Figure 1.
Distribution of Myo9b in osteoclasts.
Osteoclast podosomes and sealing zones were labeled with fluorescent phalloidin and anti-Myo9b antibodies and viewed by confocal microscopy. A, In mature osteoclasts on glass coverslips, Myo9b is at its highest levels in the perinuclear region and in the peripheral podosome belt. Nuclei are visible in blue in the merged image. B, In osteoclasts on glass coverslips, Myo9b associates closely with the podosome core protein α-actinin. C, In immature osteoclasts on glass that have not formed peripheral podosome belts, Myo9b is present, though not enriched, in internal podosome rings. D, Myo9b is mostly absent from sealing zones in osteoclasts on bone. A Z-stack image of an osteoclast on bone demonstrates that Myo9b is present throughout the cytoplasm. Scale bars = 20 µm.
Figure 2.
Knockdown of Myo9b increases cellular Rho activity.
A, Competitive RT-PCR (left panel) and Western blot (right panels) show efficient knockdown of Myo9b in mouse bone marrow-derived osteoclasts. For competitive RT-PCR, 1 pg of a synthetic RNA containing the Myo9b primer binding sites (the internal standard) was added to 1 µg sample total RNA, as described in Materials and Methods. RT-PCR then resulted in amplification of both cellular Myo9b and the internal standard, which served as a control for relative Myo9b mRNA levels. By Western analysis, knockdown of Myo9b did not significantly change protein expression of two other osteoclast myosins. GAPDH is shown as a loading control. N.S. = non-specific band. B, Protein pull-down followed by Western analysis and densitometry was used to quantify levels of active Rho or Rac in control or Myo9b siRNA-treated marrow-derived osteoclasts. Levels of active small G-proteins were normalized to levels of total Rho or Rac. The graph was compiled from three such experiments and shows that knockdown of Myo9b resulted in increased cellular levels of Rho but not Rac. #: P<0.001; N.S.: not significant.
Figure 3.
Knockdown of Myo9b causes Rho-dependent changes in podosome patterning and microtubule stability.
A, Fluorescent phalloidin labeling demonstrates that suppression of Myo9b results in loss of podosome belt formation that is reversed by inhibition of Rho with a cell-permeant C3 transferase. Scale bars = 20 µm. B, Suppression of Myo9b decreases the percentage of mature osteoclasts with podosome belts in a Rho-dependent manner. *: P<0.001; N.S. = not significant. C, Knockdown of Myo9b significantly decreases osteopontin-directed motility of marrow-derived osteoclasts. D, Knockdown of Myo9b causes loss of microtubule networks in the majority of siRNA-treated osteoclasts. Arrowheads indicate the cell periphery in an siRNA-treated cell. Scale bars = 20 µm. E, Suppression of Myo9b by siRNAs strongly diminishes acetylation of α-tubulin in osteoclasts while not affecting its total expression.
Figure 4.
Knockdown of Myo9b causes a Rho-dependent loss of bone resorptive capacity.
A, siRNA-mediated knockdown of Myo9b diminishes resorptive capacity of osteoclasts in a Rho-dependent manner, as demonstrated by photomicrographs of resorption pits on ivory slices (top). Scale bar = 50 µm. Quantification of the surface area of each pit (pit area) and total resorbed surface area (total resorption) are shown graphically below. #: P<0.05; *: P<0.005; N.S. = not significant. B, Knockdown of Myo9b does not significantly affect osteoclast number or the ability of cells to form sealing zones on bone, as indicated by photomicrographs of phalloidin-labeled sealing zones (top) and enumeration of cell number and sealing zone number and size (bottom).
Figure 5.
Overexpression of the Myo9b tail domain alters podosome and sealing zone patterning.
A, Western analysis demonstrates that transfected osteoclasts express levels of the Myo9b tail that are about 2-fold greater than control (transfected with empty vector). B, Fluorescent F-actin labeling of osteoclasts generated from control or Myo9b-transfected cells shows that F-actin distribution and spreading are severely disrupted in cells plated either on glass or bone. Scale bars = 20 µm.
Figure 6.
Knockdown of Myo9b causes mislocalization and diminished activation of Src kinase.
A, Immunocytochemical labeling of podosome belts in mature osteoclasts demonstrates that Src is mislocalized when Myo9b levels are suppressed. Arrows indicate the normal “railroad track” distribution of Src around podosome belts in control cells, while Src is present in the F-actin core of podosomes in siRNA-treated cells. B, Immunoprecipitation of total cellular Src kinase followed by Western blot with a phospho-tyrosine antibody demonstrates that siRNA treatment diminishes Src phosphorylation without affecting Src protein levels. One lane of immunoprecipitation with the mouse IgG MOPC-21 was added as a negative control. C, Immunocytochemical labeling of osteoclasts with an antibody against phosphorylated Src shows its relative absence in siRNA-treated cells. D, Inhibition of Src kinase causes loss of podosome belts in untreated and control osteoclasts, but does not further diminish podosome belts in cells with suppressed levels of Myo9b. E, Immunocytochemical analysis shows that Pyk2 distribution does not change with siRNA treatment. For all photomicrographs in this figure, scale bars = 20 µm.
Figure 7.
TNFα rescues osteoclast defects caused by loss of Myo9b expression.
A, Overnight addition of TNFα to mature control or siRNA-treated osteoclasts results in larger sealing zones, as shown by phalloidin labeling of cells on bone (left) and quantification of sealing zone perimeter (right). Scale bars = 100 µm. B, Overnight addition of TNFα to mature or siRNA-treated osteoclasts increases the average size of individual resorbed areas and strongly increases levels of total resorption.