Interaction between CO2-consuming autotrophy and CO2-producing heterotrophy in non-axenic phototrophic biofilms

As the effects of climate change become increasingly evident, the need for effective CO2 management is clear. Microalgae are well-suited for CO2 sequestration, given their ability to rapidly uptake and fix CO2. They also readily assimilate inorganic nutrients and produce a biomass with inherent commercial value, leading to a paradigm in which CO2-sequestration, enhanced wastewater treatment, and biomass generation could be effectively combined. Natural non-axenic phototrophic cultures comprising both autotrophic and heterotrophic fractions are particularly attractive in this endeavour, given their increased robustness and innate O2-CO2 exchange. In this study, the interplay between CO2-consuming autotrophy and CO2-producing heterotrophy in a non-axenic phototrophic biofilm was examined. When the biofilm was cultivated under autotrophic conditions (i.e. no organic carbon), it grew autotrophically and exhibited CO2 uptake. After amending its growth medium with organic carbon (0.25 g/L glucose and 0.28 g/L sodium acetate), the biofilm rapidly toggled from net-autotrophic to net-heterotrophic growth, reaching a CO2 production rate of 60 μmol/h after 31 hours. When the organic carbon sources were provided at a lower concentration (0.125 g/L glucose and 0.14 g/L sodium acetate), the biofilm exhibited distinct, longitudinally discrete regions of heterotrophic and autotrophic metabolism in the proximal and distal halves of the biofilm respectively, within 4 hours of carbon amendment. Interestingly, this upstream and downstream partitioning of heterotrophic and autotrophic metabolism appeared to be reversible, as the position of these regions began to flip once the direction of medium flow (and hence nutrient availability) was reversed. The insight generated here can inform new and important research questions and contribute to efforts aimed at scaling and industrializing algal growth systems, where the ability to understand, predict, and optimize biofilm growth and activity is critical.


Introduction
The accumulation of carbon dioxide in the atmosphere caused by the burning of fossil fuels is widely reported as being a leading cause of climate change. The effects of climate change are far-reaching and evident in the form of increasing global temperatures, melting of glacial ice, phototrophic biofilm. This question is highly relevant to integrated CO 2 sequestration-wastewater treatment systems in which a phototrophic biofilm may be simultaneously exposed to both inorganic and organic carbon sources. That some microalgal species can grow mixotrophically utilizing both inorganic and organic carbon, and a few are even capable of growing heterotrophically on organic carbon only [26], makes the study of autotrophic-heterotrophic interplay in phototrophic biofilms especially prudent. The insight gained through this work may contribute information for optimizing growth system design, operation, and performance.
In this study, phototrophic biofilms were grown in a CO 2 Sequestration Monitoring System (CSMS) [27]. Building upon the CO 2 Evolution Monitoring System (CEMS) described by Kroukamp and Wolfaardt [28], this system enables the real-time in situ monitoring of a phototrophic biofilm's CO 2 flux under varying conditions. This work was motivated by a series of hypotheses: i) that non-axenic phototrophic biofilms readily toggle between net-autotrophic and net-heterotrophic growth depending on the availability of labile organic carbon sources; ii) that in the presence of both CO 2 and labile organic carbon, phototrophic biofilms exhibit distinct, longitudinally discrete regions of heterotrophic and autotrophic growth with distance from reactor inflow; and iii) that this distinct longitudinal separation of autotrophic and heterotrophic metabolism is transient and reversible, based on the flow direction (and hence availability) of organic carbon nutrients.

Growth system
Biofilms were grown in a CO 2 Sequestration Monitoring System (CSMS). This system, described in detail in Ronan et al. [27], enables the real-time, in situ monitoring of CO 2 flux in phototrophic biofilms. Biofilm reactor (BR) modules were 150 cm in length and comprised a tube-within-a-tube design (Fig 1A). Biofilms grew inside a CO 2 permeable silicone tube (1.57 mm ID, 2.41 mm OD, 0.41 mm wall thickness, 2013.2 Barrer permeability; VWR International, Mississauga, ON, Canada), through which liquid growth medium flowed. The silicone tube in each BR provided approximately 74 cm 2 of colonizable surface area and was housed within a larger diameter Tygon™ tube (4.76 mm ID, 7.94 mm OD, 1.58 mm wall thickness, E-3603 formulation; VWR International, Mississauga, ON, Canada). A CO 2 -free sweeper gas (TOC grade, 24001980; Linde Canada Limited, Concord, ON, Canada) was channeled through the annular space created by the two tubes. In this configuration, CO 2 molecules could readily diffuse from the biofilm into the annular space (Fig 1B), or from the annular space into the biofilm (Fig 1C), depending on the nature of the biofilm and the prevailing CO 2 concentration gradient.
