Culturing Toxic Benthic Blooms: The Fate of Natural Biofilms in a Microcosm System

A microcosm designed for culturing aquatic phototrophic biofilms on artificial substrata was used to perform experiments with microphytobenthos sampled during summer toxic outbreaks of Ostreopsis cf. ovata along the Middle Tyrrhenian coast. This dynamic approach aimed at exploring the unique and complex nature of O. cf. ovata bloom development in the benthic system. Epibenthic assemblages were used as inocula for co-cultures of bloom organisms on polycarbonate slides at controlled environmental conditions. Biofilm surface adhesion, growth, and spatial structure were evaluated along with shifts in composition and matrix production in a low disturbance regime, simulating source habitat. Initial adhesion and substratum colonisation appeared as stochastic processes, then community structure and physiognomy markedly changed with time. Dominance of filamentous cyanobacteria and diatoms, and dense clusters of Amphidinium cf. carterae at the mature biofilm phases, were recorded by light and confocal microscopy, whilst O. cf. ovata growth was visibly limited in the late culture phases. Life-form strategies, competitiveness for resources, and possibly allelopathic interactions shaped biofilm structure during culture growth. HPLC (High Performance Liquid Chromatography) analysis of exopolysaccharidic matrix revealed variations in sugar total amounts and composition. No toxic compounds were detected in the final communities tested by LC-MS (Liquid Chromatography- Mass Spectrometry) and MALDI-TOF MS (Matrix Assisted Laser Desorption Ionization Time OF Flight Mass Spectroscopy) techniques.


Introduction
Ostreopsis cf. ovata, a toxic benthic dinoflagellate, has considerably expanded its distribution in temperate areas in the last 20 years, forming recurrent benthic harmful algal blooms (BHABs) in the Mediterranean basin [1,2]. BHABs rapidly spread along Italian Middle Tyrrhenian coast, since their first record in 1999 [3,4], with summer outbreaks sustained by the most abundant and so far widely distributed Ostreopsis Mediterranean species [5][6][7][8]. On these occasions, BHABs come in brownish, mucous cell aggregates, visible to the naked eye, proliferating on biotic and abiotic substrata of relatively shallow and sheltered coastal stretches favored by low water motion and increased input of organic nutrients [9][10][11]. Complex and diverse microphytobenthic associations co-occur with O. cf. ovata blooming cells, especially diatoms and cyanobacteria, all enmeshed in a common exopolymeric matrix that probably plays a central role in cell attachment to surfaces, colony formation, and bloom "nutrition" and dispersion [7,10,[12][13][14]. Undesirable impact of Ostreopsis BHABs includes production

Preparation of Inocula
Host samples were shaken in the storage water to allow detachment of epibenthic cells, then rinsed with filtered seawater (WhatmanGF/C filters GE Healthcare, Little Chalfont, UK) to completely remove biofilm material. Rinsing suspensions containing biofilm cells were sieved (100 µm mesh) to remove larger organisms and detritus. Bryozoans and macroalgae were weighted to determine fresh and dry weight (FW, DW).

Counting of Ostreopsis cf. ovata Cells
Aliquots of biofilm suspensions after homogenization were fixed in Lugol's iodine solution and left to settle for 24 h in 25 mL counting chambers. Ostreopsis cells were counted according to the Utermöhl method [28] using a Zeiss Axiovert 100 microscope. Cell concentrations were expressed as cells g −1 FW.

Microcosm Cultures
2 L fresh biofilm suspensions were poured into 4 L bottles containing 2 L of K/2 medium [29] to obtain a 4 L culture inoculum for each suspension. Three growth experiments, named Run 1, Run 2, and Run 3 were performed with the three different inocula at the same environmental conditions ( Table 1).
The methodological approach is shown in Figure 1. underwater to avoid epibenthic cell loss, then transferred in plastic jars in the dark and refrigerated until laboratory analysis and preparation of culture inocula on the same day.

Preparation of Inocula
Host samples were shaken in the storage water to allow detachment of epibenthic cells, then rinsed with filtered seawater (WhatmanGF/C filters GE Healthcare, Little Chalfont, UK) to completely remove biofilm material. Rinsing suspensions containing biofilm cells were sieved (100 µm mesh) to remove larger organisms and detritus. Bryozoans and macroalgae were weighted to determine fresh and dry weight (FW, DW).

