Crenarchaeal Biofilm Formation under Extreme Conditions

Background Biofilm formation has been studied in much detail for a variety of bacterial species, as it plays a major role in the pathogenicity of bacteria. However, only limited information is available for the development of archaeal communities that are frequently found in many natural environments. Methodology We have analyzed biofilm formation in three closely related hyperthermophilic crenarchaeotes: Sulfolobus acidocaldarius, S. solfataricus and S. tokodaii. We established a microtitre plate assay adapted to high temperatures to determine how pH and temperature influence biofilm formation in these organisms. Biofilm analysis by confocal laser scanning microscopy demonstrated that the three strains form very different communities ranging from simple carpet-like structures in S. solfataricus to high density tower-like structures in S. acidocaldarius in static systems. Lectin staining indicated that all three strains produced extracellular polysaccharides containing glucose, galactose, mannose and N-acetylglucosamine once biofilm formation was initiated. While flagella mutants had no phenotype in two days old static biofilms of S. solfataricus, a UV-induced pili deletion mutant showed decreased attachment of cells. Conclusion The study gives first insights into formation and development of crenarchaeal biofilms in extreme environments.


Introduction
In nature, most microbes are assumed to exist predominantly in surface-associated communities, encased in a self-produced matrix, termed biofilms [1,2,3]. Thus, the formation of biofilms reflects the native growth conditions for most microbial species. Cells within biofilms differ substantially from their planktonic counterparts, particularly with regard to an increased resistance towards numerous environmental perturbations. Thus, the mechanism of biofilm formation and its importance for microbial survival in natural habitats has attracted increasing interest in recent years.
To date, studies on microbial biofilms have mainly been conducted on bacteria, in particular with regard to pathogenic species in which biofilms play an important role in disease development [4,5]. In sharp contrast, for the archaeal domain only very limited information is available on this topic. Archaea are frequently detected in biofilm communities from many different environments [6,7], but biofilm formation by archaea has only been sparsely studied. So far, all studies have dealt with the formation of biofilms by euryarchaeotes. The first archaeal biofilm was described for the hyperthermophilic Thermococcus litoralis. The T. litoralis biofilm developed in rich media on hydrophilic surfaces, such as polycarbonate filters, and was accompanied by mannose-type extracellular polysaccharides production [8]. Archaeoglobus fulgidus biofilm formation, measured as attachment to the sides of cultivation vessels, was found to be increased in response to unfavorable environmental conditions, including high metal concentrations, pH and temperature changes [9]. Upon adhesion to (a)biotic surfaces, mediated by flagella or pili, Pyrococcus furiosus and Methanobacter thermoautotrophicus formed monospecies biofilms, respectively [10,11]. Development of P. furiosus and Methanopyrus kandlerii bi-species biofilms was shown to be established within less than 24 hours on biotic surfaces [12].However, the formation of a layered biofilm was dependent on the initial colonization of the surface by M. thermoautotrophicus cells to which P. furisous could adhere by using its flagella and establishing cell-to-cell contacts. Very recently, two distinct biofilm morphologies were described in the extremely acidophilic euryarchaeote Ferroplasma acidarmanus Fer1, a multilayered film forming on glass and pyrite surfaces and up to 5 mm-long filaments that were also found in natural environments [13]. Proteomic studies on these biofilms showed that 6 out of the 10 up-regulated proteins were involved in the adaptation to anaerobic growth indicating anaerobic zones in the multilayered Ferroplasma biofilms.
In this study, we use the crenarchaeal model organism Sulfolobus spp. to initiate comprehensive studies on archaeal biofilms. Sulfolobus species are hyperthermoacidophiles growing optimally at 70-85uC and pH 2-3 that are found worldwide in geothermically active environments such as solfataric fields. They express a variety of surface structures including flagella and type IV-like pili [14,15,16] which have been shown to be involved in motility and UV light-induced cell aggregation [16,17]. A recent study indicated that flagella and pili are also essential for initial surface attachment [18]. The same study demonstrated that Sulfolobus can attach to a variety of surfaces including glass, mica, pyrite and gold coated carbon grids. An initiation of microcolony formation by the attached cells was observed, indicating that Sulfolobus may be able to develop into structured microbial communities reminiscent to that of many eubacteria.
To further assess the ability to form biofilms, S. solfataricus, an European isolate from Italy [19], S. acidocaldarius, originally isolated from Yellowstone National Park [20] and S. tokodaii, an isolate from Japan [21], were chosen for a comparative study. Using an adapted microtitre plate assay, the impact of multiple environmental conditions on biofilm formation by these three Sulfolobus species was tested. Confocal laser scanning microscopy (CLSM) was employed to study the formed microbial communities in detail. We demonstrate that all three Sulfolobus species develop into distinct three-dimensional communities. The adapted methods will enable further detailed studies on how archaeal biofilms are formed and how their structures develop.