In this study, the CSMS consisted of three sequential BRs (Fig 2). The first, BR prod (CO 2producer), received fresh growth medium and housed a pure culture heterotrophic bacterial biofilm. The purpose of this biofilm was to provide a steady source of CO 2 to the phototrophic biofilm being studied in the two subsequent BRs downstream. The sweeper gas passed through the annular space of BR prod , collecting CO 2 produced by the heterotrophic biofilm and carrying it toward a non-dispersive infrared CO 2 analyzer (Analyzer 1) (LI-820; LI-COR Biosciences, Lincoln, NE, USA). The gas stream then travelled through the annular space of the two downstream biofilm reactors BR cons and BR cons2 (CO 2 consumer), via a second CO 2 analyzer (Analyzer 2) positioned between them. Finally, the gas passed through a third CO 2 analyzer (Analyzer 3), placed immediately downstream of BR cons2 . Each CO 2 analyzer has a precision in the range of 1 ppm and was set to log CO 2 concentrations at one-minute intervals.
The linked BR cons -BR cons2 unit received its own feed of fresh growth medium and housed a phototrophic biofilm, which was inoculated by injecting 6 mL of culture into the liquid Biofilm inoculation and growth occurs inside a highly CO 2 -permeable silicone tube, through which liquid growth medium is pumped at a constant flow rate [27]. The comparatively CO 2impermeable Tygon™ tube housing this silicone tube creates an annular space through which the sweeper gas is channeled at a constant flow rate. CO 2 molecules can readily diffuse in either direction across the wall of silicone tube according to the concentration gradient. (B) For a CO 2 -producing (i.e. heterotrophic) biofilm, CO 2 molecules diffuse from the aqueous environment inside the silicone tube into the dry annular space. (C) When CO 2 -laden gas is channeled into a BR containing a CO 2 -consuming (i.e. autotrophic) biofilm, CO 2 molecules are pulled in the opposite direction, from the annular space into the silicone tube.
https://doi.org/10.1371/journal.pone.0253224.g001  [27]. The first, BR prod , received its own feed of fresh growth medium and housed a heterotrophic pure culture bacterial biofilm, providing a consistent source of CO 2 . As CO 2 molecules diffused out of the BR prod silicone tube into the annular space, they were carried downstream by the sweeper gas to a CO 2 analyzer (Analyzer 1). The gas stream then travelled through the annular space of the linked BR cons -BR cons2 via a second CO 2 analyzer (Analyzer 2) positioned between them, before terminating at a third and final CO 2 analyzer (Analyzer 3). BR cons -BR cons2 housed the phototrophic biofilm of interest. The difference in CO 2 concentration logged by the analyzers provided a direct measure of CO 2 flux in the biofilm. Since medium flow was continuous through the linked BR cons -BR cons2 modules, they represent two halves of one biofilm system approximately 300 cm in length.
https://doi.org/10.1371/journal.pone.0253224.g002 influent tube of BR cons using a sterile syringe needle. Medium flow was paused for a period of one hour following inoculation to allow for some initial cell adherence. Given that growth medium flowed continuously through BR cons -BR cons2 , these BRs represent two halves of one biofilm system with a combined length of 300 cm.
The specific placement of the three CO 2 analyzers in the CSMS allowed for the direct measurement of the CO 2 concentration: i. entering the phototrophic BR cons -BR cons2 unit, ii. at the halfway point of BR cons -BR cons2 , and iii. exiting BR cons -BR cons2 . Therefore, at any time point it was possible to determine the CO 2 flux of the entire phototrophic biofilm (Eq 1), or for the proximal and distal halves of the biofilm separately (Eqs 2 and 3).
Combined BR cons À BR cons2 CO 2 flux ¼ Analyzer 3 À Analyzer 1 ð1Þ Using the ideal gas law, the CO 2 concentrations measured by the analyzers were converted to rates of CO 2 flux and presented in units of μmol/h. Negative CO 2 flux values denote net CO 2 uptake, while positive values denote net CO 2 production by the biofilm.
The linked BR cons -BR cons2 unit was illuminated continuously via a fluorescent plant growth light (JSV2 Jump Start T5 Grow Light System; Hydrofarm Inc., Petaluma, CA, USA), resulting in a photosynthetic photon flux density (PPFD) of approximately 141.79 μmol/m 2 /s. Both BR prod and BR cons -BR cons2 were fed their respective growth media at a flow rate of 15 mL/h. The BR cons -BR cons2 retention time and dilution rate was 23.2 minutes and 2.58 h -1 , respectively. This dilution rate exceeds the maximum specific growth rates of microalgal and bacterial strains [29][30][31], ensuring that suspended, non-biofilm-bound cells were readily washed out the system, and hence contributed minimally to the consumption or production of CO 2 . The sweeper gas flowed through the system at a constant flow rate of 150 mL/h.
