Open the LightBox software for image acquisition. To select the laser and filter sets, use parameters for an orange dye by selecting the dye used in the staining from the dye list. Using an overview, click on live to focus the cells.
Once the cells are in focus, adjust the confocal acquisition settings. Next, select the rectangular ROI button and create a region of interest around a mitochondria of interest by clicking and dragging to shape the region. To adjust detector gating for stimulated emission depletion or stead acquisition, next to the general menu, select the gating menu or click and hold to add the menu to the view.
For stead acquisition, set the excitation laser to 15 to 20%and the stead depletion laser to 20 to 25%with 10 line accumulations. Use a pixel dwell time of four microseconds and a pixel size of 20 to 25 nanometers. Perform time lapse by selecting the time dropdown menu, then setting the number of iterations to five and a time interval of 25 or 30 seconds.
A Z-stack can be taken using the volume option and adjusting the desired Z volume range and step size. Open stead image deconvolution software to deconvolute raw stead images with the software algorithm. After ensuring microscopic parameters are correct, select the express button and set the deconvolution type too fast, standard, aggressive, or conservative for varying degrees of deconvolution power.
Execute the deconvolution. Save the images in ICS2 format. In image J, click on file then open to open the ICS2 files from the deconvolution software.
Select plugins, then segmentation followed by trainable Weka segmentation to open the deconvoluted stead images in the trainable Weka segmentation plugin. In the segmentation settings, select the Gaussian blur, membrane projections and sobel filter features. Label one class as cristae and the other as background.
Next, draw a line over the structure to assign to either class. Then select the add button on the right hand side for either cristae or background. Then select the train classifier button on the left hand side to generate a map based on the information provided to the plugin.
Click the save classifier button to save the classifier settings for future segmentation. Use the cristae probability map to threshold the image in image J to generate a binary mask, and then go to analyze, then analyze particles. Finally, draw a multi-point line, adjust the line thickness to several pixels wide and spline the line to fit the mitochondria.
In this study, imaging of mitochondria in undifferentiated, retinoic acid differentiated SH-SY5Y cells with time-lapse imaging as shown. The deconvoluted stead images can be used to determine cristae periodicity in a given area and cristae size and shape measurements.
