PSICIC: Noise and Asymmetry in Bacterial Division Revealed by Computational Image Analysis at Sub-Pixel Resolution

Abstract

Live-cell imaging by light microscopy has demonstrated that all cells are spatially and temporally organized. Quantitative, computational image analysis is an important part of cellular imaging, providing both enriched information about individual cell properties and the ability to analyze large datasets. However, such studies are often limited by the small size and variable shape of objects of interest. Here, we address two outstanding problems in bacterial cell division by developing a generally applicable, standardized, and modular software suite termed Projected System of Internal Coordinates from Interpolated Contours (PSICIC) that solves common problems in image quantitation. PSICIC implements interpolated-contour analysis for accurate and precise determination of cell borders and automatically generates internal coordinate systems that are superimposable regardless of cell geometry. We have used PSICIC to establish that the cell-fate determinant, SpoIIE, is asymmetrically localized during Bacillus subtilis sporulation, thereby demonstrating the ability of PSICIC to discern protein localization features at sub-pixel scales. We also used PSICIC to examine the accuracy of cell division in Esherichia coli and found a new role for the Min system in regulating division-site placement throughout the cell length, but only prior to the initiation of cell constriction. These results extend our understanding of the regulation of both asymmetry and accuracy in bacterial division while demonstrating the general applicability of PSICIC as a computational approach for quantitative, high-throughput analysis of cellular images. Recent studies have shown that all cells, including bacteria, are highly spatially organized. However, many questions about bacterial organization remain unanswered, often due to difficulties associated with visualizing and analyzing structures within such small and variably shaped cells. We have overcome these limitations by developing a generally applicable computational method for quantitatively analyzing cellular images at subpixel resolution. Our method uses interpolation to find cell borders accurately and precisely. Using these contours as a starting point, we automated the construction of a general-purpose internal coordinate system for each cell to facilitate comparisons between differently shaped cells. We applied this new method to two unsolved problems in bacterial cell biology. We first showed that a Bacillus subtilis asymmetric-division regulator is itself asymmetrically localized, thereby demonstrating our ability to extract information previously thought inaccessible by light microscopy. We also demonstrated our newfound ability to study characteristics of large populations by studying the accuracy of the symmetric division of Escherichia coli. We discovered a new role for the Min system, which inhibits polar division, in regulating division throughout the cell length. These results deepen our understanding of two important problems in bacterial cell biology while demonstrating the utility of our approach to studying subcellular structure in a wide range of biological systems.