In this study, CO 2 was provided to the phototrophic biofilm via a heterotrophic bacterial biofilm. At steady state, the biofilm's consistent CO 2 output serves to demonstrate the supply of an inexpensive and renewable CO 2 source to facilitate the growth and subsequent study of the phototrophic biofilm downstream. While it would also be feasible to accomplish this using a gas tank with a known or controllable CO 2 concentration, this approach highlights the utility in the design of the CSMS biofilm reactor modules, where the permeability of the inner silicone tube can facilitate not only the delivery of CO 2 to a biofilm, but also the collection and subsequent shuttling of CO 2 out of a biofilm. A scaled-up CSMS-based system could also feasibly integrate CO 2 -laden industrial flue gas or fermentation gas to support algal growth and achieve CO 2 bio-sequestration, a notion which is garnering increasing attention [32][33][34][35]. Delivering CO 2 in the gas phase through a CO 2 -permeable membrane, as is the case in the CSMS, may also reduce the extent of pre-processing required for these gas streams. Additionally, through relatively minor modifications to the system, it would be feasible to re-circulate the CO 2 -laden gas in a closed-loop configuration in order to improve overall sequestration efficiency.

Test cultures and growth medium
For each experiment, an overnight culture of the heterotrophic bacterium Pseudomonas aeruginosa (PA01) (originally obtained from Prof P.V. Phibbs at the Pseudomonas Genetic Stock Center, East Carolina University) [36] was inoculated into BR prod . CO 2 produced by the resulting biofilm readily diffused through the highly permeable silicone tube and entered the BR prod annular space, where it was continuously removed by the sweeper gas and carried downstream to Analyzer 1 and beyond. The P. aeruginosa biofilm provided a consistent source of CO 2 to facilitate the growth and study of the phototrophic biofilm downstream. The bacterial culture was fed with a tryptic soy broth prepared as a 0.6 g/L solution (2% concentration relative to the manufacturer's directions for typical batch cultivation), with a final composition of 0.34 g/L casein peptone, 0.06 g/L soya peptone, 0.05 g/L glucose, 0.1 g/L NaCl, and 0.05 g/L K 2 HPO 4 . Media were prepared in distilled water and autoclaved at 121˚C for 20 minutes prior to use.
The phototrophic culture was enriched from an aerated wastewater lagoon in Dundalk, Ontario, Canada [27]. It was cultivated in a modified Bold's Basal Medium (M-BBM) with a composition of 0.22 g/L (NH 4 ) 2 SO 4 , 0.025 g/L NaCl, 0.025 g/L CaCl 2 � 2H 2 O, 0.075 g/L MgSO 4 � 7H 2 O, 0.175 g/L KH 2 PO 4 , 0.075 g/L K 2 HPO 4, 8.34 mg/L FeSO 4 . Notably, no carbon was provided in the liquid medium. Illumination was provided continuously via the same fluorescent plant growth light described above. This culture was transferred to fresh medium biweekly, ensuring that the culture used to inoculate BR cons -BR cons2 in each experiment was not more than 14 days old. Unless otherwise stated, the results reported here are from biofilms at least twenty hours old.
A confocal laser scanning microscope (CLSM) was used to confirm the non-axenic nature of the phototrophic culture. A simple two-channel procedure, described by Lawrence et al. [37] and used subsequently by Barranguet et al. [38] and others, was used to confirm and differentiate the presence of algal and non-algal fractions within the culture used for inoculation of BR cons -BR cons2 . The former was visualized based on chlorophyll autofluorescence, while the latter was distinguished using SYTO 9 green fluorescent nucleic acid stain (S34854; Thermo-Fisher Scientific, Waltham, MA, USA) [37].
A well-mixed 100 μL undiluted culture sample was vacuum filtered through a black polycarbonate filter (0.2 μm pore size, GTBP02500; ThermoFisher Scientific, Waltham, MA, USA). The filter was then immersed in SYTO 9 stain (20 mg/ml) for ten minutes in the dark, before being rinsed twice with deionized water to remove excess stain. The filter was placed on a glass microscope slide and visualized using a Nikon Eclipse 80i-C1 confocal laser scanning microscope (Nikon Instruments Inc., Melville, NY). SYTO 9-stained bacteria were excited using a 488 nm laser and visualized through a 515/530 nm filter, producing a green signal. Microalgal cells conversely, were excited using a 632 nm laser and visualized through a 650 nm long pass filter, producing a red signal.

