Biochemical and Biophysical Research Communications 404 (2011) 780–784
Contents lists available at ScienceDirect
Biochemical and Biophysical Research Communications journal homepage: www.elsevier.com/locate/ybbrc
Zyxin phosphorylation at serine 142 modulates the zyxin head–tail interaction to alter cell–cell adhesion Greg S. Call, Jarom Y. Chung, John A. Davis, Braden D. Price, Tyler S. Primavera, Nick C. Thomson, Mark V. Wagner, Marc D.H. Hansen ⇑ Physiology and Developmental Biology, Brigham Young University, 574 WIDB, Provo, UT 84602, United States
a r t i c l e
i n f o
Article history: Received 10 December 2010 Available online 17 December 2010 Keywords: Zyxin Adhesion MDCK
a b s t r a c t Zyxin is an actin regulatory protein that is concentrated at sites of actin–membrane association, particularly cell junctions. Zyxin participates in actin dynamics by binding VASP, an interaction that occurs via proline-rich N-terminal ActA repeats. An intramolecular association of the N-terminal LIM domains at or near the ActA repeats can prevent VASP and other binding partners from binding full-length zyxin. Such a head–tail interaction likely accounts for how zyxin function in actin dynamics, cell adhesion, and cell migration can be regulated by the cell. Since zyxin binding to several partners, via the LIM domains, requires phosphorylation, it seems likely that zyxin phosphorylation might alter the head–tail interaction and, thus, zyxin activity. Here we show that zyxin point mutants at a known phosphorylation site, serine 142, alter the ability of a zyxin fragment to directly bind a separate zyxin LIM domains fragment protein. Further, expression of the zyxin phosphomimetic mutant results in increased localization to cell–cell contacts of MDCK cells and generates a cellular phenotype, namely inability to disassemble cell–cell contacts, precisely like that produced by expression of zyxin mutants that lack the entire regulatory LIM domain region. These data suggest that zyxin phosphorylation at serine 142 results in release of the head–tail interaction, changing zyxin activity at cell–cell contacts. Ó 2010 Elsevier Inc. All rights reserved.
1. Introduction Proper actin dynamics and structure are required for cell adhesion and migration [1]. A number of actin regulatory proteins have been implicated in controlling actin dynamics at adhesion sites and to drive cell migration, including zyxin. Zyxin localizes to cell–cell and cell–substrate adhesions [2]. Cells from zyxin null mice exhibit increased cell migration rates [3] and inability to reorganize actin of stress fibers following stretch stimulation [4,5]. Zyxin function at cell–cell adhesions has been implicated by expression of zyxin deletion mutants that alter cell–cell adhesion [6] and cell scattering [7]. Importantly, adhesion sites are highly dynamic structures undergoing constant remodeling, suggesting that the function of actin regulatory proteins, like zyxin, must be controlled by biochemical mechanisms that increase or decrease activity as needed. Zyxin participates in regulation of actin dynamics by binding to VASP [8], an interaction that occurs through a series of four proline-rich repeats at the amino terminus of the zyxin sequence [9]. This interaction is blocked by an intramolecular interaction that is mediated the carboxyl terminal zyxin LIM domain region [10]. Interesting, the zyxin LIM domain region plays a major role ⇑ Corresponding author. Fax: +1 801 422 0700. E-mail address:
[email protected] (M.D.H. Hansen). 0006-291X/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2010.12.058
in regulating zyxin function in cell–cell adhesion; it is expression of zyxin mutants deleted of this region that results in more rapid cell–cell junction formation and strengthening [6]. Further strengthening the idea that zyxin activity is negatively regulated by the LIM domain-mediated head–tail interaction, several LIM domain binding partners have been shown to be prevented from interacting with the LIM domains in full-length zyxin, but not shorter zyxin fragments containing LIM domains [11,12]. Interestingly, these binding partners can associate with full-length zyxin when phosphorylation has occurred, specifically at serine 142 in the case of the Acinus-S interaction [11]. These reports suggest a simple model for how zyxin function in cell migration and adhesion might be achieved, namely by a phosphorylation event that alters the zyxin head–tail interaction. Such head–tail interactions have been observed in a growing number of actin regulatory systems, including vinculin, ERM proteins, and N-WASP [13–15]. Such head–tail interactions are regulated by a number of mechanisms, including protein and lipid binding [15,16], and phosphorylation [17,18]. Here we examine how phosphorylation affects the head–tail interaction of zyxin, as well as its function at cell–cell contacts. We find that phosphomimetic and phosphoresistant mutations of zyxin serine 142 alter the zyxin head–tail interaction; only the zyxin S142A mutant, and not the S142D mutant, binds a LIM domain
