Oxygen microenvironments in Escherichia coli biofilm nutrient transport channels: insights from complementary sensing approaches

Abstract

Chemical gradients and the emergence of distinct microenvironments in biofilms are vital to the stratification, maturation and overall function of microbial communities. These gradients have been well characterized throughout the biofilm mass, but the microenvironment of recently discovered nutrient transporting channels in Escherichia coli biofilms remains unexplored. This study employs three different oxygen sensing approaches to provide a robust quantitative overview of the oxygen gradients and microenvironments throughout the biofilm transport channel networks formed by E. coli macrocolony biofilms. Oxygen nanosensing combined with confocal laser scanning microscopy established that the oxygen concentration changes along the length of biofilm transport channels. Electrochemical sensing provided precise quantification of the oxygen profile in the transport channels, showing similar anoxic profiles compared with the adjacent cells. Anoxic biosensing corroborated these approaches, providing an overview of the oxygen utilization throughout the biomass. The discovery that transport channels maintain oxygen gradients contradicts the previous literature that channels are completely open to the environment along the apical surface of the biofilm. We provide a potential mechanism for the sustenance of channel microenvironments via orthogonal visualizations of biofilm thin sections showing thin layers of actively growing cells. This complete overview of the oxygen environment in biofilm transport channels primes future studies aiming to exploit these emergent structures for new bioremediation approaches.

Full text

1
Oxygen microenvironments in Escherichia coli biofilm
nutrient transport channels: insights from complementary
sensingapproaches
BeatriceBottura1,†, GailMcConnell1,†, Lindsey C.Florek2, Marina K.Smiley2, RossMartin1, ShannanFoylan1, AshEana1,
Hannah T.Dayton2, Kelly N.Eckartt2, Alexa M.Price- Whelan2, Paul A.Hoskisson1, Gwyn W.Gould1, Lars E.P.Dietrich2 and
Liam M.Rooney1,3,*
RESEARCH ARTICLE
Bottura etal., Microbiology 2025;171:001543
DOI 10.1099/mic.0.001543
Received 09 January 2025; Accepted 03 March 2025; Published 06 May 2025
Author aliations: 1Strathclyde Institute for Pharmacy and Biomedical Sciences, University of Strathclyde, Glasgow, G4 0RE, UK; 2Department of
Biological Sciences, University of Columbia, New York City, NY, 10027, USA; 3Department of Bacteriology, School of Infection & Immunity, University of
Glasgow, Glasgow, G12 8TA, UK.
*Correspondence: Liam M. Rooney, liam. rooney@ glasgow. ac. uk
Keywords: biofilm physiology; microbial communities; redox sensing.
Abbreviations: LB, lysogeny broth; PMT, photomultiplier tubes; ROS, reactive oxygen species.
†These authors share senior authorship.
Two supplementary figures and one supplementary table are available with the online version of this article.
001543 © 2025 The Authors
This is an open- access article distributed under the terms of the Creative Commons Attribution License. This article was made open access via a Publish and Read agreement between
the Microbiology Society and the corresponding author’s institution.
Abstract
Chemical gradients and the emergence of distinct microenvironments in biofilms are vital to the stratification, maturation and
overall function of microbial communities. These gradients have been well characterized throughout the biofilm mass, but the
microenvironment of recently discovered nutrient transporting channels in Escherichia coli biofilms remains unexplored. This
study employs three dierent oxygen sensing approaches to provide a robust quantitative overview of the oxygen gradients
and microenvironments throughout the biofilm transport channel networks formed by E. coli macrocolony biofilms. Oxygen
nanosensing combined with confocal laser scanning microscopy established that the oxygen concentration changes along
the length of biofilm transport channels. Electrochemical sensing provided precise quantification of the oxygen profile in the
transport channels, showing similar anoxic profiles compared with the adjacent cells. Anoxic biosensing corroborated these
approaches, providing an overview of the oxygen utilization throughout the biomass. The discovery that transport channels
maintain oxygen gradients contradicts the previous literature that channels are completely open to the environment along
the apical surface of the biofilm. We provide a potential mechanism for the sustenance of channel microenvironments via
orthogonal visualizations of biofilm thin sections showing thin layers of actively growing cells. This complete overview of the
oxygen environment in biofilm transport channels primes future studies aiming to exploit these emergent structures for new
bioremediation approaches.
INTRODUCTION
e spatial and structural heterogeneity of biolms leads to the formation of complex emergent architectural features that maintain
their durability and ability to distribute key resources. Recently discovered Escherichia coli biolm transport channels present an
example of such structures [1]. ese nutrient transport channels were discovered using novel mesoscopic imaging methods [2].
ey were found to form according to their local environmental conditions [3] as an emergent property governed by cell shape
and driven by cell- to- cell and cell–surface interactions [4]. e proven ability of these channels to translocate nutrients and small
particles within mature macrocolony biolms presents a timely opportunity to develop exploitative channel- mediated bioremedia-
tion strategies to lower the impact of biolms on public health and industry. However, the chemical microenvironment of biolm
transport channels must be elucidated prior to the development of new chemical treatments, to inform their design and delivery.
Chemical gradients in biolms are well characterized and play a key role in nutrient cycling, waste product management, secre-
tion of secondary metabolites and signalling compounds and oxygen, redox and pH homeostasis [5–9]. e oxygen prole of
OPEN
ACCESS

2
Bottura etal., Microbiology 2025;171:001543
mature biolms is perhaps the most important of these when designing new combative biolm treatments. ere are three key
reasons for this: the potential for free and reactive oxygen species (ROS) to degrade the molecular structure of antimicrobial
agents [10, 11], the synergistic role that free oxygen and ROS may play in potentiating biolm destruction when combined with
antimicrobial treatments [12, 13] and the emergence of antimicrobial tolerance fuelled by oxygen depletion in mature biolms
[14–17]. erefore, an understanding of the oxygen gradients and microenvironment of biolm transport channels is important
as it will inform the development of future mitigation strategies that may seek to exploit such channels as drug delivery pipelines.
is study aims to elucidate the oxygen gradients and microenvironment of E. coli biolm transport channels using a multimodal
approach to oxygen sensing (Fig.1). We present three complementary sensing approaches to probe the oxygen prole of these
structures, and we conclude with a potential structural mechanism for maintaining the homeostasis of oxygen gradients within
the channel networks. Based on our previous work, we initially hypothesized that the channel structures formed akin to chasms,
open to the air, permitting free diusion of atmospheric oxygen into the biolm channels. However, this study determined that
the microenvironment of biolm transport channels was depleted of oxygen, with a similar oxygen prole to the biolm cell mass
and maintained by a thin apical layer of cells at the crest of each channel structure. Bioremediation strategies can now be developed
that capitalize on these observations and our robust overview of the oxygen gradients throughout these channel networks.
