Francisella tularensis subsp. holarctica wild-type is able to colonize natural aquatic ex vivo biofilms


 Biofilms are a matrix-associated lifestyle of microbial communities, often enabling survivability and persistence of such bacteria. The objective of this study was to investigate the survival of the wild-type strain A-271 of Francisella tularensis subsp. holarctica (Fth) in a natural aquatic ex vivo biofilm. To that purpose, we allowed Fth A-271 to produce its own biofilm on solid surfaces but also to colonize naturally formed biofilms from aquatic habitats, which were infected with Francisella in the laboratory. The survival rates of the bacteria in biofilms were compared to those of planktonic bacteria as a function of the employed culture condition. It could be shown by light- and electron microscopy that Fth is able to form a complex, matrix-associated biofilm. The biofilm form of Francisella showed longer cultivability on agar plates in natural water when compared to planktonic (free-living) bacteria. Be it as a part of the existing ex vivo biofilm or free-floating above as planktonic bacteria, more than 80% of Francisella were not only able to survive under these conditions for 28 days, but even managed to establish microcolonies and areas with their own exclusive biofilm architecture within the ex vivo biofilm. Here, we can demonstrate for the first time that a Francisella tularensis wild-type strain (Type B) is able to successfully colonize an aquatic multi-species ex vivo biofilm. It is worthwhile to speculate that Fth might become more persistent in the environment when it forms its own biofilm or integrates in an existing one. Multi-species biofilms have been shown to be more resistant against stress compared to single-species biofilms. This may have an important impact on the long-term survival of Francisella in aquatic habitats and infection cycles in nature.


In contrast to Ftt, Fth is more frequently associated with aquatic habitats and is widely distributed throughout Eurasia (Larson et al., 1955;Oyston et al., 2004;Sjostedt, 2007). F. tularensis maintains viability in cold water for long periods of time and it was hypothesized that the aquatic enviroment could serve as a reservoir for F. tularensis (Parker et al., 1951;Forsman et al., 2000;Sinclair et al., 2008;Telford and Goethert, 2010;Gilbert and Rose, 2012;Telford and Goethert, 2020). The main habitat of F. tularensis in the aquatic reservoir is still unknown and so far it is unclear, if F. tularensis is able to multiply within aquatic (bacteria grazing) protozoa (Abd et al., 2003;Thelaus et al., 2009;Buse et al., 2017). However, it is well-known that in (aquatic) natural environments, biofilm formation increases the survival of bacteria. The aquatic habitat-associated species F. novicida and F. philomiragia are well-known to form biofilms Margolis et al., 2010;Verhoeven et al., 2010;Van Hoek, 2013;Hennebique et al., 2019;Siebert et al., 2020) and recently it was published that Type A and Type B isolates of F. tularensis are also able to form biofilms (Champion et al., 2019;Golovliov et al., 2021;Mlynek et al., 2021). Biofilm formation is influenced by the pH, by phase variation of LPS and capsule (Champion et al., 2019;Mlynek et al., 2021), by stress/ppGpp/relA (Dean et al., 2009;Zogaj et al., 2012), by the two-component system qseC/qseB and BfpR Dean et al., 2020) and by chitinases and antibiotic susceptibility (Chung et al., 2014;Dean et al., 2015;Biot et al., 2020;Dean et al., 2020). Furthermore, it has been demonstrated that multi-species biofilms are more resistant against stress compared to single-species biofilms (Joshi et al., 2021;Wicaksono et al., 2022) -a property that has not been investigated so far for F. tularensis.
Despite of all these attempts, we are still at the beginning to understand the role of Francisella's biofilm formation in its natural environment. We here could demonstrate that a Fth wild-type (WT) strain, isolated from a beaver deceased from tularemia, is able to form a matrix-associated biofilm. In addition, we can show for the first time that Fth is able to successfully colonize an aquatic multi-species ex vivo biofilm.
Plasmid pMP814GFP was generated by infusion cloning. The GroES promoter-Gfp gene construct was introduced from pkk289Km (Broms et al., 2011) into the PstI site of pMP814 (LoVullo et al., 2009) by In-Fusion™ assembly with the Clontech In-Fusion HD Cloning Kit 1 according to the manufacturer's instructions.
Naegleria gruberi was cultivated in PYNFH medium (ATCC 1034) at room temperature.The co-culture were performed in an infection buffer (Ac buffer) (for 1 l,

