The phosphorylation state of MRLC is polyamine dependent in intestinal epithelial cells

Abstract

Cell migration is important to the integrity of the gastrointestinal tract for the normal movement of cells from crypt to villi and the healing of wounds. Polyamines are essential to cell migration, mucosal restitution, and, hence, healing. Polyamine depletion by α-difluoromethyl ornithine (DFMO) inhibited migration by decreasing lamellipodia and stress fiber formation and preventing the activation of Rho-GTPases. Polyamine depletion increased the association of the thick F-actin cortex with phosphorylated myosin regulatory light chain (pMRLC). In this study, we determined why MRLC is constitutively phosphorylated as part of the actin cortex. Inhibition of myosin light chain kinase (MLCK) decreased RhoA and Rac1 activities and significantly inhibited migration. Polyamine depletion increased phosphorylation of MRLC (Thr18/Ser19) and stabilized the actin cortex and focal adhesions. The Rho-kinase inhibitor Y27632 increased spreading and migration by decreasing the phosphorylation of MRLC, remodeling focal adhesions, and by activating Rho-GTPases. Thus phosphorylation of MRLC appears to be the rate-limiting step during the migration of IEC-6 cells. In addition, increased localization of RhoA with the actin cortex in polyamine-depleted cells appears to activate Rho-kinase. In the absence of polyamines, activated Rho-kinase phosphorylates myosin phosphatase targeting subunit 1 (MYPT1) at serine-668 leading to its inactivation and preventing the recruitment of phosphatase (protein phosphastase, PP1cδ) to the actomyosin cortex. In this condition, MRLC is constitutively phosphorylated and cycling does not occur. Thus activated myosin binds F-actin stress fibers and prevents focal adhesion turnover, Rho-GTPase activation, and the remodeling of the cytoskeleton required for migration.

the mucosal epithelium of the alimentary tract provides a crucial barrier to a broad spectrum of damaging agents and immunogenic substances within the intestinal lumen. Impairment of the integrity of the barrier is observed in various intestinal disorders including inflammatory bowel disease (IBD), celiac disease, and intestinal infections. The mucosa of the gastrointestinal tract has the unique ability to repair itself rapidly following damage. Mucosal repair consists of two phases. Early mucosal restitution is the rapid reestablishment of epithelial integrity and continuity after superficial injury, before cell proliferation, or an extensive inflammatory response occurs (4). It is characterized by sloughing of the damaged cells and migration of remaining viable cells over the denuded lamina propria (22, 37). The second phase involves replacement of the lost cells by mitosis and does not begin until 24 h or so following injury (22). The process of early mucosal restitution was originally described for the stomach, but later Fiel et al. (12) and Moore et al. (27) have shown a similar process for the small intestine.

The importance of polyamines to cellular function has been demonstrated in both normal and cancer cells from a variety of tissues (23, 39, 53). Polyamines have been shown to be essential for various processes including cell proliferation (46–48). The intracellular levels of polyamines are highly regulated and primarily depend on the activity of ornithine decarboxylase (ODC; EC 4.1.17), which catalyzes the first rate-limiting step in polyamine synthesis, forming putrescine from the amino acid ornithine. Putrescine is then converted to spermidine and spermine through the sequential addition of propylamine groups. α-Difluoromethylornithine (DFMO) inhibits ornithine decarboxylase (ODC) and prevents the formation of polyamines. DFMO depletes intracellular putrescine levels within 6 h, spermidine within 24 h, and spermine to 70% within 96 h. Using stress and hypertonic NaCl models for mucosal injury in rats, we have shown that polyamines are essential for cell migration and the healing of gastric and intestinal lesions (46–48).

Cell migration is essential for normal development, angiogenesis, wound repair, tumor invasion, and metastasis (49), and it involves dynamic changes in the cytoskeleton. Migration represents a multistep process including the formation of membrane protrusions called lamellipodia. These structures are stabilized by adhering to the extracellular matrix or to adjacent cells via transmembrane proteins linked to the actin cytoskeleton. Adhesion to the extracellular matrix provides traction sites for the forward movement of the cell. Thus dynamic assembly and disassembly of these adhesions plays a crucial role in determining the direction and rate of cell motility. During cell migration the advancement of the leading edge and retraction of the trailing edge require actin filaments in the appropriate arrangement (8, 9, 17). The lamellipodia and stress fibers are regulated by Rac1 and RhoA, respectively (35, 36). This reorganization of F-actin is mediated by treadmilling that involves polymerization and depolymerization of actin and contraction of filaments driven by myosin II motor protein (36, 38).

Studies on the regulation of cytoskeletal remodeling have recently focused on the phosphorylation of regulatory light chain (RLC) of the motor protein myosin II. Nonmuscle myosin II (NM II), an actin-activated-ATPase plays an important role in several cellular processes that convert the energy of ATP hydrolysis into force between actin and myosin filaments. Myosin II molecules are composed of three pairs of peptides: two 230-kDa heavy chains, two 20-kDa regulatory light chains that regulate NM II activity, and two 17-kDa essential light chains that stabilize the heavy chain structure (15, 16, 45). Myosin II is activated when myosin RLC (MRLC) is phosphorylated by a Ca²⁺ and calmodulin (CaM)-dependent protein kinase myosin light chain kinase (MLCK) and is inactivated when MRLC is dephosphorylated by myosin light chain phosphatase (MLCP) (44). Thus MRLC phosphorylation in a motile cell is coordinated in time and space by MLCK and MLCP. Imbalance in the coordination of the two enzymes results in altered migration and adhesions. In addition to MLCK, Rho-associated kinase (Rho-kinase) activates MRLC by direct phosphorylation and by decreasing the activity of MLCP (13, 26).