METHODS
Strains and growth conditions
All experiments were performed using the E. coli strains outlined in Table S1 (available in the online Supplementary Material);
uorescent reporters were used during imaging as a proxy for live cells and to generate sucient contrast to resolve biolm
structures. Macrocolony biolms were grown by spreading 100 µl of a 1×104 c.f.u. ml−1 inoculum on solid lysogeny broth (LB)
(Miller) medium (10 g l−1 tryptone, 5 g l−1 yeast extract, 10 g l−1 sodium chloride and 20 g l−1 agar supplemented for solid media),
supplemented where appropriate with the required selective antibiotic. Biolms were grown at 37 °C in darkened conditions
for 18–24 h before imaging. Planktonic cultures were maintained using LB (Miller) broth and incubated at 37 °C while shaking
continuously at 225 r.p.m.
Specimen preparation for imaging
Macrocolony biolms were prepared for imaging by inoculating lawns of dilute liquid culture, as above, onto moulds lled with
solid growth media. 3D- printed chamber moulds were used for imaging applications, as described elsewhere [1, 3]. To ensure
exposure to atmospheric conditions, biolms were imaged under air immersion throughout this study; imaging of oxygen
nanosensor- supplemented biolms was performed in air immersion using an inverted confocal laser scanning microscope, as
described later, and oxygen biosensor imaging was performed using the biolm specimen preparation methods described by
Baxter et al. [18]. All experiments were conducted in triplicate.
Oxygen nanosensing in biofilms
Oxygen nanosensing was conducted using proprietary oxygen- sensing nanoparticles, OXNANO (PyroScience GmbH, Germany).
Biolms were prepared for nanoparticle uptake as previously described [1]. Briey, a suspension of mid- log phase E. coli JM105
miniTn7::HcRed1 at a density of 1×104 c.f.u. ml−1 was prepared and supplemented with a nal concentration of 10 µg ml−1 OXNANO
particles. A 100 µl aliquot of the nanoparticle culture suspension was inoculated to form discrete microcolony biolms as described
above.
Following growth, specimens were imaged using a Leica SP5 confocal laser scanning microscope (Leica, Germany). Fluorescence
excitation for HcRed1 and OXNANO particles was provided via the 543 nm and 633 nm lines of a helium–neon laser, respectively.
Fluorescence emission was detected simultaneously using two photomultiplier tubes (PMT) with spectral detection set from 650
to 690 nm for HcRed1 and 740–780 nm for OXNANO particles. Images were acquired using a 10 ×/0.4NA objective lens (Leica,
Germany).
Nanoparticle calibration datasets were acquired using OXNANO bead preparations on LB agar pads imaged under atmospheric
exposure or sealed in an anoxic ascorbic acid buer. OXNANO bead preparations were created by supplementing a 10 µg ml−1
suspension of beads over a thin LB pad constructed using a 10×10 mm Gene Frame (ermo Fisher Scientic, USA), as described
elsewhere [19]. e beads were le to dry onto the pads and imaged as described above while exposed to the air (i.e. without
a coverglass) or mounted in an anoxic ascorbic acid buer solution (10 mM ascorbic acid, 90 mM NaOH) and sealed with a
coverglass (Fig. S1).
Oxygen profiling in biofilms
A 10 µm- tip oxygen microsensor (OX- 10; Unisense, Denmark) was used to measure the oxygen concentration throughout the
depth of biolm transport channels and the adjacent interstitial cell population. e microsensor electrode was calibrated using

3
Bottura etal., Microbiology 2025;171:001543
Fig. 1. A combinatorial approach to oxygen sensing in established biofilms. (a)Oxygen nanosensing uses fluorescent nanospheres which increase
in fluorescence intensity proportionally to the local oxygen concentration, providing a ratiometric intensity- based readout of oxygen concentration in
complex samples. (b)Oxygen sensing via partial pressure measurements achieved using fine- tipped electrochemical probes permits targeted and
absolute measurements of the oxygen concentration in dierent sub- populations of the biofilm by guiding the sensor tip to dierent regions, with
the option of measuring through biofilm depth using a micromanipulator. (c)Oxygen biosensing provides insight into the local oxygen environment at
the cellular level by directly sensing the cellular response to environmental oxygen via a transcriptional fusion of gfp to the promoter of the terminal
cytochrome c oxidase, cco2.

4
Bottura etal., Microbiology 2025;171:001543
a two- point calibration method. Atmospheric calibrations were conducted using a calibration chamber (CAL300; Unisense,
Denmark) containing water bubbled continuously with air for at least 30 min before reading. Anoxic calibrations were conducted
by submersion of the microsensor tip into an anoxic ascorbic acid buer solution (10 mM ascorbic acid, 90 mM NaOH) and
incubated at room temperature for 30 min to acclimatize before registering the ‘zero’ value. e electrode was then cleaned with
and stored in distilled water until required.
Macrocolony biolms grown for 20 h on LB agar plates were placed on top of an aluminium foil surface so that ambient light was
reected through the base of the biolm channels and rendered them visible using a VHX- 1000 digital microscope (Keyence,
Japan) set on a rotational mount to 45°. e microsensor electrode was mounted on a micromanipulator (MM33; Unisense,
Denmark) to facilitate automated stepped movement of the electrode through the depth of the biolm. Proles were recorded
using a multimeter and the SensorTrace Proling soware (Unisense, Denmark), as described by Jo et al. [20]. is setup
provided means to measure the oxygen concentration through the depth of a given transport channel by incrementally stepping
the electrode from the apical to the basolateral surface of the biolm. Readings were acquired with a measurement time of 3 s
and a dwell time between measurements of 5 s. A step size of 5 µm was used to measure the entire depth of the biolm (typically
ranging from 100 to 150 µm). Both the channel and interstitial regions were measured in triplicate for each strain, using three
replicate biolms with a single channel and interstitial measurement point each.
