Research ArticleCell Biology

New Roles for the LKB1-NUAK Pathway in Controlling Myosin Phosphatase Complexes and Cell Adhesion

See allHide authors and affiliations

Sci. Signal.  30 Mar 2010:
Vol. 3, Issue 115, pp. ra25
DOI: 10.1126/scisignal.2000616

Abstract

The AMPK-related kinases NUAK1 and NUAK2 are activated by the tumor suppressor LKB1. We found that NUAK1 interacts with several myosin phosphatases, including the myosin phosphatase targeting-1 (MYPT1)–protein phosphatase-1β (PP1β) complex, through conserved Gly-Ile-Leu-Lys motifs that are direct binding sites for PP1β. Phosphorylation of Ser445, Ser472, and Ser910 of MYPT1 by NUAK1 promoted the interaction of MYPT1 with 14-3-3 adaptor proteins, thereby suppressing phosphatase activity. Cell detachment induced phosphorylation of endogenous MYPT1 by NUAK1, resulting in 14-3-3 binding to MYPT1 and enhanced phosphorylation of myosin light chain-2. Inhibition of the LKB1-NUAK1 pathway impaired cell detachment. Our data indicate that NUAK1 controls cell adhesion and functions as a regulator of myosin phosphatase complexes. Thus, LKB1 can influence the phosphorylation of targets not only through the AMPK family of kinases but also by controlling phosphatase complexes.

Introduction

Loss-of-function mutations in the LKB1 kinase tumor suppressor gene result in the rare inherited Peutz-Jeghers cancer syndrome, in which patients are predisposed to develop benign as well as malignant tumors (1). LKB1 is also often mutated in sporadic cancers, most frequently in lung adenocarcinomas (2) and papillomavirus-induced cervical cancer (3). LKB1 is activated by forming a heterotrimeric complex with the pseudokinase STRAD and the armadillo repeat adaptor protein MO25 (46). Most evidence indicates that LKB1 exerts its physiological effects by phosphorylating and activating a group of 14 protein kinases that belong to the adenosine monophosphate–activated protein kinase (AMPK) subfamily (7). Of these, the best characterized are AMPKα1 and AMPKα2, which are activated by LKB1 following increases in 5′-AMP concentrations in energy-deprived cells (8). Once activated, AMPKα1 and AMPKα2 phosphorylate various proteins to restore energy stores and stimulate the transport of glucose and other nutrients into cells (9). There are 12 other LKB1-activated kinases that are collectively termed AMPK-related kinases (10, 11). In contrast to AMPKα1 or AMPKα2, the AMPK-related kinases are not stimulated by energy stress (12). LKB1 activates AMPK and AMPK-related kinases by phosphorylating a conserved Thr residue located within the T loop of the kinase domain.

Less is understood about the regulation and function of AMPK-related kinases than about the intensively studied AMPKα1 and AMPKα2. Genetic analysis suggests that microtubule affinity-regulating kinase (MARK) isoforms regulate cell polarity (13, 14), whereas the SAD [synapses of the amphid defective; also known as BRSKs (brain-specific kinases)] enzymes control axon initiation during neuronal polarization (15, 16). The abundance of NUAK1 is reportedly increased in colorectal tumors and may promote invasiveness (17). NUAK1 has also been implicated in the regulation of cell senescence (18) and suppression of apoptosis (19). Nuak1 knockout embryos perish at day 14.5 and display omphalocele, a defect of abdominal wall closure in which the organs remain outside of the abdomen (20). Most Nuak2-deficient mice perish at day 16.5 of embryogenesis as a result of exencephaly, a defect in which the brain is located outside of the skull (21). To learn more about the roles of NUAK1, we searched for interacting partners and found that NUAK1 associates with myosin phosphatase complexes through its ability to bind protein phosphatase-1β (PP1β). Our data indicate that NUAK1 phosphorylates the myosin phosphatase targeting-1 (MYPT1) regulatory subunit of this complex, promoting its interaction with 14-3-3 adaptor proteins, which leads to a suppression of dephosphorylation of myosin regulatory light chain-2 (MLC2). We also provide evidence that inhibition of NUAK1 promotes cell adhesion. Finally, we demonstrate that dissociation of cells induces phosphorylation of MYPT1 by NUAK1, leading to 14-3-3 binding and increased phosphorylation of MLC2.

Results

The MYPT-PP1β complex interacts with NUAK1 and NUAK2

Mass spectrometry was used to identify proteins in immunoprecipitates from human embryonic kidney (HEK) 293 cells stably expressing NUAK1 (Fig. 1A and table S1). A number of proteins were identified, including the deubiquitinating enzyme USP9X (ubiquitin-specific peptidase 9, X-linked), a NUAK1 interactor (12, 22). Various myosin phosphatase complexes were also detected in NUAK1, but not in control immunoprecipitates. These included the catalytic subunit of PP1β, as well as three of its regulatory subunits, namely MYPT1, MYPT2, and MBS85 (myosin-binding subunit of 85 kD). A previous study also reported that NUAK2 interacted with MYPT1, but the mechanism of interaction was not characterized further (23).

Fig. 1

NUAK1 binds to the MYPT1-PP1β complex. (A) HEK293 T-rex cells without (Control) or with inducible overexpression of HA-NUAK1 were induced with tetracycline (0.1 μg/ml) 24 hours before lysis. HA-NUAK1 was immunoprecipitated (IP) and electrophoresed on a polyacrylamide gel that was stained with colloidal Coomassie. The gel was divided into the indicated pieces, and proteins in these pieces were identified by mass spectrometry. The data obtained are summarized in table S1. MYPT1, MYPT2, and MBS85 were identified in band 8, NUAK1 in bands 9 and 10, and PP1β in band 11. (B) Endogenous NUAK1 or NUAK2 were immunoprecipitated from HEK293 cell lysate and subjected to immunoblotting with the indicated antibodies. Preimmune IgG was used as a negative control. (C) As in (B), except endogenous MYPT1 was immunoprecipitated. (D) As in (B), except endogenous PP1β was immunoprecipitated. (E) Lysates and NUAK1 immunoprecipitates from HEK293 cells stably overexpressing shRNA directed against MYPT1 were analyzed as in (B). (F) Domain structure of MYPT family members and 53BP2. (G) HEK293 cells were transfected with expression plasmids for the indicated FLAG-tagged proteins. Thirty-six hours after transfection, cells were lysed and FLAG-tagged proteins were immunoprecipitated. Immunoprecipitates and total cell lysates were analyzed by immunoblotting with the indicated antibodies. Similar results were obtained in at least three separate experiments for all data shown in this figure.

MYPT1 and PP1β form a complex that regulates the dephosphorylation of MLC2, as well as that of other substrates, such as ezrin-radixin-moesin (ERM) proteins and Polo-like kinase-1 (PLK1) (2427). Endogenous MYPT1 and PP1β coimmunoprecipitated with endogenous NUAK1 and NUAK2, which is less abundant than NUAK1 (Fig. 1B). Endogenous NUAK1 also coimmunoprecipitated with endogenous MYPT1 (Fig. 1C). Endogenous PP1β, but not PP1α or PP1γ isoforms, coimmunoprecipitated with endogenous NUAK1 (Fig. 1B) or overexpressed NUAK1 and NUAK2 (fig. S1). Moreover, endogenous NUAK1 immunoprecipitated with endogenous PP1β (Fig. 1D). We also observed that short hairpin–mediated RNA (shRNA)–mediated knockdown of MYPT1 in HEK293 cells reduced the amount of PP1β that coimmunoprecipitated with NUAK1 (Fig. 1E), suggesting that NUAK1 binds principally to MYPT1-PP1β complex. All five members of the MYPT family contain ankyrin repeats (Fig. 1F). Four of these, namely MYPT1, MYPT2, MBS85, and TIMAP, interacted with NUAK1 as well as PP1β (Fig. 1G). In contrast, MYPT3 and 53BP2, a distinct PP1 regulatory subunit (28), interacted with PP1β but not NUAK1 (Fig. 1G).