Assessing the prevalence of autotrophic-heterotrophic toggling
The phototrophic culture was inoculated into BR cons -BR cons2 and initially fed M-BBM. CO 2 was provided continuously via the heterotrophic biofilm in BR prod . After 20 hours, the M-BBM was amended with the labile organic carbon sources glucose (0.25 g/L) and sodium acetate (0.28 g/L). After approximately 30 hours of organic carbon availability, the medium was returned to the original M-BBM composition, which lacked the added carbon sources. Illumination was provided continuously throughout the experiment.

Investigating the longitudinal arrangement of autotrophic and heterotrophic metabolism
A similar experiment was performed in which the phototrophic culture was inoculated into BR cons -BR cons2 and initially fed M-BBM, with CO 2 provided continuously via the heterotrophic biofilm in BR prod . After 30 hours, the M-BBM medium was again amended with glucose and sodium acetate for the remainder of the experiment. In this case however, these organic carbon sources were added at half the concentration used previously (now 0.125 g/L and 0.14 g/L respectively). Illumination was provided by the fluorescent plant growth lights for the duration of the experiment. The CO 2 flux within the proximal half (BR cons ) and distal half of the biofilm (BR cons2 ) are presented separately.
A series of modified light-dark shift tests [27,39] was performed on BR cons2 during the period of organic carbon availability in order to confirm that CO 2 uptake observed in this portion of the biofilm was indeed the result of photoautotrophic activity. As in the experiment described above, the biofilm was initially grown under strictly autotrophic conditions, before amending the M-BBM with glucose (0.125 g/L) and sodium acetate (0.14 g/L). BR cons2 was then covered with an opaque, light-blocking sheet for two periods of 1.5 hours and 1 hour respectively, with a total of 2.5 hours allowed to elapse between the two dark periods.

Assessing reversibility in the longitudinal arrangement of autotrophic and heterotrophic metabolism
As in the experiments described above, the phototrophic culture was grown in the CSMS under initial autotrophic conditions, with M-BBM as the liquid growth medium and CO 2 provided by the heterotrophic biofilm in BR prod . After 38 hours, the M-BBM was once again amended with glucose (0.125 g/L) and sodium acetate (0.14 g/L). However, at hour 60 (22 hours after organic carbon amendment began), the direction of medium and gas flow through BR cons -BR cons2 was reversed. In this new orientation, fresh growth medium now entered BR cons2 from what was previously the effluent end and exited the system at what was previously the influent end of BR cons .
After leaving BR prod and passing through Analyzer 1, the gas stream in this reversed-flow orientation subsequently travelled through BR cons2 and BR cons in what was previously the upstream direction, with Analyzer 3 positioned between these two BR modules and Analyzer 2 positioned immediately downstream from them. In this configuration, the CO 2 flux within the two BR modules was now calculated using Eqs 4 and 5: This experiment was repeated in the same manner (autotrophic growth followed by organic carbon amendment and then flow reversal), however the glucose and sodium acetate were added at half the concentration used previously (now 0.0625 g/L and 0.07 g/L respectively).

Results and discussion
Confocal laser scanning microscopy was used to distinguish algal and non-algal members of the phototrophic culture. Although the culture was maintained under autotrophic conditions, a heterotrophic bacterial fraction was expected to have persisted, as has been described previously for other wastewater-derived phototrophic cultures (e.g. [9]). Microalgal members were recognized by chlorophyll autofluorescence and appeared red, while non-phototrophic bacteria were distinguished via SYTO 9 nucleic acid stain and appeared green (Fig 3). The appearance of non-overlapping red and green cells indicates the presence of both a photosynthetic (algal) and non-photosynthetic (bacterial) fraction in the culture, confirming that the culture was indeed non-axenic.
Algae-bacteria interactions are ubiquitous in nature and play an important role at the base of the trophic web. While numerous types of interactions between these groups are possible, encompassing both positive and negative effects [40], the natural syntrophy surrounding their O 2 and CO 2 exchange represents an important link between trophic levels and forms the basis of healthy ecosystems. Given the co-evolution of microalgae and bacteria over millions of years, it is rare to find microalgae existing as axenic cultures in nature, without some contribution from non-algal, non-photosynthetic microorganisms [41]. While the specific interactions at play within this culture were not fully elucidated, it was nonetheless unsurprising that it appeared to comprise both an autotrophic and heterotrophic fraction (Fig 3), despite being cultivated under autotrophic conditions. In engineered algal systems, non-photosynthetic bacteria are often viewed as mere contamination, necessitating costly and energy-intensive measures to maintain a sterile environment [40]. Recently, this view has begun to shift amidst a growing recognition that natural microalgal-bacterial consortia can represent a cost-effective, efficient alternative for many biotechnological applications [42].

