Suppression of Tubulin Polymerization by the LKB1-Microtubule-associated Protein/Microtubule Affinity-regulating Kinase Signaling*

LKB1, a tumor suppressor gene mutated in the Peutz-Jeghers syndrome, encodes a serine/threonine protein kinase. Recent biochemical studies have shown that LKB1 activates 14 AMP-activated protein kinase-related kinases including MARKs (microtubule-associated protein/microtubule affinity-regulating kinases) that regulate microtubule dynamics. Here we show in vitro that LKB1 phosphorylates and activates MARK2, which in turn phosphorylates microtubule-associated protein Tau at the KXGS motif and suppresses tubulin polymerization. In cells, forced expression of LKB1 suppresses microtubule regrowth, whereas LKB1 knockdown accelerates it. We further show that the phosphorylation of Tau by the LKB1-MARK signaling triggers proteasome-mediated degradation of Tau. These results indicate that LKB1 is involved in the regulation of microtubule dynamics through the activation of MARKs.

LKB1 is a tumor suppressor gene encoding a serine/threonine protein kinase. Linkage analyses showed that mutations in LKB1 are responsible for the Peutz-Jeghers syndrome, an autosomal dominant disease characterized by gastrointestinal hamartoma, mucocutaneous pigmentation, and an increased risk of cancer (1)(2)(3). Somatic LKB1 mutations have been reported also in some sporadic cancers (4 -6). Phenotypes in Lkb1 knock-out mice further support that LKB1 is a tumor suppressor gene. Specifically, Lkb1 ϩ/Ϫ mice develop gastrointestinal hamartomas after Ͼ20 weeks of age and hepatocellular carcinomas after Ͼ50 weeks (7)(8)(9)(10).
Regarding the normal functions of LKB1, several lines of evidence indicate that LKB1 plays an important role in cell polarity. In Caenorhabditis elegans and Drosophila melanogaster, the LKB1 orthologs par-4 and lkb1, respectively, regulate cell polarity (11,12). Also in mammalian cells, the LKB1 activity can induce complete cell polarization (13), suggesting the evolutionary conservation of its role in cell polarity.
Interestingly, 6 of 14 kinases possibly downstream of LKB1, namely MARK1/2/3/4 and SAD-A/B, have been reported to regulate microtubule (MT) dynamics through phosphorylation of microtubule-associated proteins (MAPs; Tau, MAP2, and MAP4) at the conserved KXGS motifs that face the microtubule lattice (20 -22). It remains to be determined whether the other eight downstream kinases affect MT dynamics. The phosphorylation of KXGS motifs reduces the activity of microtubuleassociated proteins (MAPs) that stabilize MTs, presumably by weakening the electrostatic interaction between MAPs and MTs (20,23). Genetic analyses of D. melanogaster have suggested that Drosophila lkb1 has a functional link to the Drosophila MARK ortholog par-1 that regulates MT dynamics and cell polarity (12,24,25). Moreover, the genetic studies in C. elegans have revealed that the MARK ortholog par-1, in addition to the LKB1 ortholog par-4, is essential for cell polarization (11). Despite such circumstantial evidence, roles of LKB1 in MT dynamics have remained elusive. Here we provide both biochemical and cell biological evidence that LKB1 regulates MT dynamics through MARK, especially MARK2 that has been the best characterized in the MT-dependent cellular processes including establishment of cell polarity (26 -31).
Recombinant Adenoviruses-Recombinant adenoviruses were constructed by Adeno-X Expression System 1 (Clontech Laboratories, Mountain View, CA) according to the manufacturer's protocol. The titers of the recombinant adenoviruses were determined by the method previously reported by others (37). Every recombinant adenovirus was infected at a multiplicity between 25 and 50 plaque-forming units/cell.
