The major goal of this project is to achieve precise correlation between images in the light microscope, where specific proteins are routinely identified using fluorescent antibodies, and images in the IVEM, where protein identification is more difficult, but resolution and detection of small structures is greater. The logical extension of this is to examine the very same cells by both light and electron microscopy. We are able to obtain images of the same filamentous and pre-filamentous structures in cultured cardiac myocytes using ordinary light, epifluorescence, confocal laser scanning, and intermediate voltage electron microscopes. The sequence of examination is: 1) phase or interference microscopy of living cells, 2) epifluorescence microscopy of cells that have been lightly fixed, permeabilized, and stained with fluorescent tagged antibodies to specific filament proteins, 3) scanning confocal laser microscope images of the same, and 4) IVEM images obtained after further fixation and critical point drying. In addition to cardiac myocytes, we have studied fibroblasts, localizing alpha actinin and vinculin using monoclonal and polyclonal antibodies. A second project that requires this technology and also provides good test images to evaluate the correlation programs looks at the early expression of genes for intermediate filament proteins that have been introduced into cells using molecular biological techniques. Often what is seen in confocal fluorescence microscopy is very small and very irregularly shaped objects, and when the location of these is explored in the same cells in the IVEM, the cytoplasmic areas involved are densely packed with filamentous structures of many sizes and configurations. Clearly only a small part of what is seen in the IVEM belongs to the protein class being seen in fluorescence images using antibodies. It will be totally impossible to identify the corresponding structures in the IVEM images without detailed and precise mapping of one image onto the other. Our computer graphic programs import the two images and superimpose them, making adjustments of scale, position, and orientation to get the best fit (currently evaluated by eye) between the images. In the current year, we have explored methods for using fiducial marks for this alignment, as we have done in the past for aligning stereo images. We also are initiating exploration of the use of warping transformations to correct for any distortions that occur in the images in either microscope, after evaluating this distortion using test specimens. A third, major project in this development is the study of centractin localization in transfected PtK2 cells (described below.)
Showing the most recent 10 out of 38 publications
