Nonmuscle myosin II isoforms coassemble in living cells

Abstract

Nonmuscle myosin II (NM II) powers myriad developmental and cellular processes, including embryogenesis, cell migration, and cytokinesis [1]. To exert its functions, monomers of NM II assemble into bipolar filaments that produce a contractile force on the actin cytoskeleton. Mammalian cells express up to three isoforms of NM II (NM IIA, IIB, and IIC), each of which possesses distinct biophysical properties and supports unique as well as redundant cellular functions [2-8]. Despite previous efforts [9-13], it remains unclear whether NM II isoforms assemble in living cells to produce mixed (heterotypic) bipolar filaments or whether filaments consist entirely of a single isoform (homotypic). We addressed this question using fluorescently tagged versions of NM IIA, IIB, and IIC, isoform-specific immunostaining of the endogenous proteins, and two-color total internal reflection fluorescence structured-illumination microscopy, or TIRF-SIM, to visualize individual myosin II bipolar filaments inside cells. We show that NM II isoforms coassemble into heterotypic filaments in a variety of settings, including various types of stress fibers, individual filaments throughout the cell, and the contractile ring. We also show that the differential distribution of NM IIA and NM IIB typically seen in confocal micrographs of well-polarized cells is reflected in the composition of individual bipolar filaments. Interestingly, this differential distribution is less pronounced in freshly spread cells, arguing for the existence of a sorting mechanism acting over time. Together, our work argues that individual NM II isoforms are potentially performing both isoform-specific and isoform-redundant functions while coassembled with other NM II isoforms.

Figure 1

TIRF-SIM of cells expressing NM IIA with N- and C-terminal fluorescent tags allows identification of individual NM IIA bipolar filaments. (A) Cartoon of NM II alone, with an N-terminal EGFP reporter, or with a C-terminal mApple reporter (light chains not depicted). (B) Cartoon of a NM II bipolar filament containing a combination of the three NM II constructs pictured in (A). Below the filament is a cartoon depicting what one would expect to see when such a filament was imaged with TIRF-SIM. (C) Shown is a TIRF-SIM image of a U2OS cell expressing EGFP-NM IIA. The red numbered boxes correspond to the magnified insets to the right. The scale bars represent 2 μm for C and 300 nm for C1-C3. (D and E) Shown are TIRF-SIM images of U2OS cells expressing EGFP-NM IIA and the F-actin reporter mApple-F-tractin. The white numbered boxes correspond to the insets to the right. The scale bars represent 2 μm for D and E and 300 nm for the insets. (F) Shown is a TIRF-SIM image of a U2OS cell co-expressing EGFP-NM IIA and NM IIA-mApple. The white numbered boxes correspond to the magnified insets to the right. The scale bars represent 2 μm for F and 300 nm for F1-F3.

Figure 2

NM IIA and NM IIB form heterotypic filaments in live cells. U2OS cells co-expressing NM IIA-mApple and EGFP-NM IIB were imaged using either confocal microscopy (A) or TIRF-SIM (B-E). The white arrow in the middle panel of (A) indicates the leading edge, where NM IIB signal is reduced but not absent. For each TIRF-SIM image, the white numbered boxes correspond to the magnified insets to the right. Heterotypic filaments are apparent in regions of the cell without stress fibers (B), in lamellar extensions and transverse arcs (C), in ventral stress fibers (D), and in sub-nuclear stress fibers (E). In inset C1, the white arrow indicates a heterotypic filament while the white arrowhead indicates a filament with little or no signal for EGFP-NM IIB (which could be either a homotypic filament of myosin IIA or a heterotypic filament containing unlabelled NM IIB or too few molecules of EGFP-NM IIB to be clearly visible; see Experimental Procedures for further discussion). The scale bar in A represents 10 μm. The scale bars in B-E represent 2 μm for the larger images and 300 nm for the insets. See also Figures S1 and S2.

Figure 3

NM IIA and NM IIB form heterotypic filaments in the contractile ring of LLC-Pk1 cells. LLC-Pk1cells expressing NM IIA-mApple and EGFP-NM IIB were imaged during anaphase using TIRF-SIM. The two images are of the same cell at early anaphase (A) and mid-late anaphase (B). White arrows indicate the position of the forming cleavage furrow. The numbered boxes in the larger images correspond to the magnified insets to the right. The scale bars represent 2 μm for larger images and 300 nm for insets. See also Movie S2.

Figure 4

Endogenous NM IIA and NM IIB form heterotypic filaments in live cells. (A) Cartoon depicting the co-assembly of EGFP-NM IIA with endogenous NM IIB. When NM IIB is localized using an antibody that recognizes its C-terminus (together with a red secondary antibody; red Y), the resulting heterotypic filament will display a green-red-green pattern when imaged with TIRF-SIM. (B) Cartoon depicting the co-assembly of endogenous NM IIA and endogenous NM IIB filament. Staining with antibodies that recognize the C-termini of both NM IIA and NM IIB, together with green (green Y) and red (red Y) secondary antibodies, respectively, should result in a strong overlap between the two signals when heterotypic filaments are present. If both signals are of equal intensity and perfectly overlapping, the result will be a single yellow punctum. Any offset in the two signals will result in a punctum with a yellow center and red and green flanking regions. Signal offset could result from steric hindrance between adjacent antibodies, uneven labeling of filaments by the antibodies, and/or uneven distribution of isoforms on either side of the bipolar filament. In addition, the intensity of the yellow signal will be diminished in heterotypic filaments where the signal for one isoform is much stronger than the signal for the other isoform. (C) COS-7 cells expressing EGFP-NM IIA were fixed, immuno-stained for endogenous NM IIB (followed by a red secondary), and imaged with TIRF-SIM. The white numbered boxes correspond to the magnified insets to the right. The scale bars represent 2 μm for the larger image and 300 nm for the insets. (D-F) MDA-MB-231 cells were fixed, immuno-stained with a rabbit polyclonal antibody for endogenous NM IIA (followed by a green secondary) and a mouse monoclonal antibody for endogenous NM IIB (followed by a red secondary), and imaged with TIRF-SIM. (D and E) Shown are TIRF-SIM images of lamellar regions, transverse arcs and ventral stress fibers. The numbered boxes correspond to the row of images to the right, which present grey scale images for αNM IIA and αNM IIB and the merged image. Arrowheads indicate some of the overlapping green and red puncta indicative of heterotypic filaments. Note that yellow is not obvious in D2 because the signal for NM IIA is much stronger than the signal for NM IIB. The scale bars represent 2 μm for the larger images and 300 nm for insets. (F) Discrete individual puncta were analyzed for their relative content of endogenous NM IIA and NM IIB and the results plotted as a function of distance from the cell edge in 2 μm increments (see Experimental Procedures for details). The data, which is plotted on a log scale as geometric means with 95% confidence intervals, is a compilation of over 8, 800 puncta from 6 cells. Note that the ratios in (F) are relative and do not necessarily represent that actual ratio of NM IIA to NM IIB in individual filaments due to differences in expression and assembly levels (see Experimental Procedures). See also Figures S3 and S4.