Microbiota of Attine Ants' Gardens: Visualizing a Microbial Landscape by Scanning Electron Microscopy

Summary

We propose an optimized Scanning Electron Microscopy protocol for visualizing highly heterogeneous and delicate samples containing plant and fungal biomass, together with microbiota and biofilm. This protocol allows describing the spatial dimensions of the microbiota organization.

Abstract

In macroscale ecosystems, such as rainforests or coral reefs, the spatial localization of organisms is the basis of our understanding of community ecology. In the microbial world, likewise, microscale ecosystems are far from a random and homogeneous mixture of organisms and habitats. Accessing the spatial distribution of microbes is fundamental for understanding the functioning and ecology of the microbiota, as cohabiting species are more likely to interact and influence each other's physiology.

An interkingdom microbial ecosystem is at the core of fungus-growing ant colonies, which cultivate basidiomycete fungi as a nutritional resource. Attine ants forage for diverse substrates (mostly plant-based), metabolized by the cultivated fungus while forming a spongy structure, a "microbial garden" that acts as an external gut. The garden is an intertwined mesh of fungal hyphae growing by metabolizing the substrate, opening niches for a characteristic and adapted microbiota to establish. The microbiota is thought to be a contributor to substrate degradation and fungal growth, though its spatial organization is yet to be determined.

Here, we describe how we employ Scanning Electron Microscopy (SEM) to investigate, with unprecedented detail, the microbiota and biofilm spatial organization across different fungiculture systems of fungus-growing ants. SEM imaging has provided a description of the microbiota spatial structure and organization. SEM revealed that microbiota commonly assemble in biofilms, a widespread structure of the microbial landscapes in fungiculture. We present the protocols employed to fix, dehydrate, dry, sputter coating, and image such a complex community. These protocols were optimized to deal with delicate and heterogeneous samples, comprising plant and fungal biomass, as well as the microbiota and the biofilm.

Introduction

Ecosystems are composed of organisms interconnected by processes in a specific geographical location (i.e., the environment). Organisms interact with their environment over time, from which complex and heterogeneous spatial patterns emerge. Spatial patterning determines ecological diversity and stability and, ultimately, ecosystem functioning^1,2,3,4. In macroscale ecosystems, such as wetlands, savannas, coral reefs, and arid ecosystems, spatial patterns are correlated with resource flow and concentration. Permitting resource optimization, spatial heterogeneity and patterning result in more resilient ecosystems than homogeneous ones². The spatial localization of organisms being at the basis of community ecology is also translated to the microbial world.

Microbial ecosystems, far from organisms randomly and homogeneously mixed throughout microhabitats, exhibit spatial patterns defining much of their functioning^5,6,7. From Winogradsky columns to environmental- and host-associated microbiota, these ecosystems are heterogeneously organized in space, with spatial arrangements eliciting different phenotypic responses. Cohabiting species are more likely to interact and influence each other's physiology. Thus, the community spatial organization, more than its composition per se, delimits ecosystem properties and ecological niches^5,7,8. Illustrating these concepts, changes in spatial patterning seem correlated to the pathological progression of dental plaques, caries, gum diseases^9,10, inflammatory bowel disease¹¹, cystic fibrosis lung infections, chronic wound infections^12,13, colorectal cancer, and adenomas¹⁴.

Under the scope of microbial biogeography (the study of biodiversity distribution and patterning across space and time on a microscale), the knowledge of microbial ecosystems is vastly benefitted by comprehending their spatial patterns^{6,13,15,16,17}. We have looked into spatial patterns of an insect-built microbial ecosystem, found at the core of the charismatic fungus-growing attine ant (Hymenoptera: Formicidae: Myrmicinae: Attini: Attina) colonies. There resides a "microbial garden," centered around a basidiomycete fungus in the tribe Leucocoprinae (Basidiomycota: Agaricaceae) or in the family Pterulaceae (Basidiomycota: Agaricales)^{18,19,20,21,22}. The garden is a spongy structure emerging from an intertwined mesh of hyphae that grows by metabolizing the mostly plant-based substrate incorporated by the ants (Figure 1). These may include, according to the attine genera: dry plant parts, insect frass and carcasses, freshly cut leaves, seeds, and floral parts^23,24. Analogous to an external herbivorous gut, the garden enzymatically and chemically converts recalcitrant polymers into labile nutritional resources, providing the ants with essential amino acids, lipids, and soluble sugars^{21,25,26,27,28}.