Survival of Fth WT in infection buffer and in co-culture with amoebae
Naegleria gruberi was mechanically detached from the flask bottom, transferred to 50 ml tubes and centrifuged at 800× g for 10 min. After cell counting in a Neubauer counting chamber (C-Chip, Digital-Bio, DHC-N01), amoebae were adjusted to 5 × 10 5 cells ml −1 in Ac buffer, transferred into wells of a 24-well plate (1 ml/well) and incubated for 2 h at 25°C. Bacteria grown 3 days on MTKH-plates, resuspended in Ac buffer and adjusted to OD 600 = 1 (corresponding to ~10 9 CFU ml −1 ). The infection was done with a multiplicity of infection (MOI) of 10 and incubated for up to 28 days at 25°C. In parallel, the same number of bacteria were transferred into the wells of a second 24-well plate containing 1 ml Ac buffer per well, but without amoebae. For colony-forming unit (CFU) determination at various time points, bacteria were harvested by mechanical disruption of the amoebae or directly plated onto agar plates. The agar plates were incubated at 37°C, 5% CO2. Grown bacterial colonies were counted with a colony counter (aCOLyte, Synbiosis, Cambridge, UK) and calculated as bacteria per ml.

Biofilm formation
Biofilm formation at a solid-gas interface An over-night culture of Fth LVS and Fth A-271 in medium T was adjusted to OD 600 = 1. 10 μl (approximately 100 bacteria) were spotted onto each polyester filter (Ø 13 mm, PE02CP01300; Pieper Filter GmbH, Bad Zwischenahn, Germany) located on a MTKH agar plate. This agar plate was incubated at 37°C and 5% CO 2 for up to 3 days. Every day, one filter was removed and fixed in 4% paraformaldehyde at 4°C for 48 h. For scanning electron microcopy In a second set of experiments, the biofilm was generated on Thermanox coverslips (Nunc™ Thermanox™ coverslips, Ø13mm). Fth strain A-271 grown on MTKH agar plates at 37°C and 5% CO 2 for 2 days were used to start a culture (25 ml medium T in a 100 ml flask). After 7 h of incubation at 37°C and 250 rpm, the culture was adjusted to OD 600 = 0.3. Each well of a 12-well plate containing a Thermanox coverslip was filled with 1 ml of the diluted bacterial suspension and incubated without shaking at 37°C and 5% CO 2 for 2 days. From day three the plates were shaken at 100 rpm without additional CO 2 for two weeks, each third day the coverslip was transferred to a new 12-well plate cotaining fresh medium T (1 ml each well). Then the coverslips were transferred into 6-well plates with 2 ml medium per well. During the next 2 weeks, every third day half of the medium T (1 ml) was replaced by fresh medium. This biofilm was then used either for scanning electron microscopy or for the survival experiments in natural water.

Survival of Fth (biofilm and planktonic) in natural water
The Thermanox coverslip with biofilm (see above, 3 days after last replacement with fresh medium T) was removed into a new well and the remaining supernatant was defined as the planktonic bacterial fraction. The coverslip was washed once and filled with 1 ml sterilized natural water (SNW). The supernatant was diluted 1:1000 in SNW. The biofilm as well as the planktonic bacteria were then incubated for a further 28 days at RT (approx. 22.5°C). At day 0, 2, 7, 15, 21 and 28 both bacterial cell types were treated as follows: The biofilm was removed from the coverslip using a cell culture scraper. The whole liquid was then transferred into a 1.5 ml tube and vortexed three times for 45 s. Planctonic cells were also transferred into 1.5 ml tubes and vortexed. Dilutions of all bacterial suspensions were used for viability testing (Live-Dead staining) or plated out onto MTKH agar plates to determine the number of CFU.
In order to gauge the biofilm's viability, the Thermanox™ coverslips with Fth biofilm were stained with the LIVE/DEAD™ BacLight Bacterial Viability kit following the manufacturer's instructions, and imaged in the CLSM (LSM 780, Carl Zeiss Microscopy, Oberkochen, Germany). Five representative spots of the Thermanox™ coverslips were chosen (center, 3, 6, 9 and 12 o' clock) and images of a standardized area were taken with the 20x NA0.8 objective. The bacteria in the individual green/red channels were counted using ImageJ 2 (U. S. National Institutes of Health, Bethesda, MD, USA) and recalculated to the area of a Thermanox™ coverslip.
As for the viability of planktonic bacteria, 10 μl of the supernatant were placed in each slot after LIVE/DEAD™ staining of a disposable Neubauer counting chamber (C-Chip™, DigitalBio, NanoEnTek). For safety reasons, the slots were sealed with nail polish (Maybeline, express finish 40 s). Standardized imaging was done using the corner squares (DHC-F01 grid pattern) of each side of the counting chamber. The bacteria in the individual green/red channels were counted using ImageJ (U. S. National Institutes of Health, Bethesda, Maryland, USA, see Footnote 2) and recalculated to one milliliter.

