p70S6K1 (S6K1)-mediated Phosphorylation Regulates Phosphatidylinositol 4-Phosphate 5-Kinase Type I γ Degradation and Cell Invasion*

Phosphatidylinositol 4-phosphate 5-kinase type I γ (PIPKIγ90) ubiquitination and subsequent degradation regulate focal adhesion assembly, cell migration, and invasion. However, it is unknown how upstream signals control PIPKIγ90 ubiquitination or degradation. Here we show that p70S6K1 (S6K1), a downstream target of mechanistic target of rapamycin (mTOR), phosphorylates PIPKIγ90 at Thr-553 and Ser-555 and that S6K1-mediated PIPKIγ90 phosphorylation is essential for cell migration and invasion. Moreover, PIPKIγ90 phosphorylation is required for the development of focal adhesions and invadopodia, key machineries for cell migration and invasion. Surprisingly, substitution of Thr-553 and Ser-555 with Ala promoted PIPKIγ90 ubiquitination but enhanced the stability of PIPKIγ90, and depletion of S6K1 also enhanced the stability of PIPKIγ90, indicating that PIPKIγ90 ubiquitination alone is insufficient for its degradation. These data suggest that S6K1-mediated PIPKIγ90 phosphorylation regulates cell migration and invasion by controlling PIPKIγ90 degradation.

PIPKI␥90 is essential for cell migration, invasion, and metastasis. It is required for focal adhesion assembly and disassembly, key steps in cell migration (11). Depletion of PIPKI␥90 inhibits growth factor-stimulated cell migration in MDA-MB-231 breast cancer cells and HeLa cervical cancer cells (14,15). PIPKI␥90 knockdown also blocks the invasion of breast cancer and colon cancer cells (11,16). Furthermore, PIPKI␥90-depleted 4T1 breast cancer cells show significant reduction in tumor progression and metastasis (13). PIPKI␥90 also regulates neutrophil migration by controlling cell polarity as well as rear retraction (17)(18)(19). PIPKI␥90 is a substrate for Src, which phosphorylates PIPKI␥90 at Tyr-644, enhancing its binding to talin and reducing talin-␤ integrin interaction (20). Talin, in turn, activates integrins and initiates FA assembly to regulate cell migration and invasion. In addition, phosphorylation of PIPKI␥90 at Tyr-639 by epidermal growth factor (EGF) receptor influences tumor cell migration and metastasis (13).
It has been demonstrated that the ubiquitin proteasome pathway regulates FA assembly and disassembly and, consequently, cell migration and invasion through ubiquitinating FA proteins (16,(21)(22)(23)(24)(25)(26), and our research indicates that PIPKI␥90 is a key molecule that mediates the role of the ubiquitin proteasome pathway in this regard. Our published data indicate that PIPKI␥90 functions to regulate focal adhesion assembly and disassembly (11). We also demonstrated that PIPKI␥90 ubiquitination at Lys-97 by HECTD1, an E3 ubiquitin ligase that regulates cell migration, results in PIPKI␥90 degradation, thus controlling dynamic PIP 2 production to mediate FA assembly/ disassembly, cell migration, invasion, and metastasis (16). However, it is not clear how upstream signaling pathways control PIPKI␥90 ubiquitination or degradation during cell migration and invasion.
Ribosomal protein S6 kinase ␤ 1 (also called p70S6K1 or S6K1), a serine-threonine kinase, is one of the mTOR pathway effectors. It is well known that S6K1 regulates cell growth, survival, and metabolism (27)(28)(29)(30)(31). Recent evidence indicates that it also regulates cancer cell invasion and metastasis (32,33), but the molecular mechanisms behind these processes are less defined. In this study, we demonstrate that S6K1 phosphorylates PIPKI␥90 at Thr-553 and Ser-555 and that S6K1-mediated phosphorylation controls PIPKI␥90 degradation to regulate the development of FAs and invadopodia and, consequently, cell migration and invasion.
