Automated organelle-based colocalization in whole-cell imaging

Abstract

The use of fluorescence microscopy to investigate protein colocalization is an invaluable tool for understanding subcellular structures and their associated proteins. However, current techniques are largely limited to two-dimensional (2D) imaging and often require manual segmentation. Here, we present OBCOL, a methodology to automatically segment and quantify protein colocalization not within an image as a whole but on all individual punctuate organelles within a 3D multichannel image. A wide variety of colocalization statistics may then be calculated on the objects found, and features reported for each such as position, degree of overlap between channels, and number of component objects. OBCOL was validated on imaging of two fluorescent markers (Dextran, EGF) in 3D microscopy imaging. OBCOL's application was then exemplified by investigating the colocalization of three fluorescently tagged proteins (VAMP3, Rab11, and transferrin) on recycling endosomes in mammalian cells. The methodology showed for the first time the diversity of endosomes labeled with one or more of these proteins and quantitatively demonstrated the degree of overlap among these proteins in individual recycling endosomes. The consistent segregation of these markers provides novel evidence for the subcompartmentalization of recycling endosomes. OBCOL is a flexible methodology for 3D multifluorophore image analysis. This study clearly demonstrated its value for investigating subcellular structures and their constituent proteins.