Natural ex vivo biofilm model
In order to obtain natural ex vivo biofilms from an environmental aquatic habitat, 5-10 Thermanox™ coverslips were fixed with a distance of approx. 5 cm along a thread. These devices were placed 20-30 cm beneath the water plane of a ditch near Berlin, Germany (Coordinates: 52°, 25 min, 17,84 s north, 13°, 44 min, 58,64 s east) at different seasons (September, November, April and June). The devices were removed after 4 weeks when the coverslips showed a visible biofilm on the surface and transported to the laboratory in a waterfilled bottle. Single coverslips were rinsed with 5 ml SNW from the same water souce wherefrom the biofilm was obtained. The coverslips were placed in wells of a 24-wellplate and covered with 0.2 ml SNW. The cultures were infected with 0.2 ml 10 8 Fth A-271/pMP814GFP or as indicated in SNW on day 0. The cultures were kept at room temperature or at 4°C for the period of time as indicated. The number of colony-forming cells and of genome copies by multiplex real-time PCR targeting tul4 (lpnA gene) were determined after scraping the cells from the coverslips as previously described . Suspensions were plated onto selective "Neisseria" agar plates (see strains and growth conditions). GFP-positive clones were counted under UV light and a subset of these bacteria were verified as Fth by qPCR analysis. For microscopic analyses, the coverslips were fixed in 4% paraformaldehyde at 4°C for 48 h and processed as described in the method section (biofilm formation).

Image processing
Images have been cropped and adjusted for optimal brightness and contrast (applied to the entire image) using Adobe Photoshop ® (Adobe Systems, San Jose, CA, USA). The image D of Figure 5B was colorized with Adobe Photoshop ® .

Results
Francisella strain A-271 is a virulent wild-type strain of the subspecies Fth

Intracellular replication and virulence
Here, we investigated the Fth wild-type (WT) strain A-271, isolated from the carcass of a beaver deceased from tularemia, and without subsequent serial passaging on agar plates (Schulze et al., 2016). The complete genome sequence of strain A-271 has been determined and displays genes for all classical virulence factors of Fth (Sundell et al., 2020). Phylogenetically, this isolate belongs to the B.12 subclade (biovar Frontiers in Microbiology 04 frontiersin.org II, erythromycin resistant) of Ft. subsp. holarctica. Strains of biovar II are frequently isolated also from tularemia patients in Germany (Appelt et al., 2019). In addition, we investigated strain A-271 for intracellular replication in the macrophage-like U937 and J774.1 cells (Supplementary Figure S1), a key factor of Fth virulence mainly determined by the type six secretion system (T6SS). Survival of Fth strain A-271 in buffer and in co-culture with amoebae at 25 and 4°C.
In aquatic habitats the cultiviabilty of Francisella is increased at lower temperatures and in the presence of amoebae. We investigated if our WT strain Fth A-271 exhibits a similar phenotype. As expected, when Fth A-271 was incubated in Ac buffer, CFU was stable over a period of 22 days at 4°C, whereas at 25°C, no further bacterial growth (CFU) could be detected on agar plates after ≥14 days of incubation (Supplementary Figure S2A).
Furthermore, we checked the survival of Fth WT strain in Ac buffer, incubated at 4°C or 25°C in the presence of the amoebae N. gruberi. At 4°C the CFU values of Fth A-271 were stable during the whole experiment, whereas at 25°C no CFU could be detected after 22 to 28 days of incubation (Supplementary Figure S2B). Thus, in the presence of N. gruberi the strain was culturable 1 to 2 weeks longer than without amoebae (Supplementary Figures S2A, B). The obtained results indicated that the culturability of our Fth WT strain A-271 was increased at lower temperatures and in co-culture with amoebae as previously described for other (virulent) Francisella strains (see above).
A number of questions arose, firstly if our Fth WT strain would be able to form its own biofilm, then whether biofilm formation might be related to enhanced survival in the environment, and finally if A-271 is able to colonize natural aquatic biofilms.