G.S. Call et al. / Biochemical and Biophysical Research Communications 404 (2011) 780–784
containing protein in direct binding assays. Further, while both mutants localize to cell–cell contacts, MDCK cells expressing the S142D zyxin mutant exhibit defects in cell–cell junction breakdown during growth factor-induced scattering. Importantly, this scattering phenotype is precisely identical to that observed in cells expressing zyxin mutants lacking the regulatory LIM domain region [7]. 2. Materials and methods 2.1. Protein purification and binding studies GST-zyxin fusion proteins were produced in plasmid-transformed bacteria. After growing and inducing, with IPTG, under conditions optimized for each protein, bacteria were harvested by centrifugation at 4 °C. Pellet was resuspended in ice cold Buffer A (100 mM NaCl, 20 mM Tris, 0.5% (v/v) Triton X-100, 10 lg/ml antipain, 10 lg/ml leupeptin, 10 lg/ml pepstatin A, 1 mM PMSF, 1 mM DTT, pH 7.4) and homogenized by sonification on ice. The resulting extract was clarified by centrifugation at 50,000g for 15 min at 4 °C. GST fusion proteins were recovered from extracts with glutathione-agarose beads equilibrated in Buffer A and eluted with 10 mM glutathione in the appropriate buffer. Proteins were dialyzed into Buffer A and concentration determined by Bradford method [19]. For binding studies, proteins were mixed in the appropriate combination and incubated on ice for 1 h. Samples were then loaded onto sucrose density gradients made of Buffer A supplemented with 7.5–25% (w/v) sucrose. Gradients were spun at 28,000 rpm for 16 h in an SW28 rotor at 4 °C. Fractions collected from the gradients following centrifugation were mixed with sample buffer and were processed for Western blot analysis with antizyxin antibodies (Abcam). Antibody binding in Western blots was detected with HRP-conjugated secondary antibodies (GE Healthcare) and ECL detection reagent (Pierce). Gels and developed film were scanned with a FluoroScan and densitometry of images was performed using AlphaEase software.
781
in vitro protein binding experiments using purified recombinant fragments of zyxin protein. To test how phosphorylation at serine 142 alters LIM domain region binding, we made GST-zyxin fragments lacking the LIM domain region (GST-zyxinDLIM) and bearing mutations that alter serine 142 to either an alanine or an aspartic acid. Such alanine and aspartic acid substitutions are expected to prevent or mimic, respectively, zyxin phosphorylation at this site. In support of this reasoning, identical alanine and aspartic acid substitutions at zyxin serine 142 have been reported to prevent or permit, respectively, binding availability of the LIM domain region in full-length zyxin for Acinus-S, an interaction that requires phosphorylation of serine 142 [11]. To test whether these mutations altered the zyxin head–tail interaction, zyxin proteins bearing S142A or S142D mutations were separated by rate-zonal centrifugation in sucrose gradients, either alone or following mixture and incubation with a GST fusion protein containing the LIM domain region of zyxin. The sedimentation profile of the purified zyxin fragments in each experiment was determined by Western blot analysis of gradient fractions with zyxin antibodies that recognizes an epitope in the GST-zyxinDLIM protein fragments (Fig. 1). Alone, purified GST-zyxinDLIM domain proteins bearing either serine 142 point mutation sediment in sucrose gradients with identical profiles. While this result is expected given that both GSTzyxinDLIM proteins are the same molecular weight, the identical sedimentation profiles suggest that no dramatic conformational change occurs in these zyxin fragments as a result of one of the serine 142 point mutations. Similarly, when GST-zyxinDLIM protein bearing the S142A mutations is separated in a sucrose density gradient after mixture and incubation with the GST-zyxin LIM