Ultrastructural, enzymatic, and transcriptomic analyses carried out for gardens of the leaf-cutting genera Atta and Acromyrmex suggest these environments structure a continuum of substrate degradation and nutritional patches^{26,29,30,31,32}. Young parts of the garden tend to be darker due to freshly incorporated substrate after being fragmented. These recently added substrates are often colonized from the edges, which were cut by ant workers and inoculated with mycelial clumps. Radiating from cut edges, fungal hyphae spread over the substrate^29,32,33. Hyphal abundance increases as substrate degradation progresses, resulting in whitish and metabolically active regions^30,31,32. Older regions, with more degraded substrate and an abundant microbiota^29,32, tend to present brownish tones and higher humidity. Workers remove fragments of this region, separating them in waste piles, where they also take substrates that harm the fungal symbiont^34,35,36. Waste piles, although physically detached from the garden, are a spot of continuous substrate degradation and nutrient cycling by the abundant inhabitant microbiota^{29,32,37,38,39}.

A microbiota mainly composed of Enterobacter, Klebsiella, Pantoea, Pseudomonas, and Serratia, also inhabits the garden, apparently being shared by diverse attine fungiculture systems. Encoding metabolic pathways that could complement fungal metabolism, the microbiota potentially participates in the garden's physiological responses^{40,41,42,43,44}. Not only did metagenomic data indicate that the microbiota was there^41,42, but also Scanning Electron Microscopy (SEM) analysis of leaf-cutting ants' fungiculture showed mostly rod-shaped bacteria over plant substrate³². Though bacteria (including cellulolytic strains) were isolated from the entire garden, they were visualized only in older parts of the garden and in waste piles, as well as in the initial pellet carried by foundress queens^29,32. It was also uncertain whether the microbiota could form biofilms in vivo (i.e., in the garden and waste), as suggested by their metabolic capacity⁴² and observed in vitro⁴⁴.

Here, we employed SEM to further comprehend the microbiota spatial organization across the garden regions, detailing microbiota-substrate and microbiota-hyphae physical interactions. By providing images with larger focal depth, SEM permits observations of three-dimensional microscopic structures in high resolution, enabling a thorough analysis of the garden microbiota spatial patterns. We detail steps to fix, dehydrate, dry, sputter coat, and image such heterogeneous and delicate fungal-based samples. By removing the postfixation step using osmium tetroxide (OsO₄) and reducing the dehydration time, we simplified the protocols^32,33,45 for preparing garden and waste samples for SEM analysis. This adapted protocol preserves hyphal structural patterns, as well as the microbiota and biofilm spatial organization, and could be applied to other delicate microbial ecosystems and biofilms.

Figure 1: Attine microbial gardens. The garden is a sponge-like structure resulting from an intertwined mesh of hyphae that grows by metabolizing the mostly plant-based substrate incorporated by the ants. Also inhabiting the garden is the microbiota, which encodes metabolic pathways that could complement fungal metabolism. Metagenomic data and previous Scanning Electron Microscopy analysis indicated its presence, though we had scarce knowledge of its spatial organization and physical interactions with the substrate and fungal hyphae. We employed SEM to unveil the microbiota and biofilm spatial organization and patterning. Illustrations by Mariana Barcoto (garden and microbiota adapted from Barcoto and Rodrigues ⁹⁴), and photos by Mariana Barcoto and Enzo Sorrentino. Please click here to view a larger version of this figure.

Protocol

1. Sampling field colonies

NOTE: When collecting ant colonies, certify that all the permissions required by local legislation are obtained before collecting. In our case, the collecting permit #74585 was issued by Instituto Chico Mendes de Conservação e Biodiversidade (ICMBio). When the samples come from a lab colony, go to section 2.

1. Locate and mark the colony. Excavate a trench surrounding the nest area until the garden chamber is exposed (Figure 2A).
   NOTE: Some attine species may build their colonies under the leaf litter or inside rotten logs. In such cases, carefully revolve the litter or carefully break the logs for collecting the samples. For detailed information on locating, collecting, and maintaining living colonies of diverse attine ant species, see Sosa-Calvo et al.⁴⁶.
2. Open the garden chamber laterally to prevent the soil from falling on the garden surface. Carefully collect garden samples using either entomological forceps, a spoon, or a kitchen skimmer, depending on the garden size.
   NOTE: Be sure to sterilize the tools before collecting. When collecting leaf-cutting ants' gardens, wear thick fabric gloves to prevent (or at least attenuate) worker bites. For other attine species, fabric gloves are optional.
3. Transfer the garden samples to a clean plastic recipient containing a layer of plaster for balancing the garden's humidity. After transferring the garden and ant workers, hermetically close the recipient to avoid sample drying. Store garden samples at 23-25 °C until processing.
4. Close the trench with previously removed soil.

Figure 2: Sample preparation protocol. (A) Sampling of field colonies. (B) Sample processing. (C) Brief fundamentals and workflow for sample preparation: 1. Fixation: for strengthening and preserving sample structure. 2. Dehydration: samples' water content is exchanged for ethanol. 3. Critical point drying: liquid CO₂ replaces ethanol and is evaporated. 4. Mounting: sample displayed for analysis. 5. Sputter-coating with gold: prevent sample charging. 6. Imaging. Illustrations and photos by Mariana Barcoto. Please click here to view a larger version of this figure.