We have shown that polyamine depletion increases the phosphorylation of MRLC, which is localized with a thick actin cortex at the cell periphery. The formation of cortical actin correlated with decreased activities of Rho family GTPases and migration in the polyamine-depleted cells (32). Although, MLCK- and Rho-kinase-mediated phosphorylation of MRLC has been reported to increase migration, staurosporine, an inhibitor of serine-threonine kinases, decreased the phosphorylation of MRLC and increased migration in polyamine-depleted cells (34). However, it is unclear whether increased MLCK and Rho-kinase or decreased MRLC phosphatase activities are responsible for the constitutive phosphorylation of MRLC and thereby inhibition of migration in the absence of polyamines.

Our results show that both MLCK and Rho-kinase play important roles in the regulation of migration. However, inhibition of Rho-kinase decreased phosphorylation of MRLC, reorganized the actin cytoskeletal structure and focal adhesions, and prevented the inhibition of migration in polyamine-depleted cells. Polyamine depletion inhibits dephosphorylation of MRLC by inhibiting myosin phosphatase activity, which leads to the formation of a static actomyosin complex, stabilizes focal adhesions, and subsequently prevents the downstream activities of Rho-GTPases.

MATERIALS AND METHODS

Materials.

Cell culture medium and fetal bovine serum (FBS) were obtained from Mediatech (Herndon, VA). Dialyzed FBS (dFBS), glutathione agarose beads, and anti-vinculin and anti-actin antibodies were purchased from Sigma (St. Louis, MO), and trypsin-EDTA, antibiotics, and insulin were from GIBCO-BRL (Grand Island, NY). Protease inhibitors, phosphatase inhibitors, phosphate buffer saline (PBS), Dulbecco's PBS (DPBS), formaldehyde, bicinchoninic acid (BCA), and mammalian protein extraction reagent were purchased from Thermo Fisher Scientific (Rockford, IL). Antibodies for phosphorylated MRLC (pMRLC; pThr18/Ser19, pThr18, pSer19) and total MRLC were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Myosin IIa, myosin phosphatase targeting protein 1 (MYPT1), phospho-MYPT, and phospho-paxillin antibodies were purchased from Cell Signaling Technology (Beverly, MA). Anti-Rac1 and anti-FAK antibodies were purchased from Upstate Biotechnology (Lake Placid, NY). Alexa fluor dyes-conjugated secondary antibodies and rhodamine-phalloidin were purchased from Molecular Probes (Eugene, OR). The enhanced chemiluminescence substrate Western Lightning TM was purchased from PerkinElmer Life and Analytical Sciences (Shelton, CT). DFMO was a gift from ILEX Oncology (San Antonio, TX). ML-7, Y27632, and NSC23766 were purchased from EMD Biosciences (La Jolla, CA). All other chemicals were of the highest purity commercially available.

Cell culture.

The IEC-6 cell line (ATCC CRL 1592) was obtained from the American Type Culture Collection (Manassas, VA) at passage 13. This cell line is derived from normal rat intestine and was developed and characterized by Quaroni et al. (29). IEC-6 cells are nontumorigenic, originate from intestinal crypt cells as judged by morphological and immunologic criteria, and retain the undifferentiated character of epithelial stem cells. IEC-6 cells were maintained in T-150 flasks in Dulbecco's Minimal Essential Medium (DMEM) supplemented with 10% FBS, 10 μg/ml insulin, and 50 μg/ml gentamicin sulfate at 37°C and 10% CO₂. Stock cells were passaged once a week and medium was changed three times a week. Before an experiment cells were trypsinized, counted using a Beckman Coulter counter, and grown for 3 days in dFBS in control, DFMO (5 mM), or DFMO plus 10 μM putrescine (PUT) containing media and were serum starved for 24 h before an experiment.

Migration assay.

Cells were grown as the same protocols. On day 4, plates containing confluent monolayer with cells grown in control, DFMO, and DFMO + PUT media were marked in the center by drawing a line with a black marker along the diameter of back of the plates. Wounding of the monolayer was performed perpendicular to the marked line using a gel-loading microtip. Plates were washed once with HBSS containing Ca²⁺ and Mg²⁺ to remove damaged cells, and respective media with or without ROCK inhibitor (Y27632) were added and incubated for 7 h. The area of migration was photographed with a charged-couple device camera at the intersection of the marked line and the wound edge at 0 and after 7 h. Cell migration was calculated as wound area covered using Imagej software. Each experiment was repeated three times in triplicate.

RhoA and Rac1 assay.