Oxygen biosensing in biofilms
e oxygen reporter plasmid, pAW9 (Table S1) [21, 22], was transformed into JM105 via electroporation and maintained during
all subsequent experiments using 100 µg ml−1 ampicillin. Macrocolony biolms were grown as described for imaging above and
imaged using the Mesolens in wideeld epiuorescence mode, as described previously [1, 3]. Briey, a 490 nm LED (pE- 4000;
CoolLED, UK) provided an excitation source for GFP expressed under anoxic conditions. Fluorescence emission was captured
using the Mesolens coupled to a VNP- 29MC CCD camera with a chip- shiing modality (Vieworks, South Korea) with a triple-
bandpass lter (540 ± 10 nm) placed in the detection pathway. Mesoscopic imaging was conducted using water immersion (n=1.33)
with the correction collars set accordingly to minimize spherical aberration via refractive index mismatch.
Paran-embedding and biofilm thin sectioning
in sections of macrocolony were generated and imaged by wideeld epiuorescence microscopy to provide a sagittal view of
biolm transport channels (Fig. S2). Biolms were prepared according to the methods described by Cornell et al. [23] and Smiley et
al. [24]. Briey, JM105 miniTn7::gfp biolms were grown as described above and prepared for processing by embedding in cooled
molten 1% (w/v) agarose. Fieen millilitres of 1% agarose were delicately poured over the surface of the Petri dish containing
macrocolony biolms. Aer setting, each was excised with a clear margin surrounding the biolm as a cube of solid agar base
and agarose top. Specimens were placed into histocassettes and xed overnight in 4% (w/v) paraformaldehyde in PBS (pH 7.2)
before washing in increasing ethanol concentrations (25%, 50%, 70% and 95% in PBS and three times in 100% ethanol). e
biolms were then cleared with three washes of Histo- Clear II (National Diagnostics, USA), inltrated with molten paran wax
(Leica Biosystems, Germany), and le to solidify for 3 h. Biolms were then mounted and sectioned into 10 µm slices using an
automatic microtome (905200ER; ermo Fisher Scientic, USA). Specimens were mounted on clean glass slides and air dried
overnight before heat xing at 45 °C for 30 min. e slides were rehydrated in PBS by decreasing through the ethanol gradient
above and mounted in ProLong™ Diamond Antifade Mountant (Invitrogen, USA) before being sealed with type 1.5 coverglasses.
Specimens were imaged using an inverted IX81 microscope coupled to a FluoView FV1000 confocal laser scanning microscope
(Olympus, Japan). Fluorescence excitation for GFP was provided by a 488- nm argon laser (GLG3135; Showa Optronics, Japan).
Fluorescence emission was detected using a PMT with spectral detection set from 510 to 550 nm. Images were acquired using a
10 ×/0.4NA objective lens (Olympus, Japan).
Image analysis
All image processing and analyses were conducted using FIJI v1.54f [25]. Images were false- coloured and linearly contrast- adjusted
post- analysis for display purposes where appropriate.
For oxygen nanosensing, images were pre- processed by performing a polar transform function [26] to transform from polar to
Cartesian coordinates and simplify the analysis of radially projecting channel structures. e coordinates of the centre of the
biolm were calculated using the Measure function in FIJI, and the number of pixels per angular coordinates was set to 4096 pixels
(11 pixels per degree). Converting to a Cartesian projection of the image facilitated the selective measurement of exterior and
internalized nanosensor populations. A region of interest was cropped from the exterior of the biolm and the internal region,
and a median lter (σ=5 pixels) was used to reduce noise contributions from the analysis with minimal impact on the signal
intensity. e nanosensing beads were thresholded using automatic yen parameters, and the uorescence emission intensity of
the OXNANO sensors was measured using the Analyse Particles function to generate the mean uorescence intensity of each
nanosensor particle.

5
Bottura etal., Microbiology 2025;171:001543
Statistical analysis
Statistical analyses were conducted using Prism v8.0.2 (GraphPad, USA). Following a normality test, the intensity of
OXNANO particles between the internal and exterior populations was compared using a two- tailed Mann–Whitney test. e
emission intensity data were normalized ranging between the maximum and minimum intensity of the thresholded beads
(
x
′
=
x−x
min.
xmax.−xmin.
). e OXNANO calibration data were also compared using a two- tailed Mann–Whitney test. Prolometry data
were tted with an exponential one- phase decay function and prediction intervals were plotted using Prism.
RESULTS
Oxygen nanosensing reveals the oxygen gradients along biofilm transport channels
Biolm transport channels were observed to transport oxygen- sensing nanoparticles, thereby providing a means to quantify the
oxygen microenvironment within channels in comparison to the external surroundings. Fig.2 shows the uptake and modulation
of the uorescence emission intensity of OXNANO sensors throughout a HcRed1- expressing JM105 biolm. Biolms exhibited
the classical microsphere uptake behaviour previously reported [1], with OXNANO sensors transported towards the core of the
biolm via the channel structures. A polar transform facilitated the projection of radially propagating channels as linear structures
with a clear boundary at the periphery of the biolm (Fig.2a). Internalized nanosensors exhibited an 86% decrease in the median
relative uorescence intensity compared to the external population exposed to atmospheric oxygen concentrations, indicating
that the oxygen concentration was being attenuated along the length of the transport channels. While these data conrmed the
change in the oxygen microenvironment, they do not provide absolute quantication of the oxygen concentrations within these
emergent structures.
Calibration of the OXNANO beads was attempted by imaging under the same acquisition parameters as biolm specimens
but using lawns of beads either exposed to the atmosphere or mounted and sealed in an anoxic ascorbic acid buer (Fig. S1).
However, the calibration data did not provide a sucient dierence in the emission intensity between the two calibration points
and, therefore, positioned OXNANO data as purely ratiometric when the uorescence emission signal is captured using a confocal
microscope, as opposed to the usual detection method which routinely detects phosphorescence emission using proprietary
metres.
