, and MKK6 were generated from total cellular RNA using the Qiagen One-step reverse transcription-PCR kit. Sequenced cDNAs were cloned into bacterial and mammalian expression vectors as described in the text. cDNA expression constructs for the mammalian PI-4,5-P 2 4-phosphatases were a generous gift from Dr. Philip Majerus (Washington University, St. Louis).
Yeast Two-hybrid Screen-The PIPKII␤ cDNA was submitted to the University of Wisconsin Molecular Interaction Facility for two-hybrid screening. The kinase insert domain of PIP-KII␤ (Asp 287 -Met 365 ) was cloned into a GAL4 fusion vector and screened against human five cDNA libraries according to Molecular Interaction Facility protocols.
Expression and Purification Recombinant Proteins-For expression of recombinant proteins, bacterial expression constructs were transformed into BL21(DE3)-competent cells (Novagen). Liquid cultures were inoculated with a single colony, and recombinant protein expression was induced by the addition of 1 mM isopropyl 1-thio-␤-D-galactopyranoside. Following expression, bacteria were lysed by sonication in PBS containing 0.5% Triton X-100 and Complete mini protease inhibitor mixture (Roche Applied Science). His 6 and GST fusion proteins were purified on HisTrap HP and GSTrap FF affinity columns, respectively (Amersham Biosciences), according to the manufacturer's instructions.
GST Pulldown and Co-immunoprecipitation Assays-For GST pulldown assays, equimolar amounts of purified recombinant protein were incubated in 50 mM Tris, 150 mM NaCl, 0.5% Triton X-100, pH 8 in the presence of glutathione-Sepharose resin (Amersham Biosciences). After 4 h at 4°C, the resin was pelleted at low speed and washed three times in reaction buffer. Bound protein was eluted from the resin by the addition of 2ϫ Laemmli sample buffer, resolved by SDS-PAGE, and detected by Western blot. For co-immunoprecipitation assays, HEK293 cells were transfected via calcium phosphate precipitation by standard methods. 24 h after transfection, cells were washed in PBS and lysed by low amplitude sonication in 50 mM Tris, 150 mM NaCl, 0.5% Triton X-100, 0.1% deoxycholate, 0.5 mM EDTA, pH 8. Cell lysates were centrifuged for 15 min at 20,000 ϫ g, and supernatant was recovered. Immunoprecipitations were performed by adding 3 g of antibody and 20 l of protein G-Sepharose resin per ml of cell lysate. After incubation at 25°C for 3 h, the resin was pelleted and washed three times in lysis buffer. Bound protein was eluted in 2ϫ Laemmli sample buffer, resolved by SDS-PAGE, and detected by Western blot.
Immunofluorescence-HeLa cells were seeded on glass coverslips and transfected using FuGENE 6 transfection reagent (Roche Applied Science). Twenty four hours after transfection, cells were washed in cold PBS and fixed in methanol. Primary antibodies were diluted to 1 g/ml in PBS ϩ 0.1% Triton X-100 ϩ 3% bovine serum albumin and incubated on the coverslips at 37°C. Coverslips were washed three times with PBST and incubated with fluorophore-conjugated secondary antibodies for 1 h at 37°C. After being washed with PBST, coverslips were mounted onto microscope slides using VectaShield mounting medium (Vector Laboratories). Visualization was performed using a Bio-Rad MRC-1024 laser scanning confocal microscope (W. M. Keck Laboratory for Biological Imaging, Madison, WI).
In Vivo Ubiquitylation Assays-HEK293 cells were transiently transfected with the indicated combinations of expression vectors by calcium phosphate precipitation. 24 h after transfection, MG132 was added to the media at a final concentration of 10 M, and cells were incubated for 6 h at 37°C. Cells were washed twice with PBS, lysed in PBS with 8 M urea and 0.2% SDS, and sonicated to reduce viscosity. Lysates were incu-bated for 3 h at 25°C with nickel-Sepharose resin (Amersham Biosciences). After washing the resin three times with lysis buffer, proteins were eluted in sample buffer, resolved via SDS-PAGE, and detected by Western blot.