Counting of Ostreopsis cf. ovata Cells
Aliquots of biofilm suspensions after homogenization were fixed in Lugol's iodine solution and left to settle for 24 h in 25 mL counting chambers. Ostreopsis cells were counted according to the Utermöhl method [28] using a Zeiss Axiovert 100 microscope. Cell concentrations were expressed as cells g −1 FW.

Microcosm Cultures
2 L fresh biofilm suspensions were poured into 4 L bottles containing 2 L of K/2 medium [29] to obtain a 4 L culture inoculum for each suspension. Three growth experiments, named Run 1, Run 2, and Run 3 were performed with the three different inocula at the same environmental conditions ( Table 1).
The methodological approach is shown in Figure 1. Two identical microcosms (biofilm growth systems) were used for July parallel cultures (Run 1 and Run 2) and one for the August experiment (Run 3), with simultaneous control of irradiance, temperature, and flow velocity. Each microcosm (UFZ Centre for Environmental Research, Magdeburg, Germany) contained four separate flow lanes ( Figure 1) with integrated inlet and outlet devices at the start and the end of each lane. Each chamber had a 798 cm 2 area made up of 42 removable polycarbonate slides, utilised as growth substrata. The inoculum was pumped from the 4 Two identical microcosms (biofilm growth systems) were used for July parallel cultures (Run 1 and Run 2) and one for the August experiment (Run 3), with simultaneous control of irradiance, temperature, and flow velocity. Each microcosm (UFZ Centre for Environmental Research, Magdeburg, Germany) contained four separate flow lanes ( Figure 1) with integrated inlet and outlet devices at the start and the end of each lane. Each chamber had a 798 cm 2 area made up of 42 removable polycarbonate slides, utilised as growth substrata. The inoculum was pumped from the 4 L reservoir through the inlet device within each lane with a laminar flow of 50 L h −1 over the growth surface. A turbulence reducer and a water temperature sensor were located at the inlet device of each chamber and the flow rate was valve-regulated. Temperature was set at 25 • C and irradiance at 110 µmol photons m −2 ·s −1 . Fluorescent lamps (True-light 36 W AURALIGHT) were used in a 14/10 light/dark regime. The biofilm growth was followed for 50 days bloom duration at sampling sites, and accurately tracked on-line by the reduction in light transmittance measured in each chamber by nine light sensors positioned under the growth surface. Biofilm samples were collected at three temporal developmental phases: initial adhesion, active growth, and mature phase (corresponding to 90%, 50%, and 10% of the transmitted light, respectively) [26]. Each growth experiment ended when the recorded transmitted light was less than 10%, indicating the achievement of a stationary growth phase. Transmittance values were converted into a percentage of light absorbed by biofilms to obtain the growth curves [25,26]. The medium was changed at regular intervals (twice per week).

• Biodiversity
To assess biodiversity of phototrophs in natural and cultured communities, samples scraped from randomly selected slides were fixed with formaldehyde at 2% and glutaraldehyde at 2.5% final concentrations and then stored at 4 • C. Fixed samples were examined using a ZEISS Axioskop light microscope equipped with differential interference contrast (DIC) at 40× and 100× objectives. Samples were also observed at light microscope after staining for 10 min with Alcian Blue (AB) 1% in HCl 0.5 N (pH 0.5) or in 3% acetic acid (pH 2.5) to evaluate the presence of sulfated and carboxylic polysaccharides in the extracellular matrix [30].

• Biomass
Cultured communities sampled at the three phases of development were scraped off the polycarbonate slides to evaluate the biomass. Biofilm biomass was determined as dry weight (DW) by oven drying samples at 60 • C for 72 h. Phototrophic biomass was assessed by determining Chl a concentration, extracted overnight in 90% acetone in the dark and then quantified spectrophotometrically according to [31].
The biofilm architecture and spatial distribution of phototrophs during the different phases of development was obtained by observing intact biofilm, after non-destructive, non-invasive sampling, at the confocal laser scanning microscope (CLSM) (Olympus FV1000, IX81 using Plan-Apochromatic 60× (NA 1.42, oil) objectives) at the Centre of Advanced Microscopy "P. Albertano", Department of Biology, University of Rome "Tor Vergata", Rome, Italy. CLSM was used in a multichannel mode, whereby the different channels mapped individual biofilm components. The excitation wavelengths were in the blue (488 nm, Ar), green (543 nm, Ar/HeNe), and red (636 nm, Ar/HeNe) regions. Data consisted of a set of two dimensional (2D), cross-sectional images in the x-y plane that were captured along the z-axis. The three-dimensional images were obtained through acquisitions in the XY plane, Z step size 0.5 µm, using the software package IMARIS 6.2.0 software (Bitplane AG, Zurich, Switzerland).