Quantitative electrochemical oxygen sensing provides a direct readout of the oxygen gradients throughout
biofilm transport channels
Electrochemical oxygen proling was used to provide a high- resolution readout of the oxygen levels over the depth of biolm
transport channels and compared to the adjacent interstitial cell population. Fig.3a illustrates the precise guidance of a 10
µm- tip oxygen sensing electrode directly into the apical surface of a transport channel and the adjacent cell population. e
atmospheric oxygen concentration was measured as 266.3 µmol l−1 at a temperature of 23.3°C during calibration. Fig.3b shows
Fig. 2. Oxygen nanosensing reveals the oxygen gradients along biofilm transport channels. (a)A maximum intensity projection of an E. coli biofilm
(green), showing dark transport channel networks, merged with fluorescent oxygen- sensing nanoparticles (cyan). Comparative intensity analysis
between the exterior and interior nanosensors was conducted by performing a polar transform operation and segmenting nanoparticles from each
region for analysis. Grey arrows indicate the relative position of channel structures filled with nanosensors between the input image and the polar
transform output. (b)Nanosensors in biofilm channels exhibited an 86% reduction in median fluorescence intensity compared to exterior nanosensors,
indicating a decrease in oxygen concentration along the channel network towards the core of the biofilm (****P<0.0001) (NExternal=270, NInternal=786;
acquired over three biological replicates).

6
Bottura etal., Microbiology 2025;171:001543
the oxygen concentration through the depth of the biolm transport channels for WT, gfp and HcRed1- expressing biolms. e
oxygen concentration decreased immediately under the surface of all biolms, reaching 0.3 µmol l−1 within 20 µm of the apical
surface. e anoxic region persisted until the basal layers of the biolm, where partial recovery in the oxygen concentration
was observed in all strains. However, the channels consistently exhibited a higher recovery in the oxygen concentration in these
regions compared to the interstitial cell populations. e WT strain exhibited routinely lower oxygen recovery levels compared
to the uorescent strains.
Oxygen biosensing concurs with nanosensing and electrochemical sensing to provide a robust overview of
biofilm transport channel oxygen microenvironments
Fig.4 shows a maximum- intensity projection of a biolm expressing a uorescent Pcco2::gfp oxygen reporter that conditionally
expresses gfp under anoxic conditions with high delity. ese data show an anoxic core accompanied by diuse uorescence
Fig. 3. Oxygen concentrations decrease through biofilm depth and dier at the biofilm base. (a)Images showing the insertion of an oxygen microelectrode
probe into a biofilm transport channel (left) and biofilm cell population (right). Measurement traces of the oxygen profile through the depth of the
biofilm in channel and cell populations are presented for the (b)non- fluorescent WT, (c)GFP and (d) HcRed1- expressing strains used in this study.
Each plot shows the oxygen concentration for three replicate biofilms at each measurement position in both the channel and the adjacent biomass.
The data are fitted with a one- phase decay and plotted with prediction intervals. Both channel and cell populations in all three strains exhibited a
decrease in oxygen concentration from atmospheric oxygen (266 µmol l−1) to 0.3 µmol l−1 within 20 µm from the biofilm surface. Transport channels
exhibited higher oxygen concentrations at depth compared to the adjacent cell populations. A cropped inset is presented showing oxygen recovery at
the basal portion of each population. Measurements were obtained from three replicate biofilms for each condition, with a single channel and biofilm
cell population sampled in each.

7
Bottura etal., Microbiology 2025;171:001543
emission along the transport channel networks with the highest intensity at the centre of the biolm and a gradual decrease
towards the leading edge, interspersed with anoxic subpopulations matching the spatial pattern of canonical biolm transport
channels reported elsewhere [1, 3, 4, 9]. e spatial overview of oxygen gradients from these data concurs with the nanosensing
and electrochemical proling data, which show a change in oxygen gradients along the channel structures and similarities between
the microenvironments of the channels and interstitial cell populations, respectively.
Thin sectioning suggests a mechanism for maintaining oxygen and nutrient gradients in biofilm transport
channels
Our data show that the overall oxygen concentration does not signicantly dier between the lumen of the transport channels
and the biolm cell population. is result confounded the originally proposed structure of biolm transport channels, which
assumed that they were open to the environment, thereby permitting free diusion of atmospheric oxygen into the lumen of the
channels. However, by imaging thin sections of macrocolony biolms, we have determined that the apical surface of the biolm
remains sealed by a thin layer of capping cells at all points (Fig.5a). Even at positions where transport channels are observed,
they are not open to the atmosphere as rst presumed from lateral viewpoints. We propose that this layer of actively respiring
cells and matrix components is sucient to maintain the oxygen gradients we observe via nanosensing, electrochemical sensing
and biosensing methods. Moreover, we observed that the inferior portions of biolm transport channels are oen open to the
underlying substrate. is observation supports the nutrient transport ndings documented in other studies [1, 3] and explains
the increase in oxygen recovery in biolm channels revealed from our electrochemical sensing data. Fig.5b provides a diagram-
matic representation summarizing a working model of how biolm transport channels maintain a marked oxygen gradient,
partial recovery from the diusion of free oxygen and acquisition of nutrients from their underlying grown medium substrate.
Fig. 4. Oxygen biosensing confirms the oxygen profile of biofilm transport channels. A maximum intensity projection of a mature E. coli biofilm
expressing a cco2 promoter- gfp fusion (Pcco2- gfp). The emission intensity of GFP indicates levels of anoxia throughout the biofilm, commensurate with
a dense hypoxic core. The diuse signal from interstitial cell populations surrounding the channel structures can be observed permeating radially
from the centre. A schematic x, z cross- section illustrates the anoxic gradients shown by the oxygen sensing, which concur with nanosensing and
electrochemical sensing methods.