Production of Recombinant MARK2 Proteins in Escherichia coli-To attach GST protein to the amino terminus of MARK2 (wild type (WT)/TA), the SalI/NotI fragment of the full-length cDNA encoding the wild-type or mutant MARK2 (TA) was inserted in the SalI/NotI-digested pGEX-6P-1 vector (GE Healthcare). One liter of LB medium was inoculated with E. coli (BL21 strain) containing the recombinant pGEX-6P-1 plasmid that encodes GST-MARK2 (WT) or GST-MARK2 (TA) and was incubated at 37°C in the presence of 100 g/ml ampicillin until A 600 reached 0.8. To induce expression of the GST-tagged proteins, isopropyl-␤-D-thiogalactopyranoside was added at 100 M. After culture for 24 additional hours at 20°C, cells were harvested, and the lysates were prepared in the lysis buffer (phosphate-buffered saline, 2% (v/v) Tween 20, 1% (v/v) Triton X-100, 1 mM EDTA, 1 mM dithiothreitol). After ultrasonic DNA shearing, the debris was removed from the lysates by centrifugation. The recombinant proteins were purified on Glutathione-Sepharose 4B beads (GE Healthcare) and then eluted from the resin by incubation with PreScission Protease (GE Healthcare) that recognized the specific amino acid sequence between GST and MARK2 at 320 units/ml in the cleavage buffer (50 mM Tris-HCl (pH 7.5), 500 mM NaCl, 1% (v/v) Brij-35, 1 mM EGTA, 1 mM dithiothreitol) at 4°C for 5 h.
In Vitro Phosphorylation Assay-Phosphorylation reactions were performed at 30°C in the phosphorylation reaction buffer Tubulin Polymerization Assay-Tubulin polymerization assay kit (Cytoskeleton, Denver, CO) was used according to the manufacturer's protocol. Polymerization was performed with 30 M tubulin and 0.3 M Tau in the polymerization buffer without glycerol. MARK2 (WT; 800 nM) was incubated with LKB1⅐STRAD␣⅐MO25␣ (25 nM) in the phosphorylation reaction buffer at 30°C for 15 min. Tau (3 M) was added, and the mixtures were incubated for an additional 15 min. The control reaction containing MARK2 protein (800 nM) was incubated without ATP. The samples were boiled for 3 min to inactivate LKB1 and MARK2 and then used for polymerization reaction. Tau is a heat-stable protein and remains active in the enhancement of tubulin polymerization even after the heat treatment. The turbidity was monitored at 340 nm every minute for 60 min at 37°C with a thermostatically controlled spectrophotometer Benchmark Plus Microplate Spectrophotometer (Bio-Rad). Each sample was determined in duplicate polymerization assays in three independent experiments.
Small Interfering RNA (siRNA) Vector and Stable LKB1 Knockdown in HepG2 Cell Lines-The human hepatoblastoma cell line HepG2 was obtained from Cell Resource Center for Biomedical Research, Tohoku University, and cultured in Dulbecco's modified Eagle's medium with high glucose (Sigma) supplemented with 10% (v/v) fetal bovine serum. As previously described (13), LKB1-pTER was constructed by inserting the siRNA oligonucleotides against LKB1 (5Ј-CGAAGAGAAGC-AGAAAATG-3Ј) into pTER vector (38). LKB1-pTER was converted into LKB1-pTERlox containing the hygromycin B resistance gene and loxP sites. HepG2 cells were transfected with the LKB1-pTERlox by using Effectene Transfection reagent (Qiagen, Hilden, Germany). Stable LKB1 knockdown cell lines were established by the hygromycin B (Invitrogen) selection (250 g/ml). These clones were infected with recombinant adenovirus containing Cre recombinase or LacZ gene. Clones #15H and #31H ("H" for high level of LKB1 expression) were cultured without hygromycin B, but clones #15L and #31L ("L" for low level of LKB1) were maintained in the presence of hygromycin B (250 g/ml). To minimize the effect of hygromycin B, #15L and #31L were incubated without hygromycin B for at least 24 h before the microtubule regrowth assays.
Proteasome Inhibitors and Lithium Treatment-MG132 (10 M; Sigma) or clasto-lactacystin ␤-lactone (5 M; Calbiochem) was added to the MEF3-2 cells 4 h before the harvest. For lithium treatment cells were incubated with 20 mM LiCl or NaCl simultaneously with the recombinant adenovirus infection.