RESULTS
The Kinase Insert Domain of PIPKII␤ Is Necessary and Sufficient for Nuclear Targeting-We have demonstrated previously the presence of the type I␣ and type II␤ PIPK isoforms within mammalian nuclei, including their targeting to nuclear speckles (10). To identify the domain within PIPKII␤ required for its nuclear targeting, a panel of deletion mutants was generated, and their subcellular distributions were assessed by indirect immunofluorescence microscopy. The kinase insert domain of PIPKII␤ was required for nuclear localization, as its deletion prevented nuclear targeting of PIPKII␤ (Fig. 1A). These results are consistent with previous reports that an acidic ␣-helix within the kinase insert domain promotes nuclear targeting of PIPKII␤ (36,37). To determine whether the PIPKII␤ kinase insert domain is sufficient for PIPKII␤ nuclear targeting, the kinase insert was fused to LacZ, and its subcellular localization was examined. Whereas a LacZ control was localized entirely within the cytosol, the kinase insert fusion (Ins-LacZ) targeted to the nucleus (Fig. 1B). These results identified the kinase insert domain of PIPKII␤ as being both necessary and sufficient for its nuclear targeting.
Identification of SPOP as a PIPKII␤-binding Protein-Although the PIPKII␤ kinase insert mediates nuclear translocation, it does not contain a canonical nuclear localization signal. We therefore hypothesized that PIPKII␤ is targeted to the nucleus through its association with interacting partners. To identify proteins that interact with PIPKII␤, we performed a yeast two-hybrid screen against several human cDNA libraries using the PIPKII␤ kinase insert domain as bait ( Fig. 2A). One of the proteins identified by the two-hybrid screen was SPOP (speckle-type POZ domain protein) ( Fig. 2A), a nuclear speckle-associated protein and a substrate specificity factor for Cul3based ubiquitin ligases (18 -20).

SPOP Interacts Specifically with PIPKII␤ in Vitro and in Vivo-
The interaction between PIPKII␤ and SPOP was assessed using both in vitro and in vivo binding assays. To assess their in vitro association, GST pulldown assays were performed with recom-binant purified GST-SPOP and His 6 -PIPKII␤. PIPKII␤ was specifically retained by GST-SPOP but not by GST alone (Fig.  2B), confirming their interaction. SPOP contains two conserved domains: an N-terminal MATH (Meprin and Traf Homology) domain and a C-terminal BTB domain. To assess the contribution of each domain to the interaction between SPOP and PIPKII␤, GST pulldown assays were performed with purified recombinant GST-MATH and GST-BTB proteins. PIPKII␤ co-precipitated with GST-MATH but not GST-BTB (Fig. 2B), demonstrating that SPOP interacts with PIPKII␤ through its MATH domain. To assess the specificity of SPOP for PIPKII␤, the highly similar PIPKII␣ isoform was also tested. Despite ϳ80% primary sequence identity between the two PIPKII isoforms, SPOP did not interact with PIPKII␣ (Fig. 2C).
The interaction between PIPKII␤ and SPOP was also tested in vivo. HEK293 cells were transiently transfected with mammalian expression constructs of PIPKII␤ and SPOP. Twenty four hours after transfection, PIPKII␤ and SPOP were immunoprecipitated from cell lysates, resolved by SDS-PAGE, and probed by Western blot. PIPKII␤ and SPOP co-precipitated with each other, confirming their in vivo interaction (Fig. 3). Similar to in vitro results, PIPKII␣ did not co-precipitate with SPOP. The interaction of endogenous PIPKII␤ and SPOP was also assessed. Endogenous PIPKII␤ was immunoprecipitated from HEK293 lysates using either an ␣-PIPKII␤ antibody or normal rabbit IgG. Western blot analysis identified endogenous SPOP in the PIPKII␤ immunoprecipitates, thus confirming the interaction of endogenous PIP-KII␤ and SPOP proteins (Fig. 3C).  PIPKII␤ Co-localizes with SPOP at Nuclear Speckles-Having confirmed their interaction, we next assessed whether PIP-KII␤ and SPOP co-localize within the nucleus. HeLa cells were transiently transfected with SPOP and PIPKII␤, and their subcellular distributions were analyzed by indirect immunofluorescence microscopy. A distinct pool of PIPKII␤ co-localized with SPOP at nuclear speckles (Fig. 4). In contrast, PIPKII␣ was present exclusively within the cytosol.