Conserved Gly-Ile-Leu-Lys motifs mediate binding of NUAK1 to the MYPT1-PP1β complex

We identified a fragment of NUAK1 (residues 390 to 425) containing a conserved Gly-Ile-Leu-Lys (GILK) motif that interacted with endogenous MYPT1-PP1β (fig. S2). The C-terminal region of NUAK1 contains three evolutionarily conserved GILK motifs at residues 399 to 402 (site 1), 466 to 469 (site 2), and 523 to 526 (site 3) (fig. S3). An alanine scan analysis in which GILK motif residues and adjacent residues were mutated in a fragment of NUAK1(1–425) containing only the site 1 GILK motif revealed that mutation of the GILK motif residues, but not that of adjacent residues, impaired binding to endogenous MYPT1-PP1β complex (Fig. 2A). Changing the GILK motif to Gly-Ile-Lys-Lys (GIKK) or Gly-Lys-Lys-Lys (GKKK) ablated the interaction of NUAK1(1–425) with endogenous MYPT1-PP1β (Fig. 2B). Mutation of all three GILK motifs in full-length NUAK1 was required to prevent interaction with endogenous MYPT1-PP1β (Fig. 2C). Mutation of any two of the GILK motifs partially reduced interaction, whereas individual mutations had little effect. A biotinylated peptide encompassing the site 1 GILK motif affinity-purified endogenous MYPT1-PP1β from HEK293 cell extracts (Fig. 2D). Mutation of any of the GILK motif residues within this peptide reduced or blocked the capture of MYPT1-PP1β (Fig. 2E). Moreover, incubation of MYPT1 immunoprecipitates with site 1 GILK peptide led to the dissociation of NUAK1 under conditions in which mutant peptides were less efficient (Fig. 2F). NUAK1 and NUAK2 are the only AMPK family kinases that contain GILK motifs, and we found that endogenous MYPT1-PP1β interacted with NUAK1 and NUAK2, but not with nine other AMPK-related kinases tested (Fig. 2G) or AMPKα1 or AMPKα2 (fig. S4).

Fig. 2

GILK motifs of NUAK1 mediate interaction with the PP1β subunit of the MYPT1-PP1β complex. (A and B) HEK293 cells were transfected with expression plasmids encoding either wild type (wt) or point mutants of HA-NUAK1(1–425). Thirty-six hours after transfection, cells were lysed and HA-tagged proteins were immunoprecipitated and analyzed by immunoblotting with the indicated antibodies. (C) As in (A) and (B), except that cells were transfected with full-length NUAK1 wild type (wt) or mutants with the indicated GILK sites mutated to GKKK. Site 1 residues, 399 to 402; site 2 residues, 466 to 469; site 3 residues, 523 to 526. (D) The indicated amounts of biotinylated GILK peptide (residues 388 to 407 of NUAK1, biotin-SPSKLSSKRPKGILKKRSNS) or 10 μg of a control peptide (residues 1006 to 1023 of WNK4, biotin-SEEGKPQLVGRFQVTSSK) were conjugated to streptavidin-Sepharose (10 μl). The conjugates or empty resin were incubated with HEK293 cell lysate, washed, and subjected to immunoblotting analysis with the indicated antibodies. (E) As in (D), except that the indicated variants of the GILK peptide were coupled to streptavidin-Sepharose. (F) HEK293 cells were transfected with an expression plasmid for FLAG-MYPT1. Thirty-six hours after transfection, cells were lysed and the indicated peptides were added to the lysis buffer. FLAG-MYPT1 immunoprecipitates (IP) and supernatants (SUP) were immunoblotted with the indicated antibodies. (G) HEK293 cells were transfected with expression plasmids for the indicated HA-tagged AMPK-related kinases. Thirty-six hours after transfection, cells were lysed and HA-tagged proteins were immunoprecipitated and subjected to immunoblot analysis. (H) As in (A) and (B), except cells were transfected with MYC-PP1β wild type (wt) or the indicated mutants. (I) As in (A) and (B), except cells were transfected with constructs encoding either FLAG-MYPT1 wild type (wt) or the KVKF-MYPT1 mutant (F38A) that is unable to interact with PP1. (J) As in (F), except cells were untransfected and endogenous PP1β was immunoprecipitated from cell lysates. Similar results were obtained in at least three separate experiments for all data shown in this figure. Single-letter abbreviations for the amino acid residues are as follows: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; and Y, Tyr.

GILK motifs in NUAK1 bind directly to PP1β

Database searches for other proteins containing GILK motifs identified inhibitor-2, a PP1-binding protein. Mutational analysis (29) and the crystal structure of the PP1–inhibitor-2 complex (30) reveal that inhibitor-2 interacts with PP1 through its GILK motif. This suggests that NUAK1 might interact with the MYPT1-PP1β complex through the GILK motifs of NUAK1 binding to PP1β. Comparison of the crystal structures of the PP1γ–inhibitor-2 complex (PDB 2O8G) (30) and the PP1β-MYPT1 complex (PDB 1S70) (31) revealed that PP1 has two distinct hydrophobic pockets that bind to the GILK motif of inhibitor-2 and the Lys-Val-Lys-Phe (KVKF) motif of MYPT1. We designed mutants of the GILK-binding pocket on PP1β that prevented binding to inhibitor-2 but not to MYPT1 (Leu54→Ala and Phe118→Ala), and mutants of the KVKF-binding pocket that inhibited binding to MYPT1 but not to inhibitor-2 (Asp241→Ala and a double mutant, Phe257→Leu and Phe293→Leu) (Fig. 2H). Strikingly, mutation of the GILK-binding pocket, but not that of the KVKF-binding pocket, prevented binding of NUAK1 to PP1β (Fig. 2H). To confirm that the association of MYPT1 with NUAK1 was mediated by NUAK1 interacting with PP1β, we disrupted the PP1-binding KVKF motif on MYPT1 (32) and observed that this prevented binding to NUAK1 (Fig. 2I). Incubation of PP1β immunoprecipitates with a GILK motif peptide, but not a control peptide, induced dissociation of NUAK1 but not of MYPT1 (Fig. 2J). In vitro reconstitution studies with recombinant MYPT1, PP1β, and NUAK1 also confirmed that NUAK1 was only able to associate with MYPT1 in the presence of PP1β (fig. S5, A and B). Furthermore, a mutant NUAK1 lacking all GILK motifs was unable to associate with MYPT1 and PP1β in vitro (fig. S5B). Consistent with NUAK1 and inhibitor-2 binding to the same site in PP1β, overexpression of inhibitor-2 in HEK293 cells reduced the interaction of NUAK1 with PP1β (fig. S5C).