Microtubule Regrowth Assay-Serum-starved MEF3-2 cells on the coverslips were treated with 33 M nocodazole (Sigma) for 2 h at 37°C. After 3 washes with Dulbecco's modified Eagle's medium pre-warmed at 30°C, the cells were incubated in fresh Dulbecco's modified Eagle's medium at 30°C. In experiments using HepG2, fetal bovine serum (10%; v/v) was added to the medium through the entire process. The cells were fixed in Ϫ20°C methanol for 3 min and processed for immunofluorescence staining.
Immunofluorescence Staining-Cells fixed in Ϫ20°C methanol for 3 min were re-hydrated in phosphate-buffered saline for 15 min. After blocking with 10% (v/v) goat serum and 3% (w/v) bovine serum albumin in phosphate-buffered saline for 1 h at room temperature, the coverslips were incubated with anti-␤tubulin and anti-␥-tubulin antibodies overnight at 4°C. After rinses with phosphate-buffered saline, the coverslips were incubated with Alexa Fluor 488 goat anti-rat and Alexa Fluor 594 goat anti-mouse secondary antibodies (Molecular Probes, Eugene, OR). The coverslips were then examined with Leica DM6000B fluorescence microscope (Leica Microsystems, Wetzlar, Germany) equipped with a 40ϫ objective. Images were captured with ORCA-ER cooled CCD camera (Hamamatsu Photonics, Hamamatsu, Japan). High magnification images were collected on Leica TCS SL laser scanning confocal microscope (Leica Microsystems) equipped with a 63ϫ oil-immersion objective.
Statistical Analysis-Statistical analyses were performed by using GraphPad PRISM version 4 (GraphPad Software, San Diego, CA). Values of p Ͻ 0.05 were considered significant.

Suppression of Tubulin Polymerization by LKB1 in Vitro-To
examine the potential role of LKB1 in MT dynamics, we first tested whether LKB1 would affect Tau phosphorylation in vitro. LKB1, together with STRAD␣ and MO25␣ subunits, phosphorylated the activation loop of protein kinase MARK2 at threonine 208 (Fig. 1A, WT) but not the T208A mutant (Fig. 1A, TA), consistent with a report by others (18). We then examined whether LKB1 activity could modulate the activity of MARK2 in phosphorylating microtubule-associated protein Tau at serine 262 of the KXGS motif (39). In MT dynamics, Ser-262 of Tau is reported to be the most important residue that is phosphorylated by MARK2. More specifically, Ser-262 phosphorylation reduces its affinity to MTs by ϳ75%, thereby suppressing its effects on tubulin polymerization (39). As expected, LKB1 increased Tau phosphorylation at Ser-262 through MARK2 (Fig. 1B). In contrast, in the absence of LKB1, both wild-type MARK2 (WT) and mutant MARK2 (TA) showed only low and basal-level kinase activity on Tau. In addition, LKB1 failed to activate the mutant MARK2 (TA), and LKB1 by itself did not directly phosphorylate Tau at Ser-262. These results show that LKB1 stimulates Tau phosphorylation through MARK2 in vitro. We, therefore, examined the roles of LKB1 on the in vitro tubulin polymerization reaction in the presence of 0.3 M Tau protein (Fig. 1C). Under this assay condition, in the absence of Tau, no tubulin polymerization was observed, but the addition of Tau (0.3 M) into the reaction mixture immediately induced tubulin polymerization (data not shown). As a control, MARK2 in the absence of ATP showed a tubulin polymerization curve with a short nucleation phase followed by a fast growth phase and, finally, a steady phase with the maximum polymer mass (Fig. 1C, open circles). MARK2, in the presence of ATP, significantly inhibited tubulin polymerization due to its basal activity (Fig. 1, C-E, open squares;  see below), and further addition of LKB1 markedly enhanced the inhibitory effects of MARK2 in a dose-dependent man-ner (Fig. 1C, closed circles, open triangles, and closed  squares). LKB1 alone exerted only a slight effect on tubulin polymerization in the presence of ATP (Fig. 1C, closed triangles). To quantify the data, we determined the mean initial rate of tubulin polymerization for each curve (Fig. 1D). The basal activity of MARK2 (Fig. 1D, open square) reduced the initial rate by ϳ35% of the control (Fig. 1D, open circle). LKB1 significantly enhanced the MARK2-mediated reduction in the initial rates in a dose-dependent manner (Fig. 1D,  closed circle, by ϳ40%; open triangle, by ϳ65%; closed square, by ϳ80%), although LKB1 alone showed only a minimal effect (Fig. 1D, closed triangle). In addition, we found that these initial rates were inversely correlated with the phosphorylation levels of MARK2 (Thr-208) and Tau (Ser-262), respectively, as shown in Fig. 1E. Collectively, these results indicate that LKB1 suppresses the Tau-mediated tubulin polymerization through the phosphorylation and activation of MARK2.