8
Bottura etal., Microbiology 2025;171:001543
DISCUSSION
e oxygen prole and chemical microenvironment of biolm transport channels have remained unexplored since their discovery.
is study provides three complementary approaches to quantify the oxygen microenvironment of biolm channels to provide
insight for follow- on studies that might exploit these structures for bioremediation purposes. We used a combination of oxygen
nanosensing, electrochemical sensing and biosensing to provide a robust measurement of oxygen gradients throughout biolm
channel networks. ese data show that, while oxygen gradients change along the length of biolm transport channels, there is a
sharp oxygen gradient along their depth. Images of biolm thin sections conrm that biolm transport channels are capped by a
thin layer of cells covering the apical surface of the channel structures, while they remain open to the underlying growth substrate.
is observation provides a potential mechanism for maintaining such chemical gradients within biolm transport channels.
Key resource tracking by biolm transport channels has been well documented in recent literature [1, 3, 4, 18], but it remained
unclear if these seemingly devoid structures were either hypaethral or enclosed; thereby, the chemical microenvironment was
unknown. e channel structures we discuss here were originally hypothesized to be open chasms spreading along the radii of
macrocolony biolms due to their appearance as dark voids during optical imaging experiments. However, this study concluded
that thin layers of surface cells seal and maintain the chemical gradients within the biolm channels, unlike other larger scale
water channels reported in mushroom- shaped Pseudomonas aeruginosa [27, 28], Klebsiella [29] and Bacillus spp. [30, 31] biolms.
Previous studies have used nanosensors [32], electrochemical proling [20, 24, 33] and P
cco2
- mediated biosensing [20, 21] to
study the oxygen gradients and redox prole of P. aeruginosa biolms. While these studies have provided robust insights into
the chemical microenvironment within the bulk mass of pseudomonad biolms, the void structures of E. coli transport channels
presented a priority for this study, which would inform if the local conditions might impact future channel- exploiting treatments.
Previous optical oxygen nanosensing techniques have been achieved in planktonic cultures [34], small Pseudomonas spp.
aggregates [35, 36] and antimicrobial susceptibility screens [37]. Flamholz et al. recently documented the common pitfalls
in oxygen nanosensing applications in biofilms, where the constituent cells can produce quenching molecules that may
impact the reliability of optical oxygen quantifications [32]. Moreover, we present a new detection method for OXNANO
sensors, using a confocal laser scanning microscope to detect the fluorescence emission of the nanoparticles, rather than
the phosphorescence emission that other studies detect either using proprietary phosphorescence metres or measuring the
extremely long phosphorescence lifetimes. Our calibration data (Fig. S1) show the opposite trend of intensity versus oxygen
concentration owing to the measurement of fluorescence emission rather than phosphorescence emission. This observation
Fig. 5. Orthogonal visualization presents an explanation for chemical gradients in biofilm transport channels. (a) A thin section of a fixed gfp- expressing
E. coli JM105 biofilm is presented with arrows at the apical surface indicating regions where transport channels can be viewed sagittally. Thin layers of
cells cap the apical surface of the biofilm and enclose the top of transport channels. The inferior portion of the channels remains open to the underlying
substrate. A representative view of the section position through the centre of the colony is presented (top left). (b) A schematic proposing nutrient and
oxygen gradients through the depth of biofilm transport channels. A thin layer of apical cells seals the top of channels and maintains a strict oxygen
gradient, while nutrients and free oxygen can diuse from the basolateral surface through open channel conduits into the biomass.

9
Bottura etal., Microbiology 2025;171:001543
is explained through competition in the electronic pathway, where a decrease in phosphorescence increases fluorescence
owing to more excited- state electrons returning to the ground state via the singlet pathway as opposed to undergoing
intersystem crossing and returning from the triplet state. The platinum- based chemistry of the OXNANO probes promotes
intersystem crossing and triplet decay pathway, hence why most users measure phosphorescence quenching as a function
of oxygen concentration. However, the routinely used phosphorescence metres are best suited to studying specimens in
solution and are incapable of sub- sampling different regions of complex specimens, such as biofilms. Our study shows that
the fluorescence emission can be imaged to provide a spatial overview and proportional intensity readout to measure oxygen
gradients in complex specimens. Moreover, our calibration data and the decrease in oxygen concentration we observe by
measuring fluorescence emission are corroborated by the electrochemical measurements also performed in this study. We
provided a spatial overview of the relative change in nanosensor intensity as a function of oxygen depletion along biofilm
transport channels and supplemented this ratiometric data with high- resolution electrochemical sensing experiments.
Other studies have also used optical means to study the chemical microenvironments of microbial communities, such as
by transient state monitoring coupled with single plane illumination microscopy [38], ratiometric oxygen- responsive dye
measurements [39] or planar optodes or in situ imaging probes [40]. Each of these techniques is limited either by their
spatial resolution, attainable field of view and low throughput or by inhomogeneous illumination, leading to spurious
quantification [41]. Oxygen nanosensing coupled with conventional low- power confocal microscopy provides some means
of circumventing these challenges, although the fluorescent nanosensor approach requires calibration and supplemen-
tary methods to provide reliable quantitative measurements [32]. Overall, this approach requires complementation with
secondary methods to verify the nanosensing outputs.
Electrochemical sensing of oxygen gradients using Clark- type sensors has been well documented in microbial communities with
many examples [42] spanning environmental microbiology [43, 44], model systems [45], pathogen biology [15, 46] and wastewater
management [47]. e accuracy of this sensing method, both in terms of precision targeting of biolm sub- populations using
the micron- scale sensor tip and the sensitivity and dynamic range, provides a robust quantitative method for complex microbial
communities [48]. e complete oxygen proles of E. coli biolm transport channels we generated concluded that, although the
oxygen environment does change through the depth of the transport channels, it does not dier greatly from the prole of the
surrounding biolm mass. is is important to understand due to the leading role that free molecular oxygen and ROS can play
in the destruction of antimicrobial agents [10, 11], in turn, leading to diminished treatment ecacy. Moreover, understanding
the chemical microenvironment is important for developing ROS- mediated antimicrobial combination therapies [12, 49, 50]
and understanding the impact of anoxia- induced antimicrobial tolerance [51–53]. It is unclear why the concentration of oxygen
recovers to a greater extent in photoprotein- expressing biolms compared to the WT, but this may be linked to the requirement
for free oxygen during chromophore maturation for both GFP [54, 55] and HcRed [56], or due to the increased thickness observed
in WT biolms compared to their uorescent counterparts.
