Experimental and molecular approximation to microbial niche: trophic interactions between oribatid mites and microfungi in an oligotrophic freshwater system

Study site

The CCB (26°50′N, 102°8′W) poses a biodiversity paradox, as this extremely oligotrophic oasis harbors remarkably high levels of biodiversity (Souza et al., 2006, 2012), being regarded a model for Cambrian food webs at a stoichiometric knife-edge (Elser et al., 2006). Numerous freshwater systems containing rich microbial diversity (including transient aquatic microfungi) are distributed within this area (Souza et al., 2006, 2012; Desnues et al., 2008; Velez et al., 2016). The climate in the CCB is generally hot and arid with extreme temperatures such as 45 °C in July and temperatures below 0 °C being commonly reported in January (Conagua, 2015). Specifically, the Churince freshwater system (26°50′N; 102°8′W) comprises a perennial spring and two desiccation lagoons that are connected by short shallow streams (López-Lozano et al., 2012). This endangered system exhibits extreme and fluctuating temperature and water chemistry conditions, characterized by high concentrations of sulfates and scarce nutriments regarded as oligotrophic (Minckley & Cole, 1968; Tobler & Carson, 2010).

Sampling

We submerged 15 wood baits of Pinus sp. (16 × 10 × 2 cm) along the Churince freshwater system (Figs. 1A and 1B) during the transition from wet to dry season, from September 2014 to March 2015 (field permit FAUT-0230, emitted by Secretaría de Medio Ambiente y Recursos Naturales, Subsecretaría de Gestión para la Protección Ambiental, Dirección de Vida Silvestre). An elevation map of the CCB (Fig. 1A) was generated using layers from the Instituto Nacional de Estadística y Geografía (INEGI, 2013) with the packages sp. (Pebesma & Bivand, 2005) and maptools (Bivand & Lewin-Koh, 2015). After six months, the wood panels were recovered (Figs. 1C and 1D), placed into individual Zip-lock® plastic bags, transported to the laboratory, and processed within 24 h of arrival to the laboratory. The test blocks were processed according to the methodology described by Jones (1971), which in summary consists of incubating test blocks in plastic boxes (with sterile moist paper towels forming a moist chamber). Moist chambers were kept at room temperature (22–26 °C) under natural daylight and examined periodically for the presence of mites and fungal structures over a period of six months using a stereo-microscope (Discovery V8 Stereo; Carl Zeiss, Göttingen, Germany).

Microfungal isolation and identification

Fungi from wood baits were isolated using Potato Dextrose Agar (PDA; Fluka Analytical; Sigma-Aldrich, St. Louis, MO, USA) supplemented with 0.5 g/mL of penicillin-G and 0.3 g/mL of streptomycin sulfate to prevent bacterial development. The taxonomic assignment of axenic isolates was based on morphological and molecular characterization. The morphology of microfungal reproductive structures was examined (Samson et al., 2014; Yilmaz et al., 2016) using a transmitted light microscope “Carl Zeiss Axiostar Plus” equipped with the microscope software “Zeiss ZEN”. The universal DNA barcode marker for fungi (nuclear ribosomal internal transcribed spacer region, ITS rDNA; Schoch et al., 2012) was analyzed for each isolate. To do this, microfungi were transferred to Potato Dextrose liquid medium until adequate growth occurred. Mycelium was collected and DNA was extracted using the protocol described by Doyle & Doyle (1987). The ITS rDNA region was amplified and sequenced using ITS1 and ITS4 primers as previously described by White et al. (1990). Sanger sequencing reactions were performed in both directions by the High Throughput Genomics Center Facility, University of Washington and by the Macrogen Inc., South Korea. Isolates and total DNA were deposited in the culture collection of the Laboratorio de Evolución Molecular y Experimental, Instituto de Ecología, headed by Dr. Valeria Souza, and the Laboratorio de Ecología Molecular de Micromicetos en Ecosistemas Amenazados, Instituto de Biología, headed by Dr. Patricia Velez; both at the Universidad Nacional Autónoma de México (UNAM). Cultures and total DNA are available for research upon request.