Spectral analyses
Spectral analysis (CLSM-SA) of different autofluorescence signals was also carried out on biofilm samples by using the lambda scan (λ scan) function of CLSM exciting specific regions of interest (1 µm 2 ROIs) [32]. Lasers with the following excitation wavelengths: 488 nm; Ar, 543 nm; Ar/HeNe and red 635 nm; Ar/HeNe were used. Spectra were obtained for emission wavelengths ranging between 400 and 800 nm. • Matrix exopolysaccharides Bound/capsular exopolysaccharides (CPS) were also extracted from mature communities using 0.1 M H 2 SO 4 at 95 • C [33,34]. The carbohydrate fractions were measured spectrophotometrically using the phenol-sulphuric acid method [35]. In order to check the extraction efficiency of each eluent, the pellet obtained after the extraction was stained with Alcian Blue dyes, specific for acidic polysaccharides, to check localization of positive reaction in light microscopy [23]. CPS extracts were then analysed for their monosaccharide composition using RP-HPLC (reverse phase), using a Beckman Ultrasphere ODS (Octa Decyl Silane) column as reported in [36].

• Toxins
Ostreopsis toxins were analysed in inocula and culture material by LC-MS (Orbitrap XL). MALDI-TOF mass spectrometry for cyanobacterial toxins was performed by AnagnosTec GmBH (Postdam, Germany) on mature biofilm samples.

Results and Discussion
The dynamic approach employed here represents the first attempt to culture the whole O. cf. ovata BHAB community on substrata to better understand microphytobenthos/biofilm development and shifts under simplified, stable environmental conditions that can mimic bloom evolution in sheltered, temperate waters during summer.

Taxonomic Composition
Bryozoans and Cystoseira sp. (Phaeophyceae) hosted microphytobenthic communities with similar gross aspect and taxonomic composition irrespective of time and site of sampling. Epizoic and epiphytic associations macroscopically appeared as brownish mucous material, slightly diffluent and more or less adherent to the colonised surfaces. Light microscopy of host fragments revealed complex networks of microalgae and cyanobacteria. Epizoic associations in July samples included the dinoflagellates Ostreopsis cf. ovata, Amphidinium cf. carterae, Coolia monotis, and Prorocentrum lima. Among diatoms, solitary naviculoids and nitzschioid forms, along with tube dwelling and colonial species prevailed (Amphora spp., Ceratoneis closterium, Entomoneis sp., Grammatophora sp., Lichmophora spp., Rhabdonema sp.) over large centrics of Coscinodiscus and Thalassiosira genera. Fascicles of Nitzschia cf. martiana abundantly colonised bryozoans and had clusters of O. cf. ovata cells embedded in the mucopolysaccharidic material forming the diatom crenulated tube, as clearly evidenced by cytochemical staining in light microscopy ( Figure 2). Oscillatorialean cyanobacteria, mostly less than 3 µm in diameter, and few coccal morphotypes were also largely present. Cystoseira sp. thalli, sampled in August, appeared to host proportionally more abundant Ostreopsis cf. ovata and Coolia monotis cells.
clusters of O. cf. ovata cells embedded in the mucopolysaccharidic material forming the diatom crenulated tube, as clearly evidenced by cytochemical staining in light microscopy ( Figure 2). Oscillatorialean cyanobacteria, mostly less than 3 µm in diameter, and few coccal morphotypes were also largely present. Cystoseira sp. thalli, sampled in August, appeared to host proportionally more abundant Ostreopsis cf. ovata and Coolia monotis cells.  Recurrence of these composition patterns was also recorded during BHAB monitoring along Northern Adriatic sea coast [7] and along Catalonia coast (Spain, NW Mediterrenean sea) [37] and concentration data reflect ranges and variation observed in the Mediterranean basin [19].

