Bacteria experience physical forces in their natural environment including forces caused by osmotic pressure, growth in constrained spaces, and fluid shear. While the effects of physical forces on eukaryotic cell physiology are well studied and documented, these studies on prokaryotic cells like bacteria are very limited. The cell envelope is the primary load-carrying structure of bacteria, but the mechanical properties of the cell envelope are poorly understood. This thesis explores the mechanical stresses in bacterial cell envelope due to forces exerted by environmental and experimental stimuli and suggests a method of estimating the Young’s modulus of the bacterial cell envelope using extrusion loading and inverse finite elements Analysis.

Physical forces are experienced by bacteria in many situations including rapid changes in osmolality, adhesion to surfaces and forces generated by growth in confined spaces and biofilms. Furthermore, the primary goal of bactericidal antibiotics is to cause mechanical failure (lysis) of the cell envelope. Recent studies have demonstrated that physical forces can influence several physiologic processes in bacteria including the initiation of biofilm synthesis after adhesion to a surface and the assembly and function of multicomponent efflux pumps. But how well do these experiments emulate reality? The first part of this thesis covers some of the core concepts of mechanical engineering that could be used to model a bacterial cell and provides quantitative estimates for the mechanical stresses developed inside the bacterial cell envelope in the natural environment and under experimental conditions. The work suggests that the experimental methods generate mechanical stresses within the cell envelope that are well within that range of what would be expected in the environment. These experimental methods and others could provide more insight into the role of physical forces in bacterial physiology and antibiotic resistance.

One experimental method to mechanically stimulate bacteria is called extrusion loading. The extrusion loading approach is advantageous because it is less labor-intensive (hundreds of cells are examined at once), has more well-defined boundary conditions during loading than atomic force microscopy, and can be applied to bacteria of all shapes and sizes. Here, we combine this method with an optimization-based inverse finite element analysis to estimate the Young’s modulus of the cell envelope of four different types of bacterial cells. The inverse finite element model determines the Young’s modulus that minimizes the difference between the deformations of bacteria in experiments at 4-12 different load magnitudes and finite elements simulations of the same load magnitudes. The sample sizes for the current study are 536 untreated E. coli, 308 E. coli treated with A22, 517 V. cholerae, and 258 S. aureus. The Young’s modulus of the cell envelope was 2.06 ± 0.04 MPa for E. coli, 0.84 ± 0.02 MPa for E. coli in the absence of MreB, 0.12 ± 0.04 MPa for V. cholerae, 1.52 ± 0.08 MPa for S. aureus. We also explore how this analysis is affected by parameters used in the finite elements analysis by performing sensitivity analyses.

Lastly, we explore using Machine Learning for acquisition of data from the extrusion loading experiments. Getting data from images of bacteria undergoing extrusion loading is a tedious task, and requires direct human intervention and interpretation. Here I incorporate Deep Neural Networks into this task of image analysis could expedite this process and a well-trained algorithm can reach a human level accuracy. Here I adapt YOLOv5 a powerful, state-of-the-art Object Detection tool, for detection of cells in microfluidics-based extrusion loading of E. coli cells. The result section presents its performance in the detection task as well as compares its performance against a human annotator.