Quality assessment and assembly of the sequences was performed using the finishing tool Consed version 29.0 (Ewing & Green, 1998; Ewing & Green, 1998; Gordon, Desmarais & Green, 2001). For the taxonomic assignment of microfungal isolates, sequence homology was evaluated through the comparison to the UNITE database version 7.2 using the BLAST algorithm (with an e-value > 0.001; Abarenkov et al., 2010; Koljalg et al., 2013). Sequence similarity for defining OTUs was set as proposed by Millberg, Boberg & Stenlid (2015), with a cut-off value of 98–100% for presumed species, 94–97% for genus level and 80–93% for order level. For conflicting hits, the lowest common rank level was used (Peršoh et al., 2010; Table S1).

Mite culture, identification, and food preference bioassays

Oribatid mites were maintained on moist wood panels that were profusely colonized by the tested fungi until processing. For taxonomic identification and illustration, the specimens were mounted in lactic acid on temporary cavity slides. Images were obtained using an AxioCam MRC5 camera mounted on a Carl Zeiss AxioZoom V16 microscope. Mites were identified based on the taxonomic characteristics described for the genus Trhypochthoniellus (Balogh, 1972; Weigmann, 1999; Fujikawa, 2000; Kuriki, 2005).

Additionally, a phylogenetic reconstruction based on COI sequence data was performed. The gnatosome (bucal sections) and leg portions were dissected and collected into 0.2 mL PCR microtubes containing buffer (Phire Animal Tissue Direct PCR Kit; ThermoFisher Scientific; Waltham, MA, USA). The reactions were carried out according to the manufacturer’s protocol at a 55 °C annealing temperature. Oribatid specific primers were used (Arch1 and HCO2198; Heethoff et al., 2007) at a 0.5 μM final concentration. Negative controls were included, where the sterilized needle was dipped into the PCR buffer prior to mite dissection. Sanger sequencing reactions were performed in both directions by the Laboratorio de Secuenciación Genómica de la Biodiversidad y de la Salud, Instituto de Biología UNAM, México. The assembly of forward and reverse sequences was done using Geneious v.9.1 (Biomatters, Auckland, New Zeland). Sequences were manually edited and trimmed using BioEdit software (v7.0.5; Hall, 2005). Using a BLAST-n analysis, closely related NCBI sequences were selected to compute a phylogenetic reconstruction including analogous oribatid species and external taxonomic categories as previously reported by Klimov et al. (2018). Phylogenetic placement of the CCB mites was determined according to a Maximum-Likelihood (ML, 1,000 bootstrap) analysis run in RAxML v.8.2.10 (Stamatakis, Hoover & Rougemont, 2008) using GTRCAT model; Cryptocellus sp. AD1430 was defined as an outgroup. All ML inferences were performed in the CIPRES Science Gateway portal (Miller, Pfeiffer & Schwartz, 2010).

Mite diet preference experiments were implemented as described by Hubert et al. (2004). Oribatids were transferred from wood baits into culture chambers consisting of a 2 mm thick layer of plaster of Paris mixed with charcoal (ratio 9:1) in small Petri dishes (5 cm in diameter). Five individuals were placed in each culture chamber and starved for five days. Then, to corroborate active feeding, a fungal plug (1 cm²) of each one of the three isolated fungal taxa was offered to the mites individually. During a second experiment, food preference was tested by presenting fungi simultaneously to starved mites in the following combinations: Pleosporales sp. (a), Talaromyces sp. (b), and Aspergillus niger (c) individually; paired combinations a-b, a-c, and b-c; and simultaneously a, b, and c. Culture chambers were incubated under antiseptic conditions at 25 °C, 12 h light-12 h dark cycles, and 85% RH, and moistened regularly (3 day intervals) with 5 mL of sterile distilled water.