Growth in the Microcosm
Biofilm did grow on the artificial substrata placed on the bottom of the microcosm lanes under high temperature and low disturbance simulating conditions that support bloom development in the sampling sites. Non-destructive, real-time monitoring of biomass accumulation was obtained by recordings of the subsurface light sensors. After inoculation, during which microorganisms settled and adhered onto the slides, a lag phase followed ( Figure 3). Initial phases lasted only three days in Run 3 but up to ten in Run 1 and Run 2 cultures. Exponential growth (active phase) followed in the three growth experiments and cultures displayed different maximum growth rates, the highest registered for Run 1 and Run 2 communities. After a period of exponential increase, curves assumed a S-shape and reached a stationary phase indicating similar trends in biomass accrual. These trends are comparable to those obtained for marine biofilms of the Schelde estuary mudflats (The Netherlands), cultivated in the same system [24]. During the late development phase, biofilm sloughing occurred in Run 3 at day 38, likely due to a lower adhesive potential of the biofilm, resulting from lower exopolysaccharide content estimates in Run 3 cultures (0.012 mg cm −2 ) compared to Run 1 (0.038 mg cm −2 ) and Run 2 (0.028 mg cm −2 ). In any case, partial detachment of biofilm flocs was recorded during the experimental time in the three Runs, likely due to diffusion limitation caused by biofilm thickness. This can be considered one factor intrinsic to the biofilm causing dispersion of the BHAB cells in the water column in nature. compared to Run 1 (0.038 mg cm −2 ) and Run 2 (0.028 mg cm −2 ). In any case, partial detachment of biofilm flocs was recorded during the experimental time in the three Runs, likely due to diffusion limitation caused by biofilm thickness. This can be considered one factor intrinsic to the biofilm causing dispersion of the BHAB cells in the water column in nature.

Biofilm Observations
Macroscopic observations of biofilms were performed daily to monitor biofilm development on the slides. Initial adhesion and substratum colonisation appeared as stochastic processes in the three experiments performed. Cultures appeared as series of large patches, different in pigmentation that became more coherent with experimental time, when a more homogeneous coverage of substrata was visible. At the end of the experiments, corresponding to the late phase of development, biofilms

Biofilm Observations
Macroscopic observations of biofilms were performed daily to monitor biofilm development on the slides. Initial adhesion and substratum colonisation appeared as stochastic processes in the three experiments performed. Cultures appeared as series of large patches, different in pigmentation that became more coherent with experimental time, when a more homogeneous coverage of substrata was visible. At the end of the experiments, corresponding to the late phase of development, biofilms were approximately 1 mm thick. At this phase, microorganisms formed macroscopic, mucilaginous, bright blue-green mats, showing a coherent structure, where the filamentous forms were more or less oriented to the flow direction of the culture medium ( Figure 4).  Light microscopy observations of biofilm-fixed fragments scraped from randomly selected polycarbonate slides during the three Runs performed, showed a general reduction in taxon richness of cultures over experimental time. O. cf. ovata cells were healthy and proliferated at initial and active phases (Figure 5a,b), but then Amphidinium cf. carterae dominated among dinoflagellates and a general prevalence of filamentous, straight, or tightly coiled, non heterocytous cyanobacteria Light microscopy observations of biofilm-fixed fragments scraped from randomly selected polycarbonate slides during the three Runs performed, showed a general reduction in taxon richness of cultures over experimental time. O. cf. ovata cells were healthy and proliferated at initial and active phases (Figure 5a,b), but then Amphidinium cf. carterae dominated among dinoflagellates and a general prevalence of filamentous, straight, or tightly coiled, non heterocytous cyanobacteria (Leptolyngbya, Lyngbya, Oscillatoria, Microcoleous, Phormidium, Pseudanabaena, Spirulina spp.) was observed in later phases ( Figure 5 c-i). Among diatoms, tube-dwelling species dominated in late growth phases (Parlibellus sp.) along with the solitary Amphora spp. and Ceratoneis closterium (Figure 5c,g,i). This loss in diversity and shift to cyanobacterial dominance has also been reported in other experiments performed on phototrophic biofilms cultivated in photobioreactors [24,25]. Light microscopy observations of biofilm-fixed fragments scraped from randomly selected polycarbonate slides during the three Runs performed, showed a general reduction in taxon richness of cultures over experimental time. O. cf. ovata cells were healthy and proliferated at initial and active phases (Figure 5a,b), but then Amphidinium cf. carterae dominated among dinoflagellates and a general prevalence of filamentous, straight, or tightly coiled, non heterocytous cyanobacteria (Leptolyngbya, Lyngbya, Oscillatoria, Microcoleous, Phormidium, Pseudanabaena, Spirulina spp.) was observed in later phases ( Figure 5 c-i). Among diatoms, tube-dwelling species dominated in late growth phases (Parlibellus sp.) along with the solitary Amphora spp. and Ceratoneis closterium ( Figure  5 c, g, i). This loss in diversity and shift to cyanobacterial dominance has also been reported in other experiments performed on phototrophic biofilms cultivated in photobioreactors [24,25].

