Evaluating Bacterial Biofilms With Confocal Imaging
App Note / Case Study
Last Updated: July 1, 2024
(+ more)
Published: January 29, 2024
Credit: iStock
Bacterial biofilms are complex, adaptable multi-cellular architectures that play important roles in human and environmental health and disease. Confocal microscopy has long been a core pillar of biofilm research and model development, but these micro-environments can be difficult to study due to their complex, 3D nature.
Discover a single confocal imaging reader able to replace up to five individual devices and perform a variety of analysis outcomes common in biofilm workflows, including biomass screening, characterizing growth architecture and antibiotic susceptibility testing.
Download this FREE app note now to explore:
- An instrument that integrates fluorescence, absorbance and both confocal and widefield imaging with multiple objective magnifications
- Simple, centralized user interface and data management
- A significant cost saving over individual instruments
Application Note
Microbiology
Author
Wendy Goodrich,
Agilent Technologies, Inc.
Abstract
Bacterial biofilms are complex communities of microorganisms organized within a
self- or host-propagating extracellular substrate and can present with beneficial and/
or detrimental results in natural environments and on many critical abiotic surfaces.
Understanding the formation, development, components, and mechanisms of this
bacterial way of life continues to be important to numerous fields of study and often
requires a multidisciplinary approach that relies on several dedicated instruments
to acquire data for analysis. This, in addition to their three-dimensional (3D)
architecture, makes biofilms a particularly suitable model for evaluating the utility of
a novel confocal imaging reader for analyzing them. The Agilent BioTek Cytation C10
confocal imaging reader is specifically built to integrate the functionality of up to five
different instruments in one device, including both widefield and confocal imaging
capabilities. This permits a broad choice of assays during experimental design while
simultaneously supporting an orthogonal method. A variety of detection, imaging,
and analysis methods enabled by the Cytation C10 are demonstrated in this
application note using multiple bacterial strains and common biofilm assays.
Efficacy of a Novel Confocal
Imaging Reader for Evaluating
Bacterial Biofilms
2
Introduction
It is generally understood that bacteria prefer to live a
communal lifestyle encased in a 3D extracellular matrix
composed of a mix of extracellular polymeric substances
(EPS). This architecture has been shown to consist of
several microenvironments and is adaptable depending on
conditions found in the surrounding host.31 Bacteria undergo
distinct phenotypic changes that make them highly proficient
and opportunistic at building and maintaining these biofilm
communities in diverse, and even extreme, habitats.1, 7, 25 In
addition to their ubiquitous presence in the natural world,
human-built environments provide vast niches for bacteria
to establish biofilms, and they are researched over many
scientific disciplines due to the diverse roles biofilms play.
Such roles include pathogenic, protective, or therapeutic
actors in human and environmental health and disease,
causative and preventative spoilage mechanisms in the
cultivation and production of food products, toxic soil and
marine remediation agents, or as vehicles in biofouling and
corrosion in water- and fuel-processing infrastructures that
support our daily needs.
Over the past four to five decades, the field of biofilm science
has grown to include numerous different species and
experimental conditions. This has resulted in paradigm shifts
of understanding, notably in describing biofilm structure and
function as well as the biofilm life cycle itself.11, 17 Parallel to
this, many novel biomaterials and technologies have been,
or are being, developed or employed to gain insight into
unanswered questions or solve urgent problems related
to biofilms. However, in vitro methods using indirect and/
or direct biochemical and imaging assays that rely on
absorbance, fluorescence, luminescence detection, and/
or widefield and confocal microscopy, continue to be
fundamental in experimental design of biofilm systems,
and are widely cited in biofilm literature.1, 4, 5, 6 These tools
are particularly useful during biofilm model development—a
process that depends on the testing and optimization of
numerous and diverse variables and outcomes, starting
with the fundamental choice of bacterial species, whether it
forms a biofilm, and the conditions under which this occurs.
Additional experimental parameters may include:
– Investigating the effects of different growth nutrients and
substrates on biofilm development;
– Determining optimal concentrations of stains, dyes, and
other reagents or compounds;
– Selecting effective methods to measure response to
genetic manipulations or other exposures such as to
surfactants, pharmaceutical compounds, or toxins;
– Determining changes in phenotype between planktonic
cells and mono and polymicrobial community life;
– Determining the results of applied external pressures like
fluctuations in shear stress, temperature, pH, oxygen, or
ionic charge at any or all points along a biofilm lifecycle.
These stages of model development exist regardless of
research novelty, and can result in protocols for defining
reproducible ground truths that are important when
translating common models to pioneering ones. They are also
important when generating foundational data that can inform
on downstream experimental direction.
Figure 1. Overview of the biofilm evaluation workflow using the Agilent BioTek Cytation C10 confocal imaging
reader (upper right). Descriptive text is in white, and the described experimental assays and/or end points are
in orange.
3
Table 1. Assay materials and instrument imaging supplies.
This application note presents representative results from a
collection of biofilm case studies, performed to evaluate the
potential of the novel Cytation C10 confocal imaging reader
as a tool for biofilm characterization. An overview of the
evaluation workflow is shown in Figure 1.
Biofilms were assessed for properties such as surface affinity
and growth characteristics, in addition to compositional
features, including components and architecture of the EPS.
The Cytation C10 proved to have multiple features conducive
to biofilm analysis, as realized by the capability of the
instrument to perform the detection and imaging techniques
required by the various in vitro biofilm assays described
herein. Having this functionality available in one device
centralized both protocol definition and data management
via a single user interface while allowing the choice of a wide
range of detection and staining reagents.
Experimental
Materials
Table 1 contains detailed assay materials and instrument
imaging supplies used.
Category Description/Source/Part Number
Model Organisms
Pseudomonas aeruginosa GFP
(PAGFP)
Pseudomonas aeruginosa GFP
(ATCC; p/n 15692GFP)
Escherichia coli GFP (ECGGFP) Escherichia coli GFP (ATCC; p/n 25922GFP)
Staphylococcus aureus (SA) Staphylococcus aureus subsp. aureus Rosenbach
(ATCC; p/n 25923)
Bacillus subtilis (BSub) Bacillus subtilis subsp. subtilis (Ehrenberg) Cohn
(ATCC; p/n 35021)
Staphylococcus epidermidis (SE) Staphylococcus epidermidis (Winslow and
Winslow) Evans (ATCC; p/n 12228)
Vessels
96-well microplate Corning clear polystyrene (PS)
(Thermo Fisher Scientific; p/n 12-566-202)
96-well microplate Agilent Black PS optically enhanced clear bottom;
(p/n 204626)
96-well microplate Krystal Black PS glass bottom
(Southern Labware; p/n 324002)
96-well microplate Nunc MaxiSorp clear F8 breakaway wells
(Thermo Fisher Scientific; p/n 469957)
24-well microplate Krystal Black PS glass bottom
(Southern Labware; p/n 324042)
Agar plate 100 mm x 20 mm Style culture dish
(Corning; p/n 430591)
Dual-lock culture tubes 14 mL (Falcon; p/n 352059)
2 mL microcentrifuge tubes Screw cap (Fisher; p/n 02-682-558)
15 mL centrifuge tubes Screw cap (Corning; p/n 430790)
Media and Reagents
Tryptic soy broth (TSB) Cell culture medium (BD; p/n 211825)
Agar Bacterial isolation, viability, growth
(Sigma-Aldrich; p/n A1296)
Ampicillin Antibiotic/media additive fluorescence protein
selection (Sigma-Aldrich; p/n A5354)
Glycerol Bacteria preservative, media additive
(Sigma-Aldrich; p/n G9012)
Glucose Media additive (where noted)
(Sigma-Aldrich; p/n G9012)
McFarland latex turbidity standards Hardy Diagnostics (p/ns ML05; ML1; ML2;
ML3; and ML4)
FilmTracer SYPRO Ruby biofilm
matrix stain
Biofilm EPS stain (proteins)
(Thermo Fisher Scientific; p/n F10318)
Invitrogen SYTO 64 red fluorescent
nucleic acid stain
Bacterial cell stain
(Thermo Fisher Scientific; p/n S11346)
Invitrogen Toto-1 iodide eDNA stain
(Thermo Fisher Scientific; p/n T3600)
Presto Blue HS Cell viability reagent
(Thermo Fisher Scientific; p/n P50200)
Propidium Iodide Cell viability reagent
(Thermo Fisher Scientific; p/n P3566)
FilmTracer LIVE/DEAD Biofilm
Viability kit
Cell viability reagent
(Thermo Fisher Scientific; p/n L10316)
Calcofluor white Biofilm EPS stain (cellulose/polysaccharides)
(Sigma-Aldrich; p/n 18909)
1% crystal violet Biofilm total biomass
(Sigma-Aldrich; p/n V5265)
Resorufin Metabolic activity standard
(AdooQ biosciences; p/n A21344)
DPBS or PBS
(Dulbecco’s) Phosphate Buffered Saline
(Sigma-Aldrich; p/n D8537) or
(Thermo Fisher Scientific; p/n 14190-136)
Imaging Supplies―Objectives Part Number
2.5x Meiji Plan Apochromat 1220549
20x Olympus Plan Fluorite 1220517
60x Olympus Plan Fluorite 1220545
Imaging Supplies―Filters Part Number
CY5 1225105
DAPI 1225100
GFP 1225101
Phycoerythrin (PE) 1225113
DAPI Confocal 1945103
GFP Confocal 1945104
TRITC Confocal 1945106
Methods
See references 5, 9, 10, 15, 23, and 25-29 for context and
frameworks of many of the techniques, analysis, and methods
adapted and defined in this study. Image acquisition and
analysis details are provided in Table 2.
