Unique architecture of microbial snottites from a methane driven biofilm revealed by confocal microscopy

Microbial biofilms occur in many shapes and different dimensions. In natural and semi‐artificial caves they are forming pendulous structures of 10 cm and more. In this study a methane driven microbial community of a former medicinal spring was investigated. The habitat was completely covered by massive biofilms and snottites with a wobbly, gelatinous appearance. By using fluorescence techniques in combination with confocal laser scanning microscopy the architecture of these so far unknown snottites was examined. The imaging approaches applied comprised reflection of geogenic and cellular origin, possible autofluorescence, nucleic acid staining for bacterial cells, protein staining for bacteria and extracellular fine structures, calcofluor white for β 1 → 3, β 1 → 4 polysaccharide staining for possible fungi as well as lectin staining for the extracellular biofilm matrix glycoconjugates. The results showed a highly complex, intricate structure with voluminous, globular, and tube‐like glycoconjugates of different dimensions and densities. In addition, filamentous bacteria seem to provide additional strength to the snottites. After screening with all commercially available lectins, by means of fluorescence lectin barcoding and subsequent fluorescence lectin binding analysis, the AAL, PNA, LEA, and Ban lectins identified α‐Fuc, β‐Gal, β‐GlcNAc, and α‐Man with α‐Fuc as a major component. Examination of the outer boundary with fluorescent beads revealed a potential outer layer which could not be stained by any of the fluorescent probes applied. Finally, suggestions are made to further elucidate the characteristics of these unusual microbial biofilms in form of snottites.

• The matrix of snottites was examined by taking advantage of 78 fluorescentlylabeled lectins.
• Probing the snottite outer surface indicated an additional unknown stratum.

| INTRODUCTION
Microorganisms frequently appear in form of mobile bioaggregates or immobile biofilms.These communities are typically composed of microbial cells and their self-produced extracellular matrix.It is this matrix which makes microbial aggregates and films stable and even more the matrix can be described as a multi-functional constituent (Flemming et al., 2023;Neu & Lawrence, 2017).The matrix is crucial for biofilm internal processes as well as in the manifold interactions with the environment.The identity and consistency of the matrix may vary across a wide spectrum and as a result the numerous environmental biofilms and their associated matrices are very different in their biochemical and mechanical properties.Microbial biofilms exist in uncountable variations covering the natural environment, manmade structures, and medical issues.Biofilms develop in moderate as well as in extreme habitats.Their structure is mainly determined by the microbial species present, the nutrients available, the hydrodynamic conditions, and the presence of potential grazers.
Early studies on the structure of bioaggregates and biofilms employed transmission or scanning electron microscopy.However due to the fixation and dehydration treatment necessary, these early reports did not show the native architecture of microbial communities.Therefore the invention of confocal laser scanning microscopy (CLSM) was highly welcome for examination of microbial communities in their natural state (Neu et al., 2010;Neu & Lawrence, 2014).In the meantime CLSM became established as the method of choice for the investigation of fully hydrated biofilm communities and their extracellular matrix.
Microbial biofilms also have been found and studied in a number of caves and karst systems.A characteristic appearance of these cave biofilms is the formation of films inside or on the surface of flowing water, layered structures on humid walls as well as so-called snottites (Hose & Pisarowicz, 1999).Snottites represent pendulous, mucous structures of varying thickness and length.Such finger-like snottites up to 15 cm in length were found in a semi-artificial cave where they covered the walls and the entire ceiling of the cave.In a first manuscript we reported on the significance of methane as a major driver of biofilm formation and identified the microorganisms involved (Karwautz et al., 2018).A second manuscript describes a novel isolate, Candidatus Methylomirrabilis iodofontis, suggesting the capacity of methane oxidation, oxigenic denitritification and iodate reduction (Zhu et al., 2022).In this study we applied CLSM in combination with a number of fluorescent probes highlighting the carbohydrate components (glycoconjugates) of the microbial community in order to elucidate the structure and composition of the snottites in more detail.In contrast to the previous two publications, the focus was set on the architecture of the biofilm matrix which in many studies maybe regarded as the dark matter of microbial biofilm systems (Neu & Lawrence, 2017).