Strains and growth conditions
Sulfolobus solfataricus P2 (DSM1617), S. acidocaldarius (DSM639), S. tokodaii (DSM16993), S. solfataricus PBL2025 [22], flagella deletion mutant DflaJ [16] and the ups pili deletion mutant DupsE [17] were grown in Brock medium at 76uC, pH adjusted to 3 using sulphuric acid, and supplemented with 0.1% w/v tryptone [20]. For biofilm formation, cultures were inoculated in standing Petri dishes and grown for 2-3 days at 76uC in a metal box which was supplemented with a small amount of water to minimize evaporation of the media. For these assays Sulfolobus strains were inoculated with specific starting OD 600 of 0.03 for S. solfataricus, 0.01 for S. acidocaldarius and 0.06 for S. tokodaii.

Microtitre plate assay
The assay was performed in polystyrol 96-well tissue culture plates (flat bottom cell+, Sarstedt) to screen for the efficiency of biofilm formation under different environmental conditions. To avoid evaporation of the medium, the plates were covered with a gas-permeable sealing membrane (Breathe-Easy, Diversified Biotech, Boston, USA). After two days incubation the microtitre plate was cooled down to room temperature and the OD 600 of cell cultures from each well was measured using a luminometer (InfiniteM200, TECAN, Switzerland) at a wavelength of 600 nm. 10 ml of a 0.5% solution of crystal violet (CV) was added and incubated at room temperature for 10 minutes. Subsequently, the liquid supernatant was removed from each well and the biofilm cells attached to the well were washed with water. 100% ethanol was added to release the crystal violet from the biofilm. The absorbance of crystal violet from each well was measured at a wavelength of 570 nm. The efficiency of biofilm formation was determined by the correlation between the growth of the cells (OD 600 nm ) and the absorbance of crystal violet (OD 570 nm ).
To determine how much biomass was present as biofilm, biofilms were grown and either resuspended by prolonged vortexing to obtain the OD 600 nm , or stained with crystal violet to obtain the corresponding OD 570 nm values. This relation was used to calculate the percentage of cells within the biofilm related to the total amount of cells in biofilm and planktonic cells (see Fig.  S2 and Table S1).

Confocal laser scanning microscopy (CLSM)
For CLSM images the cells were grown for three days in uncoated plastic dishes (m-Dishes, 35 mm high; Ibidi, Martinsried). To prevent evaporation at the high incubation temperature, the lids of the dishes were closed. The medium was carefully exchanged every 24 hours to ensure aerobic growth conditions. Prior to confocal microscopy, the liquid supernatant of the biofilm, with the planktonic cells, was removed and 2 ml fresh medium was added. Images were recorded on an inverted TCS-SP5 confocal microscope (Leica, Bensheim, Germany).
DAPI (49,6-diamidino-2-phenylindole), dissolved in water to 300 mg/ml, was used to visualize the cells of the biofilm. 6 ml of the DAPI stock solution was added to the biofilm and incubated at room temperature for at least 10 minutes. Images were taken at an excitation wavelength of 345 nm and an emission wavelength of 455 nm.
Fluorescently labelled lectins were employed to visualize the EPS (extracellular polymeric substances) of the biofilms. Before adding lectins to the biofilm, the growth medium was replaced with medium adjusted to pH 5 to ensure that binding of lectins was not inhibited by low pH. Fluorescein-conjugated concavalin A (ConA) (5 mg/ml; Invitrogen, Karlsruhe, Germany), which binds to amannopyranosyl and a-glucopyranosyl residues, was dissolved in 20 mM sodium bicarbonate (pH 8) to a final concentration of 10 mg/ml. Fluorescein-conjugated ConA has an excitation wavelength of 494 nm and an emission wavelength of 518 nm.
The Alexa Fluor-conjugated lectins, which have an excitation wavelength of 591 nm and an emission wavelength of 618 nm, were used in concert with ConA. The lectin-biofilm mixtures were incubated at room temperature for 20-30 minutes in the absence of light. After incubation, the biofilm was washed with Brock media (pH 5) to remove excess label and images were taken by CSLM. Image data obtained were processed by using the IMARIS software package (Bitplane AG, Zürich, Switzerland).