2.2. Cellular imaging MDCK cells were maintained in DMEM with 10% fetal bovine serum. Cells were transfected with zyxin mutant constructs using Effectene reagent. MDCK cells were subjected to selection with G418 for 14 days and GFP-positive clones isolated. Protein localization was determined by plating cells onto collagen-coated coverslips, then fixing them with 3.7% (w/v) paraformaldehyde in PBS on ice for 15 min. GFP fluorescence was observed directly, while cadherin staining was performed using antibodies (BD Biosciences). To visualize growth factor-induced scattering, 10,000 cells were seeded into collagen-coated imaging dishes and cultured overnight, allowing the formation of small colonies of cells. A heated microscope stage using Delta T4 temperature regulation system (Bioptechs) maintained temperature at 37 °C during subsequent imaging. Conditioned medium from MRC-5 cells, containing scatter factor/hepatocyte growth factor (HGF), was added immediately prior to initiating imaging. Phase contrast images of cell scattering were taken every 2 min for 12 h. The first 10 h of each series was used for analysis as described by Sperry et al. [7], with the remaining time demonstrating continued cell viability. 3. Results and discussion Phosphorylation at serine 142 alters binding availability of the distant (located between amino acids 370 and the carboxy terminus) LIM domain region [11,12], we reasoned that zyxin phosphorylation affects the head–tail interaction mediated by the LIM domain region. We first sought to directly test this idea using
Fig. 1. A phosphomimetic point mutations at S142 blocks zyxin’s head–tail interaction. Purified GST-zyxinDLIM protein bearing either S142A (A) or S142D (B) point mutations were separated in rate-zonal sucrose gradients and their sedimentation profile determined by quantitation of band intensity following Western blot analysis of gradient fractions. Proteins were separated alone or following mixture and incubation with GST-zyxin LIM domain region protein.
782
G.S. Call et al. / Biochemical and Biophysical Research Communications 404 (2011) 780–784
domain region fusion protein, the GST-zyxinDLIM protein still sediments into a single peak in precisely the same fractions as when subjected to gradient centrifugation alone is in contrast to the behavior of the GST-zyxinDLIM protein bearing the S142A mutation. GST-zyxinDLIM S142A sediments into two peaks; one peak aligns with the peak observed for the GST-zyxinDLIM fusion proteins alone, and the second peak occurs deeper in the gradient. The appearance of a second peak demonstrates an interaction with the GST-zyxin LIM domain region fusion protein. The interaction of GST-zyxinDLIM S142A with the LIM domain region protein is consistent with the LIM domain region-mediated head–tail interaction that was previously reported [10]. That no interaction is observed between the zyxin LIM domain region and a zyxin fragment bearing the S142D mutation shows that introduction of a negative charge at this site, as occurs as a result of phosphorylation, prevents the LIM domain region-mediated head–tail interaction in the zyxin protein. This result also further suggests that the location of the zyxin head–tail interaction occurs at or near sear S142, which might explain how VASP binding to the nearby FPPPPP motifs (amino acids 50–120) in full-length zyxin is blocked by the zyxin head–tail interaction under the correct conditions [10]. Functional studies implicating zyxin in cell–cell adhesion reveal that the LIM domains act as a negative regulatory element in controlling zyxin function. Expression of full-length zyxin has no effect on cell–cell junction formation or detachment, while expression of a zyxin mutant lacking LIM domains increases cell–cell adhesion strength [6] and prevents cell–cell detachment [7]. How might the cell normally inhibit negative regulation by the LIM domain region to permit zyxin function at adhesion sites? The observation that zyxin phosphorylation at serine 142 breaks the LIM domainmediated head–tail interaction could explain how binding partner access to both the LIM domain region and VASP-binding proline rich repeats is controlled in cells during cell adhesion changes. We therefore sought to determine how zyxin phosphorylation at serine 142 might alter zyxin targeting to cell–cell junctions. MDCK cells were transfected with expression plasmids containing GFPzyxin bearing S142A or S142D mutations. In order to determine the localization of the zyxin point mutants, cells were seeded onto collagen-coated coverslips and then processed for immunofluorescence. Cadherin staining was performed to visualize cell–cell junctions, while direct GFP fluorescence was used to localize the zyxin mutants (Fig. 2). A striking difference in nuclear localization was immediately apparent, as expected from earlier studies of cells
Fig. 2. Localization of zyxin S142 phosphomimetic and phosphoresistant mutants. Immunolocalization of E-cadherin and direct fluorescence of GFP in MDCK cells expressing GFD-zyxin bearing either a S142A or S142D point mutation. Scale bar = 50 lm.