Biofilm Matrix
Biofilm microorganisms were immersed in an acidic exopolymeric matrix as revealed by light microscopy observations after biofilm staining with Alcian Blue dyes. Moreover, sulphated and carboxylated groups were present in the capsular polysaccharides of the cyanobacterial envelopes and in the tubes or diffluent mucilage produced by raphid diatoms. CPS extracted from all biofilm communities also contained at least seven different neutral sugars (Figure 6a). Findings on sugar composition in mature phase samples agreed with data obtained for cultured biofilms dominated by cyanobacteria [23,36]. Exopolysaccharide-producing microorganisms are favoured in nutrient supply and internal recycling, and this can explain their prevalence in the late development phase. EPS (Exopolymeric Substances) also provided high light screening of the outermost biofilm phototrophs, pointing to a key role of EPS production capacity in shaping the mature biofilm phase. composition in mature phase samples agreed with data obtained for cultured biofilms dominated by cyanobacteria [23,36]. Exopolysaccharide-producing microorganisms are favoured in nutrient supply and internal recycling, and this can explain their prevalence in the late development phase. EPS (Exopolymeric Substances) also provided high light screening of the outermost biofilm phototrophs, pointing to a key role of EPS production capacity in shaping the mature biofilm phase.

Biomass
Due to the heterogeneous distribution of microorganisms on substrata and the high biomass development at the late growth phases, sampling for counting biofilm members was not considered as representative of the diversity observed. Thus, biomass accumulation was estimated as dry weight (mg cm −2 ) and chlorophyll a content (µg cm −2 ), confirming trends of growth curves (Figure 6b,c, respectively).

Biofilm Structure
Confocal microscopy analyses were carried out by directly observing biofilms on the polycarbonate slides. This allowed preservation of biofilm integrity during sampling and analysis of the spatial relationships among biofilm members and with the substrata throughout biofilm development. Visualization of biofilm architecture, based on the autofluorescence signals of the different photosynthetic organisms, also allowed analysis of the inner components immersed in the biofilm thickness at the three sampled phases.
An in depth observation of the "patchwork" coverage of initial phases revealed that 10-day-old biofilms already had a complex structure and thickness that could reach 150 µm in Run 1 ( Figure   Figure 6. Biofilm capsular exopolysaccharides (CPS) monosaccharides in mature samples grown at each Run. Different proportions of 8 neutral sugars differed between Runs, data are expressed as mol % (a). Dry weight (b) and chlorophyll a content (c) values of biofilms at the three development phases, increasing trends over time were detected for both biomass indicators.

Biomass
Due to the heterogeneous distribution of microorganisms on substrata and the high biomass development at the late growth phases, sampling for counting biofilm members was not considered as representative of the diversity observed. Thus, biomass accumulation was estimated as dry weight (mg cm −2 ) and chlorophyll a content (µg cm −2 ), confirming trends of growth curves (Figure 6b,c, respectively).