Scanning electron microscopy
S. acidocaldarius was grown as a standing culture under the described biofilm conditions in Petri dishes with 30 ml brock media adjusted with 0.1% trypton together with polylysin treated glass coverslips. The cells were fixed with 2.5% glutaraldehyde and incubated for 5 min at room temperature. The coverslips were carefully removed and stored at 4uC in 24 well plates with PBSbuffer with 2,5% glutaraldehyde.
The samples were then postfixed with 1% osmium tetroxide for 1 h on ice. After washing the cells were dehydrated in ethanol and critical-point-dried from CO 2 . The samples were sputter-coated with 7 nm Au/Pd and examined at 20 kV accelerating voltage in an Hitachi S-800 field emission electron microscope.

Adaptation of microtitre plate assay
To enable rapid quantification of surface-attached biomass, we adapted the commonly used microtitre plate assay based on crystal violet binding [23] for use at high temperatures. To prevent evaporation of the medium it was essential to cover the plates with gas permeable sealing membranes. For incubation at 76uC the plates were placed into a metal container to further prevent evaporation of the medium. The requirements for adherence to abiotic surfaces can vary greatly among microorganisms, therefore, different plates with hydrophilic and hydrophobic surfaces were tested. All three strains, S. acidocaldarius, S. solfataricus and S. tokodaii, attached preferentially to hydrophilic surfaces at the well's walls (data not shown). The conditions for biofilm formation were further optimized for each of the strains. The amount of biomass detected after two days was strongly dependent on the starting OD 600 of the inoculum and differed for each strain (Fig. S1). Based on these results, for all subsequent experiments the starting OD for S. acidocaldarius was 0.01, for S. solfataricus was 0.03 and for S. solfataricus was 0.06. It was confirmed that the crystal violet values reflected the amount of biomass formed, as the obtained values correlated with the OD values measured from resuspended biofilm cells (data not shown). For the presentation of the microtitre plate assay results we show the correlation of the crystal violet release of the biofilm cells (OD 570 nm ) divided by the growth of the planktonic cells (OD 600 nm ) to emphasize the fraction of the cells which grow as biofilm under each condition in Fig. 1 and the absolute amount of surface-associated cells in Fig. S2.

