Biomechanical ordering of dense cell populations

PROJECT SUMMARY

Morphogenesis is one of the most important themes in biology, and is also central to nonequilibrium physics. The issue is understanding how local interaction of elementary components leads to collective behavior and the formation of a highly organized system. In living nature, this self-organization is found on many different scales, from single cells to schools of fish and herds of animals. The collective dynamics of populations of microorganisms has attracted significant attention in recent years both in the biological and physics communities. Collective behavior leads to significant selective advantages for bacteria. Many species of bacteria form biofilms on various surfaces in order to resist environmental stresses. Biofilms are commonly present in natural environments such as living tissues, the surface of rocks and soils, or aquatic systems, but they can also be found in human-made systems and devices, such as industrial piping, artificial implants, etc. Understanding biofilm formation involves a variety of physical, chemical and biological issues, such as microorganism chemotaxis, motility, cell-cell signaling, attachment, substrate effects, gene regulation, etc. In this work [1] we focus on the growth and ordering of dense colonies of rod-shaped bacteria Escherichia coli in well-controlled environments. Quantitative study of this process is essential for elucidation of the physical and biological properties of biofilms.

EXPERIMENTAL TECHNIQUES

We use strains of E. coli that lack flagella in our experimental studies of bacterial colony growth. Initially, small population of cells ( 10^2- 10^3 cells) is inserted in specially designed microfluidic cavities and supplied with the necessary carbon sources to facilitate growth. Growth of the cells can be controlled by regulating the concentration of the carbon source (e.g. glucose) in the nutrient flow. The colony structure is observed and recorded using high-throughput microscopy and analyzed using specially designed software. Microfluidic devices are designed and fabricated using well established techniques[5]. Microfluidic devices are generally restricted to the laminar flow regime primarily due to the fact that the driving pressure required to initiate and maintain turbulent flows are unfeasible at the microscale. We use differential pressure as a means of driving the flow, either by hydrostatic (gravity) or mechanical (syringe pump) methods. For device fabrication, we use replica molding of poly-dimethylsiloxane (PDMS), a transparent (down to light at ~250 nm) elastomeric rubber. Details of the fabrication may be found elsewhere [6].

MOLECULAR DYNAMICS SIMULATIONS

To model the proliferation of cells in a microfluidic environment, we generalized the algorithm which we developed earlier to describe the dynamics of granular rods [2]. The algorithm is based on the well-known method of molecular dynamics (MD) simulations [3] which is widely used in different areas of statistical physics. The main idea of the method is to follow the dynamics of individual ``particles'' which constitute the system under study. In the case of a gas these particles are molecules, and in the case of granular material each particle represent a particular grain. The motion of every particle is described by Newton's equations for translational and rotational degrees of freedom. Details of their motion depend on the structure of the particles, their mutual interaction, and on the external forces/torques exerted on every particle. The main difficulty of the method lies in determining all of the contacts and computing the contact forces. To compute the forces, we employ a soft-particle MD simulation technique which allows for overlaps between the particles. More details of the methods of effective contact detection and computation of the contact forces may be found in our earlier work on dynamics of anisotropic granular particles [2,4].

MD SIMULATIONS AND EXPERIMENTS

1. Open channel

Microfluidic experiment: growth of a bacterial colony from uniformly distributed randomly-oriented cells in an open channel. The panels show snapshots of the population taken at t=0min, 40min, 80min, 120min

a	b
c	d

Microfluidic experiment: growth of a bacterial colony from a small clump of cells in an open channel. The panels show snapshots of the population taken at t =2.7h, 3.3h, 4.0h, 4.7h

a	b
i
c	d

MD simulations: growth of a colony from the initial strip with random orientations in an open channel. The panels show snapshots of the population taken at t=0.0, 2.7, 4.2, 14.2. The "cells" are colorized according to their orientation with respect to walls, from blue (perpendicular) to red (parallel).

a	b
c	d

2. Confined geometry

Results of MD simulations for the colony growth in a closed container. The growth is limited by four rigid walls. The size of the square domain is Lx=Lz=136.6d where d is the cell diameter. a-c: Initially the colony is prepared by placing randomly oriented cells of different length in the middle section. The panels show snapshots of the population taken at t=5.0, 15.0, 30.0. The cells are colorized according to the value of the contact stress.

3. Open channel. The role of fluid flow

Snapshots from experiments and simulations of biofilm growth in an open channel with flow. Experiments were conducted with E. coli XL-10 Gold over 8 hrs with a constant pressure of 10 dyn/ cm^2, and imaged at 40x magnification. Simulations were performed in a periodic channel of size L_x=136.6d and L_z=2 L_x. Only half of the channel is shown. Initially the colony is prepared by placing randomly oriented cells of different length in the middle. In the simulations, the cells are color-coded according to their velocity normalized by the magnitude of the fluid flow, so that blue corresponds to the cells attached to the bottom and red corresponds to the cells passively advected by the fluid flow.

REFERENCES

poster