MYPT1 inhibits PP1 from dephosphorylating NUAK1

To determine whether NUAK1 bound to the MYPT1-PP1β complex was active, we incubated active recombinant NUAK1 purified from insect cells with the free catalytic subunit of PP1β and observed that NUAK1 was dephosphorylated at its T loop Thr211 residue and inactivated. In parallel reactions, PP1 also dephosphorylated the T loop Ser382 of the kinase WNK1 [with-no-K(Lys) kinase-1] (Fig. 3A). The addition of increasing amounts of MYPT1 to the reaction inhibited dephosphorylation and inactivation of NUAK1 without affecting dephosphorylation of WNK1 (Fig. 3A). In cells, NUAK1 bound to the MYPT1-PP1β complex was active because NUAK1 was phosphorylated at its T loop residue within the immunoprecipitated complex (fig. S6).

Fig. 3

NUAK1 phosphorylates MYPT1 at Ser445, Ser472, and Ser910. (A) PP1β was preincubated with the indicated amounts of MYPT1 and then used to dephosphorylate recombinant NUAK1 or GST-WNK1(1–661). Proteins were analyzed by immunoblotting with the indicated antibodies. NUAK1 activity was assayed with a peptide kinase assay. pT211, phosphorylated Thr211; pS382, phosphorylated Ser382. (B) HEK293 cells were transfected with either wild type (wt) or kinase-inactive (ki, D196A) HA-NUAK1. Thirty-six hours after transfection, cells were lysed and HA-NUAK1 was immunoprecipitated. Immunoprecipitates were incubated with recombinant MYPT1 and Mg2+-[γ-32P]ATP for the indicated times and subjected to electrophoresis (on a polyacrylamide gel) and autoradiography. (C) As in (B), except that MYPT1 was replaced with recombinant PP1β and the reaction mixture contained microcystin-LR (15 μg/ml) to inhibit phosphatase activity. (D) Recombinant MYPT1 was incubated with NUAK1 in the presence of Mg2+-[γ-32P]ATP for 30 min. Phosphorylated MYPT1 was digested with AspN and peptides were separated by reversed-phase high-performance liquid chromatography on a C18 column. The peaks containing the 32P-labeled phosphopeptides are labeled with the identified phosphorylated residue. (E) Summary of the mass spectrometry and solid-phase Edman sequencing data obtained after analysis of the peak fractions. The deduced amino acid sequence of each peptide is shown and the phosphorylated residue is indicated (bold). (F) Domain structure of MYPT1 with the position of residues phosphorylated by NUAK1 (Ser445, Ser472, and Ser910), previously reported ROCK phosphorylation sites (Thr696 and Thr853), and the cdc2 phosphorylation site (Ser473) indicated. All numbering refers to the human sequence. (G) HEK293 cells were transfected with either wild-type (wt) or the indicated mutants of FLAG-MYPT1. Thirty-six hours after transfection, cells were lysed and FLAG-MYPT1 was immunoprecipitated. Immunoprecipitates (IP) were incubated with His-NUAK1 and Mg2+-[γ-32P]ATP for 30 min and analyzed as in (B). The SSS/AAA mutant corresponds to a triple MYPT1 mutant in which Ser445, Ser472, and Ser910 are mutated to Ala. Similar results were obtained in at least three separate experiments for all data shown in this figure.

NUAK1 phosphorylates MYPT1 at conserved Ser445, Ser472, and Ser910 residues

Wild-type, but not catalytically inactive, NUAK1 phosphorylated MYPT1 in a time-dependent manner to 3.0 ± 0.3 mol of phosphate per mole of MYPT1 (Fig. 3B). It was also previously reported that NUAK2 phosphorylated MYPT1, but the sites were not mapped (23). In parallel reactions, NUAK1 failed to phosphorylate microcystin-LR–inactivated PP1β (Fig. 3C). [32P]MYPT1 phosphorylated by NUAK1 was digested with the endoproteinase AspN and analyzed by chromatography on a C18 column. Three major 32P-labeled phosphopeptides were observed (Fig. 3D), and solid-phase Edman sequencing and mass spectrometry identified the phosphorylation sites as Ser445, Ser472, and Ser910 (Fig. 3E). These sites are located C-terminal to the ankyrin repeats (Fig. 3F) and are conserved in vertebrate but not invertebrate species (fig. S7). Furthermore, Ser445, Ser472, and Ser910 lie in favorable AMPK family phosphorylation motifs, with conserved hydrophobic amino acids at the −5 and +4 positions and an Arg or Lys residue at the −3 position from the phosphorylation site (33, 34) (fig. S7). Individual mutations of Ser445, Ser472, or Ser910 slightly reduced the phosphorylation of MYPT1 by NUAK1, whereas the mutation of all three residues inhibited phosphorylation to a greater degree (Fig. 3G). MYPT1 was also phosphorylated by NUAK1 when it was in complex with PP1β (fig. S8). Consistent with PP1β not dephosphorylating NUAK1 when complexed with MYPT1 (Fig. 3A and fig. S6), the addition of microcystin-LR to inhibit PP1β did not enhance phosphorylation of MYPT1 by NUAK1 (fig. S8).

The LKB1-NUAK network regulates MYPT1 phosphorylation and interaction with 14-3-3 isoforms

We generated phosphospecific antibodies against the NUAK1 phosphorylation sites on MYPT1 and used these reagents to demonstrate that MYPT1 overexpressed in HEK293 cells was phosphorylated at Ser445, Ser472, and Ser910. The specificity of the phosphospecific antibodies was established by demonstrating that mutation of the phosphorylation site to Ala abolished antibody recognition (Fig. 4A). This analysis also revealed that mutation of Ser445 to Ala prevented phosphorylation of Ser472 but not that of Ser910, whereas mutation of Ser472 or Ser910 did not affect phosphorylation of other residues (Fig. 4A). It was previously reported that phosphorylation of Ser472 on MYPT1 stimulated interaction with 14-3-3 isoforms (35). Consistent with this finding, wild-type MYPT1, but not mutant MYPT1[S472A], bound to 14-3-3 isoforms, as determined by coimmunoprecipitation and overlay assays (Fig. 4A). Because mutation of Ser445 abolishes Ser472 phosphorylation, the MYPT1[S445A] mutant also failed to interact with 14-3-3 isoforms (Fig. 4A). Mutation of Ser445, Ser472, and Ser910 individually or in combination did not affect interaction with PP1β or NUAK1 (Fig. 4A).

Fig. 4

MYPT1 is phosphorylated in cells in an LKB1- and NUAK1-dependent manner. (A) HEK293 cells were transfected with expression plasmids for wild-type (wt) or the indicated mutants of FLAG-MYPT1. Thirty-six hours after transfection, cells were lysed and FLAG-tagged proteins were immunoprecipitated and analyzed by immunoblotting with the indicated antibodies and a 14-3-3 overlay assay. Total cell lysates from transfected cells were also subjected to immunoblotting. MYPT1 SSS/AAA corresponds to a triple MYPT1 mutant with Ser445, Ser472, and Ser910 mutated to Ala. (B) Lkb1+/+ (wild-type) or Lkb1−/− (knockout) MEFs were lysed and analyzed by immunoblotting with the indicated antibodies. Endogenous MYPT1 was immunoprecipitated and its ability to interact with 14-3-3 proteins was assayed by a 14-3-3 overlay assay. Endogenous LKB1 or NUAK1 was also immunoprecipitated from MEF cell lysate and its activity was assayed. (C) As in (B), except that Nuak1+/+ or Nuak1−/− MEFs were used. (D) As in (B), except that HeLa T-rex cells stably overexpressing empty vector, wild-type (wt) LKB1, or kinase-inactive (ki) LKB1[D194A] were used. (E) As in (B), except that HEK293 cells stably overexpressing shRNA against NUAK1 were used. (F) HEK293 T-rex cells with tetracycline-inducible overexpression of HA-NUAK1 wild type (wt), HA-NUAK1 with all three GILK motifs mutated to GKKK (3xIL/KK), or kinase-inactive HA-NUAK1[D196A] were induced with tetracycline (TET; 0.1 μg/ml) 24 hours before lysis. Cell lysates were analyzed by immunoblotting with the indicated antibodies. The kinase activity of NUAK1 was assayed in immunoprecipitates. Similar results were obtained in at least three separate experiments for all data shown in this figure.