To find out whether EGF or HGF stimulates PIPKI␥90 phosphorylation at residues Thr-553 and Ser-555, MDA-MB-231 cells stably expressing FLAG-PIPKI␥90 were serum-starved and stimulated with EGF, HGF, SCF, and PDGF. FLAG-PIPKI␥90 was immunoprecipitated with anti-FLAG-agarose beads, and PIPKI␥90 phosphorylation was detected with an anti-RXRXXpS/T motif antibody. EGF and HGF stimulated PIPKI␥90 phosphorylation, whereas SCF and PDGF did not (Fig. 1E). Similar results were observed in MDA-MB-468 cells (supplemental Fig. S1A). HGF and EGF stimulated Akt and S6K1 activation in a time-dependent manner, whereas SCF and PDGF had no obvious effects ( Fig. 1E and supplemental Fig. S1, B and C). Because both S6K1 and Akt were activated by HGF or EGF in MDA-MB-231 cells, we tested whether S6K1 or Akt mediate PIPKI␥90 phosphorylation. MDA-MB-231 cells that stably express FLAG-PIPKI␥90 were treated with Akt inhibitor VIII or the S6K1 inhibitors DG2 and PF4708671 and then challenged with HGF. Akt inhibitor VIII inhibited HGF-stimulated Akt, S6K1, and PIPKI␥90 phosphorylation. The S6K1 inhibitors DG2 and PF4708671 did not influence Akt and S6K1 activation but inhibited S6K1 activity (as indicated by the reduction in ribosomal protein S6 phosphorylation) and PIPKI␥90 phosphorylation (Fig. 1F). To further examine whether S6K1 phosphorylates PIPKI␥90 in cells, MDA-MB-231 cells that stably express FLAG-PIPKI␥90 were infected with lentiviruses that express S6K1 shRNAs or empty vector. The resulted cells were stimulated with vehicle or HGF. S6K1 knockdown significantly inhibited HGF-induced PIPKI␥90 phosphorylation (Fig. 1G). These results indicate that PIPKI␥90 is a substrate for S6K1.
To assess the potential role of PIPKI␥90 phosphorylation in cancer cell invasion, the Matrigel-invasive capabilities of PIPKI␥90-depleted MDA-MB-231 cells that express ZZ-PIPKI␥90, ZZ-PIPKI␥90 T553A,S555A , or ZZ-PIPKI␥90 T553E,S555E were measured. Re-expression of PIPKI␥90 WT in PIPKI␥90depleted cells restored cell invasion to an extent comparable with the invasion of cells expressing empty pLKO.1 vector, and that of PIPKI␥90 T553E, S555E partially rescued cell invasion. In contrast, re-expression of PIPKI␥90 T553A, S555A only slightly enhanced cell invasion (Fig. 3, A and B). Similar results were observed when PIPKI␥90 and the mutants were expressed in parental MDA-MB-231 cells (supplemental Fig. S2), suggesting a dominant negative function of PIPKI␥90 T553A,S555A . To explore the role of S6K1 in cell invasion, we examined the effect of the S6K1 inhibitor DG2 on the invasion of MDA-MB-231 cells. We found that S6K1 inhibition impaired invasion of the cells (Fig. 3C). In particular, 10 M S6K1 inhibitor DG2 significantly decreased the invasive potential of the cells by ϳ90% (in the absence of HGF) and 80% (in the presence of HGF). To further examine the requirement for S6K1 in cell invasion, this kinase was depleted in MDA-MB-231 cells using S6K1 shRNA (Fig. 3D). Cells transfected with S6K1 shRNA could not invade efficiently compared with cells expressing shRNA control (Fig.  3E). S6K1-depleted cells, even in the presence of HGF, could not invade normally compared with cells expressing shRNA control. Akt1, another protein kinase that potentially phosphorylates PIPKI␥90, was also depleted in MDA-MB-231 cells by using two different shRNAs. Depletion of Akt1 caused a slight reduction in the phosphorylation of S6K1 and S6 ribosomal protein (Fig. 3F). Depletion of Akt1 in MDA-MB-231 cells did not exhibit a significant reduction in invasive ability. As shown in Fig. 3G, depletion of Akt1 slightly reduced HGF-induced invasion of MDA-MB-231 cells. However, in the absence of HGF, cells expressing Akt1 shRNAs had higher number of invaded cells compared with cells with shRNA control, implying that Akt1 is not mandatory for the invasion of MDA-MB-231 cells. To further examine the role of S6K1-mediated PIPKI␥90 phosphorylation in cell invasion, the effects of the S6K1 inhibitor DG2 on the invasion of PIPKI␥-depleted cells that express ZZ-PIPKI␥90, -PIPKI␥90 T553A,S555A , or -PIPKI␥90 T553E,S555E were examined. DG2 significantly inhibited the invasion of cells expressing PIPKI␥90 but had only marginal effects on the invasion of cells expressing PIPKI␥90 T553A,S555A or -PIPKI␥90 T553E,S555E (Fig. 3H). These results indicate that S6K1-mediated PIPKI␥90 phosphorylation regulates cell invasion.
