TRPM7 regulates myosin IIA filament stability and protein localization by heavy chain phosphorylation

Abstract

Deregulation of myosin II-based contractility contributes to the pathogenesis of human diseases, such as cancer, which underscores the necessity for tight spatial and temporal control of myosin II activity. Recently, we demonstrated that activation of the mammalian alpha-kinase TRPM7 inhibits myosin II-based contractility in a Ca(2+)- and kinase-dependent manner. However, the molecular mechanism is poorly defined. Here, we demonstrate that TRPM7 phosphorylates the COOH-termini of both mouse and human myosin IIA heavy chains--the COOH-terminus being a region that is critical for filament stability. Phosphorylated residues were mapped to Thr1800, Ser1803 and Ser1808. Mutation of these residues to alanine and that to aspartic acid lead to an increase and a decrease, respectively, in myosin IIA incorporation into the actomyosin cytoskeleton and accordingly affect subcellular localization. In conclusion, our data demonstrate that TRPM7 regulates myosin IIA filament stability and localization by phosphorylating a short stretch of amino acids within the alpha-helical tail of the myosin IIA heavy chain.

Fig. 1

Fragments of the MHCIIA that served as substrate in TRPM7 kinase reactions. A schematic diagram of the MHCIIA is depicted, with the coiled-coil domain spanning amino acids 1091 to 1924. The MHCIIA ends with a short nonhelical tail piece (amino acids 1924–1960). For each myosin IIA fragment, the first and last amino acids are indicated on either side of the line. Above the line is the total number of threonine and serine residues within each fragment.

Fig. 2

TRPM7 phosphorylates the COOH-terminus of mouse MHCIIA. Different regions of the coiled-coil domain of mouse MHCIIA were expressed as GST fusion proteins. The purified recombinant proteins were incubated with WT or KD TRPM7 in the presence of [γ-³²P]ATP. The products of the kinase reaction were separated by SDS-PAGE, and the gel was stained with Coomassie brilliant blue (bottom panel). Phosphorylated proteins were detected by autoradiography (top panel).

Fig. 3

TRPM7 phosphorylates the α-helical tail but not the nonhelical tail of human MHCIIA. The COOH-termini and nonhelical tails of mouse and human MHCIIA were purified as GST fusion proteins and incubated with WT or KD TRPM7 in the presence of [γ-³²P]ATP. The proteins were separated by SDS-PAGE and visualized by staining the gel with Coomassie brilliant blue (bottom panel). Phosphorylated proteins were detected by autoradiography (top panel).

Fig. 4

Kinetics of myosin IIA phosphorylation by TRPM7. Purified TRPM7 was mixed with 2 μg of GST–myosin IIA, and the reaction was initiated by the addition of 0.1 mM ATP containing 5 μCi of [γ³²P]ATP. The reaction proceeded at 30 °C for the indicated times, after which 20 mM EDTA was added to stop the kinase reaction. Proteins were resolved on 10% SDS-PAGE gel and detected by Coomassie staining (bottom panel). Phosphorylated TRPM7 and GST–myosin IIA were revealed by autoradiography (top and middle panels).

Fig. 5

Mapping of phosphorylation sites in MHCIIA. (a) Identification of Ser1803 by Edman degradation sequencing of ³²P-labeled peptides. GST–MHCIIA COOH-terminus was incubated with TRPM7 in the presence of [γ-³²P]ATP and subjected to SDS-PAGE. The phosphorylated GST–MHCIIA fusion protein was excised from the gel and digested with trypsin, and the peptide mixture was separated by HPLC on a C18 column. The phosphopeptides were sequenced by Edman sequencing, with ³²P radioactivity being measured after each cycle of degradation. The phosphorylated residue was assigned by a combination of solid-phase Edman sequencing and MALDI–TOF. (b) Identification of phosphorylation sites by LC-MS/MS. Phosphorylated GST–MHCIIA COOH-terminus was digested with trypsin, and peptides were separated on a nano-LC C18 column, which was connected inline with a high-mass-accuracy LTQ-FT mass spectrometer. Representative MS² and MS³ spectra of a peptide phosphorylated at position Ser1808 are depicted (m/z observed of parent ion=637.8227, mass accuracy=0.82 ppm, +2 charge state; NL indicates neutral loss of H₃PO₄, which triggers acquisition of MS³ spectrum). (c) Summary table of the phosphorylation site results by LC-MS/MS.

