Fig 1.
Effects of the plasmid pCF10 on OG1RF biofilm rapid remodeling and growth under treatment with 100X MIC erythromycin.
Biofilms were grown in 24-well optic bottom plates or on glass coverslips in 24-well plates for with twice-daily fresh Todd-Hewitt medium exchanges for indicated times. For microscopy images, biofilms were stained with Syto9 and imaged by laser scanning confocal microscopy (LSCM). Bacterial numbers were determined by sonication followed by serial dilution and are reported as colony forming units (CFU) per biofilm. Erythromycin (100X MIC) was added to the biofilms for indicated times. (A) OG1RF 3-day old biofilm. (B) CFU of Day 3 OG1RF biofilms treated for 1 hr with TH or 100X MIC erythromycin for 1 hr. (C) OG1RF(pCF10) biofilms (Row 1) 1 day, (Row 2) 2 day, (Row 3) 3-day biofilms. (Column 1) 1 hr after the addition of TH only, (Column 2) 1 hour post 100X MIC erythromycin treatment, (Column 3) 2 hours post treatment, and (Column 4) 3 hours post treatment. Each image is representative of a total of 3 images for 3 biological repeats at each time point for a total of 39 biofilms. (D) Change in the size of structures in untreated biofilms between Day 1 and Day 3. (E) Change in CFU/biofilm after erythromycin treatment for Day 1, (F) Day 2, and (G) Day 3 biofilms. Each dot represents a biological repeat. Note that one star implies a p-value less than .05, and three stars implies a p-value less than .0001.
Fig 2.
Image processing steps and results.
(A) Microscopy image stack. (B) Post connection algorithm and removal of rigid base. (C) Microscopy x-slice example containing a protected region. (D) Corresponding x-slice in workflow. (E) Density-based clustering of x-slice with marked centers of mass. (F) Average volume of protective regions per replicate. Protective structures on average double during treatment. Double star significance indicates a p-value <0.01. (G) Percentage of x-slices containing a protected region per replicate, which remains unchanged during treatment. Image processing data results from 3 biological repeats with 2-3 replicates per experiment.
Fig 3.
Illustration of eDNA above complex structures.
A. Microscopy image capturing bottom layer of eDNA cloud (arrows) sitting atop complex structure. B. Microscopy image stained with propidium iodide, confirming the presence of eDNA. Microscopy images are of two independent untreated, 3-day biofilm.
Fig 4.
Illustration of the computational domain.
Interior of the computational domain Ω, is bounded by on the left and right, by
on the top, and by
on the bottom.
Fig 5.
Simulation results for different antibiotic entry conditions.
(A) Untreated microscopy x-slices selected using the quantification algorithm (Sect 4.6) to identify protected regions. (B) Initial conditions using (A) as source data. Note that in order for the cells to respond to antibiotic influx an extended vertical domain is required, since the stress distance () is 15
. Therefore, we added an extra 2 layers of uniformly distributed cells with a non-stressed carrying capacity above the complex structures as is seen in microscopy images. (C) Simulation results for scenario 1 (no reaction or diffusion barrier). (D) Simulation results for scenario 2 (diffusion barrier but no reaction barrier). (E) Simulation results for scenario 3 (diffusion and reaction barrier. (F) Post-treatment biofilms exposed to erythromycin for one hour. Note that the initial- and post-treatment biofilms are not the same samples. Scale bar has length 10
. For videos of simulations see supporting information S1 and S2 Video.
Fig 6.
Validating the computational model.
(i) Volumes of simulated protective structures confirming the doubling of protective regions. (ii)A Microscopy image slices used for image processing. Images were sourced from untreated 3-day biofilms grown according to Sect 4.2. (ii)B Initial conditions for antibiotic treatment simulations. (ii)C Simulation results showing rigid structure growth under stress.
Fig 7.
Simulation results for cases where preformed structures are absent.
(A) Examples of slices without preformed structures that would be determined to be inadequate following the process described in Sect 2.2. (B) corresponding initial conditions for simulation. (C) simulation results.
Table 1.
Definitions of the variables used in model.
Fig 8.
Dynamical map of a lattice node. Here, (0,0) indicates the absence of a cell, meaning no stress is applied at that node. (1,0) denotes the presence of an unstressed cell, while (1,1) signifies a stressed cell at the lattice node.
Fig 9.
Bacterial cell placement diagrams.
(A) Assuming the cell to cell distance at which less densely packed bacterial cells require to divide is L = 3 μm. Then the green boxes surrounding lattice node are the locations from which a divided cell placed into
can originate from. (B-E) Illustration of the probability that a bacterial cell at lattice node
divides and places a cell in empty lattice node
. The yellow boxes denote the lattice nodes
that ensure the division distance. The lattice nodes denoted by the red squares are the other nodes that are checked for possible bacterial cell occupation. Note that if
lies above or below
then we do not take into account the lattice nodes to the right or left. If
is to the left or right of
and if lattice nodes
or
are empty then
will not place the new cell in
.
will instead place the cell in
or
due to the observed growth anisotropy described in Sect 4.1.3. (B)
. (C)
. (D)
. (E)
. (F-G) Illustration of the two stochastic pathways through which stress induction of a cell can take place. (F) The blue box,
, denotes a lattice node such that
and
. The green box,
denotes a lattice node such that
. We write
. (G) Illustration of the quantities necessary to compute the probability that a lattice node,
, is stressed by signaling molecules. Let the lattice node,
, be such that
and
. Let the green box, H, be the set of all lattice nodes containing stressed cells within the neighborhood of
, i.e.
,
if
. Further, recall
is the radius of
. Then let
be the minimum distance between
and an element of H.
Table 2.
Overview of parameters and their values used in numerical simulations.
Fig 10.
Measurements of protective regions.
(A) Example of measuring the width () of rigid structures in a biofilm slice used for quantifying parameter
. Yellow lines register two rigid structures whose lengths are measured and averaged to compute
. The microscopy slice is of a 3-day, 1-hour treated biofilm using method outlined in Sect 4.2. (B) Red boxes encompass protective structures while the area between the two boxes would be the protected region. Under treatment, protective structures buffer and conserve the protected region allowing continued less densely packed, unstressed growth. (C) Example of measuring the distance between four cells that are not chained or part a rigid structure to estimate L.