Cells were grown and treated as described above. Rac1 and RhoA activities were determined by a pull-down assay using glutathione S-transferase (GST)-p21-activated kinase (GST-PAK) and GST-ROCK fusion proteins, respectively, following the method of Kranenburg et al. (20). GST-PAK and GST-ROCK fusion proteins were prepared by lysing the bacteria (Escherichia coli transformed with GST-PAK or GST-ROCK plasmid) in a buffer containing 50 mM Tris (pH 7.4), 100 mM NaCl, 5 mM Mgcl₂, 10% glycerol, and 1% Nonidet P-40 supplemented with protease and phosphatase inhibitors. The bacterial cell lysate was sonicated and clarified by centrifugation at 13,000 g for 13 min. The fusion protein was recovered by the addition of glutathione-agarose beads to the supernatant. Beads were washed three times in the cell lysis buffer and resuspended before the addition of cell lysates (200 μg). After 2 h of tumbling at 4°C, beads were washed with lysis buffer, and the amount of target proteins bound to GST-PAK and GST-ROCK were analyzed by performing SDS-PAGE (12%) and Western blot analysis using Rac1and RhoA-specific antibodies.

Western blot analysis.

The protocol for Western blot analysis has been described earlier (30–34). Briefly, cells were washed twice with ice-cold DPBS and lysed for 10 min in ice-cold cell lysis buffer containing protease and phosphatase inhibitors. Lysates were centrifuged at 10,000 g for 10 min at 4°C followed by SDS-PAGE. Proteins were transferred to Immobilon-P membranes (Millipore, Bedford, MA) and probed with the indicated antibodies overnight at 4°C in Tris buffer saline (TBS) containing 0.1% Tween-20 and 5% nonfat dry milk (blotting grade, Bio-Rad). Membranes were subsequently incubated with appropriate horseradish peroxidase-conjugated secondary antibody at room temperature for 1 h, and the immunocomplexes were visualized by the ECL detection system.

Immunocytochemistry.

Cells were seeded onto coverslips coated with poly-l-lysine (BD Labware, Bedford, MA) and grown as described earlier (30, 34). Cells were fixed with 3.7% formaldehyde for 15 min, permeabilized with 0.1% Triton X-100 in PBS for 10 min, and washed with PBS. Coverslips were blocked with 2% BSA in PBS for 20 min and then incubated with primary antibody for 2 h. Coverslips were then washed with 0.1% BSA in PBS for 20 min, followed by a 2-h incubation with an appropriate fluorescent dye-conjugated secondary antibody. Coverslips were mounted on glass slides and observed using a Nikon Eclips 80i UV epifluorescence microscope.

Statistics.

Data are means ± SE. All experiments were performed three times (n = 3). Western blots are representative of three experiments. Student's t-test (for samples with unequal variances) or one-way ANOVA (for samples with equal variances) determined the significance of the differences between means. P < 0.05 was regarded as statistically significant.

RESULTS

MLCK is essential for migration.

Since MLCK phosphorylates MRLC and polyamine depletion increased MRLC phosphorylation (34), we examined the role of MLCK using ML-7, an inhibitor of MLCK. Confluent IEC-6 cells grown in control, DFMO, and DFMO + PUT (DP)-containing medium were wounded and allowed to migrate for 7 h in the presence or absence of ML-7 (10 μM). The wound area covered was measured to quantify migration. Control cells treated with vehicle DMSO (UT) covered significantly more wound area in 7 h when compared with polyamine-depleted cells (DFMO). Cells grown in DFMO + PUT-containing medium migrated comparably to control cells indicating that the inhibition observed in DFMO-treated cells was due to the depletion of polyamines and not due to the effects of DFMO. Furthermore, ML-7 significantly inhibited migration in control and DFMO + PUT groups. ML-7 significantly decreased migration further in the DFMO group when compared with that observed in control and DFMO + PUT groups (Fig. 1A). ML-7 inhibited the phosphorylation of MRLC in control cells (Fig. 1A, inset). In addition, ML-7 inhibited the activities of RhoA and Rac1 induced in response to wounding without altering the levels of total RhoA and Rac1 proteins (Fig. 1B). These results suggest that MLCK activity is essential for migration and that it is linked to the activation of Rho-GTPases.

Fig. 1.

Myosin light chain kinase (MLCK) activity is essential for migration. IEC-6 cells were grown to confluence in control, α-difluoromethyl ornithine (DFMO), and DFMO + putrescine (DFMO + PUT) containing media for 3 days followed by serum starvation for 24 h. Confluent monolayers were wounded with a gel-loading tip in the center of the plates, washed and left untreated (UT), or treated with 10 μM ML-7. A: migration was calculated as described in materials and methods. Values are means ± SE of triplicates. *Significantly different compared with respective UT groups. #Significantly different compared with UT control and DFMO + PUT. Inset, the effect of ML-7 on the levels of phosphorylated myosin regulatory light chain (pMRLC) in control cells. B: at indicated time intervals following wounding, cell extracts were assayed as described in materials and methods to determine the levels of total and active (GTP-RhoA and GTP-Rac1) protein by Western blot analysis. Representative blots from three observations are shown.

Inhibition of Rho-kinase increases migration.