Biosensing using the Pcco2 reporter system facilitated a direct readout of the oxygen environment in dierent sub- populations of
the biolm. In this system, the cco2 promoter drives the expression of gfp in response to anoxic conditions at the cellular level,
providing higher spatial sensitivity than electrochemical sensing can aord and a simultaneous overview of the entire biolm using
the Mesolens. Previous studies have shown the high delity of this biosensor in P. aeruginosa biolms [20, 21] and demonstrated
that homologous transcription factors drive the anoxic response in P. aeruginosa and E. coli [22] priming the translation of the
biosensor for visualizing oxygen gradients in E. coli biolms. Our data revealed the canonical anoxic core of the macrocolony
biolms that other studies have reported [5, 7, 8, 57–60] but also show diuse signal emitted from channel- lining cells throughout
the biolm, thereby complementing our nanosensing and electrochemical sensing experiments.
Finally, the orthogonal visualization of transport channels using thinly sectioned biolm specimens revealed a thin crowning
layer of cells, where previously the channels were hypothesized to be open to the atmosphere. Previous studies have determined
that the generation of biolm oxygen gradients is not caused by physical barriers (i.e., cells, matrix, etc.), but instead penetration
is limited by the active respiration of apical cell layers [60–62]. We therefore propose that the thin layer of cells that we observe is
actively contributing to the underlying chemical microenvironment of the channel structures, maintaining the anoxic environment
that adjacent fermentative cell populations [6, 63–65] are also exposed to.
The combined outcome from the multimodal oxygen sensing presented in this study provides a robust overview of the
oxygen gradients and microenvironments of E. coli biofilm transport channels. We conclude that the combinatorial approach
taken to quantify biofilm chemical gradients is essential to build a full picture of microenvironments in complex living
systems. While the Clark- type electrochemical sensors offer the highest accuracy, their expense and requirements for
specialist equipment may limit accessibility to these methods. In turn, combining high- resolution oxygen profilometry
with biosensing and fluorescence nanosensing not only provides a global overview of oxygen dynamics but also provides
more accessible research tools to study these conditions in biofilms. This work primes future studies exploring the basic
physiology and translational potential of biofilm transport channels for new mitigation strategies. Our findings will provide
a priori knowledge for future studies seeking to develop channel- targeting therapeutics and exploitative drug delivery
strategies. Armed with quantitative evidence of the oxygen microenvironment, researchers could next determine, for

10
Bottura etal., Microbiology 2025;171:001543
example, the encapsulation or chemical modifications required to successfully deliver antimicrobials into dense biofilms
via biofilm transport channels.
Funding information
Funding was provided by the University of Strathclyde, the Medical Research Council (MR/K015583/1), the Biotechnology and Biological Sciences
Research Council (BB/P02565X/1, BB/V019643/1 and BB/T011602/1), the Microbiology Society, the Royal Microscopical Society, the Royal Academy of
Engineering Research (RCSRF2021\11\15), the National Institutes of Health and the National Institute of Allergy and Infectious Diseases (R01AI103369),
the Leverhulme Trust and the Scottish Universities Life Science Alliance.
Acknowledgements
The authors wish to thank Prof. Marvin Whiteley (Georgia Institute of Technology, GA, USA) for the kind gift of the pAW9 plasmid. We acknowledge
funding provided by the University of Strathclyde, the Medical Research Council (MR/K015583/1), the Biotechnology and Biological Sciences Research
Council (BB/P02565X/1, BB/V019643/1 and BB/T011602/1), the Microbiology Society, the Royal Microscopical Society, the Royal Academy of Engi-
neering Research (RCSRF2021\11\15), the National Institutes of Health and the National Institute of Allergy and Infectious Diseases (R01AI103369),
the Leverhulme Trust and the Scottish Universities Life Science Alliance. Figs 1, S1 and S2 were prepared using BioRender. com (licence number:
VF272F7XHX).
Conflicts of interest
The authors declare that there are no conflicts of interest.
References
1. Rooney LM, Amos WB, Hoskisson PA, McConnell G. Intra- colony
channels in E. coli function as a nutrient uptake system. ISME J
2020;14:2461–2473.
2. Rooney LM, Bottura B, Baxter K, Amos WB, Hoskisson PA, etal.
Addressing multiscale microbial challenges using the Mesolens. J
Microsc 2024;296:139–144.
3. Bottura B, Rooney LM, Hoskisson PA, McConnell G. Intra-
colony channel morphology in Escherichia coli biofilms is
governed by nutrient availability and substrate stiness. Biofilm
2022;4:100084.
4. Bottura B, Rooney L, Feeney M, Hoskisson PA, McConnell G.
Fractal complexity of Escherichia coli nutrient transport channels
is influenced by cell shape and growth environment. Microbiology
2023.
5. Stewart PS, Franklin MJ. Physiological heterogeneity in biofilms.
Nat Rev Microbiol 2008;6:199–210.
6. Jo J, Price- Whelan A, Dietrich LEP. Gradients and consequences of
heterogeneity in biofilms. Nat Rev Microbiol 2022;20:593–607.
7. Rani SA, Pitts B, Beyenal H, Veluchamy RA, Lewandowski Z, etal.
Spatial patterns of DNA replication, protein synthesis, and oxygen
concentration within bacterial biofilms reveal diverse physiolog-
ical states. J Bacteriol 2007;189:4223–4233.
8. Díaz- Pascual F, Lempp M, Nosho K, Jeckel H, Jo JK, et al. Spatial
alanine metabolism determines local growth dynamics of Escheri-
chia coli colonies. elife 2021;10:e70794.
9. Dayton H, Kiss J, Wei M, Chauhan S, LaMarre E, et al. Cellular
arrangement impacts metabolic activity and antibiotic tolerance in
Pseudomonas aeruginosa biofilms. PLoS Biol 2024;22:e3002205.
10. Pan M, Chu LM. Adsorption and degradation of five selected antibi-
otics in agricultural soil. Sci Total Environ 2016;545–546:48–56.
11. Agarwal P, Singh N, Farooqui A. Impact of antibiotics on agricul-
tural microbiome: emergence of antibiotic resistance bacteria. In:
Degradation of Antibiotics and Antibiotic- Resistant Bacteria from
Various Sources 231–246. Elsevier, 2023.