Autotrophic-heterotrophic toggling by the phototrophic biofilm
When the phototrophic culture was inoculated into the CSMS and grown initially under autotrophic conditions (i.e. no organic carbon), the resulting biofilm grew autotrophically and exhibited net CO 2 uptake, denoted by negative CO 2 flux values (Fig 4). When the growth medium was amended with glucose and sodium acetate after 20 hours, the CO 2 flux began to trend upward. Within approximately 4 hours, the biofilm had clearly shifted from net-  Initially, the phototrophic biofilm in BR cons -BR cons2 grew autotrophically, consuming CO 2 provided by BR prod via the sweeper gas. Once organic carbon became available in the culture medium (denoted by the grey box), the biofilm rapidly toggled to CO 2 -producing net-heterotrophic growth, as indicated by a switch to positive CO 2 flux values. When the organic carbon was no longer available in the medium, the biofilm's CO 2 production fell dramatically and eventually returned to net-autotrophic growth, as indicated by the return to negative CO 2 flux values.
It should be noted that this CO 2 production was a result of the metabolic activity within the phototrophic biofilm only. Given that it was calculated via Eq 1 (Analyzer 3 -Analyzer 1), it does not include any of the CO 2 originating from the heterotrophic biofilm in BR prod . This meant that the CO 2 production observed in the phototrophic biofilm during organic carbon availability could be confidently attributed to the utilization of the two labile organic carbon sources. After 51 hours, the M-BBM was returned to its original composition lacking these organic carbon sources, leading to a rapid decline in the biofilm's CO 2 production and an eventual return to net-autotrophic growth. This same CO 2 flux behaviour was consistently observed through several experimental repeats.
The data thus supported the underlying hypothesis of this experiment, that a non-axenic phototrophic biofilm readily toggles between net-autotrophic and net-heterotrophic growth, dependent on the availability of labile organic carbon sources. While there is obviously a strong impetus to use wastewater as an inexpensive and abundant growth medium for algal biotechnologies [7,43], this result speaks to an important consideration in this approach. If bio-sequestration of CO 2 (be it from point or diffuse sources) is a primary objective, it may be disadvantageous to grow the sequestering culture in a wastewater stream that is vulnerable to organic nutrient spikes. This is especially relevant in industrial wastewater, or municipal wastewater from plants treating industrial effluents, where wastewater composition can vary widely and transient organic carbon shock loads with concentrations two or more times higher than normal are common [44]. Depending on the robustness and stability of the treatment system, this can result in prolonged periods of elevated effluent BOD, which would impact a downstream algal bio-sequestration system and lead to considerable CO 2 production (as seen in Fig 4), an effect which is counter to the overall aim of a sequestration system.

Longitudinal arrangement of autotrophic and heterotrophic metabolism
The configuration of the CSMS (specifically the placement of Analyzer 2 between BR cons and BR cons2 ), allows for the examination of the respective CO 2 flux within the two halves of the phototrophic biofilm. This increased resolution is helpful in revealing metabolic partitioning that may be prevalent in the biofilm. When this approach was applied to the experimental result presented in Fig 4, it was noted that in the presence of the labile organic carbon sources, the vast majority of the biofilm's CO 2 production was occurring within the proximal half of the biofilm (Fig 5A). By hour 50, just prior to the end of organic carbon availability, approximately 93% of the observed CO 2 production was from BR cons , with only about 7% coming from the distal half of the biofilm contained in BR cons2 . Interestingly, when organic carbon was no longer available in the culture medium, the distal half of the biofilm needed only two hours to return to net-autotrophic growth (CO 2 uptake), while the proximal half of the biofilm contained in BR cons took slightly more than two days to return to net-autotrophic growth. The protracted return to autotrophic growth observed in BR cons could have been an EPS effect, whereby some labile organic carbon was retained and subsequently accessible to this portion of the biofilm after organic carbon availability had ceased.
At hour 22 (Fig 5A), when the labile organic carbon sources first became available to the biofilm, there appears to have been enough biomass within BR cons to metabolize and deplete nearly all the glucose and sodium acetate carbon. Very little therefore reached BR cons2 where it could be metabolized by the distal half of the biofilm, which accounts for its comparatively low CO 2 production. If the organic carbon sources were less abundant (i.e. available in the medium at a lower concentration), one would expect all the glucose and sodium acetate carbon to be depleted within BR cons . This would leave the distal half of the biofilm housed in BR cons2 without access to any of the supplied organic carbon, causing it to remain in net-autotrophic growth even while the biofilm's proximal half grew heterotrophically. This notion formed the basis of the second hypothesis, which stated that in the presence of labile organic carbon, a phototrophic biofilm of sufficient length exhibits distinct, longitudinally discrete regions of net-heterotrophic and net-autotrophic growth.