2. Reagents

NOTE: Bear in mind that the following solutions should be prepared beforehand.

1. Prepare 0.2 M sodium cacodylate buffer. For this, dissolve 42.8 g of sodium cacodylate in 800 mL of distilled water, stirring until dissolution and pH stabilization at 7.2 (if required, adjust pH using hydrochloric acid). Make up the volume of the solution to 1 L by adding distilled water. Store the solution at 4 °C (for ~1 month).
   CAUTION: Glutaraldehyde and paraformaldehyde are toxic and should be handled within a fume hood. Wear nitrile gloves and protective glasses all the time while handling such reagents.
2. Prepare Karnovsky's fixative (modified from Karnovsky)⁴⁷. For this, add 10 mL of 25% glutaraldehyde aqueous solution and 10 mL of 20% paraformaldehyde aqueous solution to 25 mL of 0.2 M sodium cacodylate buffer and mix. Add 1 mL of 0.1 M calcium chloride (CaCl₂) and make up the volume to 100 mL by adding distilled water. Store the solution at 4 °C (for up to 1 month).
   CAUTION: Karnovsky's fixative is harmful when inhaled and may cause skin and eye irritation. Thus, avoid breathing vapors, using only outdoors or in a well-ventilated area. Wear nitrile gloves and protective glasses all the time when handling the solution.
3. Using absolute ethanol (analytical grade), prepare 30%, 50%, 70%, and 90% ethanol solutions in distilled water; 100% ethanol is also required.

3. Sample fixation

NOTE: Fixatives harden and preserve samples, maintaining morphological features. Aldehydes (such as paraformaldehyde and glutaraldehyde) are non-coagulant fixatives of cross-linking type, inducing cross-links within and between proteins and nucleic acids⁴⁸.

1. Remove workers, eggs, pupae, and larvae from the garden samples using entomological forceps (Figure 2B). Set apart garden fragments no larger than 5 mm³. Add the fragments to a 2 mL tube (Figure 2C).
2. To the tubes containing the samples, use a Pasteur glass pipette to add ~1 mL of Karnovsky fixative solution (ensure the sample is completely covered). Mix by gentle agitation to help the sample to soak and incubate at 4 °C for at least 24 h before continuing sample processing (Figure 2C.1).
   NOTE: When the following dehydration steps will not be carried out immediately after the fixation, the protocol may be paused at this step, and samples may be stored for ~1 year at 4 °C.
   We suggest using a Pasteur glass pipette, as it is composed of an inert material and is easier to clean for reuse afterward. Garden material is highly hydrophobic, tending to float over the surface of the fixative solution. It usually takes up to 5 min to be completely soaked in the fixative. Make sure the fixative covers the garden fragments since its volume tends to be reduced as it enters the sample's pores. We empirically verified that after the sample was soaked and became wet, its components (in particular, the delicate fungal mycelium) became susceptible to shattering when mixed further. Thus, we recommend avoiding shaking the samples as much as possible.

4. Sample dehydration

NOTE: The ethanol washing series gradually exchanges the water in samples for ethanol. It is important to start with a low-concentration ethanol solution (see below) to avoid excessively damaging or collapsing such delicate samples⁴⁹.

1. Completely remove Karnovsky's fixative solution using a glass pipette, being careful to not disrupt the sample (Figure 2C.1).
   NOTE: Discard Karnovsky's fixative in a recipient properly labeled for toxic chemical residue management.
2. Immediately after removing the fixative, add 1 mL of 30% ethanol, being careful to not disrupt the sample, and incubate for 10 min at room temperature (Figure 2C.2).
   NOTE: Samples need to always be soaked in solution. Make sure to quickly replace solutions during the ethanol gradual washing series. As the garden loses its porous aspect and aggregates at the tube bottom, 1 mL of ethanol tends to be enough to cover the sample (when the sample is no larger than 5 mm³). However, if the samples are not completely covered, add ethanol until the sample is completely covered.
3. Completely remove 30% ethanol with a glass pipette, being careful not to disrupt the sample. Discard 30% ethanol properly.
   NOTE: For the entire ethanol gradual washing series, discard ethanol in a recipient properly labeled for toxic chemical residue management.
4. Add 1 mL of 50% ethanol and incubate for 10 min at room temperature. Completely remove 50% ethanol with a glass pipette, being careful not to disrupt the sample. Discard 50% ethanol properly.
5. Add 1 mL of 70% ethanol and incubate for 10 min at room temperature. Completely remove 70% ethanol with a glass pipette, being careful not to disrupt the sample. Discard 70% ethanol properly.
   NOTE: At the 70% ethanol washing step, the user may pause the protocol if needed, as sample tubes may be stored overnight at 4 °C, when the material will not be immediately processed.
6. Add 1 mL of 90% ethanol and incubate for 10 min at room temperature. Completely remove 90% ethanol with a glass pipette, being careful not to disrupt the sample. Discard 90% ethanol properly.
7. Add 1 mL of 100% ethanol and incubate for 10 min at room temperature. Completely remove 100% ethanol with a glass pipette, being careful to not disrupt the sample. Discard 100% ethanol properly.
8. Using a forceps and/or a spatula, carefully transfer the samples to specimen containers for the critical point dryer (CPD), containing sample identification labels (made previously in paper and pencil). For avoiding samples drying, this transfer is carried out with the container placed in a Petri dish covered with 100% ethanol.
9. Put the lids on the containers and immerse them in a graduated glass beaker containing 100% ethanol enough to submerge the containers. Cover the glass beaker and incubate for 10 min at room temperature; then properly discard 100% ethanol.
10. Transfer the specimen containers to another graduated glass beaker containing enough 100% ethanol to submerge the containers. Cover the glass beaker and incubate for 10 min at room temperature; then, transfer specimen containers to the critical point dryer.
    NOTE: After the washing series, glass pipettes and beakers should be abundantly rinsed with distilled water, and this residual water should be discarded in a recipient properly labeled for toxic chemical residue management. After rinsing, glass items may be washed with neutral detergent, rinsed with tap water, and air dried.

