Vertical growth dynamics of biofilms

Abstract

During the biofilm life cycle, bacteria attach to a surface and then reproduce, forming crowded, growing communities. Many theoretical models of biofilm growth dynamics have been proposed; however, difficulties in accurately measuring biofilm height across relevant time and length scales have prevented testing these models, or their biophysical underpinnings, empirically. Using white light interferometry, we measure the heights of microbial colonies with nanometer precision from inoculation to their final equilibrium height, producing a detailed empirical characterization of vertical growth dynamics. We propose a heuristic model for vertical growth dynamics based on basic biophysical processes inside a biofilm: diffusion and consumption of nutrients and growth and decay of the colony. This model captures the vertical growth dynamics from short to long time scales (10 min to 14 d) of diverse microorganisms, including bacteria and fungi.

Keywords: biofilms; biophysics; super-resolution.

Fig. 1.

White light interferometry imaging of developing biofilm topography. (A) Light intensity and (B) surface measurement data from the edge of an Aeromonas veronii inoculum on LB agar are shown 30 min after inoculation. (C) One-dimensional averaged profile of the surface topography is computed from the data in (B). (D) A 24-h timelapse of averaged profiles from a growing A. veronii biofilm is shown. The colony expands horizontally (x-axis) and vertically (y-axis), with some of its surface features persisting during development. The scale in the y-axis has been increased to better observe the data. The region shown in panels A–C is marked with a rectangle.

Fig. 2.

Biofilm height dynamics and geometric constraints. (A) Biofilm height vs. time is shown for a 48-h period in linear and log scales (inset). Three replicates of A. veronii are shown. Height increases even at early times, as seen on the log scale. (B) Change in height is plotted against biofilm height. There are two clear regimes: I) accelerated and II) decelerated growth, separated by a characteristic length. (C) The height of microbial colonies grown on small agar columns, thus preventing lateral diffusion of nutrients, is shown. In one set of experiments, colonies on the agar columns are replaced every 2 d (R). In a control set of experiments, colonies are not replaced (NR) and instead are allowed to continue growing. (D) The total height of colonies grown on individual agar columns over a period of 6 d is shown. These results demonstrate that colony height does not saturate due to nutrient depletion. (E) The thickness of the actively growing layer can be approximated as a simple minimum function and can be used to model the two different growth regimes. Nutrient concentration, Monod kinematics, and total growth from Monod kinematics are shown as a function of the distance from the interface. A minimum function approximation is shown to have good agreement with the full expression for total growth. Nutrient dynamics are shown for O₂ for a E. coli colony, with a value of L = 28.26 µm

Fig. 3.

Quantitative assessment of the interface model for biofilm vertical growth. (A) The mean height of A. veronii colonies, averaged over three replicates, is shown vs. time. Error bars represent SD across replicates. Best fit lines for the interface and logistic models are shown. In the inset, the differential form of the models is contrasted against experimental data, showing that the interface model captures the two linear regimes in growth. (B) Residuals for the best-fit predictions from the above models are shown as a function of time. On average, the logistic model by 10.77 μm, and the interface model by 1.24 μm. The gray region corresponds to the SD of the three replicates in relation to the mean value at each time.

Fig. 4.

Long-term growth of microbial colonies. (A) Long-time measurements of height versus time are shown for three different species. Error bars represent standard deviations across the 2-mm homeland region. Solid lines show the best-fit interface model. (B) Colony heights during the initial 48 h of growth of plotted against best-fit predictions from data taken during the 2 to 14 d range. The agreement between the data and the model is evident.

Fig. 5.

Growth of different species and strains over 48 h. The average height plus or minus the SD across three parallel colonies is shown in gray. Error bars represent SD across replicates. The best-fit interface model is shown in orange. The model RMSE and the total number of interferometry profiles analyzed are reported in each panel. While these microbial colonies differ in their composition, height, radius, and overall morphologies, the interface model accurately describes the average height dynamics over time for each one.