To test the hypothesis, a similar experiment was conducted, in which the biofilm was initially grown under autotrophic conditions with CO 2 supplied from the heterotrophic biofilm in BR prod . The M-BBM fed to BR cons -BR cons2 was then once again amended with glucose and sodium acetate, however this time at half the concentration used previously (0.125 g/L and 0.14 g/L respectively). Before this organic carbon amendment, both halves of the biofilm were exhibiting autotrophic growth and CO 2 uptake, indicated by their negative CO 2 flux values ( Fig 5B). However, when the labile organic carbon sources became available in the medium after 30 hours, the proximal half of the biofilm (BR cons ) rapidly shifted from CO 2 -consuming autotrophic growth to CO 2 -producing heterotrophic growth. The CO 2 flux in the distal half of the biofilm (BR cons2 ) conversely, remained relatively unchanged and continued to exhibit netautotrophic growth (CO 2 uptake) for the entire experiment.
In three subsequent experimental repeats, this same pattern of metabolic partitioning was observed. In the presence of the labile organic carbon sources, the biofilm's proximal half produced CO 2 , while the biofilm's distal half continued to consume CO 2 . When BR cons2 was placed in the dark for two periods of 1.5 hours and 1 hour respectively during the period of organic availability, a rapid response was observed in which the CO 2 flux in this portion of the biofilm shifted from negative (CO 2 uptake) to positive (CO 2 production) (Fig 6). This darkinduced interruption of photosynthesis thereby confirmed that the CO 2 uptake observed in BR cons2 during organic carbon availability was in fact attributable to autotrophic activity (as opposed to abiotic leakage or passive diffusion out of the system). The small amount of CO 2 production observed during this darkness is likely attributable to underlying baseline algal respiration. When illumination resumed, so too did the photoautotrophic activity in BR cons2 , When examining each half of this biofilm separately, nearly all the observed CO 2 production during organic carbon availability (grey box) occurred within BR cons , with only very little occurring in BR cons2 . When organic carbon availability ended, BR cons2 quickly returned to CO 2 -consuming net-autotrophic growth (denoted by a negative CO 2 flux), whereas BR cons took over two days to return to CO 2 -consuming net-autotrophic growth. (B) When the same experiment was performed but with the organic carbon sources provided at half the concentration as compared to Fig 5A, BR cons once again rapidly toggled from CO 2 -consuming net-autotrophic growth to CO 2 -producing net-heterotrophic growth during organic carbon availability. However, BR cons2 continued to exhibit net-autotrophic growth and CO 2 uptake throughout the duration of the experiment, leading to discrete, longitudinally separated regions of net-heterotrophic and net-autotrophic growth occurring simultaneously within the biofilm. causing this portion of the biofilm to quickly rebound to approximately the same negative CO 2 flux values seen before the dark test. Given this result, it can be concluded that in the presence of both inorganic and organic carbon (CO 2 as well as glucose and sodium acetate), the phototrophic biofilm was in fact exhibiting distinct, longitudinally discrete regions of net-heterotrophic and net-autotrophic growth.
Previous studies have described the use of non-axenic phototrophic cultures for simultaneous inorganic and organic carbon removal. van der Ha et al. [45] for example, co-cultured microalgae and methane-oxidizing bacteria to achieve concurrent microbial methane oxidation and CO 2 uptake. Vu and Loh [46] described a symbiotic consortium of C. vulgaris and P. putida, where the algal fraction grew photoautotrophically using CO 2 produced by the bacterium. This in turn eliminated the need for aeration by providing the oxygen needed by the obligate aerobe P. putida to carry out efficient glucose biodegradation. A similar co-culture of P. putida and C. vulgaris described by Mujtaba et al. [47] exhibited the synergistic removal of organic carbon by the former and inorganic nutrients by the latter.
While there is much literature pertaining to microalgal-bacterial symbiosis in co-culture or synthetic consortia, comparatively few studies have examined this behaviour in natural phototrophic consortia within the context of engineered biofilm photobioreactors. Whereas co-cultures still require onerous measures to maintain sterility and prevent contamination, natural non-axenic cultures can offer a level of robustness and functional redundancy that makes contaminating organisms much less of a concern [13,25]. Although wastewater treatment and CO 2 sequestration have been studied separately for many years, approaches for effectively linking these via non-axenic phototrophic biofilms are lacking [8].