Because of the crucial role of matrix metalloproteinase-mediated matrix degradation in cell invasion (36 -38), we set out to determine whether the S6K1-PIPKI␥90 pathway regulates matrix degradation. To examine whether the phosphorylationdeficient mutants of PIPKI␥90 influence matrix degradation, we examined the gelatin degradation activity of PIPKI␥90depleted MDA-MB-231 cells that were rescued with PIPKI␥90 WT , PIPKI␥90 T553A,S555A , and PIPKI␥90 T553E,S555E . Glass-bottom dishes were coated with Alexa 488-conjugated gelatin. The coated dishes were then dried, fixed with glutaraldehyde, and reduced with sodium borohydride. The cells were plated on dishes and treated with HGF. The cells were fixed and stained with cortactin, an invadopodium marker. Matrix deg-radation was examined by TIRF microscopy. Cells expressing PIPKI␥90 WT had similar matrix degradation activity compared with cells expressing shRNA control. However, cells with PIPKI␥90 T553A,S555A had significantly lower matrix degradation activity, whereas cells expressing PIPKI␥90 T553E,S555E showed a slight reduction in degraded areas (Fig. 4, A and B). To further corroborate these findings, we tested the effect of S6K1 inhibition on matrix degradation. Similar to invasion, S6K1 inhibition affected this function and considerably decreased the gelatin degradation (Fig. 4C). These data suggest that  S6K1-mediated PIPKI90 phosphorylation regulates matrix degradation.
To examine the possible association of the S6K1 pathway with cancer metastasis, human breast cancer tissue array slides, including primary tumors and the matched metastatic tumors of lymph node tissues (US Biomax), were stained for phospho-S6 ribosomal protein (Ser(P)-235/236), a substrate of S6K1. Among the tissues from 50 subjects analyzed, phospho-S6 staining was positive in 20 cases of metastatic tumors (40%) and in six cases of the matched primary tumors (12%) (Fig. 5, A and B). Also, phospho-S6 staining in 15 cases of metastatic tumors (30%) was significantly higher than the staining in the matched primary tumors; one case was lower (2%), and 34 cases were unchanged (68%). These data suggest that activation of the S6K1 pathway positively correlates with human breast cancer metastasis (p Ͻ 0.001).
To measure the kinase activity of PIPKI␥90, ZZ-PIPKI␥90 was transfected into CHO-K1 cells and immunoprecipitated with IgG-conjugated-agarose beads or protein A-agarose using ZZ-PIPKI␥90 K188,200R , a kinase-deficient mutant, as a negative control. The activities of PIPKI␥90 and mutants were measured by PIP2 production using PIP and [␥-32 P]ATP as substrates. PI(4,5)P 2 was separated by thin layer chromatography, imaged by autoradiography, and quantified by liquid scintillation counting. The kinase activity was detected in IgG-agarose beads that were incubated with ZZ-PIPKI␥90-transfected lysates but not in protein A-agarose beads incubated with the  same lysate; very low activity was observed in IgG-agarose beads that were incubated with ZZ-PIPKI␥90 K188,200R (supplemental Fig. S3A). To know whether mutation at Thr-553 and Ser-555 affects the activity of PIPKI␥90, ZZ-PIPKI␥90 WT , -PIPKI␥90 T553A,S555A , and -PIPKI␥90 T553E,S555E were transfected into CHO-K1 cells and immunoprecipitated with IgGagarose beads. The activities of PIPKI␥90 and mutants were measured using the same method. Substitution of Thr-553 and Ser-555 with alanine and glutamate did not affect PIPKI␥90 activity in vitro (supplemental Fig. S3B).