| Location and sampling
The semi-artificial cave is located near Sulzbrunn, Bavaria, Germany and was previously used as an iodide-rich medicinal spring.Details of the location are described in a first publication of the snottite microbial community (Karwautz et al., 2018).In addition, the first report contains a drawing of the semi-artificial cave together with several photos of the location.The snottites for CLSM analysis were collected in October 2015.Snottites were removed and stored in various ways including PFA-fixation/fridge, frozen/freezer and as fresh samples kept in the fridge.Fresh snottites from the ceiling were cut off with a scalpel and caught on a rectangular, rounded piece of parafilm (8 Â 3 cm 2 ).The sample with the parafilm stripe was carefully transferred horizontally into sterile polystyrene tubes and transported horizontally with minor agitation.

| Sample preparation and staining
Subsamples were prepared by cutting off cross-or longitudinalsections from the snottites using a scalpel.The sections were transferred into coverwell chambers with matching spacers of 1.5-2.8mm.
Staining and washing (if needed) was done in the coverwell chamber, which was then filled with water and finally covered with a high quality coverslip (No. 1.5H, Marienfeld).
Glycoconjugates were stained after fluorescence lectin barcoding (FLBC) which represents a screening with all commercially available lectins (see Supplement 1).From FLBC a collection of lectins was selected for subsequent fluorescence lectin binding analysis (FLBA) using the most appropriate lectins.The lectins were applied as reported previously.In short, the lectin stock solution (1 mg/mL) was diluted 1:10 for glycoconjugate staining.The samples were incubated with a few drops of lectin solution and incubated for 20 min in the dark at room temperature.Then samples were washed carefully three times with water in order to remove the unbound lectins.If needed, a counterstain with a nucleic acid specific fluorochrome was applied.
Finally fluorescent beads were employed in order to examine the outer border and potential porosity of snottites.The fluorescent beads used were fluoresbrite 17797 (Polysciences) with 1.9 μm diameter.

| Confocal laser scanning microscopy and visualization
For CLSM a TCS SP5X (Leica Microsystems) controlled by the software LAS AF version 2.6.1 was available.The system was equipped with an upright microscope and a super continuum light source (470-670 nm).Image datasets were collected at 8 bit using 25x NA 0.95 or 63x NA 1.2 water immersion objective lenses.Datasets were usually collected at 0.5 or 1 μm sectioning intervals depending on sample thickness and resolution needed.In order to take advantage of the full dynamic range, signal-to-noise-ratios were optimized using the lookup table "glow-over-under".The multichannel image datasets were visualized and projected with Imaris version 9.9.1 (Bitplane).Most of the images were projected as volume view.For clarification one dataset was projected as isosurface in combination with volume view.
Color allocationall datasets were recorded by the photomultipliers as 8 bit datasets in form of gray values from zero to 255, as a consequence the multi-channel datasets need a false color allocation.
The Red-Green-Blue (RGB) color model was applied in GRB order starting with the visible colors in the epifluorescence modus with green and red emission.For far red emission the remaining blue (B) was used.Thus the laboratory rules for allocation of false colors were agreed as follows with few exemptions:

| RESULTS
In order to start the examination of the gel-like snottites the samples were visualized by taking advantage of intrinsic sample properties.
Microbial samples from the environment frequently show reflection signals of geogenic origin, for example, soil or sand particles.In addition, microbial cells show a cellular reflection based on certain constituents accumulating on or within cells.The same is true for biofilm microbes embedded in an extracellular matrix.As a result, snottites tested for reflection showed bacteria of various shape and size as well as filamentous signals.At some locations, the reflection signals were clustered indicating microcolonies (Suppl.2, Figure 1A, B).Testing for autofluorescence gave no results.The addition of nucleic acid specific stains allowed localization of bacterial cell distribution (Suppl. 2, Figure 1C).Furthermore, it indicated already a spatial distribution of the bacteria with voids in between.
In a second step, a protein stain was applied for staining bacterial cell surfaces and potential extracellular proteinaceous constituents.
Similar to the nucleic acid staining it revealed bacterial cell distribution but also thin proteinaceous filaments (Suppl.2, Figure 2A).The application of Calcofluor White allowed probing a specific group of matrix polysaccharides (β 1 ! 3 and β 1 ! 4 linkages) and the possible presence of fungi.In fact, it stained the cell surface of large bacteria (Suppl.2, Figure 2B).A first test using a lectin pointed towards globular structures as a major snottite feature (Suppl.2, Figure 2C).Consequently, performing a lectin screening in order to identify additional useful lectins seemed to be the way to go.
During the screening by means of fluorescence lectin bar-coding (FLBC) we finally selected a panel of lectins for fluorescence lectinbinding analysis (FLBA).Testing all the commercially available lectins identified AAL, PNA, LEA, and Ban (see Supplement 1 for three letter code and further details) as the ones staining major features of the bacteria and their extracellular space.The AAL-Alexa568 lectin in combination with SybrGreen or Syto9 revealed bacterial cell distribution in relation to the lectin signals in form of thick capsules (Figure 1a) developing into tube-like strands of cells and glycoconjugates (Figure 1b) finally filling out the entire space (Figure 1c).However not all bacteria showed a signal with AAL indicating the presence of other matrix constituents.
A 3-dimensional data stack revealed the extended globular pattern of the AAL lectin stain (Suppl.2, Figure 3A).The close-up with globular lectin features showed often a double shell-like pattern (Suppl.2, Figure 3B).In other locations, the glycoconjugate patterns showed tube like structures, which seemed to be hollow inside (Figure 2a).If counterstained with a nucleic acid stain, the globular features harbored a bacterial cell inside indicating the producer of the shells and globules.In most cases, the shell had a single or double layer, but in one spot even a triple shell can be seen (Figure 2b).
Optimizing signal and contrast by using the "glow over under" lookup table indicated globular glycoconjugates together with nucleic acid stained filaments (Suppl.2, Figure 4A).However, when raising the sensitivity of the PMT, additional lectin stained glycoconjugate filaments became visible (Suppl.2, Figure 4B).Examination of several different snottite sections often showed lectin signals of different size and different intensity around cells (Figure 3a,b).In other areas, regions with globular AAL-lectin signals are next to regions where "naked" bacteria are present (Figure 4a).These seemingly free floating bacteria could not be stained with the AAL-lectin.However, they very likely possess other extracellular polymers which position them inside the snottite structure.In some areas the non AAL stained bacteria are forming long thread-like features (Figure 4b).
As the filamentous structures seem to be an important part of the snottites, other lectins were applied identifying additional types of glycoconjugates.For example, the lectin LEA bound to the filaments (Figure 5a).Similarly to PNA it was able to contrast the filaments which sometimes showed a longitudinal orientation (Figure 5b).The lectin Ban represented another probe revealing the thin filaments of the snottites (Figure 5c).By employing an inverse labelling using the lectin combinations AAL and Ban, the lectin double staining was approved.The result showed green tubular lectin patterns and red filamentous lectin patterns and vice versa (Suppl.2, Figure 5A,B).
The combination of AAL and PNA together with a nucleic acid stain clearly showed two different glycoconjugates in relation to the bacterial cell distribution.Tube-like glycoconjugates (AAL) made up the major volume of the snottites, whereas the filaments (PNA) seem to form a kind of rigid skeleton (Figure 6a).The large bacterial cells stained with Syto60 are located outside the AAL glycoconjugates.
However, bacterial cells are also present at the tip of the AAL tubes.
The large bacteria stained with Syto60 also show a reflection signal  indicating cellular inclusions (Figure 6b).As the bacteria are obscured inside the bright tube-like AAL signal, an isosurface dataset was projected.It shows the solid isosurface of Syto60 with bacteria, the transparent isosurface of the AAL lectin and the volume view of the filaments with the PNA lectin (Figure 6c).
The tip of snottites was also examined and revealed similar structural features as detected in other areas (Figure 7a).The dataset shows typical globular and tube-like AAL-stained glycoconjugates together with filamentous bacteria in between.The same is true for the region next to the tip (Figure 7b).
Another question coming up was about the edge of the snottite.
Examination of the edges revealed different structural features.The bacteria developed in strands towards the outer surface often covering the surface.The volume usually made up of AAL specific glycoconjugates formed the bulk of the snottite (Suppl.2, Figure 6A,B).
Application of StainsAll (Suppl.2, Figure 6C) and record of dual channel data sets showed dense bacterial cell distributions (570-620 nm emission) but also several globular features (650-720 nm emission) similar to the AAL-lectin signal.
By using fluorescent beads, further characterization of the snottite edge became possible.Interestingly, the beads detected by reflection always showed a certain distance to the snottite surface (Figure 8a-c).This finding suggests an additional, so far unidentified surface layer around the snottites.