Expression of LKB1 Suppresses MT Regrowth in Lkb1-null MEFs-Adenovirus-induced expression of LKB1 or MARK2 in MEF3-2 did not lead to marked alterations in the steady state MT cytoskeleton morphology, as monitored by the anti-␤-tubulin immunofluorescence (data not shown). We then investigated the effects of these kinases on MT regrowth. When incubated for 2 h with 33 M nocodazole, a MT-destabilizing reagent, all cells showed depolymerization of MTs (Figs. 3A and 4A, 0 min). After washing out nocodazole, MTs re-grew rapidly from centrosomes in the control cells infected with Adv-Cre alone, but infection of Adv-LKB1 (WT) suppressed MT regrowth substantially (Fig. 3, A and B). Adv-MARK2 (WT) caused similar suppression of MT regrowth, but to a lesser extent (Fig. 4, A and B). Adv-LKB1 (KD) stimulated the MT regrowth, probably because of a dominant-negative effect on MARKs in MEF3-2, whereas Adv-MARK2 (KD) did not affect the MT regrowth (Figs. 3, A and B, and 4, A and B; see "Discus-  AUGUST 10, 2007 • VOLUME 282 • NUMBER 32 sion"). Taken together, these results from Lkb1-null MEF3-2 are consistent with those of the in vitro tubulin polymerization experiments (Fig. 1, C and D) and implicate the LKB1-MARK signaling in MT dynamics in vivo as well.

LKB1 and Microtubules
Degradation of Tau Triggered by LKB1-To further investigate the mechanism by which the LKB1-MARK signaling regulates MT dynamics, we constructed a recombinant adenoviral construct containing the cDNA for Tau. As expected, MEF3-2 infected with Adv-Tau expressed Tau protein (Fig. 5A, left  panel). However, co-expression of LKB1 (WT) and Tau led to a marked decrease in the Tau protein level (Fig. 5A, left panel). Co-expression of MARK2 (WT) also reduced the Tau protein level (Fig. 5A, middle panel). We then found that expression of MARK2 (KD) restored the Tau protein level in the LKB1-expressing cells in a dose-dependent manner (Fig. 5A, right  panel). These results strongly suggested that LKB1 acted as an upstream kinase of MARK2 and decreased the Tau protein level in MEF3-2. We then examined whether the LKB1-MARK2 signaling reduced the mRNA level of Tau and found that neither LKB1 nor MARK2 kinase activity affected it (Fig. 5B). Thus, we speculated that the LKB1-MARK2 signaling might promote degradation of Tau rather than its synthesis. As anticipated, the addition of the proteasome inhibitor MG132 or lactacystin significantly increased the Tau protein level in MEF3-2 cells infected with Adv-Tau alone and in the cells co-infected with Adv-Tau and Adv-LKB1 (WT) (Fig.  5C). On the other hand, the inhibitors hardly increased the Tau level in Adv-LKB1 (KD)-infected cells. These results suggested that the LKB1-MARK-mediated phosphorylation of Tau signaled for its degradation by the proteasome. Consistent with this interpretation, the proteasome inhibitors increased the level of phospho-Tau (Ser(P)-262/ 356) in the cells co-infected with Adv-Tau and Adv-LKB1 (WT) (Fig.  5C). In addition, the Tau mutant (S2A, namely, S262A/S356A) that lacked the KXGS motifs became refractory to degradation evoked by LKB1 and MARK2 (Fig. 5D). These results collectively indicate that phosphorylation of the KXGS motifs by the LKB1-MARK2 signaling is the prerequisite to Tau degradation in MEF3-2.