Microalgal biofilm growth systems fall into two broad categories: those in which the biofilm is mobile (via movement of the attachment material), and those in which the biofilm remains stationary (no movement of the attachment material) [48]. Mobile systems include for example rotating algal biofilm reactors (RABRs), in which the biofilm is attached to a solid or fibrous support material coiled around a rotating cylindrical drum that alternatingly exposes cells to the liquid and gas phases. Such systems have been shown to achieve high inorganic nutrient removal and biomass production when operating within a raceway pond [23]. However, the energy required for their moving parts can be a complicating factor. In stationary In order to confirm that the CO 2 uptake observed in BR cons2 during organic carbon availability was indeed attributable to photosynthetic CO 2 fixation, BR cons2 was placed in the dark twice for at least one hour. Both times, CO 2 uptake stopped almost immediately. This signified the rapid cessation of photosynthesis and its related CO 2 fixation, and provided confirmation that the phototrophic biofilm was indeed exhibiting distinct, longitudinally discrete regions of net-heterotrophic and net-autotrophic growth when exposed to both inorganic and organic carbon (CO 2 as well as glucose and sodium acetate).
https://doi.org/10.1371/journal.pone.0253224.g006 growth systems conversely, biofilms form on a fixed attachment material with liquid growth medium flowing over the surface of the biofilm. Tubular photobioreactors for example, are closed, stationary systems which enable significant control over conditions and offer high surface to volume ratios [49].
Insight generated by the CSMS can be useful in informing the design and operation of similar scaled-up tubular systems. This can include for example determining what fraction of total reactor length is required to deplete nutrients. Since this length will vary with key factors such as influent nutrient load and flow rate, the utility of the CSMS lies in its innovative approach to informing such optimization. The longitudinal arrangement of discrete heterotrophic and autotrophic regions within the phototrophic biofilm (as described in Fig 5), suggests that significant simultaneous removal of organic and inorganic carbon may be possible in large-scale tubular photobioreactors using non-axenic phototrophic biofilms. Ostensibly, the former could be metabolized in the proximal portion of the biofilm through heterotrophic metabolism, with the latter being taken up and fixed through autotrophic metabolism occurring in the distal portion of the biofilm, where assimilation of inorganic nitrogen and phosphorus nutrients would also take place.
Despite the potential benefits offered by tubular photobioreactors in terms of CO 2 capture and contaminant removal, the energy required for operation, as well as the need for cleaning and maintenance, remain barriers limiting their widespread use. As such, there is still a need for further research in this area, focused on expanding our understanding of biofilm behaviour within these systems and developing strategies for improving their overall cost-effectiveness.

Reversibility in the longitudinal arrangement of autotrophic and heterotrophic metabolism
In the presence of inorganic and organic carbon, the phototrophic biofilm consistently exhibited longitudinally discrete regions of net-heterotrophic and net-autotrophic metabolism, with the location of these regions dictated by the gradient of organic carbon availability resulting from heterotrophic metabolism in the proximal portion of the biofilm. It was hypothesized that these discrete regions of net-autotrophic and net-heterotrophic growth are therefore transient and reversible based on the direction of nutrient flow and availability. That is, when the direction of flow is reversed, the locations of the autotrophic and heterotrophic regions will flip.
To test this hypothesis, a similar experiment was performed in which the biofilm initially grew autotrophically with CO 2 as the sole carbon source. When glucose and sodium acetate were added to the growth medium, BR cons once again toggled to CO 2 -producing heterotrophy, while BR cons2 continued to exhibit net-autotrophic CO 2 uptake (Fig 7A). At hour 60 however, the direction of medium and gas flow through these two BR modules was reversed, such that BR cons2 now contained the "proximal" half of the biofilm, gaining access to the fresh, undepleted organic carbon sources provided in the medium. BR cons conversely, was now downstream of BR cons2 and represented the "distal" half of the biofilm.
After the direction of flow was reversed, a sharp increase in CO 2 flux and a rapid switch from autotrophic to heterotrophic growth was observed in BR cons2 (Fig 7A). This aligned with expectations, given that this portion of the biofilm was now first to encounter and metabolize the organic carbon sources. In previous experiments (and before reversing the direction of flow in this experiment), this portion of the biofilm was only able to access what little (if any) of the organic carbon could pass through BR cons un-metabolized.
It is notable that the slope in the BR cons2 CO 2 flux after hour 60 is remarkably similar to the slope in BR cons after hour 38 (Fig 7A), suggesting that both halves of the biofilm exhibited the same metabolic response (and started producing CO 2 at nearly identical rates), when that portion of the biofilm was the first to encounter the organic carbon sources. This result speaks to the significant analytical utility of the CSMS. The ability to visualize the biofilm's rate of CO 2 consumption and/or production in real-time means that growth conditions can be finetuned to ensure that these rates are effectively optimized. Increasing the number of CO 2 analyzers positioned along the length of the biofilm would also enable greater longitudinal resolution, which can further inform optimal nutrient loading and reactor length under a given set of conditions.