Biofilm formed by in vitro grown Fth A-271 WT strain Biofilm structure
When Fth strain A-271 was grown in vitro for 72 h at 37°C with 5% CO 2 on agar plates on a polyester filter, SEM analysis revealed the presence of a macrocolony biofilm, in which the bacteria were densely embedded in a luxurious extracellular matrix ( Figure 1A). After incubation of Fth on coverslips (Thermanox™) in the presence of medium T for four weeks, a similar biofilm structure, although with less biofilm matrix was observed ( Figure 1B). We could distinguish micro-and macro-colonies as well as complex 3D biofilm structures. CLSM analysis of this biofilm demonstrated that the matrix was composed of proteins, carbohydrates and lipids ( Figure 1C) and also long strands of plaited eDNA, with which Francisella bacteria were associated ( Figure 1D). Similar biofilm structures were also observed using Fth strain LVS or F. sp. strain W12-1067 (data not shown).

A-271 in natural water
Fth biofilm (grown on a Thermanox™ coverslip, see above) was then introduced into SNW and incubated at RT for several weeks. On day 0, 2, 7, 15, 21 and 28 the complete biofilm was scraped from the coverslip and planktonic and biofilm bacteria were analyzed for CFU and viability (see methods). Culturability of Fth A-271 was measured by CFU on agar plates and viability by Live/Dead staining. Up to 2 days after the release of the biofilm into natural water approx. 85% of the bacteria were alive and the majority was culturable on agar plates  Figure 2A). Whereas the number of living bacteria, as detected with Live/Dead™ staining, did not change significantly over the whole period of the experiment, the number of culturable bacteria (CFU) decreased between week two and three to zero. Similar results were obtained with the planktonic form of Fth WT, with the exception that the loss of culturability occurred already between week one to two ( Figure 2B). This suggests that planktonic and biofilm forms of Fth A-271 survive to a similar amount on the way to a VBNC-like status, even though bacteria in the biofilm status showed a slightly increased culturability over time.
Next, we asked, if the Fth WT strain may be able to colonize an existing natural biofilm and survive within this natural habitat.

Fth wild-type is able to colonize a natural multi-species ex vivo biofilm (microcosm)
In order to test if Fth WT is able to colonize natural biofilms under aquatic habitat conditions, we established a natural biofilm on a Thermanox™ coverslip to use this biofilm in ex vivo experiments. For 4 weeks, we placed coverslips 20-30 cm beneath the water plane at different seasons in natural aquatic habitats near Berlin (see Materials and methods; Figure 3A). The coverslips were overgrown with biofilms visible with the naked eye ( Figure 3B). CLSM imaging of the ex vivo biofilm was performed to visually characterize the microbial composition of this biofilm. Microscopy revealed the presence of a bacterial biofilm layer at the bottom (first colonizers), as well as different amoebae and protozoa ( Figures 3C-I). For example, we found a wide variety of diatoms ( Figures 3C, D)  Altogether the results suggested that an aquatic biofilm was successfully established on the Thermanox™ coverslips during the incubation in the natural aquatic habitat.
Then these coverslips were infected with 10 7-8 GFP-labelled Fth A-271 bacteria and investigated the amount of culturable Fth cells in co-culture with this biofilm (in autoclaved natural water) for up to 63 days. Results are given in Figure 4. As in water, the time span of culturability of the bacteria was decreased at RT compared to 4°C ( Figure 4A). The culturability of Fth at 4°C in the presence of the biofilm was slightly increased compared to Fth in water ( Figure 4B) and the detectable genomic units were only reduced two to three log 10 levels over the whole period of incubation ( Figure 4B).
When the same experiment -ex vivo biofilm, co-incubated with Fth A-271 at 4°C -was imaged with the CLSM, Fth bacteria after 10 days of incubation were most commonly visible as single bacteria, micro-and macrocolonies in the bacterial bottom layer of the ex vivo biofilm ( Figure 5A, November-biofilm). On few occasions, Fth had even formed their own comprehensive, three-dimensional biofilm structure ( Figure 5B, July-biofilm, 10 th day of incubation). More frequently, Fth bacteria were found in amoebae ( Figure 5C, November-biofilm, 5th day of incubation), or (presumably) about to be taken up by one ( Figures 5D, E) (September, 5th day of incubation). Regularly, Fth was also spotted in a wide variety of ciliates ( Figures 5E, F) (November, 5th day of incubation), in particular in Vorticella spp. (Figures 5G, H) (November, 10th day of incubation) as single bacteria or clustered in vacuoles ( Figures 5F-H).
Taken all together, the results clearly demonstrated that Fth is able to successfully colonize different niches of an existing natural biofilm by forming macro-colonies, but also complex biofilm structures of their own.