expressing these mutant proteins [11]. Only small differences in protein localization at cell–cell junction targeting could be observed. Since any subtle difference in amount of mutant protein is likely a result of changes in cytoplasmic protein concentration that result from differential nuclear targeting, we conclude that changes in charge at serine 142 makes little, if any, difference in the cell–cell junction targeting of the zyxin protein. This result suggests that any regulation of zyxin by phosphorylation at serine 142 occurs at the level of protein function, rather than localization. We therefore sought to determine whether a zyxin mutation that mimics phosphorylation would alter cell–cell adhesion properties. MDCK cells expressing zyxin mutants lacking LIM domains, thus preventing any negative regulation by the LIM domain region and behaving as constitutively-active mutants, have been shown to establish and strengthen cell–cell junction more quickly than controls expressing full-length zyxin of wildtype sequence [6]. Further, MDCK cells expressing this constitutively active zyxin mutant are unable to disassemble cell–cell junctions when triggered to undergo scattering with growth factor treatment, while those expressing full-length zyxin of wildtype sequence scatter normally [7]. Since zyxin phosphorylation prevents the LIM domain region-mediated head–tail interaction, we expect that the zyxin S142D point mutant to generate the same phenotype as the zyxin fragment lacking the entire LIM domain region. We tested this idea by analyzing the scattering behavior of MDCK cells expressing GFP-zyxin S142D and parental MDCK cells. In order to analyze changes in cell–cell adhesion in populations of cells, MDCK cells transfected with the mutant zyxin construct were subjected to antibiotic selection and a stable cell lines generated. Analysis of GFP fluorescence reveals that 100% of the cells express the zyxin mutant. Cells were plated into collagen-coated imaging dishes and cultured, allowing the formation of small colonies of MDCK cells that could be analyzed by timelapse microscopy. Small colonies of 3–20 cells were selected for imaging in each experiment and then cells were treated with conditioned medium containing hepatocyte growth factor to initiate scattering. Three independent experiments were performed for each cell line tested and 20 colonies were imaged in each independent experiment. Still images taken from representative imaging series reveal significant alterations in scattering behavior in MDCK cells expressing GFP-zyxin S142D, compared to parental MDCK cells (Fig. 3). Parental MDCK cells in these experiments exhibit a scattering behavior that has been characterized at a high level of detail, including spreading, enhanced migration, detachment of cell–cell junctions, and the appearance of solitary, migratory cells. While MDCK cells expressing GFP-zyxin S142D exhibit spreading and increased migration following initiation of scattering, detachment of cell–cell junctions and the appearance of solitary, migratory cells is rarely observed. In fact, when single cells attempt to detach from colonies of cells, single points of attachment remain and the cell remains connected by long retraction fibers. Importantly, this cell–cell adhesion phenotype is identical to that reported for cells expressing zyxin mutants lacking the entire LIM domain region [7]. In order to confirm our empirical impression of altered scattering, we determined the rate of initiation of scattering and the frequency of failure to complete scattering for both cell lines (Fig. 4). All quantitation was performed blind and independently by two individuals. Consistent with our impression from the movies, a delay in initiation of scattering is observed in cells expressing GFPzyxin S142D. The t1/2 of scattering initiation, the point at which half of the colonies analyzed have initiated scattering, is 169.4 min for parental MDCK cells and 215.3 min for cells expressing GFP-zyxin S142D, an increase in scattering initiation of 45 min and over 25% later. As with our empirical observations, the phenotype exhibited by MDCK cells expressing GFP-zyxin S142D is nearly identical to that exhibited by MDCK cells expressing a
G.S. Call et al. / Biochemical and Biophysical Research Communications 404 (2011) 780–784
783
Fig. 3. Scattering of MDCK cells expressing a phosphomimetic S142 zyxin mutation. Parental MDCK cells or those expressing GFP-zyxin S142D were subjected to live cell imaging following induction of scattering by hepatocyte growth factor treatment at time 0:00. Arrows denote q retraction fiber maintaining cell–cell connections between cells attempting to detach from the main colony. Time stamp h:min. Scale bars = 50 lm.