Since Rho-kinase directly phosphorylates MRLC and also increases MRLC phosphorylation by inhibiting the activity of myosin phosphatase, we examined the effect of Y27632, a specific inhibitor of Rho-kinase on migration. Figure 2A shows that 25 μM Y27632 significantly increased migration in control, DFMO, and DFMO + PUT groups (2- to 2.5-fold). pMRLC localized with F-actin in all three groups (Fig. 2B, c, f, and i). However, the colocalization of pMRLC was more prominent with the thick cortical actin in the DFMO group as shown in the merged image (Fig. 2B, f). Although, some cells in untreated control and DFMO + PUT groups also had thick cortical actin at the periphery, pMRLC was not as prominent as that seen in the DFMO group. Cells in control and DFMO + PUT groups had prominent lamellipodia (Fig. 2B, a and g, arrows), whereas the cells in the DFMO group had few (Fig. 2B, d, arrows). Y27632 (25 μM) caused extensive lamellipodia formation in all three groups with concomitant decreases in pMRLC. Although, Y27632 decreased the association of pMRLC with actin stress fibers, the levels of pMRLC residing in the nuclei remained unchanged in all groups (Fig. 2B). Phosphorylation of MRLC at the threonine-18 and serine-19 residues, as well as on both the residues, was increased in the DFMO group and nearly disappeared in response to Y27632 (Fig. 3A). Furthermore, immunoprecipitated pThr18-MRLC probed for pSer19-MRLC and pThr18/Ser19-MRLC (Fig. 3B) confirmed the Western blot analysis (Fig. 3A) and immunolocalization of MRLC (Fig. 2B). Thus decreased phosphorylation of MRLC in response to Y27632 in the DFMO-treated cells dissolved the thick actin cortex and allowed the formation of lamellipodia essential to the migration.

Fig. 2.

Inhibition of Rho-kinase stimulates migration by inhibiting pMRLC. IEC-6 cells grown as described in Fig. 1 were left untreated or treated with 25 μM Y27632. A: migration was calculated as described in materials and methods. Values are means ± SE of triplicates. *Significantly different compared with respective untreated samples. #Significantly different compared with untreated control and DFMO + PUT groups (DP). B: preconfluent IEC-6 cells grown on poly-l-lysine-coated coverslips in control, DFMO, and DFMO + PUT (DP)-containing media for 3 days followed by serum starvation for 24 h were left untreated or treated with 25 μM Y27632 in respective serum-free media for 3 h. Cells were fixed and stained for the localization of pMRLC (Thr18/Ser19) and F-actin. Coverslips were mounted on glass slides and images were captured using CCD camera attachment with a Nikon microscope at ×40 magnification. Representative images from three experiments carried out in triplicate are shown. See text for more details of individual panels.

Fig. 3.

Y27632 inhibits phosphorylation of MRLC in polyamine-depleted cells. IEC-6 cells grown in control and DFMO medium as described in Fig. 2A were left untreated or treated with 25 μM Y27632 in respective serum-free media and incubated for 3 h. C, control; D, DFMO cells were washed with dPBS and lysed using mammalian protein extraction reagent (MPER) containing protease and phosphatase inhibitors. A: equal amounts of protein were subjected to SDS-PAGE electrophoresis followed by Western blot analysis using pSer19MRLC, pThr18MRLC, pThr18/Ser19MRLC, and MRLC antibodies. B: equal amounts of protein were immunoprecipitated using anti-goat pThr18MRLC antibody and separated by SDS-PAGE electrophoresis followed by Western blot analysis using anti-rabbit pSer19-MRLC, anti-goat pThr18-MRLC, anti-goat pThr18/Ser19-MRLC antibodies. Anti-goat antibodies showed light chain band (IgG LC). Representative blots from three observations are shown.

Since decreased activation of Rho-GTPases inhibited migration in polyamine-depleted cells (32), we determined whether the inhibition of Rho-kinase increased migration by activating Rho-GTPases. RhoA activity increased within 0.5 h in response to wounding and began to decline thereafter, whereas Y27632 prolonged RhoA activity for 1.5 h, which also decreased thereafter (Fig. 4A, top). Interestingly, a significant amount of actin was present when GTP-RhoA was pulled down (second panel, 0.5 h for control and 0.5 and 1.5 h for Control + Y27632) suggesting the association of active RhoA with actin. RhoA localized with the actin stress fibers in control cells (Fig. 4B, a–c) and with the thick actin cortex in polyamine-depleted cells (Fig. 4B, d–f). Y27632 increased the formation of lamellipodia and reorganized cortical actin into stress fibers, which remained associated with RhoA in control cells. In the DFMO group, dissolution of the actin cortex by Y27632 accompanied the relocalization of RhoA with the actin stress fibers.

Fig. 4.

Effect of Rho-kinase inhibition on the activation and localization of RhoA. A: IEC-6 cells grown in control medium were wounded and left untreated (minus) or treated (plus) with 25 μM Y27632 in respective serum-free media. At indicated time intervals, cell extracts were assayed as described in materials and methods to determine the levels of active (GTP-RhoA) protein by Western blot analysis. Whole cell extracts were used to determine the total amounts of RhoA and actin. Membranes from the pull-down assay were stripped and probed for the detection of actin using a specific antibody. Representative blots from three observations are shown. B: IEC-6 cells grown, treated, and stained for RhoA and F-actin as described in Fig. 2B. Images were captured using CCD camera attachment with a Nikon microscope at ×40 magnification. Representative images from three experiments carried out in triplicate are shown. See text for more details about individual panels.

Rac1 activity underwent cyclic activation and inactivation consistent with the dynamic nature of actin cytoskeletal remodeling in untreated cells (Fig. 5A). The inhibition of Rho-kinase by Y27632 increased GTP-Rac1 levels within 0.5 h. They remained elevated for another hour and began to decline thereafter without causing a change in total Rac1 protein (Fig. 5A). Furthermore, the association of actin with GTP-Rac1 was not evident (data not shown). These results indicate that the activity of Rho-GTPases correlates with Rho-kinase or dephosphorylation of MRLC. The inhibition of Rac1 by NSC23766 significantly inhibited basal as well as Y27632-induced migration (Fig. 5B). We have shown that the inhibition of Rac1 caused the formation of a thick actin cortex at the cell periphery similar to that observed in cells from DFMO group (32, 43). Unlike RhoA, Rac1 localized in the cytoplasm and was more concentrated in the perinuclear region in the DFMO group (Fig. 5C, b) and in lamellipodia of cells grown in DFMO and treated with Y27632 (Fig. 5C, e).