We next studied the phosphorylation of MYPT1 in LKB1 (Fig. 4B), as well as NUAK1 (Fig. 4C) wild-type and knockout mouse embryonic fibroblasts (MEFs). In wild-type MEFs, LKB1 and NUAK1 activity was detected, as was phosphorylation of endogenous MYPT1 at Ser445, Ser472, and Ser910 (Fig. 4, B and C). In contrast, in Lkb1−/− MEFs, no NUAK1 activity was detected and phosphorylation of MYPT1 at Ser445, Ser472, and Ser910 was reduced (Fig. 4B). Similarly, in Nuak1−/− MEFs, phosphorylation of MYPT1 was also suppressed (Fig. 4C). Moreover, the ability of MYPT1 to bind 14-3-3 isoforms in an overlay assay was also decreased in Lkb1−/− and Nuak1−/− cells (Fig. 4, B and C). In HeLa cells that lack LKB1 or that stably overexpress wild-type or catalytically inactive LKB1 (36), phosphorylation of endogenous MYPT1, as well as 14-3-3 binding, was dependent on the expression of wild-type LKB1 (Fig. 4D). In addition, stable knockdown of endogenous NUAK1 in HEK293 cells suppressed the phosphorylation of MYPT1 (Fig. 4E). Strikingly, ablating the GILK motifs in NUAK1 suppressed the ability of overexpressed NUAK1 to phosphorylate endogenous MYPT1 in HEK293 cells (Fig. 4F). Finally, overexpression of a catalytically inactive NUAK1 mutant exerted a dominant-negative effect by preventing the phosphorylation of endogenous MYPT1 (Fig. 4F).

Inhibition of the LKB1-NUAK1 pathway promotes cell adhesion

We have observed that Lkb1- and Nuak1-deficient cells adhere to tissue culture dishes more strongly than do wild-type cells. To investigate this further, we first compared the detachment of Lkb1+/+ and Lkb1−/− MEFs induced by addition of enzyme-free EDTA buffer. Treatment of Lkb1+/+ MEFs with EDTA buffer resulted in a time-dependent detachment of cells from the tissue culture dish, with 53% of cells dissociating within 20 min and most of the cells detaching within 1 hour (Fig. 5, A and B). However, in Lkb1−/− cells, the rate of detachment was reduced, with virtually no cells detaching for the first 30 min and 42% of cells still remaining attached to the dish after 1 hour (Fig. 5, A and B). Similarly, Nuak1−/− MEFs were more resistant than Nuak1+/+ MEFs to detachment in EDTA buffer (Fig. 5, C and D).

Fig. 5

Dependence of cell detachment on LKB1-NUAK1 pathway activity. (A and B) Lkb1+/+ or Lkb1−/− MEFs were washed once with PBS, then incubated in EDTA buffer to induce cell dissociation. Cells were fixed after incubation for the indicated times. The percentage of adherent cells was measured after crystal violet staining. Three sets of cells were analyzed and each time point was performed in triplicates. Representative phase images for the selected time points are shown (scale bar, 100 μm). *P < 0.01. (C and D) As in (A) and (B), except that Nuak1+/+ or Nuak1−/− MEFs were used. (E) HEK293 cells were incubated with or without 10 μM BX795 for 30 min. Cell lysates were analyzed with the indicated antibodies. (F and G) HEK293 cells were preincubated for 30 min with 10 μM BX795 or DMSO as control, washed once with PBS, and then incubated in PBS for the indicated time in the continued presence or absence of BX795. Cells were fixed and analyzed as in (A) and (B). (H and I) Lkb1+/+ or Lkb1−/− MEFs grown on glass coverslips were either fixed immediately after aspiration of the media or washed once with PBS and then incubated in EDTA buffer for 1 hour before fixing. Cells were stained with the MLC2 phospho-Ser19 (pS19) antibody, phalloidin (to label F-actin), and DAPI (4′,6-diamidino-2-phenylindole) (to label nuclei). Actomyosin contractile rings are indicated with white arrowheads. Three sets of cells were analyzed and representative images are shown. Scale bar, 10 μm. The staining of total MLC2 and MLC2 phospho-Thr18/Ser19 and the quantification of contracted cells are shown in fig. S11.

The protein kinase inhibitor BX795, which was originally developed to inhibit the kinase PDK1 [median inhibitory concentration (IC50) of 100 nM] (37), more potently inhibits NUAK1 in vitro (IC50 of 5 nM) (38). Consistent with BX795 inhibiting NUAK1, treatment of HEK293 cells with BX795 suppressed phosphorylation of MYPT1 at Ser445, Ser472, and Ser910 (Fig. 5E). In addition, treatment of HEK293 cells with BX795 suppressed the dissociation induced by incubation in phosphate-buffered saline (PBS) (Fig. 5, F and G). The inhibitory effect of BX795 on cell detachment was observed only in Lkb1+/+ and Nuak1+/+ cells, but not in their knockout counterparts (fig. S9, A and B), suggesting that BX795 promotes cell adhesion by inhibiting the LKB1-NUAK1 pathway. Furthermore, stable lentiviral shRNA knockdown of NUAK1 suppressed dissociation of HEK293 cells induced by PBS (fig. S9C). The resistance to detachment of Lkb1−/− MEFs or cells treated with BX795 was also observed when various matrix proteins (collagen type I, collagen type IV, or fibronectin) were used (fig. S10).

To study this phenomenon in more detail, we looked at the behavior of the actomyosin cytoskeleton during detachment in cells deficient in LKB1-NUAK1 signaling. After incubation with EDTA buffer, wild-type cells rounded up, a phenomenon that coincided with the disappearance of stress fibers and formation of an actomyosin “contractile ring” at the cell periphery containing polymerized actin and phosphorylated MLC2 (Fig. 5H and figs. S11A, S12A, and S13A). This ring was not formed in Lkb1−/− MEFs (Fig. 5I and fig. S11B), Nuak1−/− MEFs (fig. S12B), or HEK293 cells treated with BX795 (fig. S13B), nor did these cells round up (figs. S11C, S12C, and S13C). Consistent with these cells being more adherent, Lkb1−/− and Nuak1−/− MEFs had significantly more vinculin-stained focal adhesions, phalloidin-stained F-actin stress fibers, and a larger projected area than did Lkb1+/+ cells (fig. S14A) and Nuak1+/+ cells (fig. S14B), respectively. These observations suggest that inhibiting the LKB1-NUAK1 pathway enhances cell adhesion.