Biofilm Structure
Confocal microscopy analyses were carried out by directly observing biofilms on the polycarbonate slides. This allowed preservation of biofilm integrity during sampling and analysis of the spatial relationships among biofilm members and with the substrata throughout biofilm development. Visualization of biofilm architecture, based on the autofluorescence signals of the different photosynthetic organisms, also allowed analysis of the inner components immersed in the biofilm thickness at the three sampled phases.
An in depth observation of the "patchwork" coverage of initial phases revealed that 10-day-old biofilms already had a complex structure and thickness that could reach 150 µm in Run 1 (Figure 7a,b). O. cf. ovata cells were healthy and proliferating at initial and active biofilm phases. Spectral analysis of cells confirmed the results (Figure 8). Different regions of interest (ROI) were acquired and excited with different laser wavelength and the reported obtained spectra (Figure 8) showed active peridinin-chlorophyll a-protein complexes [38] for light absorption (Figure 8). The possibility to perform the spectral analysis of individual cells allowed discrimination between the different autofluorescent signals, providing a community fingerprint based on the properties of photosynthetic pigments.
As the biofilm matured, so did its complexity with stratification and heterogeneous distribution of the organisms on substrata, resulting in the formation of voids and "valleys" (Figures 9 and 10). Biofilm filamentous members, cyanobacteria and tube-dwelling diatoms, were mostly responsible of the layered structures observed (Figure 9c), while voids appeared especially around Amphidinium cf. carterae cell clusters, indicating exclusion of other biofilm members possibly due to the production of allelopathic compounds or mucous capsules around cell clusters (Figures 9c and 10a) [23,41]. The 'valley' aspect was due to globular microcolonies of coccal cyanobacteria, evidencing active, multiple cell divisions (Figure 10c,d).
7a,b). O. cf. ovata cells were healthy and proliferating at initial and active biofilm phases. Spectral analysis of cells confirmed the results (Figure 8). Different regions of interest (ROI) were acquired and excited with different laser wavelength and the reported obtained spectra (Figure 8) showed active peridinin-chlorophyll a-protein complexes [38] for light absorption (Figure 8). The possibility to perform the spectral analysis of individual cells allowed discrimination between the different autofluorescent signals, providing a community fingerprint based on the properties of photosynthetic pigments.   (a,b, arrows) showing numerous, elongated chloroplasts located at the cell periphery. Green signal was obtained with a 488-nm laser source, emission of frustule autofluorescence [39,40] is visible in (a, asterisk) where elongated diatoms occurr sparse on the slide, while a dense cluster of Coscinodiscus sp. lies interspersed with thin cyanobacterial trichomes in (d). White and red signals were captured after 543-and 635-nm laser excitation of accessory pigments and chlorophyll a of biofilm phototrophs. As the biofilm matured, so did its complexity with stratification and heterogeneous distribution of the organisms on substrata, resulting in the formation of voids and "valleys" (Figures 9 and 10). Biofilm filamentous members, cyanobacteria and tube-dwelling diatoms, were mostly responsible of the layered structures observed (Figure 9c), while voids appeared especially around Amphidinium cf. carterae cell clusters, indicating exclusion of other biofilm members possibly due to the production of allelopathic compounds or mucous capsules around cell clusters (Figures 9c and 10a) [23,41]. The 'valley' aspect was due to globular microcolonies of coccal cyanobacteria, evidencing active, multiple cell divisions (Figure 10c,d). Biofilm filamentous members, cyanobacteria and tube-dwelling diatoms, were mostly responsible of the layered structures observed (Figure 9c), while voids appeared especially around Amphidinium cf. carterae cell clusters, indicating exclusion of other biofilm members possibly due to the production of allelopathic compounds or mucous capsules around cell clusters (Figures 9c and 10a) [23,41]. The 'valley' aspect was due to globular microcolonies of coccal cyanobacteria, evidencing active, multiple cell divisions (Figure 10c,d).   , arrow)). White signal was obtained with a 488-nm laser source, magenta and cyan colour codes using 543-nm and 635-nm laser sources, respectively. Bars = 40 µm.