All experiments were run using negative controls (no fungus). Due to the limited availability of biological material on our wood panels, a replicate was implemented only for the first experiments (corroboration of the active consumption of a, b, and c). Individuals of each treatment were observed for several weeks (Table 1). During this period, the number of mites under, near, and far from the fungal plug was recorded. Also, the number of visits made by the mites among discs was counted. Based on this information, oribatids on discs were classified as “actively feeding,” while those which were walking around these discs were classified as “looking for food” (Hubert, Sustr & Smrz, 1999).

Gut-content analyses

Oribatids were examined for the content of their digestive tract through culture-dependent (plating) and culture-independent (direct PCR focusing on the ITS fungal barcoding region) methods. After four weeks of the food preference bioassays, two mites under each treatment were collected and superficially disinfected in 1.5 mL microtubes containing 1 mL solution of 75% methanol, and then transferred to 1.5 mL microtubes with 1 mL of sterile 0.03% Tween-80 solution. All tubes were gently vortexed for 30 s. Subsequently, under sterile conditions, the idiosome of each mite was dissected out with tungsten wire needles, squashed on concave slides, re-suspended in 300 μL of sterile water, and plated (100 μL) on Petri dishes containing PDA in triplicate per individual. To confirm the aseptic conditions during the experiment, negative controls were plated using 100 μL of sterile water.

Additionally, for the molecular analysis, the gut of superficially disinfected animals under each treatment was dissected out and collected into 0.2 mL PCR microtubes containing PCR buffer (Phire Animal Tissue Direct PCR Kit; ThermoFisher Scientific; Waltham, MA, USA). PCR reactions were performed according to the manufacturer’s protocol, using fungal-specific primers for the nuclear ribosomal ITS region (ITS1 and ITS4, 0.5uM final concentration). Two negative controls (buffer, and buffer where the sterilized dissecting needle was soaked) were also included. Sequencing reactions were carried out in both directions by the Laboratorio de Secuenciación Genómica de la Biodiversidad y de la Salud, Instituto de Biología UNAM, México. The quality assessment and sequences assembly were performed using the finishing tool Consed version 29.0 (Ewing & Green, 1998; Ewing et al., 1998; Gordon, Desmarais & Green, 2001). The taxonomic assignment for fungal gut-content sequences was based on the evaluation of sequence homology through the NCBI-BLAST algorithm (Table S1). The sequence similarity for taxonomic assignment was considered as proposed by Millberg, Boberg & Stenlid (2015).

Approximation to the niche space and dietary preferences in Trhypochthoniellus sp

In general, oribatids feed on a wide range of materials (Dhooria, 2016), having a dietary preference for microfungi with melanized cell walls (Mitchell & Parkinson, 1976; Maraun et al., 1998, 2003; Schneider & Maraun, 2005). In contrast, our results suggest that aquatic oribatids in the CCB may feed on non-dematiaceous taxa such as A. niger, Pleosporales sp. and Talaromyces sp. Though DNA may be degraded to some extent during gut passage (Renker et al., 2005), we were able to confirm these observations through the recovery of the fungal ITS genetic barcode region from the gut of our mites by direct PCR. Here it is important to mention that currently an inclusive genetic barcode for the Fungal Kingdom is still unapproachable. The best approximation has been achieved through the implementation of the ITS region as the standard fungal barcode. As this region has the highest probability of successful identification for the broadest range of fungi (Schoch et al., 2012). However, poor species-level resolution has been reported in certain taxonomic groups (e.g., Hughes, Tulloss & Petersen, 2018). Therefore, this limitation should be considered in ITS-based species delimitations.