Influence of pH, temperature, and iron concentration on Sulfolobus biofilm formation
Using the adapted microtitre plate assay the influence of a variety of conditions on the biofilm formation of the three Sulfolobus strains was tested. The pH values of a hot spring may be subject to change, for example by incoming rain or changes in the pH of fluid entering the hot spring. Therefore, the effect of pH values ranging from 2 to 7 on biofilm formation were evaluated. As expected, growth of all three strains was optimal around pH 3-4, but at pH 6 up to 80 and 70% of the total biomass of S. acidocaldarius and S. tokodaii, respectively, was present in biofilm ( Fig. 1, second column and Fig. S2, C). Based on the correlation between OD values and the amount of surface-attached biomass, it was evident that in both species biofilm formation protects cells against alkaline pH, as the optimum pH for biofilm formation was much higher than the growth optimum.
As the temperature in a hot spring may also be subject to rapid changes, biofilm formation in the microtitre plates were tested at temperatures ranging from 60-85uC. In the range from 65 to 80uC S. tokodaii formed equal amounts of biofilm, with decreased levels only observed at 60 and 85uC, although at 60uC the amount of cells present in biofilms was the highest (50%) (Fig. 1, first column and Fig. S2,B). In contrast, S. acidocaldarius and S. solfataricus displayed increased biofilm formation at both extremes of the temperature gradient; at 60uC S. acidocaldarius and S. solfataricus showed a 5-fold and 4-fold increase biofilm formation, respectively, when compared with the optimal growth temperature of 75uC.
The natural habitats of Sulfolobales are acidic geothermal springs which are rich in As, S and Fe [24,25]. In these springs hydrous ferric oxide (HFO) microbial mats are found which contain a variety of members of the Sulfolobales indicating that these microorganisms might play a role in mediating the oxidation of iron in these environments [24]. Therefore, the influence of iron concentration on biofilm formation was tested.
Whereas the normal iron concentration in the medium of Sulfolobales is 0.02 g/L, we tested 0.015 g/L to 0.065 g/L. Biofilm formation by S. acidocaldarius and S. tokodaii was not significantly influenced by the different concentrations of iron, but S. solfataricus displayed an optimum curve with the highest biofilm formation at 0.045 g/L (Fig. 1, third column and Fig. S2, D). When different pH values and iron conditions were combined, it was interesting to see that S. solfataricus was unable to resist the higher pH in the presence of high iron concentrations and, subsequently, biofilm formation was abolished. In contrast S. tokodaii and S. acidocaldarius biofilm formation was additionally stimulated (Fig. 1, last column and Fig.  S2, E). At pH 6 and 0.065 g/L iron, biofilm formation increased 4fold for S. tokodaii and 10-fold for S. acidocaldarius which compared with normal levels reaching 63 and 83% of cells, respectively, in biofilm compared to the total cell mass (Table S1 and Fig. S2, E).
In general, the amount of formed biofilm in microtitre plates is much less for S. solfataricus than for the other two species, most probably due to the more anaerobic conditions as compared to the static biofilm assay in Petri dishes.
Structural determination of static biofilms of the three Sulfolobus strains by confocal laser scanning electron microscopy All three Sulfolobus strains were inoculated in uncoated plastic mdishes and incubated without agitation at 76uC. Evaporation was prevented by placing the Petri dishes in a humidified metal box and the medium was carefully exchanged every 24 hrs with fresh, prewarmed medium to ensure nutrient and oxygen availability. After three days the formed biofilms were stained with DAPI, as described in the Materials and Methods section and analyzed by confocal laser scanning microscopy (CLSM). We employed DAPI staining to visualize cells as there is currently no fluorescent protein available that is stably expressed under the growth conditions of Sulfolobus spp.. S. solfataricus formed biofilms (20-30 mm thick) with a carpet like structure covering the whole surface of the Petri dish with a low density of cells (Fig. 2, middle column). The biofilm structure of S. tokodaii was 25-35 mm thick and also exhibited a carpet like structure but, in contrast to S. solfataricus, these had a high cell density and, occasionally, cell aggregates were visible (Fig. 3, overlay picture, last row). S. acidocaldarius readily formed biofilms (25-35 mm thick) which contained a high density of cells and large aggregates, forming towering structures above the surface of attached cells (Fig. 3, first row).
For bacteria it is well known that extracellular DNA can play an essential role in the formation and stabilization of the threedimensional structure of biofilms [26]. To examine whether extracellular DNA was present in biofilms of the three Sulfolobus strains, three days old biofilms were stained concomitantly with DAPI and the membrane-impermeable DNA-binding dye DDAO. In all three strains the DDAO signal was only present at locations in the biofilm where aggregates were also visible (Fig. 2, middle panels C and D). The weak DDAO signal was further reduced following DNase treatment indicating that the extracellular DNA was removed, but had no effect on the structure of the biofilms. Therefore, at this stage of the biofilm maturation, extracellular DNA does not appear to play a structural role in biofilms of these three Sulfolobus strains.
To estimate the amount of living and dead cells three days old biofilms were stained with the LIVE/DEAD stain. In S. solfataricus and S. tokodaii it was evident that throughout the biofilm less then ,2% of the cells were dead whereas in S. acidocaldarius up to ,8% of cells were dead (data not shown).