PIPKII␤ and PIPKII␣ share ϳ80% primary sequence identity, with the greatest regions of diversity at the N and C termini and within the kinase insert domain. The specific interaction and co-localization of SPOP with PIPKII␤ therefore suggested that the PIPKII␤ kinase insert is necessary for its co-localization with SPOP at nuclear speckles. To test this hypothesis, chimeric PIPKII mutants were generated in which the kinase insert domains of PIPKII␣ and PIPKII␤ were exchanged, and their subcellular distributions were analyzed. PIPKII␣(II␤ KI), which contains the PIPKII␤ kinase insert, co-localized with SPOP at nuclear speckles similar to wild type PIPKII␤. Conversely, PIPKII␤(II␣ KI), which contains the PIPKII␣ kinase insert, showed a cytosolic distribution similar to wild type PIPKII␣. These results demonstrate that the PIPKII␤ kinase insert domain mediates its co-localization with SPOP at nuclear speckles.
SPOP Promotes the Ubiquitylation of PIPKII␤ by Cul3-based Ubiquitin Ligases-The Cul3-SPOP ubiquitin ligase complex has been shown to mediate the ubiquitylation and subsequent degradation of Daxx (19). We therefore assessed whether Cul3-SPOP also promotes the ubiquitylation and degradation of PIP-KII␤. To assess PIPKII␤ turnover, HEK293 cells were transiently transfected with Cul3, SPOP, and the RING-box protein Rbx1, which is required for the activation of cullin-based ligases (21). Endogenous PIPKII␤ levels were assessed by Western blot either 24 or 48 h after transfection. Interestingly, no discernible change in endogenous PIPKII␤ protein levels was detected in transfected cells (data not shown), suggesting that the Cul3-SPOP ubiquitin ligase may not promote PIPKII␤ turnover.
However, as only a fraction of PIPKII␤ (ϳ20%) is nuclear, it would be difficult to detect enhanced turnover of nuclear PIP-KII␤ in the context of the total cellular PIPKII␤.
Next, the ability of Cul3-SPOP to ubiquitylate PIPKII␤ was tested. PIPKII␤ was co-expressed in HEK293 cells with Cul3, Rbx1, SPOP, and His 6 -ubiquitin. After treating cells with the proteasome inhibitor MG132, cell lysates were purified over nickel resin to capture ubiquitylated proteins (22). Co-expression of the Cul3-SPOP complex promoted ubiquitylation of PIPKII␤, detected as a ladder of high molecular weight protein by Western blot (Fig. 5A). Efficient ubiquitylation of PIPKII␤ required the Cul3-SPOP complex, as excluding either SPOP or Cul3 and Rbx1 failed to generate ubiquitylated PIPKII␤. To assess the specificity of the Cul3-SPOP ligase complex for PIPKII␤, ubiquitylation of PIPKII␤ and PIPKII␣ was compared. In contrast to PIPKII␤, no ubiquitylation of PIPKII␣ was observed (Fig. 5B), demonstrating that SPOP specifically promotes ubiquitylation of PIPKII␤ in vivo.
PIPKII␤ Ubiquitylation by Cul3-SPOP Is Stimulated by MKK6/p38 MAPK-p38 MAPK has been shown to phosphorylate PIPKII␤ in response to UV irradiation and oxidative stress, thereby repressing PIPKII␤ lipid kinase activity (12). We hypothesized that p38 might also modulate PIPKII␤ ubiquity- Twenty four hours after transfection, cells were lysed and immunoprecipitated (IP) with epitope tag-specific antibodies. Normal IgG was used as a negative control. B, endogenous PIPKII␤ was immunoprecipitated from HEK293 cell lysates using an ␣-PIPKII␤ antibody or normal rabbit IgG. Immunoprecipitates were resolved by SDS-PAGE and probed with ␣-PIPKII␤ or ␣-SPOP antibodies. HA, hemagglutinin. lation. Consistent with this hypothesis, a constitutively active mutant of the p38-activating kinase MKK6 (MKK6ϩ) dramatically enhanced PIPKII␤ ubiquitylation in HEK293 cells (Fig. 5C). The MKK6-dependent p38 MAPK activation was blocked by the p38 inhibitor SB203580, further supporting a role for p38 in PIPKII␤ ubiquitylation. Importantly, MKK6ϩ did not promote PIPKII␤ ubiquitylation in the absence of the Cul3-SPOP complex (Fig. 5C), confirming that Cul3-SPOP is required for p38-stimulated ubiquitylation.