Fig. 5.

Rac1 activation is essential for migration induced in response to Rho-kinase inhibition. A: IEC-6 cells were grown and treated as described in Fig. 4A. At indicated time intervals, cell extracts were assayed as described in materials and methods to determine the levels of active (GTP-Rac1) protein by Western blot analysis. Whole cell extracts were used to determine the total amounts of Rac1. Representative blots from three observations are shown. B: confluent monolayers were wounded with a gel-loading tip in the center of the plates, washed, and left untreated or treated with Y27632 (25 μM) in the presence or absence of 120 μM NSC23766. Migration was measured as described in materials and methods. Values are means ± SE of triplicates. *Significantly different compared with control. #Significantly different compared with Y27632. C: IEC-6 cells grown in DFMO-containing medium were treated and stained for Rac1 and F-actin as described in Fig. 2B. Images were captured using CCD camera attachment with a Nikon microscope at ×40 magnification. Representative images from three experiments carried out in triplicate are shown. See text for more details about individual panels.

Rho-kinase inhibits MLCP activity by phosphorylating the myosin phosphatase targeting subunit (i.e., MYPT). The NH₂-terminal region of MYPT directly binds with myosin to regulate its contractility. We determined the localization of MYPT, myosin, and actin in control, DFMO, and DFMO + PUT groups of cells. Myosin IIa strongly localized with the F-actin in all three groups examined (Fig. 6A, arrows, c, f, and i). The reorganized actin cytoskeleton in response to Y27632 was characterized by a strong colocalization of myosin IIa with F-actin (Fig. 6A, l, o, and r). MYPT1 localized in the nuclei (asterisk) as well as with F-actin stress fibers in control and DP groups (Fig. 6B, c and I, arrows). However, in the DFMO group little MYPT1 localized with the thick actin cortex (Fig. 6B, e). Upon inhibition of Rho-kinase, MYPT1 redistributed in the cytoplasm and eventually to the F-actin stress fibers (Fig. 6B, l, o, and r) where myosin is localized.

Fig. 6.

Effect of Rho-kinase inhibition on myosin IIa and myosin phosphatase targeting protein 1 (MYPT1) localization. IEC-6 cells grown and treated as described in Fig. 2B were stained for myosin IIa and F-actin (A) and MYPT1 and F-actin (B). Images were captured using CCD camera attachment with a Nikon microscope at ×40 magnification. Representative images from three experiments carried out in triplicate are shown. See text for more details about individual panels.

Since protein phosphatase 1δ (PP1δ) is part of the MLCP holoenzyme, polyamine depletion might decrease the activity or the expression of PP1δ and regulate MLCP activity and thereby MRLC phosphorylation (1, 24). The expression of PP1δ was not altered in control, DFMO-, or DFMO + PUT-treated cells (Fig. 7A). Furthermore, the levels of PP1α and PP2A were unaltered (Fig. 7A). Enzymatic activity of PP1 also remained unchanged in all groups (data not shown). However, the levels of MYPT1 and pSer668-MYPT were higher in the DFMO group when compared with those seen in control and DFMO + PUT groups, while the levels of myosin IIa were unaltered (Fig. 7B). Microcystine-agarose affinity binding assay showed equal amounts of PP2Ac and PP1δ in control, DFMO, and DFMO + PUT groups (Fig. 7C). Interestingly, the amount of actin captured along with the phosphatases was reduced in the DFMO group. Immunoprecipitation of actin also showed reduced binding of PP2A and PP1δ with actin in cells treated with DFMO (Fig. 7D), indicating a role for phosphatases in actin cytoskeletal organization. Calyculin A, an inhibitor of protein phosphatases, significantly increased the phosphorylation of MRLC in all these groups (Fig. 7, E and F). These results indicate that polyamines are required to target MYPT1 to the cytoskeleton and, thereby, recruit phosphatase PP1δ to the actomyosin complex, which results in the dephosphorylation of MRLC.

Fig. 7.

Serine threonine phosphatases and MRLC phosphorylation. IEC-6 cells were grown to confluence in control (C), DFMO- (D), and DFMO + PUT (DP)-containing media for 3 days followed by serum starvation for 24 h. Confluent monolayers were washed with dPBS and lysed using MPER containing protease and phosphatase inhibitors. Equal amounts of proteins were subjected to SDS-PAGE electrophoresis followed by Western blot analysis using specific antibodies to determine the levels of protein phosphatase 1c-α (PP1cα), protein phosphatase 1cδ (PP1cδ), protein phosphatase 2Ac (PP2Ac) (A), and β-actin, MYPT1, pSer668-MYPT1, myosin IIa, and β-actin (B). Equal amounts of cell extracts (200 μg) from control (C), DFMO (D), and DFMO + PUT (DP) groups were incubated with 50 μl microcystine-agarose (MC-agarose) and immunoprecipitated using anti-actin antibody (2 μg) for 2 h at 4°C. Proteins bound to microcystine-agarose (C) and actin (D) were eluted, dissolved in sample buffer, resolved by SDS-PAGE, and transferred to PVDF membranes. Western blot analysis was carried out to detect PP2Ac, PP1cδ, and β-actin. Whole cell extracts were used to determine the total β-actin content. E: confluent monolayers left untreated or treated with 1.5 nM calyculin A (Cal-A) for 3 h were washed with dPBS and lysed using MPER-containing protease and phosphatase inhibitors (E). Equal amounts of proteins were subjected to SDS-PAGE electrophoresis followed by Western blot analysis to determine the levels of pThr18/Ser19-MRLC and β-actin. F: densitometric analysis of pMRLC levels from blots from E. *Significantly different compared with respective untreated samples. #Significantly different compared with untreated C and DP groups.