When BR cons became the distal end of the biofilm and was no longer receiving the fresh, un-depleted organic carbon sources, its rate of CO 2 production began to gradually decrease ( Fig 7A). However, this portion of the biofilm did not indicate an overall shift to net-autotrophic growth as expected. As mentioned previously, EPS may have played a role in this muted response by retaining stores of organic carbon from earlier in the experiment, which could then be taken up and metabolized after the direction of flow was reversed. Interestingly, in natural, non-axenic phototropic biofilms, much of the initial EPS originates from heterotrophic bacteria [50]. It is also plausible that following flow reversal at hour 60, the portion of the biofilm in BR cons2 was not yet dense enough to fully deplete the organic carbon it now encountered, therefore leaving some available for the "distal" half of the biofilm now housed in BR cons . Although the precise explanation necessitates additional biofilm analyses which are ultimately beyond the scope of this paper, it was postulated that in either case these effects would be minimized, and the BR cons CO 2 flux after flow reversal would eventually cross zero and exhibit CO 2 uptake, if the organic carbon sources were available in the medium at a lower concentration.
To test this, the experiment was repeated but with the glucose and sodium acetate supplied at half the concentration used previously (now 0.0625 g/L and 0.07 g/L respectively). This experiment was also allowed to run for a longer period following the reversal of flow direction. BR cons2 once again exhibited an increase in its CO 2 flux values after flow reversal, eventually achieving net CO 2 production ( Fig 7B). As predicted, BR cons in this case showed a steady decline in CO 2 flux after flow reversal, reaching zero by approximately hour 70. The phototrophic biofilm initially grew autotrophically using CO 2 from BR prod . As previously observed, BR cons shifted to CO 2 -producing netheterotrophic growth after organic carbon amendment (grey box), while BR cons2 continued in CO 2 -consuming net-autotrophic growth. When the direction of medium and gas flow through BR cons -BR cons2 was reversed at hour 60, the CO 2 flux in BR cons stopped increasing and gradually decreased. The CO 2 flux in BR cons2 conversely, increased rapidly and began exhibiting net CO 2 production. (B) This experiment was repeated, but with the glucose and sodium acetate concentrations decreased by half. Once again, both halves of the biofilm grew autotrophically initially, before BR cons toggled to CO 2 -producing heterotrophy after organic carbon amendment (grey box). When the direction of medium and gas flow in BR cons -BR cons2 was reversed, CO 2 production in BR cons stopped increasing and began to decrease, ultimately crossing zero just prior to the end of the experiment. BR cons2 conversely gradually stopped consuming CO 2 after flow direction was reversed, eventually crossing over to net CO 2 production. https://doi.org/10.1371/journal.pone.0253224.g007 Given that the CO 2 flux in both BR cons and BR cons2 trended in opposite directions after flow reversal (Fig 7B), and considering the outcome of the previous experiment (Fig 7A), the upstream and downstream partitioning of net-heterotrophic and net-autotrophic metabolism appears to be reversible, with the position of these regions dictated by nutrient flow direction and hence availability, thus supporting the third hypothesis. Although a comprehensive community composition analysis would be needed to make definitive assertions, these results suggest that this rapid metabolic partitioning is likely not the result of significant changes in biofilm member composition, but rather a functional redundancy within the culture and changes in the metabolic activity of members already present before the influx of organic carbon.
The experiments described here present an elegant approach to the study of phototrophic biofilms in order to gain important insight regarding their behaviour under changing conditions. Such information can inform new and relevant research questions and contribute to efforts aimed at scaling and industrializing algal growth systems, where the ability to understand, predict, and optimize biofilm growth and activity is critical. In integrated wastewater treatment-CO 2 sequestration systems utilizing non-axenic phototrophic biofilms, the balance between organic and inorganic carbon metabolism is a major factor that dictates the biofilm's overall CO 2 flux. In this study, the phototrophic biofilm exhibited a rapid response and began growing heterotrophically when exposed to organic carbon, reaching a CO 2 production rate of 60 μmol/h after approximately 30 hours. This suggests that non-axenic phototrophic biofilms are indeed able to readily toggle between net-autotrophic CO 2 capture and net-heterotrophic CO 2 production based on the availability of labile organic carbon sources. When the organic carbon sources were less abundant (i.e. provided at a much lower concentration), the biofilm took only 4 hours to exhibit longitudinally discrete regions of autotrophic and heterotrophic growth in the proximal and distal portions of the biofilm respectively. The apparent reversibility in the biofilm's response to changing carbon conditions suggests a robustness and versatility within non-axenic phototrophic biofilms which may be exploitable in engineered phototrophic biofilm systems to achieve simultaneous organic and inorganic carbon removal.
Hypothesis-driven research aimed at generating fundamental insights about phototrophic biofilms are important as the need for effective CO 2 management and mitigation strategies increases, and the motivation for linking these to enhanced wastewater treatment grows. The study presented here therefore should represent a key step forward in this endeavour.