5. Critical point drying

NOTE: A Critical Point Dryer exchanges the ethanol in samples for liquid carbon dioxide (CO₂), which evaporates from the sample at higher temperature and pressure. Please follow the manufacturer's instructions for such procedures.

1. Switch on the equipment.
2. Open the chamber, place the sample containers inside, and add 100% ethanol until it covers the containers. Close the chamber.
3. Activate the Cool option, and wait until the temperature reaches 10 °C.
4. Open the CO₂ cylinder valve and activate the Stirrer option.
5. Activate the CO₂ in option, always checking the chamber to verify how much CO₂ has already filled it. When the chamber is almost filled, deactivate the CO₂ in option and activate the Exchange option, keeping it activated until there is enough CO₂ only to cover the containers. Make sure that the containers are always covered with CO₂ (i.e., deactivate the Exchange option before all CO₂ has left the chamber). Repeat step 5.5 6x.
6. Activate CO₂ in one last time and fill the chamber until the containers are covered.
7. Activate the Heat option and deactivate the Stirrer option. Close the CO₂ cylinder valve.
8. Wait until the temperature rises to 35 °C; then, activate the Gas out option.
   NOTE: At around 30 °C, the chamber reaches pressures of 70-80 bar, achieving the critical point, where the liquid disappears.
9. When the chamber pressure reaches 1 bar, all gas content has been removed. Open the chamber and remove the containers.
10. Switch off the equipment.

6. Mounting

1. Prepare SEM sample holders (i.e., aluminum stubs; Figure 2C.4).
   1. Wrap the stubs with a piece of aluminum foil, covering only the top, to facilitate stub cleaning after the analysis.
   2. Identify the stubs by writing the sample code/number on the holder bottom, guaranteeing the identification of what is placed on top.
   3. Cover the upper part with double-sided carbon tape. Place the stubs into a specimen holder.
2. Open the lid of the specimen container, and carefully transfer the dried sample to a glass Petri dish using forceps and spatula.
   NOTE: Critical point-dried gardens tend to agglomerate forming highly packed samples that should be carefully set apart as fragments no larger than 1 mm³in size.
3. Carefully place fragments on the sticky surface of the tape-covered stub. Once the garden fragment touches the tape, it is very difficult to (re)move it, so be careful not to place it in unwanted places or positions. Add up to nine fragments per stub.
4. Repeat steps 6.2 and 6.3 for each sample.

7. Sputter-coating with gold

NOTE: Coating the sample is required to prevent its charging. Follow the manufacturer's instructions for adjusting settings such as the operation gas pressure (0.5 × 10^-1 mm Hg of gas pressure in this protocol), the sputtering time (220 s), thickness of the gold layer (~120 Å), the current (50 mA), and the voltage supply. Sputtering tends to follow a common workflow though the equipment from different manufacturers may operate slightly differently.

1. Open up the hinged target arm and place the stubs on the specimen table.
2. Close down the hinged target arm and check whether the splinter shield of the glass vacuum chamber is properly embedded.
3. Open the argon cylinder valve and switch on the main power switch.
4. Follow the vacuum rising in the equipment display until it reaches the 0.5 × 10^-1 mm Hg mark in the display; then, activate Rinsing. Repeat the operation 5x.
5. Switch on the water circulation system, activate the option HV On, and open the gold-film lid. Confirm that the plasma color is pinkish. To follow this protocol, set up 220 s of sputtering with 50 mA voltage, which will deposit a gold layer of ~120 Å (12 nm).
6. HV On is automatically switched off. Turn off the water circulation system and close the gold-film lid.
7. Turn off the main power switch, allowing air to enter the vacuum chamber. For garden and waste samples, repeat steps 7.1-7.7 3x.
   NOTE: If the imaging will not occur immediately after sample preparation, store the stubs in a hermetic container filled with a silica layer to avoid rehydrating the samples.