Rho-kinase and focal adhesions.

Actin cytoskeletal remodeling and migration are intimately linked to the dynamic organization of focal adhesions (FAs) (5, 10). Since focal adhesion kinase (FAK), paxillin, and vinculin are integral components of FAs, localization of these proteins indicates the state of FAs and provides significant information about the migratory behavior of cells. In IEC-6 cells, colocalization of FAK with pY-paxillin and vinculin was observed in both control and polyamine-depleted cells. In control cells, prominent elongated focal plaques containing FAK, pY-paxillin, and vinculin at the cell periphery and within the cell body were observed (Fig. 8A, arrows). In polyamine-depleted cells, minute FAs were observed and were mainly formed at the cell periphery.Inhibition of Rho-kinase caused extensive reorganization of FAs as judged by the localization of pY-paxillin and vinculin (Fig. 8, A and B). The formation of lamellipodia correlated with the localization of pY-paxillin at the tips of stress fiber extensions in control cells treated with Y27632 (Fig. 8B, i, arrows). In the DFMO group, pY-paxillin formed a thin band colocalized with the F-actin cortex with few patches at the tips of rudimentary lamellipodia (Fig. 8B, f, arrows). The inhibition of Rho-kinase reorganized the actin cytoskeleton resulting in large lamellipodial extensions. These extensions were stabilized by the formation of FAs as evident by the localization of pY-paxillin at these tips in both the control and polyamine-depleted cells (Fig. 8B, i and l, arrows). Vinculin, another component protein of FAs, followed a pattern of localization similar to that observed for pY-paxillin (data not shown).

Fig. 8.

Organization of focal adhesions in polyamine-depleted cells. A: IEC-6 cells grown in control and DFMO-containing media as described in Fig. 2B were stained for focal adhesion kinase (FAK), vinculin, and pYPaxillin. B: IEC-6 cells grown in control and DFMO-containing media were treated and stained for pYPaxillin and F-actin as described in Fig. 2B. Images were captured using CCD camera attachment with a Nikon microscope at ×40 magnification. Representative images from three experiments carried out in triplicate are shown. See text for more details about individual panels.

DISCUSSION

Cell migration is a dynamic multistep phenomenon, involving actin polymerization, actomyosin contraction, and focal adhesion reorganization (21, 26, 35). The process is initiated by membrane protrusion, which is driven by actin polymerization in the direction of migration. The establishment of new adhesions stabilizes the membrane protrusions (28). These adhesions mature into stable adhesion plaques upon association with other component proteins such as vinculin and paxillin. Rear contraction follows, mediated by a contractile force generated by actin and myosin, leading to cell body movement. Finally, the rear of the cell detaches from the substratum. Effective migration requires that the membrane protrusions should be restricted to the leading edge, and the adhesions must turnover (7, 42). Although, much is known about the mechanisms by which these processes are regulated, the role of myosin activation is largely unknown in epithelial cells.

Myosin II (muscle and nonmuscle) has been shown to play a critical role in membrane protrusion and retrograde actin movement at the leading edge of migrating cells (11). Recent studies have focused on the phosphorylation of the RLC of myosin II. Phosphorylation of the RLC activates myosin II and leads to its cyclic attachment to actin filaments (6, 8, 41). Preventing the phosphorylation of RLC by inhibiting MLCK or dephosphorylation by MLCP inactivates myosin. In addition to MLCK, several other kinases have been shown to phosphorylate RLC in different cells (14, 18, 24). Rho-kinase increases the phosphorylation of RLC by direct phosphorylation and by preventing its dephosphorylation by inactivating MLCP. Rho-kinase through the phosphorylation and inactivation of the myosin binding subunit of myosin phosphatase (MYPT1) and direct phosphorylation of RLC leads to the generation of contractile forces (5, 19). Increased activation of myosin forms a rigid structure of actin bundles and stable adhesions and decreases Rac1 activity leading to the inhibition of migration (52). Totsukawa et al. (41, 42) showed that inhibition of MYPT by microinjection of inhibitory antibody increased RLC phosphorylation, resulting in thicker stress fibers, cortical fibers, and the inhibition of migration in fibroblasts. In addition, blocking MYPT1 antibody decreased membrane ruffling and blocked focal adhesion turnover, suggesting that the turnover of RLC phosphorylation is essential to migration (5, 7, 17). Furthermore, Xia et al. (52) showed that small interfering RNA-mediated depletion of MYPT1 resulted in prominent stress fibers associated with enhanced RLC phosphorylation. Recently, Bebbin et al. (3) showed that inhibition of NM IIA affected epithelial cell restitution by altering cell-matrix adhesion, F-actin organization, and intracellular signaling.