Detachment of cells induces LKB1-NUAK–dependent phosphorylation of MYPT1

We observed that agents that induced detachment of HEK293 cells from tissue culture dishes (EDTA, PBS, trypsin, or mechanical shaking) also enhanced phosphorylation of endogenous MYPT1 at Ser445, Ser472, and Ser910 (Fig. 6A). This effect was dependent on LKB1 and NUAK1, because in Lkb1−/− MEFs (Fig. 6B), LKB1-deficient HeLa cells (Fig. 6C), or Nuak1−/− MEFs (Fig. 6D), dissociation-induced phosphorylation of MYPT1 was reduced. Treatment of cells with the NUAK inhibitor BX795 suppressed basal phosphorylation of MYPT1, as well as phosphorylation induced by cell dissociation (Fig. 6E). Overexpression of kinase-inactive NUAK1 suppressed the phosphorylation of MYPT1 induced by dissociation of HEK293 cells (Fig. 6F).

Fig. 6

Cell detachment induces MYPT1 phosphorylation at Ser445, Ser472, and Ser910. (A) HEK293 cells were washed once with PBS and then incubated with either EDTA buffer (5 min), trypsin-EDTA (5 min), PBS (20 min), 1 mM EDTA in PBS (20 min), or 10 mM EDTA in PBS (20 min) or were removed from the flask mechanically by vigorous shaking. Detached cells were collected by gentle centrifugation and immediately lysed. Cell lysates (20 μg) were analyzed by immunoblotting with the indicated antibodies. (B) Lkb1+/+ or Lkb1−/− MEFs were washed once with PBS and then incubated with EDTA buffer for 15 min. Cells were lysed and cell lysates were analyzed by immunoblotting with the indicated antibodies. (C) As in (B), except HeLa T-rex cells stably overexpressing empty vector, wild-type (wt) LKB1, or kinase-inactive (ki) LKB1 were used. (D) As in (B), except that Nuak1+/+ or Nuak1−/− MEFs were used. (E) HEK293 cells were incubated with or without 10 μM BX795 for 30 min, washed once with PBS (with or without BX795), and then incubated with EDTA buffer (with or without BX795) for 5 min. Cell lysates were analyzed by immunoblotting with the indicated antibodies. (F) HEK293 T-rex cells with tetracycline-inducible overexpression of wild-type (wt) HA-NUAK1, HA-NUAK1 with all three GILK motifs mutated to GKKK (3xIL/KK), or kinase-inactive (ki) HA-NUAK1 D196A were induced with tetracycline (TET; 0.1 μg/ml) for 24 hours. Cells were then washed once with PBS, incubated with EDTA buffer (5 min), and lysed. Cell lysates were analyzed by immunoblotting with the indicated antibodies. (G) The MYPT1-PP1β complex was purified from E. coli. The complex was incubated with NUAK1, Mg2+-ATP, and 14-3-3ε for 30 min at 30°C. The activity of the phosphorylated complex was assayed with 32P-labeled MLC2. The SSS/AAA mutant corresponds to a triple MYPT1 mutant in which Ser445, Ser472, and Ser910 are mutated to Ala. (H) HEK293 or MEF cells were washed once with PBS and then incubated with EDTA buffer for 5 min. Endogenous MYPT1 was immunoprecipitated, and immunoprecipitates were subjected to either immunoblot or 14-3-3 overlay analysis. Total cell lysates were immunoblotted with the indicated antibodies. Similar results were obtained in at least three separate experiments for all data shown in this figure.

The LKB1-NUAK1 pathway suppresses the ability of MYPT1-PP1β complex to dephosphorylate myosin light chain

Binding of 14-3-3 to MYPT1-PP1β complex inhibits the interaction with myosin, thereby preventing dephosphorylation of MLC2 (35). We were able to confirm this finding by demonstrating that in vitro phosphorylation of MYPT1-PP1β complex by NUAK1 in the presence of 14-3-3ε resulted in inhibition of phosphatase activity as assayed with 32P-labeled MLC2 as a substrate (Fig. 6G). In parallel experiments, NUAK1 failed to inhibit a MYPT1-PP1β complex containing a mutant MYPT1 in which Ser445, Ser472, and Ser910 were mutated to Ala (Fig. 6G).

In nonconfluent adherent cells, the basal phosphorylation of MLC2 is not affected by loss of LKB1 (fig. S15A) or loss of NUAK1 (fig. S15, B and C). However, we observed that EDTA buffer–induced dissociation of HEK293 cells or MEFs from tissue culture dishes led to enhanced 14-3-3 binding as assessed by an overlay assay (Fig. 6H). Consistent with the notion that 14-3-3 binding suppresses MYPT1-PP1β activity, the detachment of cells was accompanied by an increase in MLC2 phosphorylation at Ser19, as well as phosphorylation at Thr18 and Ser19, in both HEK293 cells and MEFs (Fig. 6H). We also observed that treatment of HEK293 cells with BX795 prevented the induction of MLC2 phosphorylation induced by cell detachment (Fig. 6E).

AMPKα1 and AMPKα2 do not phosphorylate MYPT1

Ser445, Ser472, and Ser910 lie in an optimal consensus motif for phosphorylation by AMPK (33, 34) (fig. S7). In vitro, AMPK, as well as other AMPK-related kinases, phosphorylated MYPT1 at Ser445, Ser472, and Ser910 (fig. S16). However, we observed that treatment of MEFs (Fig. 7A) or HEK293 cells (Fig. 7B) with AMPK-activating agonists (AICAR or A769662) stimulated phosphorylation of AMPK substrates acetyl–coenzyme A (CoA) carboxylase and Raptor, but not that of endogenous MYPT1 at Ser445, Ser472, or Ser910. Moreover, in double Prkaa1 and Prkaa2 knockout MEFs, which lack both AMPKα isoforms, cell detachment induced normal phosphorylation of MYPT1 under conditions in which phosphorylation of acetyl-CoA carboxylase and Raptor are abolished (Fig. 7C). Thus, AMPK does not regulate phosphorylation of MYPT1 under the conditions we have studied. AMPK (fig. S4) and other AMPK-related kinases apart from NUAK1 or NUAK2 (Fig. 2G) did not interact with MYPT1-PP1β complex. Thus, the ability of NUAK isoforms to specifically phosphorylate MYPT1 in vivo is likely mediated through docking interactions involving the GILK motifs that are not present in other AMPK isoforms.

Fig. 7

AMPK does not phosphorylate MYPT1 in cells. (A) MEFs were treated with the indicated compounds for 1 hour or in EDTA buffer for 10 min before lysis. Cell lysates were analyzed by immunoblotting with the indicated antibodies. (B) As in (A), except HEK293 cells were used. (C) Wild-type, Lkb1−/−, and Prkaa1−/−Prkaa2−/− (knockout of both AMPKα1 and AMPKα2 isoforms) MEFs were incubated with EDTA buffer for 10 min. Cells were lysed and lysates were analyzed by immunoblotting with the indicated antibodies. Similar results were obtained in at least three separate experiments for all data shown in this figure.