12. Vatansever F, de Melo WCMA, Avci P, Vecchio D, Sadasivam M, etal.
Antimicrobial strategies centered around reactive oxygen species-
-bactericidal antibiotics, photodynamic therapy, and beyond. FEMS
Microbiol Rev 2013;37:955–989.
13. Van Acker H, Coenye T. The role of reactive oxygen species
in antibiotic- mediated killing of bacteria. Trends Microbiol
2017;25:456–466.
14. Borriello G, Werner E, Roe F, Kim AM, Ehrlich GD, etal. Oxygen limi-
tation contributes to antibiotic tolerance of Pseudomonas aerugi-
nosa in biofilms. Antimicrob Agents Chemother 2004;48:2659–2664.
15. Kowalski CH, Morelli KA, Schultz D, Nadell CD,
Cramer RA. Fungal biofilm architecture produces hypoxic micro-
environments that drive antifungal resistance. Proc Natl Acad Sci
USA 2020;117:22473–22483.
16. Beebout CJ, Sominsky LA, Eberly AR, Van Horn GT,
Hadjifrangiskou M. Cytochrome bd promotes Escherichia coli biofilm
antibiotic tolerance by regulating accumulation of noxious chemi-
cals. NPJ Biofilms Microbiomes 2021;7:35.
17. Ciofu O, Moser C, Jensen PØ, Høiby N. Tolerance and resistance of
microbial biofilms. Nat Rev Microbiol 2022;20:621–635.
18. Baxter KJ, Sargison FA, Fitzgerald JR, McConnell G,
Hoskisson PA. Time- lapse mesoscopy of Candida albicans and
Staphylococcus aureus dual- species biofilms reveals a structural
role for the hyphae of C. albicans in biofilm formation. Microbi-
ology 2024;170.
19. de Jong IG, Beilharz K, Kuipers OP, Veening J- W. Live cell imaging
of Bacillus subtilis and Streptococcus pneumoniae using automated
time- lapse microscopy. J Vis Exp 2011;3145:3145.
20. Jo J, Cortez KL, Cornell WC, Price- Whelan A, Dietrich LE. An orphan
cbb3- type cytochrome oxidase subunit supports Pseudomonas
aeruginosa biofilm growth and virulence. 2017;30.
21. Wessel AK, Arshad TA, Fitzpatrick M, Connell JL,
Bonnecaze RT, et al. Oxygen limitation within a bacterial aggre-
gate. mBio 2014;5:e00992.
22. Winteler HV, Haas D. The homologous regulators ANR of Pseu-
domonas aeruginosa and FNR of Escherichia coli have overlapping
but distinct specificities for anaerobically inducible promoters.
Microbiology 1996;142 (Pt 3):685–693.
23. Cornell WC, Morgan CJ, Koyama L, Sakhtah H, Mansfield JH,
etal. Paran embedding and thin sectioning of microbial colony
biofilms for microscopic analysis. J Vis Exp 2018;57196:57196.
24. Smiley MK, Sekaran DC, Forouhar F, Wolin E, Jovanovic M, etal.
MpaR- driven expression of an orphan terminal oxidase subunit
supports Pseudomonas aeruginosa biofilm respiration and devel-
opment during cyanogenesis. mBio 2024;15:e0292623.
25. Schindelin J, Arganda- Carreras I, Frise E, Kaynig V, Longair M,
et al. Fiji: an open- source platform for biological- image analysis.
Nat Methods 2012;9:676–682.
26. Donnelly E, Mothe F. Polar transformer; 2007. https://imagej.net/
ij/plugins/polar-transformer.html
27. Drury WJ, Characklis WG, Stewart PS. Interactions of 1 µm latex
particles with Pseudomonas aeruginosa biofilms. Water Res
1993;27:1119–1126.
28. Kempes CP, Okegbe C, Mears- Clarke Z, Follows MJ, Dietrich LEP.
Morphological optimization for access to dual oxidants in biofilms.
Proc Natl Acad Sci USA 2014;111:208–213.

11
Bottura etal., Microbiology 2025;171:001543
29. Stoodley P, Debeer D, Lewandowski Z. Liquid flow in biofilm
systems. Appl Environ Microbiol 1994;60:2711–2716.
30. Asally M, Kittisopikul M, Rué P, Du Y, Hu Z, et al. Localized cell
death focuses mechanical forces during 3D patterning in a biofilm.
Proc Natl Acad Sci USA 2012;109:18891–18896.
31. Wilking JN, Zaburdaev V, De Volder M, Losick R, Brenner MP, etal.
Liquid transport facilitated by channels in Bacillus subtilis biofilms.
Proc Natl Acad Sci USA 2013;110:848–852.
32. Flamholz AI, Saccomano S, Cash K, Newman DK. Optical O2 sensors
also respond to redox active molecules commonly secreted by
bacteria. mBio 2022;13:e0207622.
33. Schiessl KT, Hu F, Jo J, Nazia SZ, Wang B, etal. Phenazine produc-
tion promotes antibiotic tolerance and metabolic heterogeneity in
Pseudomonas aeruginosa biofilms. Nat Commun 2019;10:762.
34. Ladner T, Flitsch D, Schlepütz T, Büchs J. Online monitoring of
dissolved oxygen tension in microtiter plates based on infrared
fluorescent oxygen- sensitive nanoparticles. Microb Cell Fact
2015;14:161.
35. Jewell MP, Galyean AA, Kirk Harris J, Zemanick ET,
Cash KJ. Luminescent nanosensors for ratiometric monitoring
of three- dimensional oxygen gradients in laboratory and clin-
ical Pseudomonas aeruginosa biofilms. Appl Environ Microbiol
2019;85:e01116- 19.
36. Smriga S, Ciccarese D, Babbin AR. Denitrifying bacteria respond
to and shape microscale gradients within particulate matrices.
Commun Biol 2021;4:570.
37. Jusková P, Schmitt S, Kling A, Rackus DG, Held M, etal. Real- time
respiration changes as a viability indicator for rapid antibiotic
susceptibility testing in a microfluidic chamber array. ACS Sens
2021;6:2202–2210.