Tau protein can be phosphorylated not only by MARKs but also by other kinases including cyclindependent kinase 5 and glycogen synthase kinase-3␤ (GSK-3␤) (41), and GSK-3␤ plays a critical role in Tau degradation (42). In addition, several reports suggest crosstalks between the LKB1-MARK pathway and the Wnt pathway in both directions (43)(44)(45)(46)(47). We, therefore, evaluated the possible involvement of GSK-3␤ in the LKB1-mediated degradation of Tau. In the presence of a GSK-3␤ inhibitor, LiCl, the slowly migrating bands of Tau disappeared, probably due to inhibition of phosphorylation by GSK-3␤ (Fig. 5E), whereas LiCl treatment did not abolish Tau degradation induced by LKB1. These results suggest that the degradation of Tau promoted by LKB1 is independent of the GSK-3␤ activity in MEF3-2.
Accelerated Tubulin Polymerization in LKB1 Knockdown Cells-In the gain-of-function experiments using the recombinant adenoviruses, we demonstrated that the LKB1-MARK signaling regulates MT dynamics and Tau degradation in MEF3-2. However, loss-of-function experiments are equally important in elucidating the biological relevance of the LKB1-MARK signaling, because reduction in the LKB1 gene dosage shows serious effects. Specifically, Lkb1 haploinsufficiency in the heterozygous knock-out mice causes gastrointestinal hamartomas, whereas loss of heterozygosity results in hepatocellular carcinomas (7-10). Thus, we set up experiments using a siRNA against LKB1 in HepG2, a human hepatoblastoma cell line. To minimize the clonal variation, we designed an siRNA construct (LKB1-pTERlox; Fig. 6A) based on the a strategy shown in Fig. 6B. Briefly, we first established two stable LKB1-knockdown HepG2 clones (#15 and #31). Transient expression of Cre recombinase by Adv-Cre removed the cassette expressing the siRNA from LKB1-pTERlox in these cells, generating two rescued clones (#15H and #31H). LKB1 expression of #15H and #31H clones were recovered to nearly equal levels to those in the parental HepG2 cells. We also obtained control cell clones with Adv-LacZ that maintained low LKB1 level (#15L and #31L; ϳ5% of those in #15H and #31H, respectively; Fig. 6C). Using these clonal cell lines, we then performed MT regrowth experiments. The MT network almost disappeared by a 2-h treatment with 33 M nocodazole in both #31H and #31L clones (Fig. 6D, 0 min). After washing out nocodazole, the LKB1 knockdown clone #31L showed faster tubulin polymerization from the centrosomes than the control clone #31H (Fig. 6D, 4, 8, and 16 min). We obtained similar results using clones #15H and #15L as well (data not shown). These results from the loss-of-function experiments indicate that the endogenous LKB1 activity suppresses MT regrowth in the HepG2 cells, which is consistent with the results of the in vitro tubulin polymerization assay and of the gain-of-function experiments with forced expression of LKB1 in MEF3-2 cells.

DISCUSSION
In this report we have provided evidence that LKB1 suppresses tubulin polymerization through activation of MARK in vitro and in cultured cells. First, we showed in a cell-free system that LKB1 enhanced Tau phosphorylation through MARK2 (Fig. 1B). Second, we demonstrated that activation of the LKB1-  MARK2 pathway suppressed tubulin polymerization through Tau phosphorylation (Fig. 1C). Third, expression of LKB1 or MARK2 suppressed MT regrowth in cultured cells (Figs. 3 and  4), whereas LKB1 knockdown accelerated it (Fig. 6). To our knowledge, this is the first experimental demonstration that LKB1 regulates MT dynamics in vitro and in vivo.
Biological roles of MARK2 have been well documented especially in the MT-dependent cellular processes including intracellular trafficking and cell polarity (26 -31). Studies using Madin-Darby canine kidney (MDCK) cells have shown that MARK2 localizes to the lateral domains where it reorganizes MTs that serve as tracks for the intracellular delivery (26,30). Inhibition of MARK2 activity impairs formation of the apical domain in MDCK cells, probably because of disordered apical delivery (29,30). LKB1 and its orthologs are considered as key regulators of cell polarity (11)(12)(13). Our findings that LKB1 regulates MT dynamics through MARK2 are consistent with the possible role of MARKs in LKB1-induced cell polarization.