Discussion

We have described a previously unknown phosphatase complex in which PP1β interacts with MYPT1 and NUAK1. PP1 binds its regulators through two major clefts, namely the Lys/Arg-Val-X-Phe (K/RVXF)–binding pocket, which binds to targeting subunits, and the GILK-binding pocket, which interacts with molecules such as inhibitor-2 (30, 39). Although inhibitor-2 does not have a K/RVXF motif, the crystal structure shows that a KSQKV sequence in inhibitor-2 occupies the K/RVXF-binding pocket on PP1. Furthermore, substituting a GILK motif in the K/RVXF motif in NIPP1 (the nuclear inhibitor of PP1) does not affect the ability of NIPP1 to bind and inhibit PP1 (40). Intriguingly, several newly characterized PP1-binding proteins have conserved GILK or related Ser-Ile-Leu-Lys (SILK) motifs in addition to the canonical K/RVXF motifs (41). These motifs may bind PP1 in the same way as inhibitor-2, occupying both the GILK- and K/RVXF-binding pockets. In contrast, NUAK1 and NUAK2 do not have obvious K/RVXF motifs. We envisage that in the MYPT1-PP1β-NUAK1 complex, the PP1β-GILK–binding pocket will interact with the GILK motifs of NUAK1 or NUAK2 and that the PP1β-K/RVXF–binding pocket will interact with the KVKF motif of MYPT1 (Fig. 8). Database analysis indicates that NUAK1 and NUAK2 are the only human proteins that have more than one GILK motif. We speculate that the three GILK motifs may facilitate the recruitment of PP1β by NUAK1 or NUAK2. It is also possible that NUAK isoforms could interact with multiple molecules of PP1β.

Fig. 8

Schematic representation of the mechanism by which the tumor suppressor LKB1 regulates MLC2 phosphorylation. LKB1 phosphorylates and activates the kinase NUAK1, which interacts with the MYPT1-PP1β myosin phosphatase complex through its ability to interact with the GILK-binding pocket on the PP1β catalytic subunit. In response to agonists that induce cell detachment, NUAK1 phosphorylates MYPT1, thereby inducing binding to 14-3-3. This inhibits interaction with myosin and leads to increased phosphorylation of MLC2 and activation of myosin II.

A missense mutation has been characterized in a human cancer patient that changes the site 3 GILK motif in NUAK2 to GILR (42). We have introduced this mutation in NUAK2 and NUAK1 and also made a biotinylated peptide encompassing the wild-type and the GILR mutant sequence of NUAK2. However, mutation of site 3 GILK to GILR did not affect the ability of NUAK1, NUAK2, or the biotinylated peptide to interact with endogenous MYPT1-PP1β complex (fig. S17).

The interaction of NUAK1 with PP1β appears to be specific for myosin phosphatase complexes because NUAK1 interacted with four of the five myosin phosphatase complexes tested, but not with 53BP2 (Fig. 1F) or six other K/RVXF-containing PP1 targeting subunits (fig. S18). How specificity of the interaction between GILK motifs of NUAK1 and myosin phosphatases is achieved requires further investigation. The accessibility or conformation of the GILK-binding pocket on PP1 may be controlled by its interaction with MYPT family members. Secondary stabilizing interactions may also occur between the myosin phosphatase subunits and NUAK1. Our observations do not rule out the possibility that NUAK1 interacts with other PP1-targeting subunits through its ability to bind PP1β.

Our findings indicate that the association of MYPT1 with PP1β inhibits dephosphorylation and hence inactivation of NUAK1 (Fig. 3A), thus enabling NUAK1 to remain active to phosphorylate MYPT1 (fig. S8). NUAK1 phosphorylates MYPT1 on Ser445, Ser472, and Ser910 sites that are conserved throughout vertebrate evolution (fig. S7). It was previously shown that phosphorylation of MYPT1 at Ser472 induces its interaction with 14-3-3 isoforms, suppressing its binding to myosin and thereby promoting MLC2 phosphorylation (35). Although it was suggested that phosphorylation of Ser472 might be regulated by Rho-associated kinase (ROCK) (35), in our hands, phosphorylation of MYPT1 at Ser472, Ser445, or Ser910 was not affected by ROCK inhibitors under conditions in which they inhibit phosphorylation of MYPT1 at Thr696 and Thr853, which are well-characterized ROCK phosphorylation sites (25) (fig. S19, A and B). Moreover, in vitro phosphorylation analysis suggested that ROCK does not efficiently phosphorylate Ser472, Ser445, or Ser910 (fig. S16). Our findings that phosphorylation of endogenous MYPT1 was dependent on the presence of LKB1 and NUAK1 and was also sensitive to inhibition of NUAK1 with BX795 (Fig. 5E) suggest that NUAK1 is one of the major kinases mediating the phosphorylation of these residues. However, because basal phosphorylation of Ser445, Ser472, and Ser910 was detected in the Lkb1- and Nuak1-deficient cells (Fig. 6, B to D), it is possible that there is another LKB1-independent kinase that phosphorylates MYPT1 at these sites.

Overexpression of NUAK1 in cancer cells was proposed to enhance cell invasiveness (17, 43), whereas the overexpression of NUAK2 promoted cell detachment (44), and our data also suggest that inhibition of the LKB1-NUAK pathway increases cell adhesion (Fig. 5), which is independent of cell type (MEF or HEK293) or extracellular matrix (fig. S10). In various cell types, the actomyosin cytoskeleton provides contractile forces to both assemble and disassemble cell-cell and cell-surface contacts (45). We suggest that suppression of the LKB1-NUAK1 pathway during cell detachment results in decreased phosphorylation of MLC2 and thus repressed activation of myosin II, thereby enabling the formation of the actomyosin “contractile ring” that is necessary for cell detachment (Fig. 5, H and I, and figs. S11 to S13). Cell detachment also involves the destabilization of adhesion structures such as focal adhesions and stress fibers. This may explain why Lkb1 and Nuak1 knockout MEFs have more stress fibers and focal adhesions than do their wild-type counterparts (fig. S14). We also found that the dissociation of cells stimulates the phosphorylation of endogenous MYPT1 at Ser445, Ser472, and Ser910 and is accompanied by increased 14-3-3 binding and increased phosphorylation of MLC2, consistent with the notion that MYPT1 phosphorylation and 14-3-3 binding suppress the activity of myosin phosphatase [Fig. 6G and (35)]. Phosphorylation of MYPT1 and MLC2 by kinases such as ROCK plays well-established roles in the regulation of cell adhesion (46). Therefore, it would be interesting to investigate the relationship between NUAK1- and ROCK-mediated phosphorylation events. We have also observed that endogenous NUAK1 is always active when immunoprecipitated from cellular extracts, and cell dissociation, despite triggering MYPT1 phosphorylation, does not enhance total immunoprecipitated cellular NUAK1 activity (fig. S20). It would be interesting to study how cell dissociation triggers phosphorylation of MYPT1 by NUAK1.

The Caenorhabditis elegans NUAK1 and NUAK2 ortholog has been implicated in the regulation of myosin-containing thick filaments (47). UNC-82 loss-of-function mutants are unable to maintain the organization of myosin filaments and internal components of the M-line during cell growth (48). It would be interesting to investigate whether this effect is mediated through impaired regulation of the C. elegans ortholog of myosin phosphatase. In addition, NUAK1 protein abundance increases during the differentiation of myoblast cells into muscle fibers (49), suggesting a role for this enzyme in muscle development. A hallmark of gastrointestinal polyps in patients with Peutz-Jeghers syndrome is overgrowth of smooth muscle. It would be of interest to study whether lack of NUAK1 or NUAK2 activity is involved in this phenotype.