4
Cell culture
Working stocks of model organisms were prepared in tryptic
soy broth, supplemented with 10% glycerol and filter-sterilized
ampicillin, at 300 µg/mL for PAGFP, or 100 µg/mL for ECGFP.
Glycerol working stocks were kept at –80 °C. Culture media
was TSB supplemented with or without strain-specific
ampicillin. Tryptic soy agar (TSA) was prepared by adding
agar at a 1.5% (w/v) to TSB. Strain-specific ampicillin was
added when TSA cooled to between 49 and 55 °C following
autoclaving. Bacteria were isolated, grown, and cultured
following one of two methods: (A) In the first method, a loop
of working stock was streaked on strain-specific TSA and
colonized at 37 °C. The next day, one to two isolated colonies
were suspended in 6 mL strain-specific TSB in Falcon tubes,
and were grown aerobically for 16 to 24 hours at 37 °C
with agitation at 180 rpm on a Thermo Fisher Scientific
MAXQ 4450 benchtop orbital shaker to obtain a final bacterial
culture. (B) In the second method, a loop of working stock
was suspended directly into 6 mL strain-specific TSB, then
grow aerobically at 37 °C with shaking at 180 rpm for 16 to
24 hours.
Cell density calibration
Three 200 µL replicates of each McFarland latex turbidity
standard, DPBS, water, TSB, and/or 95% ethanol were
dispensed into a 96-well Corning or Agilent microplate and
read at an optical density (OD) of 625 nm to obtain a final
turbidity calibration curve, plotted using linear regression.
The mean of the DPBS, water, media, and ethanol were
calculated as the zero standard. Limit of blank (LoB) and
limit of detection (LoD) for 32 replicates of 200 µL of DPBS
were calculated using the parametric option described by the
Clinical Laboratory Standards Institute EP17-A2 standard.2
The calibration curve experiment file was subsequently used
in all experiments to determine both starting overnight growth
density and starting experimental cell densities following
dilution of the overnight stock.
Cell seeding
Two 500 µL aliquots of overnight bacterial culture were set
aside on ice to pause growth, and bacterial cell density was
determined on the remaining overnight growth suspension
by centrifuging at 6,000 g for 10 minutes in 2 mL tubes or at
5,000 g for 5 minutes in 15 mL tubes then resuspended in
5 mL DPBS. A seven-log serial dilution series was performed
in DPBS, resulting in a final eight-log serial dilution series.
Three replicates of 200 µL of each bacterial dilution were
dispensed to a 96-well Corning microplate and the plate
was read at an OD of 625 nm using the same calibration
detection experiment file described above. The 500 µL
aliquots of starting bacterial growth suspension were then
adjusted to final density based on the results extrapolated
from the McFarland standard curve by centrifuging, removing
supernatant, and resuspending the bacteria in fresh TSB.
Experimental seeding densities were generally diluted to a
final ~ 1.5 to 3 x 105
/mL.
Cell viability
0.047 grams of magnesium chloride (MgCl2
) was dissolved in
5 mL dH2
O to make a 100 mM starting stock solution. PAGFP
TSB media (TSB + 300 µg/mL ampicillin) was supplemented
with 100 mM MgCl2
at final 0, 1, and 10 mM media solutions.
A loop of PAGFP glycerol stock was resuspended in 6 mL of
each of the media solutions and grown aerobically overnight
at 35 °C with shaking at 180 rpm. The next day, the 0 mM
overnight growth stock was divided into three 2 mL aliquots.
One aliquot was designated for the air-liquid interface
adherence assay (described later), and the remaining two
aliquots were designated to test for viability of cells grown in
the different media.
Metabolic viability was determined by interpolating samples
from a resorufin calibration curve prepared in low-light
conditions from a 1 mg/mL starting stock of resorufin in
DMSO. A seven-point 2x dilution from 5 to 0 µg/mL resorufin
in 95% ethanol was dispensed at 200 µL/well in triplicate
to a black-sided 96-well microplate, and read kinetically
every 10 minutes for 2 hours using fluorescence detection
at Ex:Em 550:590 nm. A standard curve was generated for
each time point using a nonlinear, four-parameter regression
analysis. Data from the 30-minute time point was used
as the final standard curve. Metabolic viability of samples
was determined by plating 180 µL of a three-log dilution in
triplicate for each of the MgCl2
media formulations to a
black-sided, clear-bottom 96-well microplate. A 20 µL
volume of Presto Blue HS—a resazurin-based reagent—was
dispensed to each well and read kinetically at 10-minute
intervals up to 40 minutes using fluorescence detection at
Ex:Em 550:590 nm. Concentration of sample resorufin for
the 30-minute read point was interpolated from the resorufin
standard curve.
Membrane viability of each MgCl2
dilution of the overnight
stock was determined using a live/dead assay20 with some
modifications. A 15 µL volume of a 20 mM solution of
propidium iodide (PI) in DMSO was suspended in 5 mL dH2
O
and kept in the dark. An 18-month stock of PAGFP, which had
been kept in oxygen-starved conditions in stale media at 4 °C
and was determined to be nonviable, was used as the dead
cell control. The 0 mM overnight PAGFP growth stock was
used as the live cell control. The turbidity of the dead cell
control was determined from 200 µL stock solution, then
centrifuged and resuspended in 1x balanced saline. The
5
live cell control was diluted to the same cell density as the
nonviable stock by diluting the three-log dilution of the 0 mM
stock 2.5x in balanced saline.
A live/dead standard curve was calibrated as a 100, 90, 50,
10, and 0% volume of live cells to the reverse of dead cells. A
three-log dilution of each of the MgCl2
samples was prepared
in balanced saline. A 100 µL volume of each standard and
sample were dispensed in triplicate into the inside wells of
a 96-well black-sided, clear-bottom microplate. A 100 µL
volume of balanced saline was dispensed to three wells as a
no-cell control. A 100 µL volume of the PI solution was added
to all wells and the plate was incubated in the dark at room
temperature for 15 minutes. The plate was then read using
fluorescence detection at Ex:Em 485:530 nm for live cells
(green GFP), and at Ex:Em 485:630 nm for dead cells (red, PI).
A ratio of the fluorescence emission values for all wells was
calculated by dividing the green by the red emission value.
Ratio values for the standards were multiplied by the dilution
factor (2.5) to normalize standards to sample cell density.
Ratio values were then plotted against percent live cells.
Percent live cells for the sample ratio values were interpolated
from the standard curve. To account for sample values
interpolated as > 100% live, ratio values were plotted for each
of the MgCl2
concentrations against the ratio values of the
100% live standard (0 mM overnight growth stock).
Total biomass
A 200 µL volume of ~ 1.5 to 3 x 105
/mL TSB suspensions of
PAGFP, ECGFP, SA, and SE bacterial cells were dispensed in
triplicate to four sets of Nunc breakaway wells. Each set of
bacterial inoculum (time 0) was incubated at 37 °C at intervals
of 3, 6, 16, or 24 hours. Total biomass of formed biofilms at
each time point was detected in triplicate for each species
using a crystal violet assay. Briefly, media were aspirated from
the wells, and the well was washed twice with 200 µL DPBS
and allowed to dry at room temperature with the lid off in a
biosafety cabinet (≥ 30 minutes). A 200 µL volume of 99%
ethanol was added to each well, and biofilms were fixed and
permeabilized for 15 minutes at room temperature. Ethanol
was aspirated and wells were allowed to dry completely (≥ 30
minutes). A 1% bulk crystal violet solution was diluted to 0.1%
in dH2
O. A 200 µL volume of 0.1% crystal violet solution was
dispensed to each well, followed by a 15-minute incubation
at room temperature. Stain was aspirated and wells were
washed twice with 200 µL sterile water, leaving the wells
empty following the last wash. Wells were left to dry for at
least 30 minutes. Stain was eluted from the biomass using
200 µL per well of 99% ethanol and shaking at 125 rpm for
30 minutes. A 200 µL volume of ethanol was dispensed into
three wells as a negative control and total biomass on the
plate was then detected, unless noted, at absorbance 590 nm.
Biomass over time for each species was plotted against both
the starting inoculation density before dilution and the mean
crystal violet value.
Susceptibility
A 100 µL volume of diluted BSub cells (~ 1.5 x 105
/mL) were
resuspended in 1x TSB containing ± 150 µg/mL ampicillin and
dispensed in triplicate into wells containing 100 µL of the
same media (final cell density of ~ 0.75 x 105
/mL). Cell
growth with and without antibiotic was analyzed over an
18-hour kinetic time course reading turbidity at 625 nm
absorbance every 30 minutes. In parallel, 100 µL cells
± 150 µg/mL antibiotic were dispensed in triplicate to Nunc
breakaway wells into 100 µL of corresponding TSB (final cell
density ~ 0.75 x 105
/mL). The plate was incubated at 32 °C
for 48 hours followed by a crystal violet assay (described
previously) up to the elution step. Wells were left to dry
overnight then imaged at 2.5x in color brightfield. The crystal
violet elution step was then performed, and the plate was
read at 590 nm OD. Biofilms were analyzed to compare
growth and total biomass of antibiotic- and nonantibiotictreated cells using turbidity at absorbance 625 nm and
total biomass at 590 nm OD from the crystal violet assay.
