To take advantage of the benefits of microscopy, it was important to preserve the overall morphology of the specimen while also obtaining an optimal signal for myosin light chain phosphorylation. We first used overall cell morphology as the qualitative criteria to screened several fixation techniques. We tested four different fixation reagents: ethanol (EtOH), methanol (MeOH), acetone (Ac), and paraformaldehyde (PFA). PFA best preserved cell morphology while the organic solvents ethanol, methanol, and acetone, gave variable results as fixatives and in some cases caused cells to round up upon addition of any one of the three solvents (data not presented). PFA was selected as the fixation reagent, and four different permeabilization methods were screened to select the optimal permeabilization strategy. After fixation by PFA at room temperature, cells were permeabilized with methanol (PFA+MeOH, 4°C), ethanol (PFA+EtOH, 4°C), acetone (PFA+Ac, 4°C) or triton X-100 in PBS (PFA+TX100, room temperature). We also tested a simultaneous fixation and permeabilization procedure by including 0.5% triton X-100 in the PFA solution (PFATX, room temperature). The presence of phospho-MLC antibody staining along well-defined f-actin stress fibers was used as the criterion for evaluating the permeabilization method.
Each of the different permeabilization techniques resulted in distinct differences in the staining patterns. MeOH permeabilized cells showed long filaments of phospho-MLC staining organized in a manner reminiscent of f-actin stress fibers, but very low intensity f-actin stress fiber staining by Alexa 594-phalloidin (Fig. ). EtOH permeabilized cells showed an expected pattern of staining typical of both phospho-myosin filaments and f-actin stress fibers but did not produce the same intensity of staining as the PFATX method (Fig. ), while acetone permeabilized cells did not show staining for either phospho-MLC or stress fibers (Fig. ). Longer exposure times were required for the Alexa 594-phalloidin images for cells permeabilized with MeOH (10 seconds), EtOH (10 seconds), and Ac (5 seconds) than for cells permeabilized with PFA+TX or PFATX (1 second). The two TX-100 permeabilization methods yielded comparable stress fiber morphology (Fig. ), but PFATX resulted in the best overall staining of stress fibers and phosphorylated MLC (Fig. ), and the greatest fold change in MLC-phosphorylation signal upon inhibition by Y27632 (data not presented). In each of these fixation conditions where phospho-MLC signal could be observed, treatment of the cells with 10 μM Y27632 for 1 h before fixation dramatically reduced the signal from phospho-MLC (Fig. ). Upon Y27632 treatment, even when most of the MLC phosphorylation signal was inhibited, there still remained well organized parallel strands of f-actin (compare Fig. to ). Based on these results, PFATX was found to be optimal among the methods tested to preserve myosin light chain phosphorylation and actin stress fibers.
Figure 1 Results of fixing and permeabilizing conditions on phospho-MLC and actin filament staining. Images in the left column show cells that were fixed in the absence of any drugs to inhibit MLC phosphorylation. Images on the right depict cells fixed after treatment (more ...)
We quantified myosin light chain phosphorylation intensity in individual cells by employing Texas-Red maleimide (Tx-Red) labeling of each cell as a mask for the entire cell area [15
], and applying that mask to the same field examined with appropriate filters for Alexa 488 immunofluorescence. We found no systematic contribution to the phospho-MLC signal from the Tx-Red cell mask as discussed in Methods. We examined the contribution of non-specific binding of secondary antibody to the measured phospho-MLC signal and found a small constant contribution to the signal in secondary antibody only controls. This background signal was corrected for as described in Methods.
Figure shows the effect of increasing concentrations of the ROCK inhibitor on the intensity of phospho-MLC staining. The amount of diphospho-MLC in cells decreased with increasing concentration of Y27632. By examining individual cells, it is clear that MLC diphosphorylation varies between individual cells in the population. At 10 μM of Y27632, a concentration which is commonly reported in literature for use with cells, about 85% of the cellular diphospho-MLC was inhibited (Fig. ). Using the means of the distributions, the Y27632 concentration for half maximal inhibition of phospho-MLC, or IC50 value, was estimated from the inhibition curve to be 2.1 ± 0.6 μM (n = 3, Fig. ). This result was not significantly different from that obtained by Western blotting, 1.3 ± 1.2 μM (n = 3, Fig. ).
Figure 2 Dose response of phospho-MLC to Y27632 concentration. A) MLC diphosphorylation decreased with increasing concentration of Y27632 as shown by representative histograms of the range of phospho-MLC fluorescence in 150 cells measured. Histograms labels refer (more ...)