8. Imaging

NOTE: Follow the manufacturer's instructions for adjusting SEM settings, such as objective aperture diameter, operating voltage, alignment of the electron beam system, axial alignment, and stigmators.

1. Place the stubs in the sample holder, taking notes on each sample's position.
   NOTE: Wear gloves when inserting or removing the sample holder, and keep the holder as clean as possible.
2. Launch the operation software from the desktop.
3. Select the instrument settings. Visualize garden samples with the objective aperture diameter of 30 µm (i.e., in the second stage), operating in high vacuum, detecting signals from secondary electrons (SED), accelerating voltage of 20 kV, working distance of 15 - 20 mm, probe current of 40.0 (in high current mode), and varying the magnification according to the sample.
4. Following the navigation instructions, press the Vent icon and wait for the specimen chamber to be vented. A progress bar indicates the vacuum status.
5. When atmospheric pressure has been reached, open the specimen chamber and carefully insert the sample holder.
6. Gently close the chamber door and press the Evac icon to evacuate the instrument, following the vacuum status through the progress bar. The navigation system will show the moving stage position and provide a holder graph when the moving is completed. Press the Camera icon to record a photo of the holder to get a top view that will help to navigate between samples while imaging.
7. Press the Observation icon to turn on the electron gun and wait for the image to form. Use the Manual User Interface (manual control) to manually move the Z-axis to the proper height (based on the sample height and the determined working distance).
   NOTE: For prolonging the tungsten filament lifetime, pause the Observation function whenever not actively imaging.
8. Use either the options shown on the screen or the manual control to change the raster rate and image position. Move the stage to get a thorough view of the sample, employing the RDC function to focus on specific areas. When observing a target (or an interesting) structure, adjust the magnification, focus, brightness, contrast, and stigmation accordingly. Correct stigmation by using the Manual User Interface to move the stage to X and Y directions.
   1. Adjust the magnification between 100x and 700x to visualize general garden aspects (such as hyphal density, substrate, and colonization patterns).
   2. Adjust the magnification between 700x and 1,500x to visualize the microbiota spatial patterns.
   3. Adjust the magnification between 1,500x and 3,000x to observe the microbiota and biofilm physical interactions.
   4. Adjust the magnification between 3,000x and 4,000x to focus on specific microbial clusters.
9. To save an image, use the Freeze function, click on the Photo icon, and set up the file path. As a standard, analyze at least three garden fragments, imaging each one in all magnification ranges mentioned in step 8.8. This results in at least 12 images per sample, though more images may support more detailed descriptions. We suggest that 15-25 images per sample, ranging all the magnifications in step 8.8 tend to provide fine details for sample description.
   NOTE: Rastering and saving an image takes a few seconds during which any vibration on the SEM floating table should be avoided.
10. After finishing imaging, press the Vent option to vent the chamber, following the vacuum status through the progress bar. Carefully remove the holder, gently close the chamber door, and press Evac to evacuate the chamber.
11. As soon as the evacuation is completed, exit the operation software.
12. Adjust the brightness and contrast to improve visualization using an image editor.

Results

Here, we presented a simplified protocol to visualize the components of attine garden and waste samples, such as fungal hyphae, substrate, microbiota, and biofilms. SEM has enhanced our understanding of how the garden and waste scaffold the microbiota structural patterns (Figure 3). In attine gardens, fungal hyphae are branch-like structures covering portions of the substrate surface. Since fungal hyphae tend to be very sensitive to dehydration and rupture, the user may be guided by the hyphae aspect to evaluate the outcomes of sample preparation. As also observed in previous works^32,33,45, hyphae present a tubular aspect; intact, tube-shaped hyphae indicate an effective sample preparation (Figure 3A and Figure 4C). When hyphae are ribbon-shaped, such flattening is possibly an artifact of the technique (Figure 4C).

Plant substrate surfaces may be recognized by cuticle and epidermal layers, as well as stomata and trichomes. Such structures are more recognizable when substrates are not yet colonized by the fungal hyphae and the microbiota (Figure 3A and Figure 4C). When the substrate is colonized, the microbiota and/or fungal hyphae cover its surface such that specific plant structures are not easily discernible (Figure 3A). The microbiota may colonize either the fungal hyphae or the substrate, each of these in seemingly distinct spatial organization. In Figure 3, such spatial patterns were schematically illustrated and boxed in respective SEM images where they were identified; this may facilitate the link between SEM results and their interpretation.