Qualitative assessment of total biomass was provided by the
color brightfield widefield images.
Adherence—air-liquid interface
Seven replicates of 250 µL of a five-log dilution
(~ 1.5 to 3 x 105
/mL) of each of a 0, 1, and 10 mM MgCl2
cultured overnight stock (described by the cell viability assay)
were dispensed into 21 wells of two glass-bottom 24-well
plates, one lying flat and the second propped at a 45° angle
against a plastic block. Duplicate wells of the same volume
of TSB without bacteria, and a single well of dH2
O, were also
dispensed to each plate. Plates were incubated at 35 °C for
24 hours in the middle rack of the incubator either at a 45°
angle or flat to the incubator tray insert. Spent media and
nonadhered and loosely adhered cells were aspirated from
all wells and replaced with 250 µL of respective fresh MgCl2
supplemented media, then incubated either at a 45° angle or
flat for another 36 hours to continue biofilm formation. Wells
were washed with 250 µL PBS, then an additional 250 µL
PBS was added to the wells and a 3 x 3 image montage was
acquired at 2.5x in widefield, using a GFP filter to capture
cell mass. To confirm cell mass, a single 20x, 40x, and 60x
image in addition to acquisition of a 20x Z-stack in GFP was
performed on one area of interest along the border of the
visible biofilm (not shown). Biofilms were then fixed and
permeabilized in 250 µL of 99% ethanol for 15 minutes. The
remaining protocol for the crystal violet assay was performed
as described by the total biomass assay up to the point
6
before elution. Plates were left to dry overnight then imaged
using widefield color brightfield to capture total biomass at
2.5x in a 3 x 3 montage. The crystal violet assay was then
continued through the elution step and detected at 590 nm
OD. Finally, stain was aspirated from all wells, washed with
sterile water until the residue was clear, then imaged again
in a 3 x 3 montage at 2.5x using color brightfield and Cy5 to
capture residual biomass staining. Image analysis comparing
results of the plate grown at a 45° angle to the one grown flat
was done for total intensity and area of cell mass (GFP) and
OD of total biomass.
Adherence―surface affinity
A 5 µL volume of a ~ 1.5 to 3 x 106
/mL suspension of SA was
dispensed to 18-well quadrants of both a glass-bottom and
optically enhanced polystyrene microplate and incubated
at 37 °C for 3 hours. Three additional inoculations of 5 µL
were added to the first at 2-hour intervals. At the end of the
third incubation, 80 µL media was added to each inoculated
well and the plates were incubated for an additional 16 to
24 hours. A 100 µL volume of media was dispensed to wells
in rows D and H prior to the overnight incubation. For this
experiment, overnight media was replenished with 100 µL
fresh media every 16 to 24 hours over 8 days. On day 8,
biofilms were imaged both before and after media
replenishment. For unperturbed SA biofilms, 10 µL of 100 µM
Syto 64, prepared in saline from an intermediary 5 mM
concentration in DMSO, was added directly to the SA in
growth media, and perturbed cells were aspirated normally
and replenished with fresh media containing ± 10 µM Syto
64. Biofilms were then imaged at 2.5x using fluorescence
to capture cell mass (TRITC). Analysis included comparing
growth of the unperturbed biofilm to growth retained on
perturbed biofilms for both microplate surfaces using the
confluence calculation within the Agilent BioTek Gen5 Image
Prime software. Briefly, percent confluence was calculated
for each of the nine replicates of unperturbed and perturbed
biofilms. The percent confluence of the perturbed biofilms
was divided by the confluence values of the unperturbed
biofilms and multiplied by 100, resulting in a percent retained
growth end point.
EPS components and biofilm architecture
A 5 µL volume of a ~ 1.5 to 3 x 106
/mL suspension of SA or
ECGFP was dispensed to 18-well quadrants of both a glassbottom and optically enhanced polystyrene microplate and
incubated at 37 °C for 3 hours. Two additional inoculations
of 5 µL were added at 3-hour intervals before adding a final
85 µL fresh media to each well and incubating for 16 to 24
hours. Wells in rows D and H were inoculated with 100 µL
media only. For this experiment, 18 biofilm wells were
assessed at each of three time points over a total of 3 days,
and again at day 8. At each time point, Syto 64 at a final
concentration of 10 µM/well and Toto-1 at a final
concentration of 2 µM/well were dispensed directly to the
SA wells and left to incubate for 15 minutes before adding
one "drop" (~ 10 µL) of calcofluor white directly preceding the
imaging step. For ECGFP-inoculated wells overnight, media
was aspirated and replaced with 100 µL Ruby red matrix stain
for 30 minutes at room temperature. The stain was aspirated,
replaced with 50 µL of 1x calcofluor white for 1 minute, and
then 50 µL sterile water was added and the wells were
imaged. For remaining wells on the plate, spent media was
exchanged with 100 µL fresh TSB and plates were returned
to the incubator until the next time point. After the image
analysis on the third time point, fresh media was replaced in
remaining wells every 24 hours up to day 8, when the staining
assay was repeated. This was done either using the staining
method from previous time points on nine wells, or direct
staining into the media without aspiration using 10 µL of 1x
Ruby red biofilm stain and 10 µL of 1x calcofluor white for
ECGFP, or using 10 µM/well of Syto 64, 2 µm/well Toto-1, and
~ 10 µL calcofluor white determined from an independent
optimization experiment for direct staining. A Z-stack was
acquired in confocal mode at 20x at 4.2 µm intervals over
13 stacks for SA and 25 stacks for ECGFP using the DAPI
(calcofluor white), GFP (ECGFP and Toto-1), and TRITC (Syto
64) filter cubes. The Ruby red matrix stain was imaged in
parallel to the confocal imaging at 20x in widefield using a
phycoerythrin (PE) filter cube with excitation and emission
spectrum (Ex:Em 469:593 nm) not available in confocal
mode. Intracellular DNA (iDNA) and extracellular DNA (eDNA)
were quantitated for SA biofilms, and microcolony growth
was quantitated for ECGFP biofilms using cell analysis object
counting and subpopulation metrics as defined by Table 2
and described in the "Results and discussion" section.
Qualitative assessment for each of SA and ECGFP biofilm
composition was compared on representative wells from
each time point.
Instrumentation and software
All data was obtained using the Agilent BioTek Cytation 10
confocal imaging reader, equipped with programmable quad
monochromators for absorbance and fluorescence detection,
luminescence detection (not used), widefield inverted
imager, and confocal inverted imager. The instrument was
controlled by Gen5 Image Prime software version 3.12 or
higher. Descriptions and the configuration of the detection
and imaging supplies and acquisition settings are shown
by Tables 1 and 2. Results analysis used either the Agilent
BioTek Gen5 Image Prime software, Microsoft Excel for
Microsoft 365 MSO (version 2208, build 16.0.15601.20540),
and/or GraphPad Prism software, version 9.5.1 (733).
7
Assay Imaging Mode Imaging Method Image Acquisition Image Processing Image Analysis (Parameters)
Properties
Susceptibility Widefield Color brightfield 96-well plate; 2.5x objective autofocus height
2784.1 µm Qualitative
Surfface Affinity Widefield Fluorescence
96-well single image; TRITC 556:600; 2.5x
objective focal height 3000; LED intensity 10;
Camera gain 32; integration 207 ms
Background flattening
rolling ball 1132 µm
Image statistics (confluence;
lower threshold 2000)
Air-liquid Interface Widefield Fluorescence
24-well plate 3 x 3 montage; GFP 469:525; 2.5x
objective focal height 3000; LED intensity 10;
camera gain 32; integration 780 ms
Image stitching linear
blend; Background
flattening rolling ball
6000 µm
Image statistics (total area; total intensity;
lower threshold 5500)
Color brightfield 24-well plate 3 x 3 montage; 2.5x objective
focal height 3140.8
Image Stitching Linear
Blend Qualitative
Composition
SA Biofilms, Global Confocal 40 µm
spinning disk Fluorescence
96-well 20x objective Z-stack with interval 4.2 µm;
GFP (green) 472:520; DAPI (blue) 405:442;
TRITC (red) 556:600
ZProjection maximum
method of in-focus
stacks
Qualitative:
Cell analysis (cell count eDNA, cell count iDNA,
cell count colocalized eDNA–iDNA, cell count
colocalized iDNA–eDNA. On total in-focus stack
Z-projection days 1–3, on total and individual
in-focus stacks day 8. eDNA strand analysis Day 3)
Day 1
15 Z-stacks; two slices below focal height; focal
height red and green 3000 µm; integration times
red 125, green 625; gain red 31.2, green 31.9
ZProjection red and
green stacks 3–15; background flattening red
and green 3 µm
Cell analysis (eDNA count green threshold 7750
object size 0.6–7 µm, subpopulation colocalized
red mean ≥ 12000 and ≤ 38321; iDNA count red
threshold 12000 object size 0.6–5 µm,
subpopulation colocalized green mean
≥ 7750 and ≤ 50000)
Day 2
28 Z-stacks; one stack below focal height; focal
height 3000 µm; integration times: blue 700, green
4000, red 2175; gain: blue 31.9, green 32, red 32
ZProjection stacks
16–20; background
flattening rolling ball
blue 126 µm, green
and red 3 µm
Cell analysis (total eDNA objects and eDNA strand
objects: GFP threshold value 5000, object size
1.2–11, subpopulation analysis objects size ≥ 2.2
and circularity ≤ 0.25, subpopulation
colocalized red mean ≥12000 and ≤ 40000)
Day 3
50 Z-stacks; 10 images below focal height; focal
height 3000 µm; integration times blue 1675, green
875, red 75; gain blue 32, green 31.9, red 30
ZProjection stacks
14–50; background
flattening rolling ball
green and red 3 µm
Cell analysis (total eDNA objects threshold 7750,
object size 0.6–7, subpopulation co-localized objects red ≥ 12000 and ≤ 38321. Total iDNA objects
threshold 12000, size 0.6–5, subpopulation
colocalized green objects ≥ 7750 and ≤ 50000)
Day 8 13 Z-stacks, focal height 3,000 µm
ZProjection stacks
9–13, and on 7–13.