To determine whether PIPKI␥90 phosphorylation regulates its degradation, CHO-K1 cells were transfected with FLAG-PIPKI␥90 WT , FLAG-PIPKI␥90 T553A,S555A , and FLAG-PIPKI␥90 T553E,S555E and treated with DMSO and carfilzomib, a specific proteasome inhibitor. As shown in Fig. 6A, PIPKI␥90 T553A,S555A was not efficiently degraded and was more resistant to degradation than PIPKI␥90 WT and PIPKI␥90 T553E, S555E . To further confirm the stability of the T553A,S555A mutant, we determined the time course of PIPKI␥90 degradation. Avitagged PIPKI␥90 WT and mutants were transfected into CHO-K1 cells with stable expression of BirA, and then labeled with biotin. Then, biotin was washed away and cells were split into dishes with media containing avidin. PIPKI␥90 and mutants were detected using Dylight 680 Streptavidin by harvesting the cells at different time points. PIPKI␥90 T553A,S555A was more resistant to degradation in comparison to WT and PIPKI␥90 T553E,S555E mutant (Fig. 6B) and had a significantly longer half-life than the WT and PIPKI␥90 T553E,S555E (Fig. 6C).
To further demonstrate the role of S6K1-mediated PIPKI␥90 phosphorylation in PIPKI␥90 degradation, CHO-K1 cells were transfected with Dendra2-PIPKI␥90, -PIPKI␥90 T553A,S555A , and -PIPKI␥90 T553E,S555E and plated on fibronectin-coated glass-bottom dishes. The cells were irradiated by a 408-nm laser to convert the Dendra2 fusion protein into its red fluorescence form. The red fluorescence protein degradation was recorded by time-lapse imaging at 10-min intervals. Dendra2-PIPKI␥ T553A,S555A was more stable/resistant to degradation, with a half-life of Ͼ4 h, in comparison with the WT and T553E,S555E mutant of PIPKI␥90, which both showed a relatively higher rate of degradation, with half-lives of 2.5 and 3 h, respectively (Fig. 6, D and E). To examine the role of S6K1 in regulating PIPKI␥90 degradation, CHO-K1 cells that expressed Dendra2-PIPKI␥90 were treated with the S6K1 inhibitors DG2 (10 M) or PF4708671 (10 M), and the degradation of Den-dra2-PIPKI␥90 was analyzed. As shown in Fig. 6F, S6K1 inhibition caused a significant increase in the stability of Dendra2-PIPKI␥ WT compared with the control. However, DG2 had no effect on the degradation of Dendra-PIPKI␥90 T553E,S555E (Fig.  6G). These results further support the concept that S6K1-mediated phosphorylation of PIPKI␥90 facilitates its degradation.
This prompted us to examine the ubiquitination of PIPKI␥90 and these mutants. To this end, Avi-ubiquitin was cotransfected with ZZ-PIPKI␥90, -PIPKI␥90 T553A,S555A , or -PIPKI␥90 T553E,S555E into CHO-K1 cells expressing BirA, labeled with biotin, and immunoprecipitated with IgG-agarose. Ubiquitination was detected with Dylight 680 streptavidin. Substitution of Thr-553 and Ser-555 with Ala caused an increase in PIPKI␥90 ubiquitination, whereas substitution with Glu had no significant change compared with the WT protein (Fig. 7A), indicating that PIPKI␥90 ubiquitination is not sufficient for its degradation.
To compare the roles of S6K1 and Akt1 in PIPKI␥90 degradation, we examined the steady-state levels of PIPKI␥90 in S6K1-depleted MDA-MB-231 cells. The level of PIPKI␥90 in S6K1-depleted cells was significantly higher than that in cells expressing a control shRNA (Fig. 7B). Treatment with carfilzomib resulted in a significant increase in PIPKI␥90 level in cells expressing control shRNA but not in S6K1-depleted cells. However, depletion of Akt1 by expressing its shRNA had no significant effect on the steady-state levels of PIPKI␥90 (Fig.  7C). These results suggest that S6K1-mediated phosphorylation facilitates PIPKI␥90 degradation.