Phosphorylation of MLC2 plays essential roles in the regulation of cell polarity, adhesion, and other contractile functions in nonmuscle cells (45). In Drosophila, the LKB1-AMPK pathway controls polarization and cell structure through the ability of LKB1 to activate AMPK, which then directly phosphorylates MLC2 (50). However, in mammalian systems, AMPK does not efficiently phosphorylate MLC2 in vitro (51). The Drosophila NUAK1 ortholog CG11870 does not have GILK motifs; furthermore, the NUAK1 phosphorylation sites on mammalian MYPT1 are not conserved in the Drosophila MYPT1 ortholog DMBS. This suggests that the mechanism of regulation of MYPT1 through NUAK1 does not operate in Drosophila. It is possible that the LKB1 pathway in vertebrates has evolved an alternative mechanism to control MLC2 phosphorylation by controlling the activity of the myosin phosphatase through NUAK1 and NUAK2. The MYPT1-PP1β complex not only controls dephosphorylation of MLC2, but is also likely to function as a major regulator of the dephosphorylation of many key signaling proteins (26). Thus, regulation of MYPT1 by NUAK1 might have effects beyond the control of MLC2 phosphorylation. This indicates that LKB1 could control the phosphorylation of target proteins not only through AMPK family kinases, but also through the ability to regulate certain phosphatase complexes.

Materials and Methods

Buffers

Lysis buffer contained 50 mM tris-HCl (pH 7.5), 1 mM EGTA, 1 mM EDTA, 1% (v/v) NP-40, 1 mM sodium orthovanadate, 50 mM sodium fluoride, 5 mM sodium pyrophosphate, 0.27 M sucrose, 1 mM dithiothreitol (DTT), 1 mM benzamidine, and 0.1 mM phenylmethanesulfonylfluoride (PMSF). SDS lysis buffer used for the detection of MLC2 was lysis buffer supplemented with 1% (w/v) SDS instead of NP-40 and adjusted to pH 6.8. Buffer A was 50 mM tris-HCl (pH 7.5), 0.1 mM EGTA, and 1 mM DTT. Tris-buffered saline (TBS)–Tween (TBS-T) buffer contained 50 mM tris-HCl (pH 7.5), 0.15 M NaCl, and 0.25% (v/v) Tween-20. SDS (1×) sample buffer was 1% (w/v) SDS, 1% (v/v) 2-mercaptoethanol, 50 mM tris-HCl (pH 6.8), and 6.5% (v/v) glycerol.

In vitro phosphorylation of recombinant MYPT1 and PP1β

Hemagglutinin (HA)–NUAK1 wild type (wt) or kinase-inactive (D196A) transiently overexpressed in HEK293 cells were immunoprecipitated from 0.5 mg of cell lysate with 5 μl of antibody against HA-agarose. The immunoprecipitates were washed three times with 1 ml of lysis buffer containing 0.5 M NaCl and twice with 1 ml of buffer A and incubated in a total volume of 25 μl in buffer A containing 5 μg of E. coli–purified MYPT1, 10 mM MgCl2, and 0.1 mM [γ-32P]ATP (~300 cpm/pmol) at 30°C. Reactions were terminated at the indicated time points with SDS sample buffer with 10 mM DTT. The samples were electrophoresed on a polyacrylamide gel, which was stained with Coomassie, dried, and autoradiographed. The MYPT1 Coomassie bands were excised and the incorporation of 32P radioactivity was quantified by Cerenkov counting. Phosphorylation of E. coli–purified PP1β was performed as above for 30 min in the presence of microcystin-LR (15 μg/ml). For the phosphorylation of MYPT1 mutants (Fig. 3G), FLAG-MYPT1, wild-type, or the indicated mutants were transfected into HEK293 cells and immunoprecipitated from 1 mg of cell lysate. Immunoprecipitates were washed as above and incubated with 0.5 μg of Polyhistidine-tag (His)-NUAK1 purified from insect cells. In fig. S16, recombinant MYPT1 (5 μg) was phosphorylated with 1 μg of the indicated kinase.

Phosphatase reaction

For the assay in Fig. 3A, 5 ng of recombinant PP1β was preincubated with the indicated amounts of recombinant MYPT1 for 30 min at 4°C. The phosphatase was then added to 100 ng of recombinant His-NUAK1 or glutathione S-transferase (GST)–WNK1(1–661) and incubated in buffer A with 1 mM MnCl2 for 30 min at 30°C (in a final volume of 20 μl). For immunoblot analysis, reactions were terminated by the addition of SDS sample buffer. For NUAK1 kinase assays, reactions were stopped by the addition of microcystin-LR (15 μg/ml), and 5 μl of the reaction (25 ng of NUAK1) was used in a peptide kinase assay as described below. For assays in Fig. 6G, 2 μg of recombinant MYPT1-PP1β complex was phosphorylated with 0.5 μg recombinant His-NUAK1 in the presence of 10 mM MgCl2, 0.1 mM ATP, and 3 μg of recombinant 14-3-3ε in a reaction volume of 30 μl. Samples were diluted 100 to 500 times in buffer A with 1 mM MnCl2 and bovine serum albumin (3 mg/ml). Diluted complex (10 μl) was incubated with agitation for 10 min with 32P-labeled GST-MLC2 (10 μM in final volume of 30 μl). Reactions were stopped by the precipitation of proteins with 100 μl of 20% trichloroacetic acid. Samples were vortexed briefly, and the precipitate was pelleted by centrifugation (13,000 rpm, 10 min). 32P radioactivity in the supernatants and pellets was measured by Cerenkov counting, and the release of phosphate to the supernatant was calculated. Recombinant GST-MLC2 was phosphorylated with recombinant ROCKII in the presence of 10 mM MgCl2 and 0.1 mM [γ-32P]ATP (~300 cpm/pmol) at 30°C for 3 hours. Subsequently, 32P-labeled GST-MLC2 was bound to glutathione-Sepharose and extensively washed to remove [γ-32P]ATP. 32P-labeled GST-MLC2 was then eluted with glutathione, aliquoted, and stored at −80°C.

Peptide affinity purification and competition assays

For affinity purification, 1 mg of HEK293 cell lysate was incubated with 10 μl of streptavidin-Sepharose conjugated to the indicated amount of biotinylated peptide (Fig. 2, D and E). After incubation for 1 hour at 4°C, the beads were washed twice with 1 ml of lysis buffer containing 0.15 M NaCl and twice with 1 ml of buffer A and resuspended in SDS sample buffer. The samples were subjected to electrophoresis on a polyacrylamide gel before immunoblot analysis. For peptide competition assays, a biotinylated peptide was added to 1 mg of HEK293 cell lysate to a final concentration of 1 mg/ml and incubated for 1 hour at 4°C. Subsequently, cell lysates were subjected to immunoprecipitation. GILK corresponding to residues 388 to 407 of NUAK1 (biotin-SPSKLSSKRPKGILKKRSNS) as well as its indicated variants were used. As a control, the “RFQV peptide” was used, which corresponds to residues 1006 to 1023 of WNK4 (biotin-SEEGKPQLVGRFQVTSSK) (52).