Discussion
Many studies about Francisella were performed with the Fth strain LVS, which is a virulence attenuated strain and its laboratory handling is much easier. Although strain LVS is used as a surrogate for F. tularensis, this strain is not fully virulent and may not behave as a Fth WT strain (Rohmer et al., 2006;Biot et al., 2020;Mlynek et al., 2021). Thus, we used the Fth WT strain A-271, isolated from a carcass of a beaver deceased from tularemia, in our experiments (Schulze et al., 2016;Sundell et al., 2020). Here we demonstrated that this strain is a virulent Fth WT strain and able to replicate in macrophage-like cell lines. In addition, in silico analysis of the genome of this strain confirms the presence of all common virulence factors of Fth (Sundell et al., 2020). Furthermore, using this isolate (Fth A-271), we corroborate earlier findings showing that the culturability of Francisella over time in water is improved at lower (4°C) temperatures rather than at higher ones (RT, 22.5°C). Furthermore, using the amoeba Naegleria gruberi  we here can confirm prior publications, demonstrating that Francisella is not able to multiply in amoebae (Acanthamoeba castellanii, A. polyphaga, Vermamoeba vermiformis) (Buse et al., 2017;Hennebique et al., 2021). However, lower temperatures and the presence of amoeba increased the survival and the long-term culturability of the bacterium (Abd et al., 2003;El-Etr et al., 2009;Duodu and Colquhoun, 2010;Verhoeven et al., 2010;Gilbert and Rose, 2012;Ozanic et al., 2016;Buse et al., 2017;Golovliov et al., 2021;Hennebique et al., 2021). Therefore, we chose Fth A-271 WT strain to investigate its ability to form biofilms and to colonize and survive in a natural aquatic multispecies biofilm.