Analysis of Sulfolobus matrix components
The extracellular matrix that connects the cells and enables three-dimensional structuring of the communities is thought to be composed of (glyco)proteins, lipids, extrallular DNA (eDNA), and polymeric carbohydrates [27]. We therefore tested, whether eDNA, proteins, and polysaccharides play an important structural role in Sulfolobus biofilms. Experiments to inhibit biofilm formation by the addition of Proteinase K or DNase did not give conclusive results. Irrespective of the time at which DNase or Proteinase K was added, no effect on biofilm formation was observed although both enzymes showed enzyme activity under the conditions tested (data not shown).
Recently, it has been described that S. solfataricus cells, particularly when surface-attached, produce extracellular polymeric substances (EPS) containing glucose, mannose, galactose and N-acetyl-glucosamine [18,28]. To test whether cells also produce EPS during biofilm formation, three days maturated biofilms of all three Sulfolobus species were stained with DAPI, and two different fluorescently labeled lectins (Fig. 3). The lectins selected were concanavalin A (ConA), specific for terminal glucose and mannose residues and either IB4 (specific for galactosyl residues) or GSII (specific for N-acetylglucosamine residues). In all three strains it was observed that the ConA signal (green signal) corresponded to the DAPI signal (blue signal, Fig. 3). Sulfolobus cells are covered by an S-layer and is has been described that the S-layer proteins are glycosylated and can be purified by ConA affinity chromatography [29,30,31]. Whereas in S. solfataricus the ConA-derived signal did, in fact, completely overlap with the DAPI signal, in S. tokodaii and S. acidocaldarius GSII and IB4 lectin (yellow channel) stained matter was observed on top of the cell layer and, as no DAPI signal was found in this accumulated material, we concluded that these two strains secrete EPS. In both strains, these clouds of EPS also reacted with the other two lectins indicating the presence of not only mannose and glucose, but also galactose and N-acetylglucosamine (Fig. S3). In S. solfataricus, only marginal GSII-and IB 4 -mediated staining of cell attached sugar residues was observed, indicating a different glycosylation of extracellular proteins than in the other two strains. A detailed analysis of a biofilm formed by S. tokodaii and S. acidocaldarius showed extensive cell-cell connections. These connections became visible when the S. tokodaii sample was incubated with ConA and analysed by CLSM (Fig. 4). The connection might be a string of sugars or flagella/pili in which the subunits are glycosylated. In S. acidocaldarius and S. solfataricus biofilms the lectin GSII also stained thin connections between the cells (Fig. 4) which were clearly visible also in SEM pictures (Fig. 5D).

Maturation of S. acidocaldarius over a range of seven days
All experiments described so far were performed using 2-3 days old biofilms of Sulfolobus spp.. In order to monitor further community development under static conditions, biofilms of S. acidocaldarius were allowed to develop for seven days. Each day one sample was treated with DAPI and analyzed by CLSM. The thickness of the biofilm increased from 30 mM in height on day three to 150 mM (including EPS structures) on day seven (Fig. 5AB). For a more detailed analysis of the maturation of biofilm formation by S. acidocaldarius the cells were inoculated in large Petri dishes in which polylysine covered glass slides had been placed. These slides were then analyzed by scanning electron microscopy (SEM). Only 15 minutes after the addition of the cell suspension a few cells attached to the surface, and some budding of vesicles was visible (Fig. S4A). After two hours there was not an apparent increase in the number of attached cells, but nearly all attached cells had formed filamentous structures adhering the cells to the surface or neighboring cells (Fig. S4B). After 36 hours, microcolonies started to form with only a few cells remaining on the rest of the surface whereas after 48 hours the surface of the glass plates was completely covered with cells. In the microcolonies, cells appeared to be connected by a network of filamentous structures as was observed previously following lectin staining (Fig. 5D). These connections grew denser and also increasing extracellular material accumulated in the later stages of the biofilm formation (Fig. 5D). Interestingly, on the seventh day the layer of cells at the surface of the glass slide apparently disappeared and the density of cells in the detailed view was reduced compared with the sixth day ( Fig. 5C/D). To test whether the extracellular material visible in the closer SEM view of the towering structures did indeed consist of EPS, S. acidocaldarius biofilms were incubated for seven days, stained with lectins, as described above, and analyzed by CLSM (Fig. 6). Towering structures were formed which were initiated by the secretion of EPS and then colonized by cells.
The secretion of certain sugars apparently progressed in a sequential manner: initially the ConA (glucose and mannose) derived signal was much stronger, but in later stages both the GSII (galactose) and the IB4 (N-acetylgucosamine) signal increased ( Fig. 6 and Fig. S5), after which the secretion of mannose-rich sugars increases again as detected by ConA. This indicates that these sugars play an important role in biofilm maturation.