NUAK1 and LKB1 activity assays

NUAK1 or LKB1 were immunoprecipitated from the indicated amount of cell lysate. Immunoprecipitates were washed twice with 1 ml of lysis buffer containing 0.5 M NaCl and twice with 1 ml of buffer A. Assays were set up in a total volume of 50 μl in buffer A containing 10 mM MgCl2, 0.1 mM [γ-32P]ATP (~300 cpm/pmol), and 200 μM peptide substrate. After incubation for 20 min at 30°C, the reaction mixture was applied onto P81 phosphocellulose paper; the papers were washed in phosphoric acid, and the incorporation of 32P radioactivity into the peptide was quantified by Cerenkov counting. One unit (U) of activity was defined as the amount of enzyme that catalyzed the incorporation of 1 nmol of 32P per minute into the substrate. The peptide substrate for LKB1 was LKBtide (SNLYHQGKFLQTFCGSPLYRRR, residues 241 to 260 of human NUAK2 with three additional C-terminal Arg residues) (10). Sakamototide substrate for NUAK1 was ALNRTSSDSALHRRR, corresponding to residues 165 to 176 of human CREB (cyclic AMP response element–binding protein)–regulated transcription coactivator 2 with three additional C-terminal Arg residues. This peptide encompasses Ser171, which is phosphorylated by SIK2 (53). We observed that Sakamototide was 10 times as efficient a substrate as the previously used AMARA peptide (10).

Cell detachment assays

Lkb1+/+ and Lkb1−/− MEFs were provided by T. Mäkelä (Helsinki). For cell dissociation, cells were washed once in Ca2+-Mg2+–free PBS and then incubated at 37°C for the indicated time in either EDTA buffer (enzyme-free PBS-based cell dissociation buffer containing EDTA, glycerol, and sodium citrate; Invitrogen, cat. no. 13151-014), PBS (Ca2+-Mg2+–free PBS; Invitrogen, cat. no. 14190-094), trypsin [0.05% (w/v) trypsin in 0.53 mM EDTA; Invitrogen, cat. no. 25300-054], 1 or 10 mM EDTA in Ca2+-Mg2+–free PBS, or detached mechanically by vigorous shaking of the flask. Detached cells were rapidly resuspended in medium, collected by centrifugation (3 min, 70g at room temperature), and immediately lysed. For the assays shown in Fig. 5, A to G, and fig. S9, cells were grown on plastic 24- or 48-well plates. In fig. S10, 48-well plates were coated with fibronectin (10 μg/ml), collagen type I, or collagen type IV in PBS overnight and washed twice with PBS before the cells were plated. For detachment, MEFs were washed once with Ca2+-Mg2+–free PBS and incubated in EDTA buffer at 37°C for the indicated times. HEK293 cells were grown on plastic 24-well plates washed once with Ca2+-Mg2+–free PBS and then incubated in Ca2+-Mg2+–free PBS at 37°C for the indicated times. If an inhibitor was used, it was added to the medium 30 min before the PBS wash and maintained during the wash and incubation in PBS. As a control, the equivalent volume of dimethyl sulfoxide (DMSO) was used. Cells were fixed at the indicated time points with 4% (v/v) paraformaldehyde for 10 min at room temperature and washed with PBS. For quantification, cells were washed once with water, stained in 0.1% (w/v) crystal violet in 10% (v/v) ethanol for 15 min, and then washed three times with water. Crystal violet was then extracted in 10% (v/v) acetic acid and the absorbance at 590 nm was measured. Samples were diluted in water so that the absorbance at 590 nm (A590) was in the linear range between 0.1 and 1. After the subtraction of the blank, the percentage of adherent cells was calculated as the percentage of decrease, where 100% is A590 corresponding to the amount of adherent cells before detachment and 0% is A590 of an empty well (no cells attached). Phase images of fixed cells were taken on a Zeiss Axioskop microscope with a ×10 Plan Neofluar NA (numerical aperture) 0.3 objective. The images were captured on the Axiocam with Axiovision software.

Data analysis

Data were analyzed with the Student’s t test or one-way analysis of variance (ANOVA) followed by multiple pairwise comparisons (*P < 0.01). Unless otherwise indicated, error bars show SD.

Acknowledgments

Acknowledgments: We thank T. Cohen, G. Hardie, T. Mäkelä, K. Sakamoto, B. Viollet, and E. Zeqiraj for helpful discussions and generous provision of essential reagents, as well as T. Macartney for help in the mutagenesis of PP1 and subcloning of inhibitor-2. We acknowledge the Sequencing Service (School of Life Sciences, University of Dundee, Scotland) for DNA sequencing coordinated by N. Helps, the Post Genomics and Molecular Interactions Centre for Mass Spectrometry facilities (School of Life Sciences, University of Dundee) coordinated by N. Morrice, and the protein production and antibody purification teams [Division of Signal Transduction Therapy (DSTT), University of Dundee] coordinated by H. McLauchlan and J. Hastie for generation of antibodies. Funding: A Wellcome Trust Studentship funded A.Z. We thank the Medical Research Council and the pharmaceutical companies supporting the DSTT Unit (AstraZeneca, Boehringer-Ingelheim, GlaxoSmithKline, Merck KgaA, and Pfizer) for financial support. Author contributions: A.Z. performed all of the experiments; M.D. undertook cloning; D.G.C. undertook mass spectroscopy analysis; S.B. generated the myosin phosphatase complex; M.H. and S.A. provided Nuak1−/− knockout MEFs; and A.R.P. undertook microscopy analysis. A.Z. and D.R.A. designed the experiments, analyzed results, and wrote the manuscript.

Supplementary Materials

www.sciencesignaling.org/cgi/content/full/3/115/ra25/DC1

Materials and Methods

Fig. S1. Overexpressed NUAK1 and NUAK2 interact with endogenous MYPT1-PP1β complex.

Fig. S2. Identification of a small region in NUAK1 containing the GILK motif that binds to the MYPT1-PP1β complex.

Fig. S3. Alignment of GILK motifs in vertebrate orthologues of NUAK1 and NUAK2 and C. elegans UNC-82.

Fig. S4. Endogenous AMPKα1 and AMPKα2 do not immunoprecipitate the MYPT1-PP1β complex.

Fig. S5. Further characterization of binding of NUAK1 to PP1β.

Fig. S6. NUAK1 coimmunoprecipitating with MYPT1 is phosphorylated at its T loop.

Fig. S7. Sequence alignment of NUAK1 phosphorylation sites on MYPT1.

Fig. S8. Phosphorylation of the MYPT1-PP1β complex by NUAK1.

Fig. S9. Effect of the NUAK1 inhibitor BX795 and NUAK1 shRNA on cell detachment.

Fig. S10. Effect of different substrata on cell detachment.

Fig. S11. Localization of F-actin and MLC2 during detachment of Lkb1+/+ or Lkb1−/− MEFs.

Fig. S12. Localization of F-actin and MLC2 during detachment of Nuak1+/+ or Nuak1−/− MEFs.

Fig. S13. Localization of F-actin and MLC2 during detachment of HEK293 cells treated with BX795.

Fig. S14. Inhibition of LKB1-NUAK1 signaling increases number of focal adhesions, stress fibers, and cell area in MEFs.

Fig. S15. Basal MLC2 phosphorylation in LKB1- or NUAK1-deficient cells.

Fig. S16. Comparison of the phosphorylation of MYPT1 by ROCK, NUAK1, AMPK, BRSK1, or MARK2.

Fig. S17. Analysis of the NUAK2 K503R mutant.

Fig. S18. Analysis of the NUAK1 interaction with PP1-targeting subunits.

Fig. S19. The effect of the NUAK1 inhibitor BX795 and the ROCK inhibitor Y-27632 on phosphorylation of MYPT1.

Fig. S20. Cell dissociation does not affect NUAK1 activity.

Table S1. Proteins identified by mass spectrometry in NUAK1 immunoprecipitates.

References

References and Notes

View Abstract

Navigate This Article