Our previous published results indicate that PIPKI␥90 ubiquitination at lysine 97 and subsequent degradation are necessary for breast cancer cell invasion (16). To examine the role of PIPKI␥90 degradation in matrix degradation, we compared the matrix degradation activities of PIPKI␥90-depleted MDA-MB-231 cells that express codon-modified ZZ-PIPKI␥90 or ZZ-PIPKI␥90 K97R using normal and PIPKI␥90-depleted MDA-MB-231 cells as controls (Fig. 7D). PIPKI␥90 K97R is an ubiquitination-and degradation-resistant mutant. Depletion of PIPKI␥90 inhibited matrix degradation, and re-expression of PIPKI␥90 restored matrix degradation in PIPKI␥90-depleted cells whereas that of PIPKI␥90 K97R did not (Fig. 7, E and F), further supporting the hypothesis that dynamic PIPKI␥90 degradation is essential for extracellular matrix degradation.

Discussion
The ubiquitin proteasome pathway regulates FA assembly and disassembly and, consequently, cell migration and invasion by ubiquitinating FA proteins (16,(21)(22)(23)(24)(25)(26), and we recently demonstrated that PIPKI␥90 ubiquitination and subsequent degradation control FA dynamics to regulate cell migration and invasion (16). In this study, we demonstrated that S6K1-mediated PIPKI␥90 phosphorylation regulates PIPKI␥90 degradation to control the development of FAs and invadopodia and, consequently, cell migration and invasion.
We demonstrated that PIPKI␥90 is a substrate for S6K1. We showed that S6K1 phosphorylated PIPKI␥90 when they were co-transfected into CHO-K1 cells (Fig. 1B) and that substitution of the Thr-553 and Ser-555 sites with alanine abolished PIPKI␥90 phosphorylation by S6K1 in vitro and in cells (Fig. 1,  C and D). We also revealed that PIPKI␥90 phosphorylation was stimulated by HGF and EGF and that HGF-stimulated phosphorylation was inhibited by the S6K1 inhibitors DG2 and PF4708671, Akt inhibitor VIII, as well as S6K1 knockdown (Fig.  1, E-G). The S6K1 inhibitors DG2 and PF4708671 caused 68% and 45% reduction in PIPKI␥90 phosphorylation in HGF-stimulated MDA-MB-231 cells, respectively. Akt inhibitor VIII suppressed 85% of PIPKI␥90 phosphorylation. The related higher efficiency of Akt1 inhibitor is probably due to its inhibition of both Akt and S6K1 activation. Thus, we estimated that S6K1 mediated approximately 50 -70% of Thr-553 and Ser-555 phosphorylation in HGF-stimulated MDA-MB-231 cells. Endogenous PIPKI␥90 phosphorylation has not been examined because of reagent limitation. Nevertheless, these results indi-  cate that PIPKI␥90 is a substrate for S6K1 in the system we used.

S6K1 Regulates PIPKI␥90 Degradation and Cell Invasion
When we started writing this manuscript, Le et al. (39) reported that Akt1 phosphorylated PIPKI␥90 at Ser-555. Indeed, PIPKI␥90 was phosphorylated when it was co-transfected with Akt1 (Fig. 1B), and HGF-stimulated PIPKI␥90 phosphorylation was inhibited by Akt inhibitor VIII (Fig. 1F), suggesting that Akt1 is also a potential protein kinase that phosphorylates PIPKI␥90. However, depletion of Akt1 did not significantly inhibit the invasion of MDA-MB-231 cells (Fig.  3G). This result is consistent with previous reports showing that Akt activation potentially blocks carcinoma motility, including migration and invasion in breast cancer cells (40 -43). Therefore, although both S6K1 and Akt1 phosphorylate PIPKI␥90, S6K1 is functionally more relevant than Akt1 in regulating PIPKI␥90 phosphorylation and cell invasion in breast cancer cells.