Over the years, we have shown that polyamine depletion decreased the activity of Rho-GTPases, inhibited FAK autophosphorylation, greatly increased the phosphorylation of MRLC, and inhibited migration (10, 30, 34). Our previous studies have shown that the transfection of IEC-6 cells with constitutively active Rac1 restored migration of polyamine-depleted cells, while constitutively active RhoA failed to do so. Furthermore, dominant negative Rac1 caused actin cytoskeletal changes similar to those seen in the polyamine-depleted cells (32). Recently, we found that the inhibition of FAK prevents Tiam1-induced Rac1 activation during migration (10). These studies indicate that the assembly and disassembly of FAs modulates the activation of Rho-GTPases. Based on some recent studies showing that the increased phosphorylation of MRLC modified the structure of F-actin and FAs similar to that observed in polyamine-depleted cells (10), we speculate that the inhibition of migration in polyamine-depleted cells might be attributed to extensive and constitutive phosphorylation of MRLC. Inhibition of MRLC dephosphorylation, in the absence of polyamines, prevented actin cytoskeletal remodeling resulting in a rigid cytoskeletal structure unable to signal downstream activation of Rho-GTPases to support migration.

Since polyamine depletion significantly decreased Rac1 and RhoA activity (31, 32), it indicates that RhoA-mediated Rho-kinase activity may not be involved in the phosphorylation of MRLC. Moreover, decreased RhoA and, thereby, Rho-kinase activity in polyamine-depleted cells should increase MLCP activity. Therefore, increased MRLC phosphorylation could have resulted from increased MLCK activity. Thus the inhibition of MLCK by ML-7 would be thought to decrease MRLC phosphorylation and allow the reorganization of the actin cytoskeleton. Although, ML-7 decreased MRLC phosphorylation in control cells (Fig. 1A, inset), contrary to this prediction, ML-7 significantly inhibited migration in all groups (Fig. 1A). However, the extent of inhibition was significantly less in the DFMO group. These findings indicate that MLCK activity is essential for the normal cycling of MRLC phosphorylation and, therefore, migration of IEC-6 cells. However, MLCK activity may not be sufficient to account for increased pMRLC in polyamine-depleted cells. Hirano et al. (14) have reviewed the role of protein kinases in the regulation of phosphorylation and dephosphorylation of smooth muscle myosin light chain. The kinases catalyzing the activating phosphorylation of MRLC include Rho-kinase, integrin-linked kinase, Zip kinase, and p21-activated kinase. Other kinases shown to phosphorylate MYPT1 at the inhibitory site include Zip-like kinase, Zip kinase, myotonic dystrophy protein kinase, and Raf-1 (14). The inhibition of Rac1 activity (Fig. 1B) and migration by ML-7 also suggests that myosin activity is important in the regulation of Rho-GTPases. We have previously shown that inhibition of FAK prevented Rac1 activity (10). Therefore, MLCK might modulate FAK and thereby Rac1 activity. Takaishi et al. (40) showed that inhibition of MLCK by ML-7 decreased the formation of stress fibers and FAs. Since the preexisting pMRLC in polyamine-depleted cells did not undergo dephosphorylation upon inhibition of MLCK, it also indicates that additional mechanisms are involved in the dephosphorylation of MRLC. Previously, we found that the formation of lamellipodia and migration in polyamine-depleted cells increased following the inhibition of serine/threonine kinase by staurosporine, and this was associated with the decreased phosphorylation of MRLC (34). Since protein phosphatases-1c and -2Ac are inactivated by phosphorylation, it is unclear whether the observed decrease in pMRLC was due to the inhibition of a specific kinase or the activation of phosphatases.

Although, polyamine-depleted cells had ∼40% less active RhoA (31), the major fraction of RhoA protein strongly localized to the F-actin cortex (Fig. 4B) and GTP-RhoA associated with actin (Fig. 4A). Therefore, RhoA might activate Rho-kinase constitutively in the actin cortex. Consistent with the above, inhibition of Rho-kinase significantly enhanced migration in control, DFMO, and DFMO + PUT groups (Fig. 2). The inhibition of MRLC phosphorylation by staurosporine has been shown to increase lamellipodial extension at the wound edge by decreasing the levels of pMRLC at the leading edge of cells (34). Unlike control cells, polyamine-depleted cells had high levels of pMRLC localized at both the leading and trailing edges of cells with few prominent lamellipodia (34). In polyamine-depleted cells, pMRLC strongly colocalized with the thick actin cortex as seen in the merged image (Fig. 2B, f, arrows) This colocalization disappeared after treatment with Y27632, which increased the formation of lamellipodia and reorganized the actin cytoskeleton (Fig. 2B, o). Immunoprecipitation data clearly show that polyamine depletion increased the phosphorylation (Thr18/Ser19) of MRLC and that inhibiting Rho-kinase decreased it (Fig. 3). This suggests that Rho-kinase plays an important role in maintaining the dynamic turnover or cycling of the phosphorylation of MRLC required for migration. It is important to understand the events following polyamine depletion leading to increased levels of pMRLC given the background of low RhoA activity. Myosin IIa and pMRLC strongly colocalized with F-actin in control, DFMO, and DFMO + PUT groups (Figs. 2B and 6A). Although, these proteins had similar locations, cells in control and DFMO + PUT groups had lamellipodia and F-actin stress fibers, which were greatly diminished in the DFMO group (Figs. 2B and 6B). Furthermore, lamellipodia formation correlated with decreased phosphorylation of MRLC in both untreated cells and those in which Rho-kinase was inhibited. These data indicate that the turnover of MRLC phosphorylation is essential for the reorganization of the actin cytoskeleton.