Role of surface appendages on static biofilm formation in S. solfataricus
Attachment of S. solfataricus cells from shaking cultures to different surfaces is mediated by flagella and UV induced pili. Deletion mutants in which either the flaJ gene, encoding the integral membrane protein of the flagella operon, or the uspE  gene, encoding the ATPase of the UV-induced pili operon were incapable of attachment to glass surfaces, gold-coated carbon grids or mica [18]. In bacterial biofilm development, filaments such as type IV pili and flagella have a strong effect on the dynamics of biofilm formation. Therefore, we tested the DflaJ and DupsE mutants and the wild type S. solfataricus PBL2025 strain for their ability to form static biofilms in three days. PBL2025 is a S. solfataricus strain which lacks 50 genes predicted to be involved in sugar metabolism and uptake and is the only currently available S. solfataricus strain which can be genetically manipulated [22].
The PBL2025, DflaJ and DupsE strains were grown in petri dishes and tested in the microtitre plate assay. After three days the matured biofilms in the petri dishes were stained with DAPI and analyzed by CLSM. The biofilms of PBL2025 and the DflaJ strain were comparable to three day old biofilms of S. solfataricus in height, density and structure, and showed mainly a carpet like appearance ( Fig. 7A and B). However, in biofilms from the DupsE strain, the surface of the petri dish was not as completely covered with cells and the biofilm was less dense as compared to PBL2025 and the DflaJ mutant strain. Furthermore, slightly more aggregates were visible in the biofilms of the DupsE strain. These observations were supported by the results of the microtitre plate assay showing that only the DupsE strain produced significantly less biofilm than S. solfataricus, PBL2025 and DflaJ. We therefore concluded that cell appendages do not to play an important role in the early development of static biofilm formation in S. solfataricus strains.