It is generally believed that protein polyubiquitination is sufficient for protein degradation (44,45), but our findings indicate that PIPKI␥90 ubiquitination alone is insufficient for its degradation. The phosphorylation-deficient mutant PIPKI␥90 T553A,S555A cannot be degraded efficiently compared with the WT and T553E,S555E mutant (Fig. 6, B-E). Moreover, the S6K inhibitors DG2 and PF4708671 inhibited the degradation of PIPKI␥90 but not that of PIPKI␥90 T553E,S555E . However, substitution of Thr-553 and Ser-555 with alanine did not sup-  DECEMBER 2, 2016 • VOLUME 291 • NUMBER 49 press but, instead, enhanced PIPKI␥90 ubiquitination (Fig. 7A). Our data show that PIPKI␥90 binds to 14-3-3 proteins, a family of adaptor proteins that regulate protein degradation (46 -48), in a phosphorylation-dependent manner. 3 However, although a role for this interaction with 14-3-3 proteins may be involved, it remains unknown how S6K1-mediated phosphorylation regulates PIPKI␥90 degradation. The suppressive role of the phosphorylation-deficient mutant PIPKI␥90 T553A,S555A in cell migration provides a new evidence for the role of PIPKI␥90 degradation in cell migration. Previous studies have demonstrated the essential role of PIPKI␥90 in the regulation of cell migration (14 -16). Our recent study indicates that PIPKI␥90 ubiquitination by HECTD1 and subsequent degradation control FA dynamics and cell migration. Here we show that the phosphorylationdeficient mutant PIPKI␥90 T553A,S555A was resistant to degradation and inhibited migration behavior by suppressing directionality and net distance from origin in comparison with PIPKI␥90 WT and PIPKI␥90 T553E, S555E (Fig. 2). Because of the central role of FAs in cell migration, the FA defect in cells expressing PIPKI␥90 T553A,S555A may contribute to its inhibition of cell migration (Fig. 2, D and E). The effect of PIPKI␥90 T553A,S555A on FA formation is probably caused by its enhanced stability, which interferes with talin binding to ␤ integrins and integrin activation. This is consistent with our previous finding that PIPKI␥90 K97R , a degradation-resistant mutant, had a diminished FA assembly rates (16).

S6K1 Regulates PIPKI␥90 Degradation and Cell Invasion
As a downstream target of mTOR, the role of S6K1 in regulating cell growth, survival, and metabolism has been well documented, whereas its role in cancer cell invasion and the downstream targets that mediate this process remain to be defined. Previous studies have established a crucial role of PIPKI␥90 in cancer cell invasion (11,14,16). In this study, we demonstrated that S6K1-mediated PIPKI␥90 phosphorylation at Thr-553 and Ser-555 is indispensable for breast cancer cell invasion. PIPKI␥90 T553A,S555A -expressing cells had a remarkably decreased capability to invade through Matrigel. On the other hand, cells expressing the WT and PIPKI␥90 T553E,S555E mutant had similar invasive abilities (Fig. 3, A and B). This discrepancy may, in part, be due to the negative charge of the carboxyl group on the glutamate side chain, which could mimic the negative charge on a phosphorylated threonine/serine of PIPKI␥90. However, alanine with a neutral methyl side chain could not restore normal function of PIPKI␥90 in cell invasion. Inhibition of S6K1 by the S6K1 inhibitor DG2 or depletion of S6K1 using shRNAs considerably diminished the invasion of MDA-MB-231 cells (Fig. 3, C and E). Furthermore, inhibition of mTOR using rapamycin also inhibited cell invasion (49). However, depletion of Akt1 had a minimal effect on this function (Fig.  3G). Based on these findings and previous reports of the negative role of Akt1 in cell migration and invasion, we conclude that, although both S6K and Akt1 can phosphorylate PIPKI␥90, only S6K has a major positive role in regulating breast cancer cell invasion.
Matrix metalloproteinases-mediated matrix degradation is critical for cell invasion (36 -38). However, the molecular mechanisms that regulate this process are not entirely understood. Our data show that PIPKI␥90 T553A,S555A , a degradationresistant mutant, had a significantly limited cellular ability to mediate gelatin degradation. In contrast, cells expressing the WT or PIPKI␥90 T553E,S555E mutant had similar abilities to digest gelatin (Fig. 4, A and B). Moreover, PIPKI␥90 K97R , which is an ubiquitination site mutant and is resistant to proteasome degradation, was unable to restore the matrix degradation in PIPKI␥90-depleted cells (Fig. 7, E and F). Furthermore, depletion of S6K1 by shRNA enhanced the stability of PIPKI␥90 (Fig.  7B) but significantly reduced the cellular capability to degrade the gelatin matrix (Fig. 4C). These data suggest that the S6K1-PIPKI␥90 pathway controls PIPKI␥90 degradation to regulate matrix degradation and cell invasion, probably through modulating the secretion of matrix metalloproteinases (13).