Background
flattening green and
red 3 µm
Cell analysis (eDNA green threshold 7000, size
1-10, subpopulation colocalized red objects ≥ 5000.
iDNA red threshold 5500, size 0.6–5,
subpopulation colocalized green objects ≥ 8190)
ECGFP Biofilms, Global
Confocal, 40 µm
spinning disk Fluorescence 96-well 20x objective Z-stack with interval 4.2 µm;
GFP (green) 472:520; DAPI (blue) 405:442 ZProjection maximum
method of in-focus
stacks
Qualitative:
Cell analysis (microcolony enumeration) Widefield Fluorescence PE (red) 469:593 20x objective Z-stack with
interval 4.2 µm
Day 1
25 Z-stacks, one image below focal height, focal
height 3086.9 µm; integration time green 4000,
blue 100, red 2125; gains green 32, blue 32, red 32
ZProjection stacks 1–25,
background flattening
green 1 µm fine results,
blue 126 µm red 10 µm
fine results
Qualitative
Day 2
28 Z-stacks, two images below focal height, focal
height 3000 µm; integration time green 4000, blue
75, red 128; gains green 32, blue 31.1, red 32;
illumination red 10
ZProjection stacks 1–28,
background flattening
green 1 µm fine results,
blue 25 µm red 126 µm
Qualitative
Day 3
50 Z-stacks, 10 images below focal height, focal
height 3000 µm; integration time green 1025, blue
150, red 107; gains green 32, blue 31.2, red 32;
illumination red 10
ZProjection stacks 1–50,
background flattening
green 1 µm fine results,
blue 12 µm fast speed
red 12 µm fast speed
Qualitative
Day 8
36 Z-stacks; focal height 3000 µm; integration
times blue 250, green 150, red 6; gain blue 31.4,
green 31.7, red 31.7; illumination red 10
ZProjection stacks
7–17 and stacks 7–13;
background flattening
rolling ball green 50 µm,
red 60 µm
Qualitative:
Cell analysis (individual cells green threshold
12000, size 0.8–11 µm. Cell colonies green
threshold 8000, size 15–50 µm, subpopulation
exclude colocalized protein ≤ 5000. Protein
colonies red threshold 5000, size 15–50 µm,
subpopulation exclude colocalized cells ≤ 8000)
Table 2. Image acquisition and analysis settings. See Table 1 for imaging supply descriptions.
8
Results and discussion
Cell density and total biomass―UV-Vis absorbance
detection
The Cytation C10 is equipped with a custom-programmable
UV-Vis absorbance quad monochromator, allowing the
ability to tailor absorbance measurements over a range of
233 to 999 nm. This was useful in calibrating a cell density
method using turbidity, for measuring relative total biomass
of biofilms using a crystal violet assay, and for monitoring
planktonic bacteria susceptibility to antibiotic over time.
Representative results are discussed below and shown in
Figures 2A to 2D and 5D.
Measuring turbidity of a bacterial suspension is a standard
method for determining the cell density of a sample in a
reproducible way that can be applied across laboratories and
instrument platforms capable of absorbance detection. An
important part of this procedure is to calibrate absorbance
values to cell density per mL. One way to achieve this is by
calculating a standard curve using turbidity standards. The
cell density of a sample can then be interpolated from the
curve for any number of subsequent experiments. There are
a few considerations for implementing this method, such as
differences between absorbance detection system sensitivity,
the influence of cell suspension volume in the well on optical
density value, and the influence of size and morphology
variation between species on turbidity. For example, larger
bacterial cells, like BSub, would be expected to have a higher
OD at the same density as smaller cells like SA. Additionally,
SA is known to form clusters during culture and growth that
even vortexing or shaking may not disperse. Therefore, some
species may warrant results from an orthogonal method like
a hemacytometer or colony counting on an agar plate to be
calibrated back to OD. Although methods can be developed
to work around this, such as using sonication to disrupt SA
clusters, it is generally understood that the same turbidity
result from an SA bacterial suspension represents a cell
density that is more likely to be one log higher than reported
for other species like E. coli. In the experiments performed
in this study, final cell densities were potentially biased, as
cell density calculations were not corrected for cell size or
morphology.
An additional constraint of the turbidity method is found in
relation to the limit of blank (LoB) and limit of detection (LoD)
at the volumes and detection setting used. Although less
susceptible to saturating absorbance values at the high end,
turbidity is not a reliable method for verifying cell dilutions
below those of the LoD―a value on the Cytation C10 found
to be just above that of the zero standard on the turbidity
standard curve, and just below the McFarland 0.5 standard
(1.5 to 3 x 108
/mL) at 200 µL. Having an accurate and
reproducible determination of a starting density is therefore
also important to increase confidence of subsequent dilutions
that will result in ODs below the LoD value and may not be
traceable below the LoB. The LoB and LoD can also introduce
bias in lag times during kinetic bacterial growth monitoring,
as registered ODs below those values, such as when using
low starting seed densities, will not increase until the LoB
threshold is met and exceeded within the cell suspensio,n
even if the cells have started dividing. An example of the
calculated turbidity standard curve used in these studies,
including the LoB and LoD, and representative results for
calculated values of a bacterial suspension, is shown in
Figure 2A. An example of kinetic growth monitoring using
absorbance is shown in Figure 2D.
Another common assay using absorbance detection is the
crystal violet assay assessing total biomass of a biofilm. This
assay is one of the most universally cited assays in biofilm
publications due largely to its simplicity and a workflow
allowing quick turnaround for high-throughput biofilm
analysis. This assay was useful to screen distinct species
for biofilm properties such as qualitative assessment of
morphological characteristics when grown in a 96-well plate
using a ring assay (Figure 2B) and quantitative assessment
of biofilm growth over time (Figure 2C). Figure 2B reveals a
distinct ring formation for PAGFP, indicative of a preference
to grow where oxygen is readily available at the air-liquid
interface. Figure 2C reveals prolific growth of PAGFP
biofilm compared to other species, with little to no growth
of SE. These results were expected as PAGFP is known to
overexpress two polysaccharides that contribute to formation
of a "pellicle" during biofilm formation. This factors into
increased total biomass compared to non- or less-pellicleforming species, and the strain of SE used in the experiment
was chosen as a negative control as it is a known nonbiofilmforming species.12-14 Figure 2D presents turbidity and crystal
violet absorbance data in concert with widefield color
brightfield imaging for verifying antibiotic inhibition of BSub
biofilm formation when exposed to 150 µg/mL ampicillin.
Empirically, the primary limitations of the assay were found to
be:
– It stains total attached biomass—a property that
eliminates the potential to differentiate cell mass
from other biofilm components and potentially stains
extraneous substances that are not part of the biofilm.
– There are a number of aspirate and dispense steps in
the assay that warrant care as biofilms can be disrupted,
resulting in the elimination of final total biomass. These
9
steps also make the assay less compatible to biofilms
formed as nonattached aggregates.
– The elution step of the assay should be optimized as
residual dye may remain in dense and strongly adhered
biofilm, resulting in higher variability and lower accuracy
of results (see Figure 5E).
– The crystal violet stain has a very low saturation point
when detected by absorbance, requiring dilution of some
results to come within the detection sensitivity of the
reader (OD ≤ 4.0).
Although diluting any given biomass at the end of an assay
is relatively straightforward, it is an extra step, and the
dilution factor should be applied to the absorbance value.
Although this potentially increases the relativity of the result,
as absorbance values on 1:10 dilutions of pure crystal violet
in water were not linear (data not shown). One workaround
lies in reading eluted crystal violet by fluorescence using the
extended gain mode available in Gen5 Image Prime software
to increase the range of detection. Regardless of these
findings, the assay is suitable for any biofilm application,
particularly as a biofilm screening method for species biofilm
formation, treatment response, or as a comparative model for
other quantitative procedures.
Figure 2. Examples of UV-Vis absorbance detection assays used in this biofilm evaluation. (A) Turbidity standard curve calculated
to obtain cell density values for biofilm seeding dilutions showing LoB and LoD values. Example of calculated densities from
the curve for a bacterial serial dilution is shown (right). (B) Qualitative assessment of biofilm growth at two time points for two
of five species in (C) using the crystal violet assay. (C) Quantitative results of the crystal violet assay on biofilm growth of five
bacterial species over time in relation to the inoculated cell density calculated from the turbidity standard curve. (D) Results using
absorbance detection and widefield microscopy to assess susceptibility of BSub to the antibiotic ampicillin during planktonic
growth and then following biofilm formation.