Cuatro Ciénegas mites demonstrated their ability to feed on the usually avoided and unsuitable fungus A. niger, an apparently previously empty niche space for mites, according to previous investigations (Hubert et al., 2004). The consistent ingestion of A. niger in all the treatments of our study disagrees with previous findings classifying this species as non-palatable and unsuitable for mite development and population growth (Hartenstein, 1962; Sinha, 1966; Hubert et al., 2004). However, our results confirm other study findings where spores of this fungus have been detected in mite guts (Behan & Hill, 1978; Hubert, Kubátová & Šárová, 2000; Bandyopadhyay, Khatun & Chatterjee, 2009). In addition, we suggest that mite feeding on Pleosporales sp. may be associated with the wide distribution and abundance of this taxon in the CCB (Velez et al., 2016), representing a continuous and accessible food source. Similarly, the appeal of A. niger for our oribatids might be related to fast growth rates, abundant biomass and conidia production.

Even though multiple-choice tests may be biased by the spectrum of food items offered to individuals, the tested microfungi and mites were observed to coexist in our wood baits’ mesocosms (by definition a simplification of the environment), representing ecosystem dynamics. Hence, in order to delimit the total niche space of this oribatid, further experiments are needed evaluating of additional food sources, possible predators, and the mimicry of CCB distinctive physicochemical variables.

Identification of core factors associated with fungal palatability in microarthropods remains elusive (Koukol et al., 2009). Conventionally, feeding preferences have been related to fungal morphology and physiology. For example, mites avoid fast-growing microfungi, which may capture individuals in their long hyphae. Thus, species with short-hyphae are preferentially grazed (Mills & Sinha, 1971). Despite, the fact that A. niger, Pleosporales sp. and Talaromyces sp. may be considered as fast-growing taxa, our observations revealed that the mycelia of these microfungi allowed free movement of mite individuals on the wood panels and during in vitro bioassays, enabling active feeding.

Additionally, we approximated a Relative Width of Digitus Mobilis (RWD: relationship between length l and base width of the digitus mobilis w) of 4.5–5 μm in our oribatids (Fig. 6). This estimate is an indirect indicator of food items used by oribatid species (Kaneko, 1988), as mites with smaller RWD (around 2–4 μm) hypothetically feed selectively on minute matter (Wallwork, 1958). Therefore, conidia dimensions might be considered as a key trait for microarthropod ingestion (Koukol et al., 2009). This result suggests that spherical/ovoid, small (3–5 μm) conidia produced by fungal species in the present work are potentially consumable by mites, enabling the ingestion of all fungal components.

Mite-mediated fungal dispersal

Fungal spore dispersal by animals (vertebrates and invertebrates) has been documented in some systems (Allen, 1987; Warner, Allen & MacMahon, 1987; Klironomos & Moutoglis, 1999; Gehring, Wolf & Theimer, 2002), yet information on microfungal dispersal agents remains scarce. Our culture-dependent results on the gut content analysis suggest that, based on their capabilities to digest fungal tissues (cell walls and contents), these oribatids may be considered more as grazers (Siepel & de Ruiter-Dijkman, 1993) than spore dispersers (via feces), as we could not recover fungal isolates from our culture-based approach. This agrees with previous work where thorough digestion of fungal cell content by mites is reported (Van der Drift, 1965; Ponge & Charpentie, 1981; Hubert, Zilova & Pekar, 2001), highlighting its alimentary value (Hubert et al., 2004).

Nevertheless, based on their fungal-based diet, the assessed mites may still represent important dispersers for transient aquatic fungi through some other means (Blackwell & Malloch, 1991; Renker, Alphei & Buscot, 2003; Rantalainen et al., 2004). Despite conidia dispersal via mite feces being unlikely, dispersal on the body surfaces of individuals may be significant (Wallwork, 1983; Bernini, 1986; Paine, Raffa & Harrington, 1997; Hubert, Kubátová & Šárová, 2000; Ocak, Dogan & Ayyildiz, 2008). As has been it was observed in other systems, this oribatid-aided conidia transport may represent an advantage for the colonization of new substrata, and faster recovery after disturbances under harsh, fluctuating environmental conditions inherent to the CCB (Hanlon & Anderson, 1979; Hanlon, 1981; Visser, Whittaker & Parkinson, 1981).