PIPKII␤ is phosphorylated at Ser 326 by p38 MAPK, and phosphorylation is sufficient to inhibit PIPKII␤ lipid kinase activity in vitro (12). To determine whether phosphorylation of PIPKII␤ at Ser 326 also modulates its ubiquitylation, the ability of MKK6ϩ to stimulate ubiquitylation of a PIPKII␤(S326A) point mutant was assessed. Interestingly, PIPKII␤(S326A) ubiquitylation was stimulated by MKK6ϩ similar to wild type PIPKII␤ (Fig. 5D). Thus p38 MAPK stimulates PIPKII␤ ubiquitylation by a mechanism independent of Ser 326 phosphorylation.
PI5P Activates Cul3-SPOP Ubiquitin Ligase Activity toward Multiple Substrates through a Pathway Inhibited by SB203580-Because PIPKII␤ associates with Cul3-SPOP, we hypothesized that PIPKII␤ ubiquitylation might be modulated by changes in local concentrations of either PI5P or PI-4,5-P 2 . Previous studies demonstrate that inhibiting PIPKII␤, for example by p38-dependent attenuation of PIPKII␤ activity or RNA interference-mediated knockdown of endogenous PIPKII␤, or expression of the kinase-dead PIPKII␤, causes an increase in cellular PI5P levels (11,12). To test the effect of elevated PI5P levels on PIPKII␤ ubiquitylation, the ubiquitylation of wild type PIPKII␤ and a well characterized kinase-dead PIPKII␤ point mutant, PIPKII␤(D278A), were compared side-by-side. Ubiquitylation of PIPKII␤(D278A) was dramatically enhanced when compared with the wild type protein (Fig. 6A); furthermore, inhibiting p38 MAPK with SB203580 reduced PIPKII␤(D278A) ubiquitylation (Fig. 6B). These results suggested that increased cellular PI5P levels stimulate PIPKII␤ ubiquitylation by Cul3-SPOP, and that this stimulation is transduced by p38 MAPK.
As an independent method to test the hypothesis that PI5P generation stimulates the Cul3-SPOP ubiquitin ligase complex, the ubiquitylation of wild type PIPKII␤ by Cul3-SPOP was analyzed by co-expressing either of two recently characterized PI-4,5-P 2 4-phosphatases (23). Both PI-4,5-P 2 4-phosphatases caused a strong Cul3-SPOP-dependent increase in PIPKII␤ ubiquitylation, similar to the results seen with co-expression of MKK6ϩ (Fig. 6B). This effect was blocked by SB203580, illustrating a requirement for p38 MAPK downstream of PI5P. Importantly, neither MKK6ϩ nor the PI-4,5-P 2 4-phosphatases caused a detectable change in total cellular ubiquitylation (data not shown), reinforcing the specificity of this pathway for ubiquitylation of Cul3-SPOP substrates.
Our observations that PI5P and p38 stimulate the ubiquitylation of PIPKII␤ suggested that these signals may cause a general enhancement of Cul3-SPOP activity, thereby promoting the ubiquitylation of numerous Cul3-SPOP substrates. To test this possibility, the effects of PI5P and p38 MAPK on the ubiquitylation of the Fas receptor binding protein Daxx and the pancreatic transcription factor Pdx1 were analyzed. Daxx is a confirmed Cul3-SPOP substrate (19,24); Pdx1 has not previously been shown to be ubiquitylated by Cul3-SPOP but is a known SPOP-interacting protein, and its transcriptional activity is negatively regulated by SPOP (25,26). As we observed with PIPKII␤, both Daxx and Pdx1 were ubiquitylated in vivo by Cul3-SPOP, and their ubiquitylation was stimulated by MKK6ϩ in an SB203580-sensitive manner (Fig. 7A). Furthermore, co-expression of the type I PI-4,5-P 2 4-phosphatase stimulated the ubiquitylation of both Daxx and Pdx1 (Fig. 7B). Ubiquitylation was attenuated by SB203580, but not by the JNK-specific inhibitor SP600125, reinforcing a specific role for p38 MAPK. As a complementary approach, the ubiquitylation of Daxx was also assessed in cells overexpressing the kinase-dead PIPKII␤(D278A) mutant. Kinase-dead PIPKII␤ enhanced the activity of the Cul3-SPOP complex toward Daxx as shown in Fig. 7C. This was also p38 MAPK-dependent, as SB203580 attenuated the effect.