| DISCUSSION
With respect to the extracellular matrix of microbial biofilms, the biochemistry is highly divers with polysaccharides, different protein classes, extracellular nucleic acids, amphiphilic constituents as well as microbial derived refractory compounds (Flemming et al., 2023).Consequently, their analysis and visualization remains a challenge.There are a few fluorochromes specific for some of these matrix compounds, but they often bind to cellular structures as well, making a differentiation complicated.Nevertheless, establishing the application of fluorescently labeled lectins became the method of choice for examination of biofilm-related glycoconjugates (Neu et al., 2001).Lectins will bind preferentially to one of the major compounds, the polysaccharides in the matrix, but they will also bind to other glycoconjugates such as glycoproteins or glycolipids.In any case, the lectin approach virtually sheds more light into the intercellular space of microbial communities and their biochemical makeup.Lectins have been employed in many biofilm studies in order to examine the glycoconjugates in pure cultures (Neu & Kuhlicke, 2017) as well as in biofilms from environmental habitats (Neu & Kuhlicke, 2022).In fact, in the latter ones, the lectin approach suggests itself as an ideal tool as it does not require the production of antibodies.Thus, the application of fluorescently labeled lectins and other fluorescent probes in combination with CLSM allowed a multiparameter assessment of cellular and extracellular constituents in snottites.
Assessment of the confocal laser scanning microscopy findings in terms of the snottite structure revealed a highly complex and intricate spatial organization.The architecture composed of globular, tube-like and filamentous glycoconjugates suggests an important role in the metabolism and function of the microbial community.The  extracellular matrix determines the space between cells, which varied considerably in different sections of the snottites.Furthermore, the massive amount of gel-like carbohydrate matrix forms a protective layer against environmental impacts.Thereby the biofilm creates its own environment with altered import and export of gases, nutrients and waste products.However, it also may serve as an electron sink due to an unbalanced nutrient supply.The production and excretion of proteins, biopolymers, soluble metabolites and lipids by methanotrophs is widely established (Strong et al., 2015).In fact a sort of spilling reaction of the methane driven microbial community towards extracellular glycoconjugates could provide reduced carbon to other heterotrophic bacteria living in this extreme habitat (Carere et al., 2019;Russell, 2007).Thus the presence of thermogenic methane and volatile iodine may in part explain the matrix structure and attribute a specific functionality to the voluminous and gelatinous matrix.Nevertheless, some questions remain with respect to potential links between cycling of methane and iodine in this particular habitat (Karwautz et al., 2018;Zhu et al., 2022).
In course of the study additional filamentous features showed up.The filaments stained with several probes such as protein-and nucleic acid-specific fluorochromes but also with several lectins.These thin filaments may represent filamentous bacteria which are embedded in the snottite structure.Consequently, the overall architecture and composition is determined by: (1) the bacteria of various shape, size and densities, (2) the presence and absence of globular and tube-like glycoconjugates.In other words, the spacious and extended glycoconjugates making up the major volume whereas the filamentous bacteria may provide additional stability of the snottites.
A specific discussion with respect to the extracellular matrix in other snottites remains elusive as there is hardly any report available throughout the scientific literature.An early review elaborated on the various types of caves and their geomicrobiology (Northup & Lavoie, 2001).Most information found on snottites of acidic biofilms in caves had a focus on the microbial community.The main location  investigating the phylogeny of bacteria was in Italy sampling the Frasassi caves (Jones et al., 2012;Macalady et al., 2006;Macalady et al., 2007).Later studies compared the results from Frasassi with acidic biofilms found in Mexican caves (Jones et al., 2016;Jones et al., 2023).None of the studies mentioned investigated the extracellular matrix of acidic biofilms.However, another study on a pyrite leaching community in an abandoned pyrite mine examined the bacterial community as well as the polysaccharide matrix.It turned out that arabinose was the major carbohydrate with rhamnose, galactose and mannose as minor components (Ziegler et 2009).Despite this biochemical result there were no detailed conclusions drawn.
Considering the geochemistry of the snottite habitat in the present study, the molecular analysis of the microbial community with methanotrophic and methylotrophic populations (Karwautz et al., 2018) as well as the metagenome assembled genome of an isolate (Zhu et al., 2022) together with the known extracellular polysaccharide production capacity of methanotrophs (Strong et al., 2015), it may be suggested that these microorganisms are in fact responsible for the over production and massive amount of gel-like glycocojugate material as identified by FLBA.
Finally, some additional approaches for further experiments can be suggested taking advantage of fluorescence techniques in combination with laser microscopy techniques.In terms of new staining techniques other non-commercial lectins could be tested.Another option are so-called carbohydrate-binding modules (CBMs) having a variety of different specificities (Boraston et al., 2004).In fact a CBM was applied to investigate the extracellular polysaccharides of Escherichia coli biofilms (Nguyen et al., 2014).Very recently new contrasting agents for biofilm matrix constituents have been developed.For example, by means of optotracing using luminescent conjugated oligothiophenes the extracellular matrix of Salmonella biofilms was examined (Choong et al., 2016).Due to the molecular analysis presented in the first publication on the microbial community of the snottites their identity is known (Karwautz et al., 2018).Consequently it is obvious to combine fluorescence in situ hybridization (FISH) or catalyzed reporter deposition FISH (CARD-FISH) with FLBC and FLBA as presented in a marine study (Bennke et al., 2013).This will allow to localize and identify the major producers of different lectin-specific glycoconjugates.With respect to nutrient availability and environmental impacts on snottite metabolism it might be appropriate to measure diffusion into and within the snottites.For this purpose fluorescence recovery after photo-bleaching (FRAP) and fluorescence correlation spectroscopy (FCS) are available as powerful techniques.