Hyphal-colonizing microbiota are less commonly observed and when found, are composed of rod or coccoid cells occurring isolated or in small groups attached to the hyphal surface. Biofilm exopolysaccharidic matrix (EPS), looking like a "lace fabric" structure, may be observed surrounding hyphae in small clumps (Figure 3A). Fairly more common is the substrate-colonizing microbiota, found in gardens of different attine ant species. Colonization patterns seem defined to some extent by the substrate deterioration stages. Thus, while colonizing yet non-degraded substrates, the microbiota may either occur as individual cells or gather in small groups. The microbiota may layer the substrate when it presents microscopic signs of deterioration (i.e., when plant cuticular/epidermal layers and other plant structures are not discernible, and furrows and indentations are abundant).

In these "microbial mats," cells' morphology is more recognizable when the EPS matrix is absent. In cases where microbial cells are embedded in the EPS matrix, their morphology is not easily distinguishable. According to their spatial organization, we characterized the biofilms as "spread" or as "3-dimensional." When the biofilm is spread, microbial cells and the EPS matrix intermittently cover the substrate surface, forming a flat 1-3 cells-thick layer (Figure 3B). When the biofilm is 3-dimensional, microbial cells and the EPS matrix pile up, forming clusters with several-cell thicknesses. Garden regions constituted by deteriorated substrates (such as old garden and waste) shelter the microbiota in its furrows and indentations, where they form thick communities (Figure 3C). Particularly, the extensively degraded substrates in waste samples house a profuse microbiota, the most morphologically diverse among all regions. Yeasts add more complexity to this community, physically interacting with the prokaryotes via biofilm. In Atta sexdens and Acromyrmex sp. waste samples, the EPS matrices were observed forming a conspicuous fibrillar meshwork surrounding rod-shaped and coccoid prokaryotes as well as fusiform and spherical yeasts (Figure 3C).

Resulting from an intertwined mesh of hyphae, the garden's spongy structure requires additional sputtering to coat the sample more evenly. Such steps avoid the electron scattering that distorts the image and hampers its resolution (Figure 4A). Sample preparation steps preserved most of the garden structure, unveiling the spatial distribution and organization of the microbiota. It is worthy to reiterate, however, that structures such as fungal hyphae and EPS matrix are very sensitive to dehydration and very difficult to preserve, especially for SEM. The hyphal structure was relatively well preserved, yet it is reasonable to consider that the biofilm visualized in some garden samples could lack some structural components of the EPS matrix. Although the smooth outermost layer of polysaccharides enclosing the microbiota was preserved in some samples (Figure 3A, B), some of them had preserved microbiota morphological features but had no EPS matrix. However, it is not clear whether the EPS matrix was indeed absent or whether it was removed during sample preparation.

On the one hand, removing the biofilm smooth layer impairs the understanding of the precise biofilm structure. On the other hand, it reveals the morphology, spatial organization, and physical interactions of cells dwelling underneath the layer, which is our main research focus. Previous studies reported the preservation of both fungal hyphae and the biofilm matrix by employing osmium tetroxide (OsO₄) vapor as the primary fixative^{44,50,51,52,53,54}. Then, we carried out a comparative experiment fixing fresh garden samples with OsO₄ vapor, which apparently preserved the smooth aspect of biofilms (Figure 4B). This method, however, caused extensive flattening of hyphae and microbiota. This flattening resulted in the loss of general spatial aspects of the garden described previously^29,33,32,45, thus limiting further studies on garden structural patterns. Fixing with OsO₄ made it difficult both to distinguish the microbial morphology and distribution and to differentiate cell structures from technical artifacts, thus complicating data interpretation.

Figure 3: Microbiota architectural patterns. The microbiota colonizes either the fungal hyphae and the substrate in distinct spatial organization. Such spatial patterns were schematically illustrated and were boxed at respective SEM images where they were identified (delimited in the figure by white squares): (A) Hyphal-colonizing microbiota are composed of isolated or small groups of cells, with biofilm exopolysaccharidic matrix detectable in some samples. (B) Substrate-colonizing microbiota spread throughout the substrate, with the EPS matrix covering the cells such that their morphology is not easily distinguishable. (C) Substrate-colonizing microbiota forming 3-dimensional groups, constituting mono or multi-morphological biofilms, in some cases evidencing the EPS matrix. Illustrations by Mariana Barcoto. Abbreviation: EPS = exopolysaccharidic matrix. Please click here to view a larger version of this figure.

Figure 4: Fruitless imaging. Low-quality images rendering ineffective results and/or hampering the interpretation of architectural patterns. (A) Electron scattering distorts the image, hampering its resolution and, consequently, morphology identification. This was observed frequently in samples sputter-coated only once and was reduced by additional coating. (B) Fixing fresh garden samples with osmium tetroxide (OsO₄) vapor as primary fixative^{44,50,51,52,53,54} apparently preserved the smooth aspect of biofilms. However, it caused extensive hyphae and microbiota flattening, damaging garden architectural patterns. (C) Sample preparation using Karnovsky's fixative yielded higher-quality images when compared to preparations using OsO₄, preserving most of the hyphal structure, microbiota distribution, and gardens' spatial patterns. Please click here to view a larger version of this figure.