SPOP Is Also Ubiquitylated Downstream of PI5P and MKK6-
Several F-box and SOCS proteins have been shown previously to be degraded by Cul1 and Cul2 ubiquitin ligases, respectively (27)(28)(29). The C. elegans BTB protein MEL-26 also appears to be degraded by Cul3 ligases (14,15), suggesting a general mechanism whereby substrate adaptor proteins like SPOP are  degraded along with their recruited substrates. To determine whether SPOP is ubiquitylated in vivo, myc-SPOP was expressed in HEK293 cells with His 6 -Ub, and its ubiquitylation was assessed. Ubiquitylated myc-SPOP was readily detected in these assays (Fig. 8B), confirming its in vivo ubiquitylation.
SPOP recruits substrate proteins through its N-terminal MATH domain and interacts with Cul3 through its C-terminal BTB domain. To assess if SPOP is ubiquitylated specifically within either of these conserved domains, several SPOP truncation mutants lacking either their N-or C-terminal domains were constructed (Fig. 8A) and expressed in HEK293 cells. Deletion of the C-terminal domain of SPOP did not prevent its ubiquitylation (Fig. 8B). In contrast, deletion of the N-terminal domain of SPOP prevented ubiquitylation, demonstrating that SPOP is ubiquitylated specifically within its N-terminal domain.
Finally, we assessed the effects of Cul3/Rbx1 and p38 MAPK on SPOP ubiquitylation. Compared with expression of SPOP alone, expression of Cul3/Rbx1 caused an increase in SPOP ubiquitylation, whereas expression of MKK6ϩ had no effect in the absence of Cul3-SPOP (Fig. 8C). Similar to the results observed with PIPKII␤, coexpression of Cul3, Rbx1, and MKK6ϩ caused a strong increase in SPOP ubiquitylation that was attenuated by SB203580. Similar results were observed when the PIPKII␤, Daxx, and Pdx1 Western blots in Figs. 5-7 were re-probed with ␣-SPOP antibodies (data not shown). Thus we conclude that SPOP is ubiquitylated along with its cargo in a pathway enhanced by PI5P and p38 MAPK and inhibited by the p38 MAPK inhibitor SB203580.

DISCUSSION
The regulated turnover of signaling proteins is critical for both the induction and attenuation of pathways crucial for cell growth, proliferation, and survival. Modular E3 ubiquitin ligases provide a mechanism by which specific proteins are recruited for ubiquitylation through their interaction with substrate-specific adaptor proteins. Here, we have demonstrated that the nuclear phosphatidylinositol phosphate kinase PIP-KII␤ interacts specifically and co-localizes with the Cul3 adaptor protein SPOP. The Cul3-SPOP ubiquitin ligase complex ubiquitylates PIPKII␤ in vivo, and ubiquitylation is potently stimulated by the MKK6/p38 MAPK pathway. Consistent with previous reports, SPOP itself is also ubiquitylated within the Cul3-SPOP complex. Most significantly, we have demonstrated that PI5P, the lipid substrate of PIPKII␤, stimulates Cul3-SPOP activity toward multiple substrates, including the SPOP-binding proteins Daxx and Pdx1. The ability of the p38-specific inhibitor SB203580 to prevent Cul3-SPOP stimulation further suggests that PI5P functions by activating the p38 MAPK pathway. Together, these results define a novel mechanism whereby phosphoinositide signaling may modulate the functions of key signaling proteins by regulating the activity of a nuclear ubiquitin ligase. Based on our experimental data, we propose the model illustrated in Fig. 9.
p38 MAPK has been shown previously to directly inhibit the lipid kinase activity of PIPKII␤ through phosphorylation of Ser 326 . Here we have shown that p38 MAPK also stimulates PIPKII␤ ubiquitylation independently of Ser 326 phosphorylation. Furthermore, we have shown that p38 stimulates the ubiquitylation of multiple Cul3-SPOP substrate proteins, including Daxx and Pdx1. These data demonstrate that p38 stimulates the  Cul3-SPOP ubiquitin ligase activity, although currently the mechanism by which this occurs is unknown. One possibility is that either p38 or a p38 effector kinase phosphorylates Cul3-SPOP, thereby stimulating its catalytic activity. In support of this hypothesis, p38 has been reported to phosphorylate the RING E3 ubiquitin ligase Siah2, increasing its activity toward its substrate PHD3 (30). Further characterization is required to determine how p38 activation results in stimulation of Cul3-SPOP activity.