Spatial and temporary production of PIP 2 is crucial for cell migration and invasion. This highly regulated PIP 2 production is controlled by PIPKI␥90 ubiquitination and subsequent degradation. However, PIPKI␥90 ubiquitination alone is insufficient for its degradation; instead, the new data presented here show that S6K1-mediated PIPKI␥90 phosphorylation is also necessary for the degradation of ubiquitinated PIPKI␥90. S6K1 phosphorylates PIPKI␥90 at Thr-553 and Ser-555 to mediate the dynamic degradation of PIPKI␥90, thus controlling FA dynamics and matrix degradation and, consequently, cell migration and invasion. Our findings uncover a new paradigm for control of protein degradation, implying that a similar mechanism may also occur in other systems and processes.
Cell Culture and Transfection-CHO-K1 cells, MDA-MB-231 and MDA-MB-468 human breast cancer cells, and 293T human embryonic kidney cells were from the American Type Culture Collection and were maintained in DMEM (Sigma) containing 10% FBS, penicillin (100 units/ml), and streptomycin (100 g/ml). CHO-K1 and 293T cells were transfected with Safectine RU50 according to the protocol of the manufacturer.
Preparation of Viruses and Cell Infection-293T cells were transfected with the pBabe retroviral or pLKO1 lentiviral sys-tem using Safectine RU50 transfection reagent according to the protocol of the manufacturer. The virus particles were applied to overnight cultures of breast cancer cells for infection. Cells that stably express pLKO1 lentiviral shRNAs were obtained by selecting the infected cells with 1 g/ml puromycin, and cells that were infected with pBabe retroviruses were stabilized by growing infected cells in the presence of 0.7 mg/ml neomycin for 10 days.
PIPKI␥90 Phosphorylation-FLAG-PIPKI␥90 (or mutants) was co-transfected with an empty vector or a plasmid expressing active kinase into CHO-K1 cells. The cells were lysed with radioimmune precipitation assay buffer (50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1% IPEGAL, 0.5% deoxycholate, and 5 mM EDTA) containing protease inhibitor mixture and phosphatase inhibitor mixture. FLAG-PIPKI␥90 was immunoprecipitated with anti-FLAG-agarose beads. The immune complexes were analyzed by SDS-polyacrylamide gel electrophoresis and transferred to a nitrocellulose membrane. PIPKI␥90 phosphorylation was detected with an anti-RXRXXpS/T motif antibody. To detect PIPKI␥90 phosphorylation in breast cancer cells, cells stably expressing FLAG-PIPKI␥90 were treated with Akt or S6K1 inhibitor and then stimulated with growth factors. FLAG-PIPKI␥90 was immunoprecipitated, and PIPKI␥90 phosphorylation was detected as described above.
Live Cell Imaging and Dendra2-PIPKI␥90 Degradation-CHO-K1 cells were transiently transfected with Dendra2-PIPKI␥ WT , -PIPKI␥ T553A,S555A , and -PIPKI␥ T553E,S555E and cultured in fibronectin-coated glass-bottom dishes. Time-lapse live cell imaging was conducted on a Nikon A1 R microscope. Before excitation, there should not be any red Dendra2-emission signal visible. Photoconversion was performed at ϫ100 magnification with near-UV irradiation (408 nm) for 120 s. Green-to-red photoconversion was monitored in real time using a 561-nm channel. Images were captured at 20-min intervals and analyzed using NIS-Elements software.
Ubiquitination Assays-Avi-ubiquitin was co-transfected with ZZ-PIPKI␥90, -PIPKI␥90 T553A,S555A , and -PIPKI␥90 T553E,S555E and co-transfected with an ubiquitin ligase or an empty vector into CHO-K1 cells stably expressing EGFP-BirA (50). 24 h posttransfection, cells were incubated with 500 M biotin, 1 M bortezomib, and 1 M carfilzomib for 6 h and then scraped in PBS. The cells were spun down, lysed with 150 l of 1ϫ SDS sample buffer (without 2-mercaptoethanol) containing protease inhibitor mixture and bortezomib/carfilzomib and boiled immediately. The lysates were cleared, diluted to 1 ml, and incubated with rabbit IgG-Sepharose beads at 4°C for 2 h to precipitate ZZ-tagged PIPKI␥90 (or the mutants). The beads were washed and analyzed by SDS-PAGE and Western blotting as above. The ubiquitination of the ZZ domain fusion protein was detected with Dylight 680-Streptavidin, whereas the expression of the ZZ domain fusion protein was probed with Dylight 680-rabbit IgG.