Fig. 6

Ser1803 is the major phosphorylation site in the COOH-termini of the coiled-coil domains of both human and mouse myosin IIA. (a) Mutations of Thr1800, Ser1803 and Ser1808 in human MHCIIA reduce phosphorylation to background levels. The phosphorylated residues mapped in GST–human MHCIIA COOH were mutated to alanine individually or in combination. Since substrate recognition by α-kinases may be influenced by the secondary structure, Thr1810, which was not identified by MS or predicted to be phosphorylated, was also mutated as a negative control. GST–human MHCIIA proteins were incubated with TRPM7 in the presence of [γ-³²P]ATP and subjected to SDS-PAGE. Equal loading of the GST fusion proteins was verified by Coomassie staining (top panel), and phosphorylated proteins were detected by autoradiography. The phosphorylation of the different GST–human MHCIIA proteins by TRPM7 is depicted in the middle panel. The presence of equal levels of kinase activity in each sample was determined by monitoring TRPM7 autophosphorylation (bottom panel). (b) Quantification of the degree of phosphorylation of the different GST–human MHCIIA fusion proteins by TRPM7 using phosphorimager analysis. The level of ³²P incorporation into the WT GST–human MHCIIA fusion protein was set to 1, and phosphorylation of all other proteins is reported relative to this value. (c) Ser1803 is the major phosphorylation site in the helical region of mouse myosin IIA. GST–mouse MHCIIA was mutated at positions Ser1803 and Thr1931 either individually or in combination. Phosphorylation of these GST–mouse MHCIIA mutants by TRPM7 was compared with WT GST–mouse MHCIIA by in vitro kinase assay as described in (b). (d) Phosphorylation of different mouse MHCIIA proteins was quantified as described in (b).

Fig. 7

Conservation of phosphorylation sites between mouse, human and Dictyostelium myosin II. (a) Alignment of the mouse MHCIIA and human MHCIIA, MHCIIB and MHCIIC COOH-termini. The phosphorylation sites are underlined, and conserved amino acids are shown in boldface. Note that two of three residues in the conserved stretch of amino acids within the coiled-coil domain of human MHCIIAwere verified in mouse MHCIIA by LC-MS/MS. Alignment with human MHCIIB and MHCIIC reveals that the major phosphorylated residue in human MHCIIA, Ser1803, is conserved in MHCIIB but not in MHCIIC, whereas Ser1808 in mouse and human MHCIIA is conserved in both human MHCIIB and MHCIIC.

Fig. 8

Mutation of phosphorylation sites to aspartic acid affects myosin IIA filament stability. (a) The assembly of WT and mutant rods was monitored using a standard sedimentation assay. (●) WT; (○) 3×A; (▪)3×D. The solid lines represent the best fit to the Hill equation. (b) Midpoint measurement for WT and mutant myosin IIA rods. Data are reported as the mean±SEM for two to three independent experiments.

Fig. 9

Mutation of phosphorylation sites to aspartate increases Triton solubility of myosin IIA. YFP–MHCIIA constructs were transiently transfected into COS7 cells. The cells were fractionated into Triton-soluble and -insoluble fractions. Equal amounts of these fractions were separated by SDS-PAGE and immunoblotted. Exogenous myosin IIA was detected using anti-GFP antibodies, followed by IRDye 800-coupled antimouse immunoglobulin G secondary antibodies. Fluorescence was measured using an Odyssey Infrared Imaging System. (a) Representative Western blot depicting the presence of YFP–MHCIIA mutants in the different fractions. (b) Quantification of the degree of solubility of each YFP–MHCIIA fusion protein. The data are presented as the percentage of soluble MHCIIA protein (mean±SEM, n=5). *p<0.05.

Fig. 10

Relocalization of YFP–MHCIIA 3×D mutant away from the cortex. (a) Representative images of COS7 cells expressing myosin IIA mutants. YFP–MHCIIA constructs were transiently expressed in COS7 cells. Subsequently, the actin cytoskeleton was stained using Texas Red-conjugated phalloidin and cellular localization of YFP–MHCIIA and F-actin was visualized by fluorescence microscopy. (b) Cortical index measurement of the myosin IIA mutants. The cortical index for 10 representative cells was measured as previously described. Data are presented as mean±SEM. *p<0.001.