Using Tx-Red staining of cells to determine the spread area of individual cells, we examined the relationship between phospho-MLC intensity and cell morphology. We found a correlation between cell area and myosin light chain diphosphorylation (Fig. ) indicating that larger cells tended to contain more activated MLC. When cells were treated with Y27632, the diphospho-MLC signal was decreased, but the relation between MLC phosphorylation and cell area remained linear (Fig ) over the full range of cell areas. We also observed no significant change in the mean area or the distribution of cell areas measured across the population in absence or presence of Y27632 (Fig ). The mean cell areas and their standard deviations from three replicate measurements were 3465.4 ± 15.7 μm2 in the absence of Y27632 and 3523.6 ± 90.6 μm2, 3877.4 ± 55.04 μm2, 3673.1 ± 193.1 μm2, and 3505.8 ± 63.3 μm2 in the presence of 0.5, 2.5, 10 and 100 μM Y27632 respectively. The coefficient of variation (CV) of cell sizes within the population was also measured for each treatment, in order to assess the range of cell sizes present. The CVs were calculated from the standard deviation divided by the mean cell area within the population. Standard deviations between CVs determined from three replicate experiments were also calculated. The CV in cell area in the absence of Y27632 was 50.4 ± 11.2 and was 49.2 ± 3.2, 57.2 ± 7.1, 72 ± 14.3, 58.9 ± 11.9 in the presence of 0.5, 2.5, 10, and 100 μM Y27632 respectively. The absence of change in cell spreading even as diphospho-MLC signal decreased suggests that at least in this cell type, MLC diphosphorylation is not directly responsible for cell spread area. It should be noted that even though the cell spread area did not change with drug treatment, there appeared to be changes in cell shape and the roughness of cell edges. At the highest level of Y27632 treatment, 100 μM, there was a statistically significant increase in cell perimeter compared to untreated cells (p < 0.05, n = 3) (Fig. ). Images of cells in the absence of Y27632 (Fig. ), 10 μM Y27632 (Fig. ) and 100 μM Y27632 (Fig. ) are shown.
Figure 3 The relationship between cell shape and phospho-MLC in individual cells. A) In the absence of any drug treatment, the area of individual cells was linearly correlated with total diphospho-MLC content per cell. The data shown are from a representative (more ...)
In the analysis presented thus far, we selectively considered single cells that were not in direct contact with other cells. The automated microscopy and staining techniques employed in this study allow us to determine if MLC diphosphorylation levels in single cells are different from levels in cells that are in contact with each other. Cell-contact mediated through cadherins has been reported to change RhoA activity in several different cell types [16
]. Since RhoA is an upstream activator of ROCK, we examined possible downstream effects of cell-cell contact on cell spreading and MLC diphosphorylation. In the absence of Y27632 the average spread area per cell of pairs of cells that contacted each other appeared to be significantly less than that of isolated single cells (Fig. , p < 0.01, n = 3). However, this difference was not observed upon treatment with Y27632 (Fig. , p > 0.1, n = 3). Furthermore, there was no apparent difference in MLC diphosphorylation per cell in pairs compared to single cells at any of the concentrations of Y27632 used (Fig. ). We conclude that under these conditions, cell-cell contact does not affect MLC diphosphorylation in vascular smooth muscle cells.
Figure 4 Comparison of phospho-MLC response in individual cells with cells in contact with other cells. Means and standard deviations (n = 3 replicate experiments) are shown. Number of cells measured is about 150 single cells and 30 cells in pairs. A) Comparison (more ...)
To further validate the method, we employed additional ways to perturb myosin light chain phosphorylation. ECM density influences myosin light chain phosphorylation by changing the functional coupling between Rho and ROCK (Bhadriraju et al, 2007). Hence we examined the effect of lowering ECM density on MLC phosphorylation. The coating concentration of fibronectin was lowered from 10 μg/ml to 0.2 μg/ml. Myosin phosphorylation on 0.2 μg/ml FN surfaces was 20.7% ± 1.44% of controls (10 μg/ml fibronectin) as measured by microscopy (Fig. ) and 8.47% ± 5.51% of controls as measured by western blotting (Fig. ). Correspondingly, lowering ECM density also reduced the extent of cell spreading (Fig. ).
Figure 5 Phospho-MLC response to ECM density. Comparison of MLC phosphorylation (A, B) and cell spreading (C, D) between cells on substrates coated with 10 μg/ml FN and 0.2 μg/ml FN. Error bars indicate standard deviations of 3 replicate experiments. (more ...)
We also examined the effect of perturbations designed to increase MLC phosphorylation relative to controls. Cells were treated with the protein phosphatase inhibitor calyculin A to activate MLC phosphorylation, and MLC phosphorylation was measured by microscopy and western blotting. Addition of 5 nM calyculin A to cells for 3 minutes increased MLC phosphorylation to 172.1% ± 7.42% of untreated controls as measured by microscopy (Fig. ) and 160.5% ± 30.1% of controls as measured by western blotting (Fig. ). There was a small but statistically significant decrease in cell area within 3 minutes after drug addition detected by quantitative microscopy (Fig. ). It was observed that cells completely rounded up at approximately 10 minutes after the addition of the drug (data not presented).
Figure 6 Phospho-MLC and cell spreading in response to the addition calyculin A. Comparison of phospho-MLC (A,B) and cell spreading (C, D) in response to the addition of 5 ng/ml of calyculin A. Error bars indicate standard deviations of 3 replicate experiments. (more ...)
We finally examined if the method could measure MLC phosphorylation in other cell types by using Y27632 to inhibit MLC phosphorylation in NIH 3T3 cells. 10 μM Y27632 was found to inhibit MC phosphorylation by 85% relative to non-drug treated controls, both by quantitative microscopy and western blotting (Fig. ) without a statistically significant change in cell area (Fig. ).
Figure 7 Measurement of ML phosphorylation in NIH 3T3 cells. A) Quantitative Microscopy measurement of phospho-MLC in the absence or in the presence of 10 μM Y27632 in 3T3 cells, B) corresponding western blotting data and C,D) cell spreading. Error (more ...)