| CONCLUSIONS
1.The snottite matrix has a highly complex, intricate structure mainly formed by voluminous, globular and extended, tube-like glycoconjugates responsible for the wobbly, gelatinous appearance.
2. Filamentous bacteria as detected with various probes form a dense network and seem to provide additional stability to the snottite architecture.
3. After screening all commercially available lectins by FLBC, the lectins AAL, PNA, LEA, and Ban were identified as useful probes for subsequent FLBA.
Glycoconjugates of different dimension and density.(a) AAL-Alexa568, Syto9.Showing strongly stained capsules and extended weakly stained glycoconjugates.Grid size 50 μm.(b) AAL-Alexa568, Syto60.Showing strongly stained capsules and extended weakly stained glycoconjugates.Grid size 50 μm.Please notice the single bacteria in capsules and the groups of bacteria inside the weakly stained glycoconjugates.
Filaments imaged by using various lectins.(a) LEA-FITC.Showing glycoconjugates of the filamentous bacteria.Grid size 50 μm.(b) PNA-TRITC.Showing glycoconjugates of the filamentous bacteria partly oriented in longitudinal direction.Grid size 20 μm.(c) Ban-Fluo.Showing glycoconjugates in a mesh of filamentous bacteria.Grid size 50 μm.

4.
Application of these four lectin specificities indicate the major glycoconjugates as α-Fuc, β-Gal, β-GlcNAc, and α-Man with α-Fuc as a main component.5.Examination of the snottite boundary suggests additional constitu-ents forming an outer layer with unknown identity.6. Very likely, the snottite architecture stands in close relation to the geochemistry of the habitat with methane and iodine as major drivers as well as the microbial metabolism of methano-and methylotrophic microorganisms present.