10
Cell viability―fluorescence detection mode
A common limitation generally reported about the use of
fluorescence detection assays is that specific excitation
and emission filters are required that provide the specificity
and sensitivity for detection of a given fluorophore. The
fluorescence detection module onboard the Cytation C10 has
a variable-bandwidth quad monochromator from 9 to 50 nm
in 1 nm increments and a wavelength range of 250 to 700 nm,
allowing detection of a diverse range of fluorescent assays
regardless of the excitation and emission limitations often
found in filter-based systems. The fluorescence detection
module was used in this evaluation to compare measures
of cell fitness in overnight growth stock following exposure to
different concentrations of a supplement within the
culture media.
Assessing cell viability in starting biofilm cultures is important
to gauge relative cell health of the culture and provide a basis
for normalization of biofilm results in experiments that use
viability as an end point to measure the effects of applied
variables. In this case, an experiment was done to measure
the viability of PAGFP planktonic cells after overnight growth
in culture media supplemented with different concentrations
of MgCl2
, before propagating the stock in a subsequent
air-liquid interface assay. This was done to investigate the
same on biofilm growth, as it has been shown that MgCl2
, in
addition to acting in a concentration-dependent manner as
a stimulant of bacterial cell division, is also reported to have
antibacterial properties, potentially due to actions on the
bacterial cell membrane.22-24
There are many cell viability fluorescent reagents that are
amenable to bacterial cells. The advantage of these assays is
that they are relatively inexpensive, widely available, reliable,
and easy to perform. A combination of these assays on a
single culture can also be useful to inform on both metabolic
and cell membrane viability in parallel, as described here.
Two viability assays were reviewed. The first assay was a
direct mix-and-read reagent that measures relative aerobic
respiration activity of cells via an intracellular resazurin
to resorufin reduction, where higher relative fluorescence
indicates more metabolically active cells. Results for this
assay were obtained by first calculating a standard curve for
resorufin. Resorufin is innately fluorescent and was detected
using Ex:Em 550:590 nm. The resorufin concentration of
samples was then interpolated from the standard curve.
A calibration protocol feature available in Gen5 Image Prime
software allows a standard curve to be read and calculated
on the first plate, and successive plates can be read that
use the same curve to interpolate resorufin concentration of
samples. Two considerations are worth noting: (A) As with all
fluorescence detection assays, it is important to determine
the gain value for the excitation and emission pair used for
detection. Gen5 Image Prime software has different options
for determining the gain value, and data collected for this
assay used the gain returned for the high-resorufin standard
to use on all subsequent reads. (B) The resazurin reagent
is time sensitive, and although it can be detected within 10
minutes of addition, over longer periods of time, an increasing
assay window between signal-to-noise values is reported.21 In
this experiment, RFU values were obtained for the standard
curve at 10-minute intervals over a 40-minute kinetic time
course. A curve and assay window were calculated for each
time point. Although reliable data with similar signal-to-noise
values resulted at each time point (data not shown), the
time point equating to 30 minutes after plating was used for
calibration, and sample interpolation was therefore also done
using the same time point following addition of the resazurin
reagent. Figure 3A shows the resorufin calibration curve that
was used to interpolate resorufin concentrations of PAGFP
exposed to different concentrations of MgCl2
in the culture
media. PAGFP cultured cells retained high metabolic activity
regardless of MgCl2
experimental dose, indicating that the salt
did not result in loss of cell respiratory viability in overnight
cultures of planktonic cells.
The second viability assay reports on relative permeability of
the cell membrane using a dual-dye system, where both intact
and membrane-permeable cells are targeted by one reagent
(green RFU), while the second reagent displaces the first in
membrane-permeable cells (red RFU). Each of the two dyes
is detected with the same excitation but a different emission
wavelength. A higher relative fluorescent ratio between
the emission values indicates more viable cells with intact
membranes in the culture compared to compromised cells.21
The assay was adapted using total PAGFP expression from
a 2.5x dilution of the PAGFP without MgCl2
(0 mM) as the live
cell control in place of the green dye.
11
An expired PAGFP cell culture in stale media that is kept
as a nonviable control was used as the "dead" cell control
and dispensed at an inverse percent to the live cells within
each standard. A live/dead cell curve was calculated from
the ratio between the green signal for live cells and the red
emission signal for dead cells plotted against the percent
live cells within each standard. Ratio values obtained from
the standard curve were multiplied by the cell dilution factor
to normalize results to the tested samples (a 2.5x dilution of
the live cell control resulted in the same starting cell density
as the dead cells). Figure 3C shows the live/dead curve, and
Figure 3D shows the viability ratio results for four replicates
of PAGFP from each concentration of MgCl2
compared to
the ratio of the 100% live cell standard. Results from the live/
dead assay indicate that the growth culture exhibits MgCl2
dose-dependent membrane competence, with the increase
in viability likely due to the growth stimulant properties of the
salt, as previously reported.25 This would therefore increase
the live cell population in those samples.
Figure 3. Examples of fluorescence detection assays used in this biofilm evaluation. (A) Resorufin standard curve, calculated to
obtain metabolic viability values for biofilm seed stocks. (B) Resorufin concentration values interpolated from the standard curve
to assess aerobic respiration activity of samples cultured with different concentrations of MgCl2
+
in the growth media. (C) A live/
dead assay standard curve, using the 0 mM MgCl2
PAGFP bacteria stock as the live cells, and PAGFP cells from an expired lot
as the dead cells. (D) Membrane competence of the same sample cells assessed in (B) compared to the 100% live cell standard
(STD100) prepared from the 0 mM stock.
12
Adherence―color brightfield and fluorescence widefield
microscopy
A recent study reported evidence that up to 80% of bacteria
and archaea on Earth exist as biofilms.7
Although relatively
nondisruptive and even synergistic in an organic context,
biofilms can also cause persistent disease and infection
in natural systems and irreparable damage in inorganic
environments. Technically, biofilms are defined as “an
aggregate of microorganisms, like bacteria, in which cells
are frequently embedded within a self-produced matrix of
EPSs and adhere to each other and/or to a surface”.8 There
is extensive science done to understand biofilm adherence
properties, including:
– Mechanistic/molecular changes in the bacteria
– Variables influencing their preference for different
surfaces or bacterial species
– The effects of environmental conditions on attachment
– Whether different surface or other treatments—either
applied or integrated into an experimental surface
directly, or to biofilms growing on a surface—can prevent,
promote, or eliminate biofilm attachment.
Under circumstances where there can be numerous assay
parameters to optimize, it is useful to have high-throughput
screening methods that can efficiently produce data that is
relevant to answering macro questions such as how much
and/or what parts of a biofilm biomass may remain or be
eliminated by experimental variables, where microanalysis of
individual cells or biofilm constituents is less informative.
The widefield microscopy module on the Cytation C10 is a
useful tool for performing high-throughput macroanalysis
of biofilms. An entire well of a 96-well plate can be captured
using a low-magnification 2.5x objective or, using a 3 x 3
image montage. A low-magnification 2.5x objective can
capture an entire well of a 96-well plate in one image or, using
a 3 x 3 image montage, each well of a 24-well plate. This can
be accomplished within a single protocol that defines several
imaging modes including brightfield and fluorescence. In
addition to facilitating differentiation of individual components
of a biofilm from the total biomass for example, this lowmagnification widefield imaging technique also benefits from
both fast image acquisition times, enabling high throughput
screening of multiple variables in parallel, and efficient use
of computer memory by decreasing storage requirements
for saving images and data, that may result from screening
different nutritional factors or growth techniques during
biofilm development. An image statistic interface within
Gen5 Image Prime software can be used to calculate
numerous high-level metrics that can be useful for
quantitating larger representations of biology, such as areas
of attached cell mass. The low-magnification and image
statistics screening technique is demonstrated here by two
experiments.
One experiment compared biofilm surface affinity between
two different microplates. The goal of the experiment was
to investigate whether a less expensive bottom substrate
designed for high-quality imaging (optically enhanced
polystyrene (OEPS)) compared to a more costly one with
a known high optical resolution (glass). This was done by
evaluating cell mass confluence on both surfaces following
perturbation from media aspiration in live biofilms. The
variable of aspiration was chosen because it is a common
occurrence in biofilm assay workflows, whether for media
exchanges, staining, or washing of biofilms. The aspiration
step can also challenge the use of automated liquid handling
devices for biofilm assays, which in turn, can constrain the
application of high-throughput workflows that are designed
to reduce laboriousness and variability of manual methods.
Representative wells of unperturbed and perturbed SA
biofilms in Figure 4A are shown in a view highlighting cell
mass confluence (red) using a threshold outlier feature
(white) in image statistics. Gen5 Image Prime software
calculates confluence in the Image Statistics interface as
the number of pixels within a user-defined intensity range,
divided by the total pixels of the image, multiplied by 100.