Cell Migration Assays-Cells were treated with trypsin and resuspended in DMEM containing 1% FBS and 10 ng/ml EGF, plated at low densities on glass-bottom dishes (Cellvis) coated with 5 g/ml fibronectin, and cultured for 3 h in a CO 2 incubator. Cell motility was measured with a Nikon Biostation IMQ. Cell migration was tracked for 6 h. Images were recorded every 10 min. The movement of individual cells was analyzed with NIS-Elements AR (Nikon) as described previously (16).
Focal Adhesion Staining-MDA-MB-231 cells were infected with lentiviruses that express PIPKI␥ shRNA (A1) to deplete endogenous PIPKI␥, infected with retroviruses that express pBabe-FLAG-PIPKI␥90 WT or FLAG-PIPKI␥90 T553A,S555A , and selected with neomycin (0.7 mg/ml). The cells were trypsinized and plated on glass-bottom dishes that had been precoated with fibronectin (5 g/ml). The cells were cultured for 4 h. The cells were fixed with 4% paraformaldehyde for 15 min, permeabilized for 15 min with 0.5% Triton X-100, and then blocked with 5% BSA in PBS for 1 h. The cells were then incubated with a rabbit polyclonal anti-PIPKI␥ antibody and a mouse monoclonal anti-paxillin antibody, washed with PBS, and then incubated with a Dylight480-labeled goat anti-rabbit and a Dylight550-labeled goat anti-mouse secondary antibody. After washing with PBS, the images of PIPKI␥ and paxillin were acquired with a Nikon Eclipse Ti TIRF microscope equipped with a ϫ60, 1.45 numerical aperture objective, CoolSNAP HQ2 charge-coupled device camera (Roper Scientific). Focal adhesion area distribution was analyzed with Nis-Elements.
Invasion Assays-One hundred microliters of Matrigel (1:30 dilution in serum-free DMEM) was added to each Transwell polycarbonate filter (6-mm diameter, 8-m pore size, Costar) and incubated with the filters at 37°C for 6 h. Breast cancer cells were trypsinized and washed three times with DMEM containing 1% FBS. The cells were resuspended in DMEM containing 1% FBS at a density of 5 ϫ 10 5 cells/ml. The cell suspensions (100 l) were seeded into the upper chambers, and 600 l of DMEM containing 50 ng/ml HGF were added to the lower chambers. The cells were allowed to invade for 12 h (or as indicated) in a CO 2 incubator, fixed, stained, and quantitated as described previously (11).
Gelatin Degradation Assays-Gelatin degradation assays were performed as described previously (52). Briefly, glass-bottom dishes were coated with warm Alexa 488-conjugated gelatin (0.2 mg/ml) in PBS containing 2% sucrose. The coated dishes were dried, fixed with prechilled glutaraldehyde solution (0.5%), washed with PBS, and then reduced with 5 mg/ml of sodium borohydride in PBS. The dishes were washed extensively with PBS and then incubated with DMEM containing 10% FBS and antibiotics for 1 h. Cells were plated at low density to the dishes and cultured for 12 h, fixed with 4% paraformaldehyde, permeabilized with 0.5% Triton X-100 and stained with cortactin or Alexa 647 phalloidin. Images were acquired using a TIRF microscope and analyzed with NIS Elements software.
Gel Data Quantification-Gel data were quantified by analyzing inverted images using ImageJ as described previously (21). Data from different experiments were normalized to controls. If values from different experiments had a high variation, then datasets were further normalized by dividing the numbers in a dataset with a factor (e.g. 2) so that the biggest values from different experiments were similar.
Author Contributions-N. J., Q. Z., L. L., W. L., L. Q., and J. X. performed experiments and data analysis. T. G. contributed reagents and participated in discussions. N. J. wrote the paper. C. H. directed the research, performed experiments, and wrote the paper.