Our results suggest that phosphorylation of the KXGS motifs (Ser-262/Ser-356) of Tau by the LKB1-MARK signaling causes proteasome-mediated degradation of Tau (Fig. 5). Recent studies indicate that Tau, only when phosphorylated (42), is recognized and ubiquitinated by the E3 ligase CHIP (carboxyl terminus of the Hsc70-interacting protein) in its degradation process (42,48). In addition, various lines of evidence suggest that phosphorylation of the KXGS motifs may be an early step toward both the polarization of neurons in neural development and the formation of hyperphosphorylated Tau aggregates in Alzheimer disease (21,32,36,49,50). It remains to be investigated whether Tau degradation induced by LKB1 plays essential roles in neuronal development and Alzheimer disease.
Regarding the LKB1-MARK signaling in living cells, four issues are worth noting. First, adenovirus-mediated expression of LKB1 (WT/KD) or MARK2 (WT/KD) had no detectable effects on MT regrowth in MDCK epithelial cells and IEC-6 (rat intestinal epithelial cells) endogenously expressing LKB1 (data not shown). The unresponsiveness of these cell lines to the introduced genes requires further studies. Second, in Lkb1-null MEF3-2, Adv-LKB1 (WT) suppressed MT regrowth substantially as did Adv-MARK2 (WT), although to a lesser extent. The weaker response of MEF3-2 to the exogenous MARK2 than to LKB1 also remains to be investigated further. Third, in Lkb1-null MEF3-2, Adv-LKB1 (KD) accelerated MT regrowth substantially as compared with the Adv-Cre control (Fig. 3). We also found that LKB1 (KD) induced the accumulation of Tau to a level much higher than that induced by the Adv-Cre control (Fig. 5). One possible interpretation is that LKB1 (KD) may have a domi-  (Fig. 5). In the right panel, increasing amounts of Adv-MARK2 (KD) were used for infection together with LKB1 (WT)-and Tau-expressing viruses. Cell lysates were analyzed by Western blotting for Tau, LKB1, and MARK2 as well as ␤-actin as a loading control. The V5 antibody recognized only the exogenously expressed MARK2. Arrows, prominent Tau bands. B, neither LKB1 nor MARK2 affects the Tau mRNA level. Total RNA samples were isolated from MEF3-2 cells, and the mRNA levels for Tau and glyceraldehyde-3-phosphate dehydrogenase (GAPDH, internal control) were analyzed by PCR, respectively. We did not detect any bands in the reverse transcription (RT)-minus controls (data not shown). C, treatment with proteasome inhibitors restores the Tau protein level in MEF3-2 co-infected with Adv-LKB1 (WT) and Adv-Tau. At  nant-negative effect on the basal activity of MARK2. Our reverse transcription-PCR analysis showed that another kinase upstream of MARK, MARK kinase (MARKK), was expressed in MEF3-2. 4 MARKK, a member of the Ste20 kinase family, was isolated from pig brain during a search for the kinases upstream of MARK2 (34). MARKK and LKB1 phosphorylate MARK2 at the same residue (Thr-208) (18,34), whereas it remains to be determined whether MARKK can phosphorylate the other 13 AMPK-related kinases in addition to MARK2 (20). LKB1 (KD) may stably bind to MARK2 and interfere with the phosphorylation of MARK2 by MARKK in MEF3-2 cells. Finally, regarding Tau degradation, we also found that LKB1 (KD) increased the levels of phosphorylated Tau (Ser(P)-262/356) in the presence of proteasome inhibitors (Fig. 5C). The dominant negative effect of LKB1 (KD) fails to explain this part of the results. Another possible interpretation is that LKB1 (KD) may suppress recruitment of Tau to the proteasome or the proteasome activity in MEF3-2 cells. Further investigation is needed to explain the mechanism behind this phenomenon.
In conclusion, we have shown that LKB1 suppresses tubulin polymerization through the activation of MARK both in vitro and in vivo and that LKB1-MARK signaling is involved in Tau degradation. These results may help elucidate the roles of LKB1 in the MT-dependent cellular processes such as intracellular trafficking and cell polarity.