Discussion

SEM uses an electron beam to scan the sample, generating an enlarged image of it such that one can visualize three-dimensional microstructures in high resolution. As SEM operates under high vacuum, the removal of up to/more than 99% of water from samples is required. Inside the SEM vacuum chamber, partially hydrated samples may dehydrate and collapse, besides scattering electrons. For high-resolution imaging in SEM, sample preparation should include procedures for removing water while keeping the changes in volume and morphology at a minimum, as well as coating to prevent charging during imaging^55,56,57.

First, the sample is chemically fixed by crosslinking its proteins and nucleic acids, achieved by diverse techniques. A widely used method involves combining glutaraldehyde and formaldehyde (or paraformaldehyde) in a buffer, such as piperazine-1,4-bis [2-ethanesulfonic acid] (PIPES); N-[2-hydroxyethyl]-piperazine-N′-2-ethanesulfonic acid (HEPES); phosphate-buffered saline (PBS); or sodium cacodylate. Chemical fixation strengthens the sample, preserving its structure for drying. Depending on their nature, samples may need a secondary fixation (in osmium tetroxide) and even a tertiary fixation (in uranyl acetate). Second, samples must be dehydrated. Since air drying causes extensive collapse of fine structures, a common alternative technique is critical point drying (CPD). In CPD, the sample is not submitted to surface tension, preventing artifacts from forming. On the one hand, water reaches the critical point at 228,5 bar and 374 °C, which would easily destroy the sample. On the other hand, the CO₂ critical point is 73,8 bar and 31 °C, justifying its use as a transitional fluid. When increasing the temperature, CO₂ is converted to gaseous form and removed from the sample^{58,59,60,61,62,63}. Third, non-conductive samples should be coated to prevent their charging and increase the secondary-electron yield, which consequently contributes to obtaining sharp image signals⁶⁴.

The protocol presented here combines procedures for preserving fungal-based samples^{32,33,45,65,66,67,68} and biofilms^69,70, though simplifying the process by removing postfixation steps and reducing the dehydration time. Yet, it preserved the delicate garden structure, further detailing the microbiota's physical interactions and unveiling biofilm spatial structure. Microbiota spatial assembling is essential both for establishing interactions and setting their regulatory patterns. Spatial distribution shapes the microbial social dynamics, from which intrinsic properties may emerge. These are required to deal with the variability and complexity faced in microbial ecosystems^71,72,73. Our findings reinforce the microbiota integrating garden structural patterns, supporting their presumed participation in physiological responses^32,42,43,44. We provided the in situ evidence that biofilms are relevant for microbial interactions in the attine fungiculture^42,44. Biofilms are complex and heterogeneous multicellular systems, where microorganisms embedded in a self-secreted EPS matrix form social networks. Cooperatively and/or antagonistically, interactions within the biofilm determine niches opening, occupation, and modification^71,72,73,74. SEM allowed further comprehension of the microbiota spatial organization across the garden regions, detailing microbiota-substrate and microbiota-hyphae physical interactions, while simplifying the methods for forthcoming research on this area.

Research on the garden microbiota is receiving growing attention, in particular, focusing on processes that could shape the composition of these unique communities^75,76, as well as their function and interactions^42,43,44. Bacterial-fungal crop interactions could assist the fungus in solubilizing phosphate, producing siderophores, and degrading cellulose and chitin. Such interactions, which might occur as a mutualism-antagonism continuum, could regulate the garden's functioning⁴⁴. The methodological approach we present here enables one to glance, in situ, at the structural dimension in which interactions could take place. It emphasizes that, as most of the already studied microbial ecosystems, attine gardens are patterned, heterogeneously organized in space^5,7,8. SEM already provided in situ spatial perspective on microbiota dynamics in diverse host-associated communities, such as cyanosponges⁷⁷, humpback whale skin⁷⁸, mushrooms⁶⁶, cnidarians⁷⁹, human gut⁸⁰, lichens⁸¹, also contributing for describing the plastisphere (a novel anthropogenically derived ecosystem)⁸². Each of these studies, however, employed a slightly distinct stepwise method for sample preparation.