Since Rho-kinase regulates the activity of MLCP, we determined the levels of PP1cδ protein and its enzymatic activity. The levels of PP1cδ, PP1cα, and PP2Ac were similar in control, DFMO, and DFMO + PUT groups (Fig. 7A). The enzymatic activities of PP1cδ were also similar in all three groups (data not shown), but the activity of PP2A decreased in polyamine-depleted cells (33). Increased MRLC phosphorylation following treatment with calyculin A confirmed that phosphatases play an important role in maintaining the activity of myosin (Fig. 7, E and F). Microcystin-LR binds PP1 and PP2 and inhibits their activity (25). When conjugated with agarose (MC-agarose) in a pull-down assay, PP1 and PP2A were found in equal amounts in all three groups. Interestingly, actin was pulled down along with phosphatases, but the levels of actin were lower in the DFMO group. Furthermore, immunoprecipitation of actin confirmed that polyamine-depletion decreased the association of PP2A and PP1δ with actin (Fig. 7D). This suggests that protein phosphatases are not targeted to the actin cytoskeleton in the absence of polyamines (Fig. 7C). In eukaryotic cells, targeting and regulatory proteins control PP1 activity. The catalytic subunit (PP1c) of PP1 is generally associated with tissue-specific regulatory and targeting proteins (1, 51). Furthermore, the catalytic activity of PP1 is regulated by the phosphorylation state of the regulatory proteins (1, 13). MYPT1, a PP1cδ targeting protein, strongly localized with the actin cytoskeleton as shown in the merged image in control cells (Fig. 6B, c) and in the nuclei of cells in all three groups (Fig. 6B, asterisks). The inhibition of Rho-kinase had little effect on the distribution of MYPT1 in control and DP groups. However, inhibition of Rho-kinase increased the association of MYPT1 with stress fibers in the DFMO group. In addition, MYPT was highly phosphorylated at serine-668 in polyamine-depleted cells compared with cells in the control and DFMO + PUT groups (Fig. 7B). The phosphorylation of MYPT by Rho-kinase has been shown to inhibit phosphatase activity in vitro (2, 18, 24, 50). The decreased localization to the actin cortex and increased phosphorylation of MYPT1 suggest that the greater phosphorylation of MRLC in polyamine-depleted cells might be due to the lack of, or inactivation, of MYPT1 at the actin cortex. This would prevent targeting of PP1cδ to the cortex and, thereby, inhibit the dephosphorylation of MRLC. Although most studies have shown that increased phosphorylation of MRLC enhances migration (2, 8, 9), this is the first study to demonstrate that static phosphorylation of MRLC can prevent remodeling of the actin cytoskeleton in a specific physiological context.

Actin cytoskeletal remodeling accompanied the reorganization of FAs as indicated by the localization of FAK, pYpaxillin, and vinculin (Fig. 8). In control cells, adhesion proteins formed elongated patches at the cell periphery, within the cells, and at the site of lamellipodia. In contrast to control cells, polyamine-depleted cells had small adhesions localized at the cell periphery with the F-actin cortex. The inhibition of Rho-kinase was not only associated with cellular protrusions but also the reorganization of FAs as evident by the distribution of FAK, paxillin, and vinculin (Fig. 8B). This redistribution of the FAs caused the activation of Rho-GTPases. The inhibition of Rac1 activity by NSC23766 prevented migration induced in response to Rho-kinase inhibition (Fig. 5B) suggesting that the activation of Rho-GTPases occurs downstream of MRLC activation and the reorganization of FAs.

We speculate that in the presence of polyamines active RhoA dissociates from stress fibers allowing the depolymerization of F-actin and formation of new stress fibers required to direct the migration of stationary cells. As depicted in Fig. 9, in polyamine-depleted cells activated RhoA predominantly localized with F-actin stress fibers resulting in the activation of Rho-kinase, which in turn phosphorylate MYPT resulting in its inactivation. Inactivated MYPT fails to bind the phosphatase PP1cδ, thereby, preventing the dephosphorylation of MRLC. Since both MLCK and Rho-kinase phosphorylate MRLC, it remains phosphorylated and myosin is constitutively activated in polyamine-depleted cells. In its activated state, myosin binds F-actin forming the heavy actomyosin cortex stabilizing focal adhesions. This stabilization prevents the downstream activation of Rac1, which is essential to the formation of lamellipodia, stress fibers, and migration.

Fig. 9.

Role of myosin activation during migration of intestinal epithelial cells. Schematic representation of the signaling events leading to the constitutive phosphorylation of MRLC, formation of actomyosin cortex, and inhibition of migration in polyamine-depleted IEC-6 cells.

GRANTS

This publication was made possible by National Institute of Diabetes and Digestive and Kidney Disease Grant DK-052784 and ARRA-USPHS-GR-DK052784-13S1.

DISCLOSURES

Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Institutes of Health. This work was also supported by the Thomas Gerwin endowment. No conflicts of interest, financial or otherwise, are declared by the author(s).