Discussion
It is well known that archaea and bacteria coexist in natural biofilms, playing essential roles in the Earth's biogeochemical cycles as well as in human disease [2]. The formation of bacterial biofilms has been very well documented. Studies have been carried out on euryarchaeal biofilm formation whereas we presented here the first detailed insights into crenarchaeal biofilm formation.
We chose the thermoacidophilic crenarchaeotes Sulfolobus spp as a model to establish methods for the analysis of hyperthermophilic archaeal biofilm formation. Sulfolobus spp. exist in acidic, mostly muddy, hot springs all over the world in which the hydrological dynamics result in rapid variations in temperature, pH and geochemical conditions. Therefore, these organisms must quickly adapt to these changing conditions or exist in a state that enables them remain undisturbed by such changes. As fully maturated biofilms protect their inhabitants from environmental disturbances, this form might be a way for Sulfolobus spp to survive in their habitats. The three selected strains were originally isolated from well separated geographical locations and each of the strains did, in fact, behave differently following the initiation of biofilm formation. Of the three strains, S. acidocaldarius most readily engaged in community formation either in microtitre plate assays or in static biofilm conditions when compared with the other Sulfolobus strains. In particular, when challenged with low temperature (60uC) or the combination of near neutral pH and low iron concentrations, S. acidocaldarius responded with highly increased biofilm formation demonstrating the ability of this strain to adapt to rapid changes in temperature and pH.
The maturation of bacterial biofilms proceeds via defined steps including initial attachment and further development into microcolonies secreting extrapolymeric substances [32]. During maturation multilayered biofilm structures are shaped and kept together by the secretion of EPS, extracellular DNA, and proteins [27,33]. Cells can be released from matured biofilms at any time point to proceed with a planktonic life style. Very recently, we have shown that S. solfataricus displays flagella and pili-dependent attachment to abiotic surfaces [18]. After two days of attachment to a glass slide in a shaking culture the cells started to produce EPS which contained glucose, mannose, galactose and N-acetylglucosamine demonstrating the first phase of biofilm formation. Similar to bacterial biofilm formation, it is evident that after initial attachment Sulfolobus cells start to form microcolonies that are surrounded by an extracellular matrix, containing EPS and, most probably, proteins. The function of EPS formation in these Sulfolobus strains may serve a variety of purposes. A natural deletion mutant of S. solfataricus which lacks 50 genes overproduced EPS when attached to a glass slide [18]. The deleted region contains genes mainly involved in sugar degradation and transport and these were shown to be upregulated in attached S. solfataricus cells, implying that they play an important role in the modulation of the EPS. One might speculate that the EPS is used as a carbohydrate reservoir which might also be the case when the cells are engaged in biofilm formation. When the S. acidocaldarius biofilm was incubated for seven days it was evident that the different sugars were produced in a consecutive manner implying that they may serve different purposes. Moreover, a layer of EPS was produced which enabled the formation of three-dimensional tower-like structures, especially in S. acidocaldarius. It appeared that after seven days the S. acidocaldarius biofilms detached by releasing attached cells to the planktonic phase.
In bacteria, cell appendages such as type IV pili and flagella are very well known for their influence and importance in the dynamics and development of static and hydrodynamically grown biofilms [34]. Like in Vibrio cholerae and Shewanella oneidensis MR-1, the S. solfataricus pili mutant DupsE exhibited decreased biofilm formation in the microtitre plate assay [35,36]. Also more dense aggregates were observed as in Pseudomonas aeruginoasa and S. oneidensis MR-1 type IV pili mutants [36,37]. However, the flagella mutant showed no obvious differences in static biofilm formation to the wild type, As the S. solfataricus flagella and pili mutant could not attach to several different surfaces in shaking culture [18], it will be interesting for future studies whether flagella and pili have a greater impact on biofilm formation in flow chamber systems.
Taken together, we demonstrated that Sulfolobus species can engage in biofilm formation and developed methods to study these in more detail. Of the three strains, S. acidocaldarius formed the largest amounts of biomass and was able to evade unfavorable conditions most successfully by choosing this life style. Interestingly, these data support the observation that S. acidocaldarius is mainly sampled from the crusts surrounding acidic hot springs and mud holes (Karl-Otto Stetter, personal communication) whereas S. solfataricus and S. tokodaii are primarily isolated from the midst of these types of hot springs, where the hot fluids are bubbling up to the surface [38] (Christa Schleper, personal communication).
In the future, it will be interesting to study how Sulfolobus strains behave in mixed biofilms and even in communities including other inhabitants of these acidic hot springs.

Supporting Information
Table S1 Found at: doi:10.1371/journal.pone.0014104.s001 (0.12 MB DOC) Figure S1 Optimization of inoculation conditions for biofilm formation of S. acidocaldarius, S. solfataricus and S. tokodaii. The strains were inoculated with different OD 600 and incubated in a microtitre plate for three days. The correlation of the measured crystal violet absorbance of the formed biofilm and the OD600 values of the planktonic cells is presented. Each bar represents the mean of 8 different samples.  Figure S3 CSLM analysis of three day old static biofilms of S. acidocaldarius, S. solfataricus and S. tokodaii by lectins. After three days the biofilms of S. acidocaldarius (first row), S. solfataricus (second row) and S. tokodaii (last row) were incubated with DAPI and different lectins and were analyzed by CSLM. The first column shows the DAPI signal (blue), the second column the Con A signal (green), the third column the IB4 signal (yellow) and the last column the overlay of the other three channels. Bars are 20 mm in length. CLSM: confocal laser scanning microscopy.