Biofilm formation
Strain Fth A-271 grown on agar plates or in medium T on Thermanox™ coverslips was able to form micro-and macro-colonies, as well as a 3D biofilm structure with a large amount of matrix material, containing strands of eDNA, carbohydrates, proteins and lipids. These results demonstrate the ability of this strain to form its own biofilm ( Figures 1A, C). Investigating the survival of bacteria in natural water, the experiments demonstrated that the survival of bacteria in biofilms and planktonic bacteria seem to be similar, but the time-span of culturability of biofilm bacteria was increased Frontiers in Microbiology 07 frontiersin.org compared to their corresponding planktonic form (Figure 2). This indicates that bacteria within the biofilm stay culturable for a longer time as also shown for F. novicida and F. philomiragia or for Francisella in co-culture with amoebae Ozanic et al., 2016;Buse et al., 2017;Siebert et al., 2020;Hennebique et al., 2021;Mlynek et al., 2021). The results suggest that biofilm formation may also enhance Francisella persistence within an aquatic habitat.
Colonization of a natural aquatic biofilm (microcosm) We mentioned above that Fth WT strain A-271 in our experiment slowly lost its culturability and that they formed cells in a VBNC-like state (Figure 4). VBNC has been described for Francisella, but to our knowledge, the ability to resuscitate these forms has not been published for Francisella so far (Forsman et al., 2000;Thelaus et al., 2009;Duodu and Colquhoun, 2010;Backman et al., 2015). For Legionella pneumophila in contrast, it has been shown that resuscitation of bacteria in the VBNC status is possible in amoebae (Steinert et al., 1997;Epalle et al., 2015). However, Francisella is able to survive and persist over long time periods in natural aquatic environments in a temperature-dependent manner (Parker et al., 1951;Forsman et al., 2000;Sinclair et al., 2008;Berrada and Telford Iii, 2011;Broman et al., 2011;Gilbert and Rose, 2012;Golovliov et al., 2021).
Furthermore, it was demonstrated that co-cultures of Francisella and amoebae increase the culturability but not the virulence of Francisella, as Francisella is not able to multiply within amoebae (Abd et al., 2003;El-Etr et al., 2009;Verhoeven et al., 2010;Ozanic et al., 2016;Buse et al., 2017;Hennebique et al., 2021). In addition, microcosm experiments revealed that Francisella did not survive the grazing of the ciliate Tetrahymena pyriformis, while a nanoflagellate was found to favor Fth survival (Thelaus et al., 2009). In addition, F. noatunensis survival was shown to be higher in sterile than in nonsterile microcosms (Duodu and Colquhoun, 2010) and survival seems to be dependent on the number and species of bacterial grazing protozoa (Mironchuk Iu and Mazepa, 2002;Thelaus et al., 2009;Duodu and Colquhoun, 2010); and own unpublished results). Thus, the survival of Francisella in the aquatic habitat is influenced by a high amount of different biotic and abiotic factors and therefore, we investigated the ability of Fth to colonize a multispecies biofilm.
It has been demonstrated that multi-species biofilms are more resistant against stress compared to single-species biofilms (Joshi et al., 2021;Wicaksono et al., 2022), but so far this has not been shown for Francisella. As mentioned above, the strain Fth A-271 was isolated from a dead beaver living in an aquatic habitat. Thus we analyzed the ability of this strain to survive and to colonize a natural aquatic multi-species biofilm, in a microcosm-like experiment. The coverslips were incubated for 4 weeks in a natural aquatic habitat at different seasons. We characterized these biofilms as a multi-species biofilm, exhibiting diatoms, Vorticellidae, Rotifera, as well as different amoebae and bacteria. Interestingly, these biofilms could be colonized by our Fth WT strain and we could observe micro-and macro-colonies, as well as mature biofilm-like structures of Fth within the natural biofilm. Furthermore, as demonstrated for the co-culture with amoebae, the presence of the natural biofilm increased the culturability of Fth; and the effect was found also to be temperature-dependent (4°C > RT). Recently, research started to investigate general interspecies interactions (e.g., an biofilm adapted metabolism) within multi-species biofilms (Joshi et al., 2021;Wicaksono et al., 2022) which may be also important for the persistence of Francisella in the environment.
In this study, we demonstrated that Fth strains are able to successfully colonize natural-like aquatic multi-species biofilms, a behavior which seems to be important for their observed long-term survival within aquatic habitats.
To our knowledge, we are the first to investigate if a Fth WT strain is able to colonize and survive within a natural aquatic multi-species biofilm. We can demonstrate that amoebae, biofilm formation and low temperatures increase the culturability (survival) of Fth in the aquatic environment and that Fth is, indeed, able to successfully colonize a multi-species biofilm. This may have impact on the longterm survival of Francisella in aquatic habitats and should be further investigated. (C) The same experiment as shown in (B), but detection of Fth DNA in biofilm, in planktonic supernatant from the biofilm, and in natural water without previous biofilm formation (control). The Thermanox™ coverslips overgrown with or without biofilm in NSW were infected with 4×10 7 F. tularensis on day 0. At several time points as indicated, CFU and GE were determined from the NSW or the supernatant of the coverslips, and of the scraped biofilm (see Material and Methods).

Data availability statement
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation. Visualization of Fth' survival in natural ex vivo biofilm. (A) (CLSM image) GFP expressing Fth formed microcolonies (blue-greenish signal) within the natural biofilm (DAPI, dark blue) -diatom: yellowish autofluorescence). (B) (CLSM image) Fth formed a comprehensive, 3D biofilm (green) within the natural biofilm -diatoms: red autofluorescence. (C) (CLSM image) In the natural biofilm, Fth bacteria (red) were found ingested by amoebae (bright blue). (D) (SEM image) A bacterium, with the characteristic shape of Fth (arrow, compare with Figure 2B), is about to be ingested by an amoeba (bright blue). This scene enfolds between matrix (green) embedded bacteria (dark red) and various diatoms (ocker and orange). (E,F) (CLSM image) GFP expressing Fth bacteria within different types of ciliates (arrows). (G,H) Many Vorticella (red autofluorescence) in the natural biofilm showed accumulated Fth, which formed dense clusters (green).