A key discovery from our experiments is the ability of PI5P to stimulate Cul3-SPOP activity through a p38-dependent signaling pathway. Although relatively little about the physiological functions of PIPKII␤ is currently known, one recurring theme is the role of its substrate, PI5P, as a key regulatory molecule. Previous reports have linked PIPKII␤ and PI5P to the modulation of insulin signaling and of a nuclear stress-response pathway (12,31,32). In both of these pathways, the primary function of PIPKII␤ appears to be regulating PI5P levels by converting it to PI-4,5-P 2 . These observations provide an intriguing contrast to the canonical Type I PIPK signaling pathways, in which PI-4,5-P 2 is the key regulatory species modulating effector proteins. Our results now identify a third signaling pathway in which PI5P elicits a potent regulatory effect, and suggest that PIPKII␤ regulates the system by maintaining low PI5P levels under resting conditions. If PIPKII␤ is primarily a regulator of Cul3-SPOP rather than a substrate, this might explain our observation that endogenous PIPKII␤ is not degraded upon overexpression of Cul3-SPOP in cultured cells.
Although our data provide significant insight into a novel phosphoinositide-regulated signaling pathway, numerous questions remain. Physiological stimuli that activate the PI5Psensitive stimulation of Cul3-SPOP have yet to be identified. Because activation of Cul3-SPOP is p38-sensitive, it is possible that Cul3-SPOP mediates a stress-response pathway. The previous characterization of PI5P as a transducer of a p38-dependent stress-response pathway supports this hypothesis (12). Additionally, a stress-response model would be consistent with the observation that Daxx, which modulates survival and apoptosis signaling pathways, is itself ubiquitylated by Cul3-SPOP. The mechanism by which PI5P activates p38 MAPK is also currently unknown, for example whether PI5P directly stimulates an upstream kinase of the p38 cascade. Additionally, the role of cytosolic versus nuclear PI5P in p38-dependent activation of Cul3-SPOP remains to be defined. Distinguishing between these two pools of PI5P will help to more thoroughly characterize the signaling pathway leading to activation of Cul3-SPOP. Finally, the physiological effects of Cul3-SPOP remain poorly understood, e.g. whether Cul3-SPOP is primarily involved with stress-response pathways or if it modulates numerous other pathways in the cell. This uncertainty is due in part to the lack of known Cul3-SPOP substrates. As new substrates continue to be identified and characterized, the principal physiological function(s) of Cul3-SPOP will become increasingly clear. Further investigation of these and other aspects will be critical for a more complete understanding of how phosphoinositide signaling contributes to the regulation of Cul3-SPOP and its substrates, and how PI5P, p38 MAPK, and Cul3-SPOP contribute to cell physiology.
Our observation that SPOP is ubiquitylated by the Cul3 ubiquitin ligase complex is consistent with previous reports that substrate adaptors for Cul1, Cul2, and Cul3 E3 ligases are ubiquitylated in an autocatalytic mechanism. Our data specifically identify the N-terminal half of SPOP containing the MATH domain as the major ubiquitylation target. Structural analyses of cullin E3 ligases indicate that cullin E3s form rigid scaffolds that orient E2-conjugating enzymes in close proximity to their substrates, and that the E2s only transiently associate with the E3 complex (33)(34)(35). Therefore, proper orientation is critical for efficient ubiquitylation of a substrate protein. Because BTB proteins such as SPOP interact with their substrates through their MATH domains, the MATH domain would likely be in close proximity to the E2-conjugating enzyme. In contrast, the BTB domain of SPOP, through which it interacts with Cul3, would be more distal to the E2 enzyme and therefore a less efficient target for ubiquitylation. Other cullin specificity factors may be similarly ubiquitylated within their substrate-binding domains. A more detailed analysis would be beneficial to more thoroughly understand the function of cullin specificity factors as well as the activities of cullin-based ubiquitin ligases as a whole.
In summary, the data presented in this study define a novel mechanism by which phosphoinositide signaling regulates a nuclear ubiquitin ligase complex. The interaction of PIPKII␤ with the Cul3-SPOP ubiquitin ligase, coupled with the ability of the PIPKII␤ substrate PI5P to modulate Cul3-SPOP activity through a signaling pathway inhibited by the p38 inhibitor SB203580, identifies a new function for phosphoinositide signaling within the nucleus. Identification of stimuli that enhance or attenuate this pathway, as well as delineation of the downstream physiological effects, will be invaluable to more thoroughly understand the roles of phosphoinositides within the nucleus.