The signal threshold defined for this experiment is shown in
Table 2. In this experiment, cell confirmation within an area
of the total cell mass was done using a 20x confocal image
Z-stack, an example of this can be seen in the assay results
represented by Figures 6 and 7. Using data from the mean
of nine replicates of both aspirated and nonaspirated wells,
Figure 4B indicates that, likely due to differences in ionic
forces of the bottom surfaces, the glass surface resulted in
statistically significant higher growth. However, the OEPS
surface had a higher percent biofilm retention, indicating that
cells may be more strongly adhered to the OEPS surface.
Additionally, although both surfaces were subject to cell mass
loss from gentle aspiration, irreversibly attached persister
cells remained, and the biofilm was not entirely eliminated
from either surface.
13
Figure 4. Widefield microscopy image analysis comparing SA biofilm
surface affinity to different substrates. (A) Threshold images of live biofilm
cell mass confluence on different surfaces before (growth) and after
(retained) media aspiration on day 8 SA biofilms. (B) Biofilms grew more but
retained less cell mass on glass than on the composite surface at day 8.
The goal of the second experiment was to find optimal
conditions for growing biofilms on the bottom of a highdensity vessel that could be more conducive to imaging by
reducing or eliminating biofilm growing outside the focal
range of the microscope, such as vertically along the sides
of the wells. This followed evidence from prior experiments
(see Figure 2B) showing that PAGFP preferred to grow as a
biofilm at an oxygen-rich air-liquid interface. This can be seen
by the prominent ring of biofilm around the microplate wells
from where the top of the inoculum volume was in direct
contact with oxygen, and along the sides of the wells where
the pellicle drapes down as a function of media aspiration.
This pattern renders much of the biology incompatible to
microscopic analysis within the vessel. An additional problem
had been found from empirical data collected by independent
experiments revealing that, after dispensing small volumes
of bacterial inoculum onto the bottom of several different
microplates, a quick migratory response of the inoculum
towards the sides of the well bottom that were in contact with
the walls of the vessel occurred. An experiment was therefore
designed to screen for conditions favorable to the adherence
of cells to the bottom of a microplate that could be captured
within the widefield imaging field of view of the Cytation C10
in a high-throughput manner, allowing multiple replicates of
multiple conditions to be tested in parallel. Two techniques
were investigated. The first was to grow biofilm at an angle,
where the air-liquid interface was in direct contact with the
bottom of the vessel, resulting in an area of biofilm that
could be captured by inverted microscopy, and to compare
results from that to biofilms grown in vessels lying flat. The
second technique was to use different concentrations of
MgCl2
as a supplement in the culture media, as it has been
reported that in addition to the antibacterial and bacterial
growth stimulating properties of MgCl2
23, 25, the salt has also
been investigated in a species-, concentration-, and media
composition-dependent manner for possible effects on
bacterial attachment through electrostatic and physiologydependent adherence processes.23 This could possibly
mitigate the migration pattern observed empirically by
manipulating the ionic charge within the vessel, making the
sides along the bottom of the well less attractive to the cells.
The workflow of the assay was to first verify whether MgCl2
influenced planktonic growth and cell viability. These results
were discussed previously and shown in Figure 3. The same
growth stock from the viability experiments was prepared in
fresh MgCl2
-supplemented media and inoculated in replicates
of seven to either the angled or flat air-liquid interface plates
to also gauge the effects of MgCl2
for each method.
Three metrics were used to compare the growth methods.
For each well, cell mass within the biofilm was differentiated
using fluorescence imaging of GFP at 2.5x in a 3 x 3 montage
(Figure 5A). Image statistics was used to calculate total signal
intensity and total area of the cell mass (Figures 5B and C).
Complete image acquisition and analysis settings are defined
in Table 2. Several signal threshold settings were evaluated.
Higher threshold values resulted in targeting areas of high
signal intensity of the biofilm cell mass representing greater
cell density, whereas lower signal threshold values captured
more overall cell mass within the biofilm, including areas that
were less dense. As slight differences in the experimental
variables may be detected more at the low end than the
high—for example high cell mass would be detected even
14
at lower thresholds, but lower cell mass areas could be lost
from calculation at higher thresholds—a lower value that was
representative of more cell mass was used for data analysis.
The data shows that the total area and intensity values
correlate in both a growth method and dose-dependent
manner, with the 10 mM MgCl2
biofilms grown flat having
statistically significantly more total intensity and area of cell
mass than biofilms grown with no MgCl2
either flat or at an
angle. The 10 mM angle replicates were more variable and
somewhat less dense than those grown flat, but also still
higher in both area and intensity than cell mass grown at 0
mM MgCl2
; however, these replicates were not significantly
different than any of the other methods. The lower P values
for the total intensity comparisons could indicate that the
cell mass was also denser for the 10 mM MgCl2
replicates,
indicating more strongly adhered cell mass. This would
coincide with the finding that MgCl2
stimulates cell division
and, in this case, even as a biofilm phenotype.25 Of note, the
1 mM MgCl2
biofilms presented with spurious and
inconclusive results requiring further study and were therefore
left out of this analysis.
Following GFP imaging, the biofilms were stained with crystal
violet to assess total biomass within the wells. Following
fixation, staining, and drying overnight, the total biomass of
each well was imaged at 2.5x as a 3 x 3 montage in color
brightfield mode (Figure 5A). The crystal violet was then
eluted as described previously and detected at 590 nm OD
(Figure 5D). The two most notable results from the total
biomass data were that the 10 mM flat replicates presented
with lower overall total biomass than the 0 mM flat replicates,
even though they had statistically higher total intensity and
total area as calculated from the image analysis; and the
0 mM flat replicates were statistically different from both
groups of replicates from the angle plates―findings also not
supported by the image analysis. From past empirical data
it was observed that residual crystal violet stain remained
in biomass after elution and that this may be a cause of
variability with the method. Following the detection step,
therefore, stain was aspirated from the wells and the wells
were washed until the liquid ran clear. The residual biomass
was then dried, and the wells imaged again in color brightfield
to capture any residual staining remaining in the well. Two
examples of residual staining from the 0 mM angle and
0 mM flat plate are shown in Figure 5E (upper left and upper
right of the quadrant respectively) compared to blank wells
from the same plates. These biofilm wells were chosen
because they were the furthest outliers from the crystal
violet assay of each respective replicate set; both these
replicate sets had the most significant variability and change
from the cell mass data, and these replicate groups had the
highest residual staining of all replicate groups as determined
qualitatively. This may reinforce the previous findings that
would suggest residual noneluted crystal violet may be
a cause of increased variability in biofilms. Less residual
staining in the 10 mM flat replicates, and a lower overall total
biomass readout for those replicates than for the 0 mM flat
plate, could indicate that the presence of MgCl2
may result
in less pellicle formation even with higher cell density, as has
been previously reported.25 This could also be supported by a
qualitative assessment from the 10 mM angle representative
well in Figure 5A that visually presents with thinner and lighter
staining of the noncell mass areas of the total biomass where
the pellicle would stain, as compared to the 0 mM angle well,
for example.
In summary, the collective data suggest that PAGFP biofilms
supplemented with 10 mM MgCl2
and grown lying flat within
a microplate vessel produced biofilms with higher cell mass,
but potentially less pellicle formation in proportion, compared
to those grown at an angle with or without supplement, or flat
with no supplement. All methods resulted in robust biofilm
formation containing high-density cell mass conducive to
inverted imaging through the bottom of the vessel using a
low-magnification, high-throughput approach. Image analysis
proved a sufficient and enhanced method for screening data
on individual components of a biofilm as represented by cell
mass than the comparative crystal violet assay that does not
discriminate individual biofilm components.
15
Figure 5. Widefield image analysis comparing PAGFP biofilm adherence using different growth methods and media
concentrations of MgCl2. (A) Images of biofilm cell mass in live biofilms and total biomass postfixation and crystal
violet staining when grown flat or at a 45° contact angle to the bottom of a microplate with and without MgCl2. (B) Total
cell mass area for each condition. (C) Total cell mass signal intensity for each condition. (D) total biomass detected
using absorbance detection of eluted crystal violet staining. (E) Examples of residual crystal violet staining after elution,
demonstrating artifact of the method for a 0 mM angle (top left) and a 0 mM flat (top right) well compared to blank wells
from the respective test plates (bottom). All error bars are the mean ± 95% confidence interval.
16
EPS components and architecture― confocal microscopy
Biofilms are characterized by the development of an EPS
that is propagated when bacteria develop a biofilm lifestyle.
This gives them an inherent three-dimensional quality that
can be observed and reconstructed using optical sectioning
with a confocal microscope. Although a widefield imager can
also capture biofilm compositional properties, confocal laser
scanning microscopy (CLSM) is frequently used in biofilm
studies due to the ability to reduce out-of-focus light and
improve image quality. The reduced background resulting
from this method is especially useful when imaging live
biofilms over a range of focal heights in host or experimental
media, as background such as media autofluorescence
can often interfere with acquiring clear images during
more detailed microanalysis of very small objects such as
bacteria or other biofilm components. This can become more
pronounced when moving deeper into the biofilm volume.9, 10,
15
The Cytation C10 confocal imaging reader enables capture
of fluorescent images across the blue to far-red spectrum
with a choice of seven filter cubes and two spinning-disk
sizes to accommodate both lower and higher magnification
from 20 to 60x. The Gen5 Image Prime software Z-stacking
tool can be defined for step intervals as low as 0.1 µm over a
total stack height of up to 3,333 slices depending on interval
step size. Numerous different stage adapters are available
to accommodate a variety of labware vessels suitable for 3D
image analysis of biofilms.