Garden samples are highly delicate and heterogeneous, with fungal hyphae thought to be adapted for not losing water²⁶, besides a diverse set of incorporated substrates²⁴, the microbiota and its biofilm. Yet, the sample preparation protocol we adapted preserved most of such structures and allowed us to expand the characterization of the microbiota spatial structure. We consider fixation with Karnovsky's solution to be the critical step in the protocol, as the 3-dimensional aspect of such a delicate sample can collapse and/or flatten at this point. Complex fungal-based samples are usually prepared for SEM by fixing with glutaraldehyde and postfixing with OsO₄^65,66,67,68. Combining aldehydes (paraformaldehyde and glutaraldehyde), however, has been reported to improve biofilm preservation for SEM^69,70. Methods for preserving biofilms may also include 0.15% Alcian Blue or 0.15 % Ruthenium Red to the initial Karnovsky's fixative, as well as 1% tannic acid to 1% osmium for postfixation. Fixing attine gardens and waste with Karnovsky's and postfixing with OsO₄ preserved bacteria in old garden and waste samples, though not evidencing the biofilm³². Since our samples had the garden structure, microbiota, and biofilms preserved without the osmium postfixation step employed in previous protocols^32,33,69,70, we preferred to simplify the protocol by removing this step. OsO₄ is an oxidizing agent used for fixing (especially) lipidic cell components⁸³. As it is highly corrosive, OsO₄ may cause severe burns when in contact with skin and mucosa, besides damaging the respiratory or gastrointestinal tract when inhaled or ingested^84,85. By removing the OsO₄ post-fixation, the protocol consequently reduces the number of toxic regents to be manipulated.

We also suggest the ethanol washing series instead of using acetone, which may be too harsh for delicate samples⁸⁶. The ethanol washing series with 10 min-long steps also simplifies former SEM protocols⁴⁵ without hampering image resolution. Besides, we empirically verified that additional sputtering steps rendered higher-quality images, which was consistent for garden samples of all species we tested. Such additional steps possibly coated the samples more evenly, as electron scattering was much more frequent in samples coated only once (Figure 4A). Thus, we encourage future work testing additives and conditions that could render an improved protocol for biofilm visualization in garden samples.

SEM is a powerful tool for detailing the microbiota cellular morphology and EPS spatial arrangement in surfaces, revealing the topography and texture of garden samples in high resolution and magnification^70,87. Yet, it does not provide taxonomic discrimination for the community members, and consequently, we do not have a distinction of who is where. Further, despite SEM's large focal depth, it does not inform on the 3-dimensional features of the sample. For a complete description of the attine microbial garden, including taxonomic biogeography and 3-dimensional spatial resolution, SEM analysis could be considered a first procedure in stepwise imaging techniques for garden samples. For taxonomic resolution, SEM could be coupled with fluorescence in situ hybridization (FISH)^88,89 and its variants, such as high-phylogenetic-resolution FISH (HiPR-FISH)⁹⁰, combinatorial labeling and spectral imaging FISH (CLASI-FISH)^9,91, and 3-D projections⁹². Three-dimensional microstructure analysis could be further detailed by X-ray microtomography (X-ray µCT)⁹³. So far, SEM has enabled us to observe in detail the microbiota spatial organization surrounding hyphae and in the substrate surface, furrows, and indentations. Such microbial patterning is building our knowledge of putative microbial interactions in attine gardens. We believe that the preparation steps we presented could be widely applied to refine studies on microbial spatial organization in other delicate microbial ecosystems.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
2 mL tube	Axygen	MCT-200-C-BRA	To fix and dehydrate samples
Calcium chloride anhydrous	Merck	C4901	CaCl2 anhydrous to prepare Karnovsky’s fixative
Critical point dryer	Leica	EM CPD 300	For critical point drying
Double Sided Carbon Conductive Tape, 12 mm (W) X 5 M (L)	Electron Microscopy Sciences	77819-12	For mounting samples
Entomological forceps	No specific supplier		To manipulate garden samples
Ethyl alcohol (=ethanol), pure (≥99.5%)	Sigma-Aldrich	459836	For dehydration
Forceps	No specific supplier		To manipulate garden samples
Glass beaker	No specific supplier		For dehydration
Glass Petri dish	No specific supplier		To manipulate garden samples
Glass pipette	No specific supplier		To fix and dehydrate samples
Glutaraldehyde (Aqueous Glutaraldehyde EM Grade 25%)	Electron Microscopy Sciences	16220	To prepare Karnovsky’s fixative
Gold target	Ted Pella, Inc.	8071	To sputter coat with gold
Hydrochloric acid	Sigma-Aldrich	320331	For adjusting solutions pH
Image editor	Photoshop	any version	To adjust images
Paraformaldehyde (Paraformaldehyde 20% Aqueous Solution EM Grade)	Electron Microscopy Sciences	15713	To prepare Karnovsky’s fixative
Propilene recipient	No specific supplier		For maintaining alive ant colonies
Scanning Electron Microscope	JEOL	IT300 SEM	For sample imaging
Sodium cacodylate trihydrate	Sigma-Aldrich	C0250	For preparing sodium cacodylate buffer
Spatula	No specific supplier		To manipulate garden samples
Specimen containers with 15 mm dia. x 10 mm high	Ted Pella, Inc.	4591	For critical point drying
Sputter coater	Baltec	SCD 050	To coat with gold
Stub (Aluminium mount, flat end pin) 12.7 mm x 8 mm	Electron Microscopy Sciences	75520	For mounting samples