zyxin mutant lacking the entire regulatory LIM domain region [7]. Quantitation also reveals that 75% of colonies of MDCK cells expressing the S142D mutant failed to complete scattering during the 12 h imaging period, compared to only 25% of colonies of parental MDCK cells. This supports our observation that cells expressing GFP-zyxin S142D fail to detach cell–cell contacts and is a nearly identical phenotype to that reported for MDCK cells
expressing a zyxin fragment lacking the LIM domain region [7]; a single point mutation mimicking zyxin phosphorylation at serine 142 generates a zyxin mutant that behaves identically to a zyxin fragment lacking the entire LIM domain region of over 150 amino acids residues. Importantly, we cannot rule out how release of the head–tail interaction generates increased strength of cell–cell adhesions,
784
G.S. Call et al. / Biochemical and Biophysical Research Communications 404 (2011) 780–784
members of the Hansen lab, whose insight and suggestions were important in the preparation of the manuscript, and to Dr. Ye of Emory University, who provided constructs for mammalian expression of GFP-zyxin point mutants. References
Fig. 4. Quantitation of the zyxin S142D scattering phenotype. (A) Scattering initiation curves (the cumulative percentage of colonies initiating scattering as a function of time) for parental MDCK cells or MDCK cells expressing GFP-zyxin S142D. (B) Percentages of colonies of either parental MDCK cells of MDCK cells expressing GFP-zyxin S142D failing to complete scattering.
but we favor a model where zyxin is more able to participate in actin dynamics. Release of the head–tail interaction allows increased availability for VASP, strengthening cell–cell contacts. In support of this idea, VASP, like zyxin, is required for establishment of strong epithelial cell–cell adhesions [19,20]. The LIM region binds several proteins and its sudden availability after phosphorylation could impact cell–cell adhesion. However, the observation that zyxin S142D phenocopies a zyxin mutant lacking the entire LIM domain region argues against this possibility. Similarly, zyxin binds a-actinin and has been postulated to participate in cell–cell adhesion via this interaction. But again, prior work targeting the binding site for a-actinin in zyxin indicates that it does not participate in cell–cell junction formation [21]. Taken with these previous observations, results presented here demonstrate how zyxin phosphorylation prevents zyxin head–tail and suggests how the loss of the regulatory head–tail interaction alters zyxin function at cell–cell junctions. It remains to be seen whether this regulation alters zyxin activity in other cellular functions, such as cell–substrate adhesion. Acknowledgments G.S.C. and N.C.T were supported by undergraduate summer research internships from the BYU Cancer Center. Thanks to
[1] J.T. Parsons, A.R. Horwitz, M.A. Schwartz, Cell adhesion: integrating cytoskeletal dynamics and cellular tension, Nat. Rev. Mol. Cell Biol. 11 (2010) 633–643. [2] A.W. Crawford, M.C. Beckerle, Purification and characterization of zyxin, an 82,000-dalton component of adherens junctions, J. Biol. Chem. 266 (1991) 5847–5853. [3] L.M. Hoffman, C.C. Jensen, S. Kloeker, C.L. Wang, M. Yoshigi, M.C. Beckerle, Genetic ablation of zyxin causes Mena/VASP mislocalization, increased motility, and deficits in actin remodeling, J. Cell Biol. 172 (2006) 771–782. [4] H. Hirata, H. Tatsumi, M. Sokabe, Mechanical forces facilitate actin polymerization at focal adhesions in a zyxin-dependent manner, J. Cell Sci. 121 (2008) 