Two assays, one using a gram-positive and the other a
gram-negative species, were performed to evaluate biofilm
composition using the confocal microscope on the Cytation
C10. In one assay, SA biofilms were used to monitor the
presence and localization of eDNA in relationship to iDNA
over time. eDNA is a key component of bacterial biofilm EPS
with a role in biofilms that continues to be investigated, but
has been shown to potentially represent a mechanism for
horizontal gene transfer in bacteria and/or provide structural
definition resulting in channels that enable moving, diffusing,
and/or assimilating constituents within the biofilm, such as
nutrients, waste, or therapeutics.3, 4, 11, 16
A lack of colocalization of eDNA and iDNA within the biofilm,
in addition to the presence of both free eDNA strands
and transitionary states of eDNA availability in the matrix
following loss of cell viability, could represent support for
the theories that eDNA plays both a structural role and
gene transfer mechanism with SA biofilms. This hypothesis
was investigated qualitatively and quantitatively using two
nucleic acid stains that differ in their cell permeability to
differentiate and determine colocalization of iDNA and eDNA.
Polysaccharides within the biofilm EPS were visualized using
a counterstain for cellulose.
Characteristics of an eight-day lifecycle of SA biofilms was
assessed using qualitative analysis on optimal Z-stack
images obtained from confocal microscopy at 20x. Results
are shown in Figure 6.
Figure 6. Qualitative assessment of SA biofilm development over time.
iDNA (red), eDNA (green), polysaccharides (predominately cellulose, blue).
Although most eDNA staining is due to membrane permeability of intact
nonviable cells, examples of eDNA strands are highlighted in the early growth
matrix (zoom insets). Images are taken from live biofilms.
17
Cell analysis was then used to compare percent eDNA, iDNA,
and colocalized objects within the biofilm over the same time
periods using multiple Z-stacks. The cell analysis steps taken
to differentiate and quantitate the percent of eDNA, iDNA, and
colocalized objects on each in-focus Z-stack compared to a
Z-projection on all in-focus stacks is shown in Figure 7.
Figure 7. Quantitative analysis of eDNA and iDNA in live SA biofilms using the Agilent BioTek Cytation C10 confocal imaging
reader and Gen5 Image Prime software confocal Z-stacking image analysis. (A) Steps for enumerating eDNA and iDNA in
SA biofilms (left). Subpopulation analysis is used to define colocalized eDNA and iDNA objects (step 4, right). The orange
border area in the step 3 image represents the zoomed imaged area in step 4. Blue arrows identify colocalized objects. (B)
Percent of eDNA, iDNA, and colocalized objects calculated from total eDNA and iDNA objects are graphed. (Lower left) the
mean percent is plotted on n = 6 from Z-projections of total in-focus stacks for each individual day. (Lower middle) Percent
objects for each in-focus stack compared to the Z-projection of all stacks is shown for one day 8 biofilm. (Lower right) The
thumbnail image accesses a rotating 3D rendition of stacks 10 to 13 represented by the middle graph.
18
The Z-projection comparative analysis containing all infocus stacks was done based on a premise that iDNA and
eDNA may be more likely to overlap in a compressed Z-plane
then within any individual stack, a property of Z-projection
on multiple stacks that could bias both the enumeration
and localization of individual objects. A distinct geography
of eDNA and iDNA within the biofilm was found from this
analysis. The data from both the Z-projections of all in-focus
stacks from all days and individual stack data extrapolated
for day 8 shows a compelling lack of overlap (≤ 1.6%)
between eDNA and iDNA in both individual and multistack
Z-projections. Although this finding could be partially biased
based on the choice of signal threshold defining a colocalized
object, a possible interpretation could be that iDNA and
eDNA objects remain in a sessile mode within the biofilm,
adding architectural features such as structural support or,
as reported, channels within the biofilm, for example. The
data also show that an inverse relationship of iDNA to eDNA
developed during growth of SA biofilms, as indicated by a fivetime increase in percent eDNA and a 2.5x decrease of percent
iDNA from days 1 to 3 to day 8. It could be inferred from this
that bacteria may enter a prolonged stationary stage over
time due to growth density or other stress that results in slow
or no viable cell proliferation within the biofilm, parallel to a
loss of sustainable viability of existing cells. The quantitative
cell analysis also resulted in a finding that, although a majority
of eDNA was represented as a homogenous population of
membrane-permeable nonviable cells, a distinct population of
eDNA objects could be characterized as eDNA strands, such
as those shown in Figure 6. Size and shape qualities of these
strands was applied to enumerate them using subpopulation
analysis, and the percent of these objects compared to the
total eDNA masked within the image was calculated using a
scatter chart analysis.
Figure 8 represents an example of the subpopulation analysis
for a 48-hour biofilm, resulting in a calculated 4.5% of total
eDNA objects defined as eDNA strand objects at that growth
time point.
In the second assay, qualitative and quantitative analysis
using confocal microscopy was used to evaluate microcolony
development in ECGFP biofilms over time. These well-defined
aggregates of cells and other EPS components have been
hypothesized to form due to cellular quorum sensing as a
way of conserving energy and resources while supporting an
environment favorable for producing progeny.5, 6, 18, 19 Findings
from the qualitative assessment are characterized in Figure
9, illustrating observed stages of microcolony growth over
four time points shown at 20x, with 60x images of individual
channels from a single Z-slice from another replicate at day 8
shown for comparison.
Figure 8. Subpopulation analysis of eDNA strand objects using the
Agilent BioTek Gen5 Image Prime software. (Top) criteria for defining a
subpopulation on eDNA strand objects; (center) example of eDNA object
(left) identified by the subpopulation criteria with a pink mask (right);
(bottom) scatter chart showing the percent of all objects meeting the
subpopulation criteria in the image (pink) versus total eDNA objects
enumerated (black). The full image for this representative data can be seen
in Figure 6, shown with examples of other strand objects zoomed.
Figure 10 describes a process for quantitating microcolonies
using cell analysis steps to differentiate cellular and EPS
protein components and enumerate well-formed and
proliferated microcolonies, over each layer of a Z-stack
acquired from the minimum and maximum height of infocus cellular and EPS protein objects, from day 8 of biofilm
growth. From the images at this time point the in-focus
microcolony biology covered a total Z-height of 37.8 µm
(Z-stacks were acquired at 4.2 µm focal height intervals).
The graph of enumerated microcolony objects for each stack
indicates correlation of cell and EPS colocalization over
all in-focus stacks. A movie from the bottom to top stacks
of the microcolony analyzed provides a visual rendition of
this. In addition to showing a change in cell density over the
Z-height of the microcolonies, which corresponds to a peak
of enumerated cellular aggregates in the middle four stacks
of the Z-plane, dense cell aggregates exist even in lesspopulated stacks. This can also be seen in Figure 9 from an
earlier time point characterizing colony division, where there
are few cells in the surrounding biofilm but densely coalesced
cells colocalized to protein aggregates.
19
Figure 9. Qualitative assessment of ECGFP microcolony development in live biofilms. Bacteria (green), EPS proteins (red), EPS
polysaccharides (predominately cellulose, blue). A) 20X evaluation over time. B) 60X at day 8 for comparison.
20
Figure 10. Steps for identifying and enumerating microcolonies using the Agilent BioTek Cytation
C10 confocal imaging reader and Agilent BioTek Gen5 Image Prime software confocal Z-stacking
image analysis.
In contrast, protein aggregates remain uniformly enumerated
over nine stacks, including the stacks with peak cellular
aggregates, even continuing into higher elevation within the
biofilm, where many of the protein aggregates (~15 out of
25) are no longer colocalized with cellular objects. Due to the
number of microcolonies in the imaged area, one possible
explanation for the increased individual cell density in some
stacks may be the result of some microcolonies outside the
imaged area collapsing and releasing progeny into the biofilm
matrix. This, in turn, may lead to increased microcolony
development, or more individual cells joining established
microcolonies, for example. The finding that there are more
protein aggregates more uniformly enumerated over the
entire Z-plane, and prominently at the higher elevations where
they no longer colocalize with cellular aggregates, could imply
a structural scaffolding role for the proteins in microcolony
development, acting as a cohesive element both layered with
and enveloping the cells, for example.
21
Conclusion
The Agilent BioTek Cytation C10 confocal imaging reader
has capabilities conducive to integrating data acquisition
and analysis from different detection and imaging methods,
as demonstrated using biofilm assays as a model. Multiple
bacterial strains, different experimental variables, and a
variety of stains and vessels were analyzed by means of
numerous configurations of absorbance and fluorescence
detection and both widefield and confocal imaging modes,
using multiple objective magnifications. This resulted in
a variety of analysis outcomes informing on fundamental
experimental optimization variables common to biofilm in
vitro workflows, including determining cell density and viability
of starting cultures, biofilm formation and total biomass
screening, assessing qualities of biofilm substrate adherence,
characterizing EPS and growth architecture from live biofilm
imaging, and comparing antimicrobial tolerance of planktonic
and biofilm bacterial cells. The instrument eliminates
constraints that may be imposed by a single-purpose
instrument. It has a configurable design with the potential to
replace up to five individual devices and is controlled using
one centralized user interface and data management format,
all at a price point that offers significant value.