38. Karampatzakis A, Sankaran J, Kandaswamy K, Rice SA, Cohen Y,
etal. Measurement of oxygen concentrations in bacterial biofilms
using transient state monitoring by single plane illumination
microscopy. Biomed Phys Eng Express 2017;3:035020.
39. Larsen M, Borisov SM, Grunwald B, Klimant I, Glud RN. A simple
and inexpensive high resolution color ratiometric planar optode
imaging approach: application to oxygen and pH sensing. Limnol
Oceanogr Methods 2011;9:348–360.
40. Kühl M, Rickelt LF, Thar R. Combined imaging of bacteria and
oxygen in biofilms. Appl Environ Microbiol 2007;73:6289–6295.
41. Sankaran J, Karampatzakis A, Rice SA, Wohland T. Quantitative
imaging and spectroscopic technologies for microbiology. FEMS
Microbiol Lett 2018;365.
42. Saccomano SC, Jewell MP, Cash KJ. A review of chemosensors
and biosensors for monitoring biofilm dynamics. Sens Actuators
Rep 2021;3:100043.
43. Sørensen K, Řeháková K, Zapomělová E, Oren A. Distribution of
benthic phototrophs, sulfate reducers, and methanogens in two
adjacent saltern evaporation ponds in Eilat, Israel. Aquat Microb
Ecol 2009;56:275–284.
44. Hicks N, Ubbara GR, Silburn B, Smith HEK, Kröger S, etal. Oxygen
dynamics in shelf seas sediments incorporating seasonal vari-
ability. Biogeochemistry 2017;135:35–47.
45. Guimerà X, Moya A, Dorado AD, Illa X, Villa R, etal. A minimally inva-
sive microsensor specially designed for simultaneous dissolved
oxygen and pH biofilm profiling. Sensors 2019;19:4747.
46. Kiamco MM, Atci E, Mohamed A, Call DR, Beyenal H. Hyperosmotic
agents and antibiotics aect dissolved oxygen and pH concen-
tration gradients in Staphylococcus aureus biofilms. Appl Environ
Microbiol 2017;83:e02783- 16.
47. Qi C, Zhou Y, Suenaga T, Oba K, Lu J, etal. Organic carbon determines
nitrous oxide consumption activity of clade I and II nosZ bacteria:
genomic and biokinetic insights. Water Res 2022;209:117910.
48. Revsbech NP. Simple sensors that work in diverse natural envi-
ronments: the micro- Clark sensor and biosensor family. Sens
Actuators B Chem 2021;329:129168.
49. Akbari MZ, Xu Y, Lu Z, Peng L. Review of antibiotics treatment by
advance oxidation processes. Environ Adv 2021;5:100111.
50. Qi W, Jonker MJ, de Leeuw W, Brul S, ter Kuile BH. Reactive oxygen
species accelerate de novo acquisition of antibiotic resistance in E.
coli. iScience 2023;26:108373.
51. Schaible B, Taylor CT, Schaer K. Hypoxia increases antibi-
otic resistance in Pseudomonas aeruginosa through altering the
composition of multidrug eux pumps. Antimicrob Agents Chem-
other 2012;56:2114–2118.
52. Jensen PØ, Kolpen M, Kragh KN, Kühl M. Microenvironmental
characteristics and physiology of biofilms in chronic infections of
CF patients are strongly aected by the host immune response.
APMIS 2017;125:276–288.
53. Olsen I. Biofilm- specific antibiotic tolerance and resistance. Eur J
Clin Microbiol Infect Dis 2015;34:877–886.
54. Reid BG, Flynn GC. Chromophore formation in green fluorescent
protein. Biochemistry 1997;36:6786–6791.
55. Barondeau DP, Putnam CD, Kassmann CJ, Tainer JA, Getzo ED.
Mechanism and energetics of green fluorescent protein chromo-
phore synthesis revealed by trapped intermediate structures. Proc
Natl Acad Sci USA 2003;100:12111–12116.
56. Wilmann PG, Petersen J, Pettikiriarachchi A, Buckle AM, Smith SC,
etal. The 2.1A crystal structure of the far- red fluorescent protein
HcRed: inherent conformational flexibility of the chromophore. J
Mol Biol 2005;349:223–237.
57. de Beer D, Stoodley P, Roe F, Lewandowski Z. Eects of biofilm
structures on oxygen distribution and mass transport. Biotech
Bioeng 1994;43:1131–1138.
58. Depetris A, Tagliavini G, Peter H, Kühl M, Holzner M, etal. Biophys-
ical properties at patch scale shape the metabolism of biofilm
landscapes. NPJ Biofilms Microbiomes 2022;8:5.
59. Zhang Y, Cai Y, Zeng L, Liu P, Ma LZ, etal. A microfluidic approach
for quantitative study of spatial heterogeneity in bacterial biofilms.
Small Sci 2022;2:2200047.
60. Tokunou Y, Toyofuku M, Nomura N. Physiological benefits
of oxygen- terminating extracellular electron transfer. mBio
2022;13:e0195722.
61. Barraud N, Kelso M, Rice S, Kjelleberg S. Nitric oxide: a key medi-
ator of biofilm dispersal with applications in infectious diseases.
CPD 2014;21:31–42.
62. Stewart PS, Zhang T, Xu R, Pitts B, Walters MC, et al. Reaction-
diusion theory explains hypoxia and heterogeneous growth
within microbial biofilms associated with chronic infections. NPJ
Biofilms Microbiomes 2016;2:16012.
63. Serra DO, Klauck G, Hengge R. Vertical stratification of matrix
production is essential for physical integrity and architecture
of macrocolony biofilms of Escherichia coli. Environ Microbiol
2015;17:5073–5088.
64. Létoé S, Chalabaev S, Dugay J, Stressmann F, Audrain B, et al.
Biofilm microenvironment induces a widespread adaptive amino-
acid fermentation pathway conferring strong fitness advantage in
Escherichia coli. PLoS Genet 2017;13:e1006800.
65. Martín- Rodríguez AJ. Respiration- induced biofilm formation
as a driver for bacterial niche colonization. Trends Microbiol
2023;31:120–134.
Edited by: D. JV Beste and J. Connolly