2795–2804. [5] M. Yoshigi, L.M. Hoffman, C.C. Jensen, H.J. Yost, M.C. Beckerle, Mechanical force mobilizes zyxin from focal adhesions to actin filaments and regulates cytoskeletal reinforcement, J. Cell Biol. 171 (2005) 209–215. [6] M.D. Hansen, M.C. Beckerle, Opposing roles of zyxin/LPP ACTA repeats and the LIM domain region in cell–cell adhesion, J. Biol. Chem. 281 (2006) 16178– 16188. [7] R.B. Sperry, N.H. Bishop, J.J. Bramwell, M.N. Brodeur, M.J. Carter, B.T. Fowler, Z.B. Lewis, S.D. Maxfield, D.M. Staley, R.M. Vellinga, M.D. Hansen, Zyxin controls migration in epithelial–mesenchymal transition by mediating actin– membrane linkages at cell–cell junctions, J. Cell Physiol. 222 (2010) 612–624. [8] J. Fradelizi, V. Noireaux, J. Plastino, B. Menichi, D. Louvard, C. Sykes, R.M. Golsteyn, E. Friederich, ActA and human zyxin harbour Arp2/3-independent actin-polymerization activity, Nat. Cell Biol. 3 (2001) 699–707. [9] B. Drees, E. Friederich, J. Fradelizi, D. Louvard, M.C. Beckerle, R.M. Golsteyn, Characterization of the interaction between zyxin and members of the Ena/ vasodilator-stimulated phosphoprotein family of proteins, J. Biol. Chem. 275 (2000) 22503–22511. [10] J.D. Moody, J. Grange, M.P. Ascione, D. Boothe, E. Bushnell, M.D. Hansen, A zyxin head–tail interaction regulates zyxin–VASP complex formation, Biochem. Biophys. Res. Commun. 378 (2009) 625–628. [11] C.B. Chan, X. Liu, X. Tang, H. Fu, K. Ye, Akt phosphorylation of zyxin mediates its interaction with acinus-S and prevents acinus-triggered chromatin condensation, Cell Death Differ. 14 (2007) 1688–1699. [12] T. Hirota, T. Morisaki, Y. Nishiyama, T. Marumoto, K. Tada, T. Hara, N. Masuko, M. Inagaki, K. Hatakeyama, H. Saya, Zyxin, a regulator of actin filament assembly, targets the mitotic apparatus by interacting with h-warts/LATS1 tumor suppressor, J. Cell Biol. 149 (2000) 1073–1086. [13] R. Gary, A. Bretscher, Ezrin self-association involves binding of an N-terminal domain to a normally masked C-terminal domain that includes the F-actin binding site, Mol. Biol. Cell 6 (1995) 1061–1075. [14] L. Molony, K. Burridge, Molecular shape and self-association of vinculin and metavinculin, J. Cell Biochem. 29 (1985) 31–36. [15] R. Rohatgi, L. Ma, H. Miki, M. Lopez, T. Kirchhausen, T. Takenawa, M.W. Kirschner, The interaction between N-WASP and the Arp2/3 complex links Cdc42-dependent signals to actin assembly, Cell 97 (1999) 221–231. [16] J. Weekes, S.T. Barry, D.R. Critchley, Acidic phospholipids inhibit the intramolecular association between the N- and C-terminal regions of vinculin exposing actin-binding and protein kinase C phosphorylation sites, Biochem. J. 314 (Pt 3) (1996) 827–832. [17] T. Matsui, M. Maeda, Y. Doi, S. Yonemura, M. Amano, K. Kaibuchi, S. Tsukita, Rho-kinase phosphorylates COOH-terminal threonines of ezrin/radixin/ moesin (ERM) proteins and regulates their head-to-tail association, J. Cell Biol. 140 (1998) 647–657. [18] C. Schwienbacher, B.M. Jockusch, M. Rudiger, Intramolecular interactions regulate serine/threonine phosphorylation of vinculin, FEBS Lett. 384 (1996) 71–74. [19] J.A. Scott, A.M. Shewan, N.R. Den Elzen, J.J. Loureiro, F.B. Gertler, A.S. Yap, Ena/ VASP proteins can regulate distinct modes of actin organization at cadherinadhesive contacts, Mol. Biol. Cell 17 (2006) 1085–1095. [20] V. Vasioukhin, C. Bauer, M. Yin, E. Fuchs, Directed actin polymerization is the driving force for epithelial cell–cell adhesion, Cell 100 (2000) 209–219. [21] M.D. Hansen, M.C. Beckerle, Alpha-Actinin links LPP, but not zyxin, to cadherin-based junctions, Biochem. Biophys. Res. Commun. 371 (2008) 144– 148.