References
1. Ghannoum, M.; Parsek, M.; Whiteley, M.; Mukherjee,
P. Editors, Microbial Biofilms, 2nd Edition. ASM Press,
October 2015. ISBN:9781555817459. Microbial Biofilms,
2nd Edition | Wiley.
2. Clinical Laboratory Standards Institute EP17-A2 standard
Evaluation of Detection Capability for Clinical Laboratory
Measurement Procedures; Approved Guideline-Second
Edition, June 2012.
3. Montanaro, L.; Poggi, A.; Visai, L.; Ravaioli, S.; Campoccia,
D.; Speziale, P.; Arciola, C.R. Extracellular DNA in biofilms.
Int J Artif Organs. 2011 Sep; 34(9):824-31. DOI: 10.5301/
ijao.5000051
4. Azeredo, J.; Azevedo, N.; Briandet, R.; Cerca, N.; Coenye,
T.; Costa, A.R. et al. Critical review on biofilm methods.
Critical Reviews in Microbiology 2017, Vol 43, pgs 313-
351. DOI: 10.1080/1040841X.2016.1208146
5. Wilson, C.; Lukowicz, R.; Merchant, S.; Valquier-Flynn, H.;
Caballero, J.; Sandoval, J.; Okuom, M.; Huber, C.; Brooks,
T.D.; Wilson, E.; Clement, B.; Wentworth, C.D.; Holmes,
A.E. Quantitative and Qualitative Assessment Methods
for Biofilm Growth: A Mini-review. Res Rev J Eng Technol.
2017 Dec; 6(4) PMID: 30214915
6. Guzman-Soto, I.; McTiernan, C.; Conzalez-Gomez, M.;
Mah, T-F.; Griffith, M.; Alarcon, E. et al. Mimicking biofilm
formation and development: Recent progress in in vitro
and in vivo biofilm models. iScience 24, 2021 May 21,
102443, pgs 1-51. DOI: 10.1016/j.isci.2021.102443
7. Flemming, H-C.; Wuertz, S. Bacteria and archaea on Earth
and their abundance in biofilms. Nat Rev Microbiol 17,
247-260 (2019). DOI: 10.1038/s41579-019-0158-9
8. Vert. M. et al. Terminology for biorelated polymers and
applications (IUPAC recommendations 2012). Pure
Appl. Chem. (2012) 84, 377-410. DOI: 10.1351/PACREC-10-12-04
9. New, T.R.; Manz, B.; Volke, F.; Dynes, J.J.; Hitchcock,
A.P.; Lawrence, J.R. Advanced imaging techniques for
assessment of structure, composition, and function in
biofilm systems. FEMS Microbiol Ecol. 2010 Apr; 72(1):1-
21. DOI: 10.1111/j.1574-6941.2010.00837.x
10. Schlager, S.; Meyer, R.L. Confocal microscopy imaging
of the biofilm matrix. Journal of Microbiological Methods
138, 2017, pgs 50-59. DOI: 10.1016/j.mimet.2016.03.002
11. Penesyan, A.; Paulsen, I.T.; Kjelleberg, S.; and Gillings, M.R.
Three faces of biofilms: a microbial lifestyle, a nascent
multicellular organism, and an incubator for diversity. npj
Biofilms and Microbiomes 2021, Volume 7 Article Number
80, pgs 1-9. DOI: 10.1038/s41522-021-00251-2
12. Zhang, Y-Q; Ren, S-H; Li, H-L, et al. Genome‐based
analysis of virulence genes in a non‐biofilm‐forming
Staphylococcus epidermidis strain (ATCC 12228).
Molecular Microbiology 2003 DOI: 10.1046/j.1365-
2958.2003.03671.x
13. Marmont, L.S.; Whitfield, G.B.; Rich, J.D.; Yip, P.;
Giesbrecht, L.B.; Stremick, C.A.; Whitney, J.C.; Parsek,
M.R.; Harrison, J.J.; Howell, P.L. PelA and PelB proteins
form a modification and secretion complex essential
for Pel polysaccharide-dependent biofilm formation
in Pseudomonas aeruginosa. J Biol Chem. 2017 Nov
24;292(47):19411-19422. DOI: 10.1074/jbc.M117.812842
14. Thi, M.T.T; Wibowo, D., Rehm, BHA. Pseudomonas
aeruginosa Biofilms. Int J Mol Sci. 2020 Nov
17;21(22):8671. DOI: 10.3390/ijms21228671
15. Reichhardt, C.; Parsek, MR. Confocal Laser Scanning
Microscopy for Analysis of Pseudomonas aeruginosa
Biofilm Architecture and Matrix Localization.
Front Microbiol. 2019 Apr 2; 10:677. DOI: 10.3389/
fmicb.2019.00677
16. Okshevsky, M.; Meyer, R.L. The role of extracellular DNA
in the establishment, maintenance, and perpetuation of
bacterial biofilms. Critical Reviews in Microbiology, 2015,
41:3, 341-353. DOI: 10.3109/1040841X.2013.841639
17. Sauer, K.; Stoodley, P.; Goeres, D.M.; Hall-Stoodley, L.;
Burmolle, M.; Stewart, P.S.; and Bjarnsholt, T. The Biofilm
life cycle: expanding the conceptual model of biofilm
formation. Nature Reviews, Microbiology 2022, 20, 608-
620. DOI: 10.3109/1040841X.2013.841639
18. Flemming, HC.; Wingender, J. The biofilm matrix. Nature
Reviews Microbiology 2010, 8, 623-633. DOI: 10.1038/
nrmicro2415
19. Sharma, G.; Sharma, S.; Sharma, P.; Chandola, D.;
Dang, S.; Gupta, S.;Gabrani, R. Escherichia coli biofilm:
development and therapeutic strategies. J Appl Microbiol.
2016. Aug;121(2),309-19. DOI: 10.1111/jam.13078
20. Live/Dead BacLight Bacterial Viability Kits, Product
Information Manual. Revised: 15-July-2004. Document
Connect (thermofisher.com)
21. PrestoBlue HS Cell Viability Reagent Product Information
Manual Rev. B.0. Document Connect (thermofisher.com)
22. Song, B.; Leff, L.G. Influence of magnesium ions
on biofilm formation by Pseudomonas fluorescens,
Microbiological Research, 2006, 151, Issue 4, 355-361.
DOI: 10.1016/j.micres.2006.01.004
23. Oyarzúa Alarcón, P.; Sossa, K.; Contreras, D.; Urrutia,
H.; Nocker, A. Antimicrobial properties of magnesium
chloride at low pH in the presence of anionic bases.
Magnes Res. 2014 Apr-Jun;27(2):57-68. DOI: 10.1684/
mrh.2014.0362
24. Webb, M. The Influence of Magnesium on Cell Division
3. The Effect of Magnesium on the Growth of Bacteria in
Simple Chemically Defined Media. Microbiology, 1949,
3(3), 418-424. DOI: 10.1099/00221287-3-3-418
25. Costerton JW, Lewandowski Z, Caldwell DE, Korber
DR, Lappin-Scott HM. Microbial biofilms. Annu Rev
Microbiol. 1995, 49:711-45. DOI: 10.1146/annurev.
mi.49.100195.003431
26. Merritt, J.H.; Kadouri, D.E.; O’Toole, G.A. (2006),
Growing and Analyzing Static Biofilms. Current
Protocols in Microbiology, 00:1B.1.1-1B.1.17. DOI:
10.1002/9780471729259.mc01b01s00
27. Allkja, J.; Charante, van F.; Aizawa, J.; Reigada, I.; GuarchPerez, C.; et al. Interlaboratory study for the evaluation
of three microtiter plate-based biofilm quantification
methods. Nature Scientific Reports 2021 11:13779. DOI:
10.1038/s41598-021-93115-w
28. O’Toole, G.A. Microtiter dish biofilm formation assay. J
Vis Exp. 2011 Jan 30; (47):2437. DOI: 10.3791/2437
29. Crouzet, M.; Le Senechal, C.; Brozel, V.S.; Costaglioli, P.;
Barthe, C.; Bonneu, M.; Garbay, B.; Vilain, S. Exploring early
steps in biofilm formation: set-up of an experimental
system for molecular studies. BMC Microbiol. 2014 Sep
30; 14:253. DOI: 10.1186/s12866-014-0253-z
30. Okshevsky, M.; Meyer, R.L. Evaluation of fluorescent
stains for visualizing extracellular DNA in biofilms.
Journal of Microbiological Methods 2014 105, 102-104.
DOI: 10.1016/j.mimet.2014.07.010
31. Flemming, HC.; Van Hullebusch, ED.; Neu, TR.; Nielsen,
PH.; Seviour, T.; Stoodley, P.; Wingender, J.; Wuertz, S. The
biofilm matrix: multitasking in a shared space. Nat Rev
Microbiol. 2023 Feb;21(2): 21: 70-86. DOI: https://doi.
org/10.1038/s41579-022-00791-0
www.agilent.com/lifesciences/biotek
DE87094723
This information is subject to change without notice.
© Agilent Technologies, Inc. 2023
Published in the USA, December 11, 2023
5994-6764EN
Brought to you by
Download This App Note for FREE Now!
Information you provide will be shared with the sponsors for this content. Technology Networks or its sponsors may contact you to offer you content or products based on your interest in this topic. You may opt-out at any time.
Experiencing issues viewing the form? Click here to access an alternate version