Low-magnification observations carried out to evaluate the structure and proportions of the different phototrophic components at active and mature communities showed biofilms strongly characterized by the dominance of cyanobacteria and the abundant presence of only a few diatoms taxa (Figure 9a and Figure 10a,b,c). Cyanobacteria represented the best competitors in the advanced phases of community development probably due to their capacity to thrive in the low light microenvironments established across the biofilm thickness as also shown in [36]. This could imply a risk of massive cyanobacterial development in benthic habitats and toxicity outbreaks, although our simplified conditions excluded zooplankton and invertebrate predators that would also shape the algal structure and relative abundance of each taxa. In any case, MALDI-TOF analysis of mature biofilms cultured in this study did not evidence bioactive compounds. Occasionally, loosely attached filamentous phototrophs, streamers, were observed floating in flow direction in biofilms at Networks of diverse, more or less oriented, filamentous cyanobacteria lie close to clusters of Amphidinium cf. carterae cells, apparently excluding other biofilm organisms (a, arrow), and of large raphid diatoms (b, arrow). 3D renderings evidencing microcolonies of actively dividing coccal, colonial cyanobacteria (c arrow), forming channels (c, double arrow) and valleys (d, arrow)). White signal was obtained with a 488-nm laser source, magenta and cyan colour codes using 543-nm and 635-nm laser sources, respectively. Bars = 40 µm.
Low-magnification observations carried out to evaluate the structure and proportions of the different phototrophic components at active and mature communities showed biofilms strongly characterized by the dominance of cyanobacteria and the abundant presence of only a few diatoms taxa (Figures 9a and 10a,b,c). Cyanobacteria represented the best competitors in the advanced phases of community development probably due to their capacity to thrive in the low light microenvironments established across the biofilm thickness as also shown in [36]. This could imply a risk of massive cyanobacterial development in benthic habitats and toxicity outbreaks, although our simplified conditions excluded zooplankton and invertebrate predators that would also shape the algal structure and relative abundance of each taxa. In any case, MALDI-TOF analysis of mature biofilms cultured in this study did not evidence bioactive compounds. Occasionally, loosely attached filamentous phototrophs, streamers, were observed floating in flow direction in biofilms at the mature phase. These structures could reflect an increase in surface area and a reduction in surface-boundary layer and secondly could be a mode of dispersal as these small communities could have better chance of survival when inhabiting a new environment (Figure 11a,b). This study contributed to describing the fate of an O. cf. ovata BHAB on artificial substrata under microcosm simplified, stable environmental conditions. It is often difficult to investigate biofilms in nature because of the large range of biotic and abiotic interactions that occur, the consequent ambiguity of experimental conditions, and the impossibility to maintain biofilm integrity during sampling procedures. Results indicated that at the low disturbance conditions tested, life-form strategies, competitiveness for resources, and possibly allelopathic interactions shaped the structure of the cultures during the biofilm growth. Toxin production did not occur in culture, although toxins were recorded in the inocula used for the experiments (0.137, 0.049, and 0.044 pg cell −1 of ovatoxin a in Run 1, Run 2, and Run 3, respectively). Phototrophs with a filamentous growth habit and surface-associated motility, such as tube-dwelling diatoms and cyanobacteria (thin, light orientated, trichomes), out-competed flagellated life-forms. Exploitation of more favourable compartments was obvious in development of streamers floating in the flow direction. These accounted for an increase in surface area and could be a favourable mode of dispersal. In the hydrodynamic conditions tested, O. cf. ovata, possibly relying on vertical migrations to reduce surface-boundary layers and resource limitation, was limited in its vegetative growth. This study contributed to describing the fate of an O. cf. ovata BHAB on artificial substrata under microcosm simplified, stable environmental conditions. It is often difficult to investigate biofilms in nature because of the large range of biotic and abiotic interactions that occur, the consequent ambiguity of experimental conditions, and the impossibility to maintain biofilm integrity during sampling procedures. Results indicated that at the low disturbance conditions tested, life-form strategies, competitiveness for resources, and possibly allelopathic interactions shaped the structure of the cultures during the biofilm growth. Toxin production did not occur in culture, although toxins were recorded in the inocula used for the experiments (0.137, 0.049, and 0.044 pg cell −1 of ovatoxin a in Run 1, Run 2, and Run 3, respectively). Phototrophs with a filamentous growth habit and surface-associated motility, such as tube-dwelling diatoms and cyanobacteria (thin, light orientated, trichomes), out-competed flagellated life-forms. Exploitation of more favourable compartments was obvious in development of streamers floating in the flow direction. These accounted for an increase in surface area and could be a favourable mode of dispersal. In the hydrodynamic conditions tested, O. cf. ovata, possibly relying on vertical migrations to reduce surface-boundary layers and resource limitation, was limited in its vegetative growth.