Nematode-free agricultural system of a fungus-growing termite

Fungus-growing termites forage dead plant materials from the field to cultivate symbiotic Termitomyces fungi in the nest. Termite foraging behavior and the entry of symbiotic arthropod inquilines may transfer nematodes into a nest and adversely affect fungus production. To test whether nematodes were transferred to fungus gardens by termites and inquilines, we examined the occurrence of nematodes in fungus gardens, five termite castes, and nine species of inquilines of a fungus-growing termite, Odontotermes formosanus. Our results revealed that nematodes were commonly carried by foraging termites and beetle inquilines. Numerous nematodes were found under the beetle elytra. No nematodes were found on termite larvae, eggs, and wingless inquilines. In addition, nematodes rarely occurred in the fungus garden. By observing the response of nematodes to three species of Termitomyces spp. and the fungus gardens, we confirmed that the fungus and fungus gardens are not actually toxic to nematodes. We suggest that nematodes were suppressed through grooming behavior and gut antimicrobial activity in termites, rather than through the antimicrobial activity of the fungus.

genic to fungi are potential pests in agricultural systems and are potentially transferred to fungus gardens when (1) insect foragers carry materials with microbes from extranidal environments and contaminate the cultivating substrates and fungus; (2) numerous "guests" (inquilines), such as myrmecophiles, associated with ants, and termitophiles, associated with termites [34][35][36][37] enter nests of social insects through chemical mimicry or chemical insignificance 38,39 , and may carry pathogenic microbes into the agricultural system; and (3) pathogens are vertically transmitted through colonizers, such as the alates of ants and termites, carrying pathogens from parental colonies 1 to newly founded colonies.
Pest management strategies have also been reported in agricultural systems of insects. For example, in fungus gardens of leaf-cutter ants, the parasitic microfungus Escovopsis (Hypocreales: Hypocreaceae) are suppressed by antimicrobial chemicals produced by actinomycete bacteria, a symbiont of leaf-cutter ants [40][41][42] . Um, et al. 43 also observed that a strain of Bacillus in fungus gardens of termites suppressed non-Termitomyces microbes. Allogrooming behavior among nest mates, which has been reported in termites and ants, may remove pathogens from body surfaces 2,44,45 . In termites, partitioning of foraging and nest-caring tasks among individuals has been suggested to inhibit the transfer of pathogens in colony 46 . Fungus-growing termites were also hypothesized to suppress non-Termitomyces microbes by passing fungus garden materials through their guts, which generally have high levels of antimicrobial activity 47 . purpose of this study. We aimed to understand the transmission and management of nematodes in an agricultural system of a fungus-growing termite, Odontotermes formosanus (Shiraki). We investigated three transmission pathways of nematodes in the agricultural system: (1) vertical transmission via alate, (2) horizontal transmission via foragers, and (3) horizontal transmission via inquilines. In addition, we assessed the potential of fungus gardens as media for nematodes. We also examined factors that may suppress nematode populations in termite nests.

Materials and Methods
Hosts of nematodes. Five nests of Odontotermes formosanus from different localities in Taiwan were located by searching for fruiting bodies (mushrooms) of Termitomyces on the ground and were excavated to investigate their potential as hosts of nematodes in nests, including fungus gardens, termites, and inquilines (Figs 1 and 2; Supplementary Information, Table S1); 3-7 fungus gardens were collected from each nest. Since the compositions of microbes may vary between the upper and the lower parts of fungus gardens 47 , 6 g from the upper and lower parts were collected separately (Supplementary Information, Table S2) to investigate the presence of nematodes. The upper parts of fungus gardens are dark-colored fresh substrates, comprising partially-digested plant material. By contrast, the lower parts are whitish aged substrates, comprising dense fungal hyphae and highly decomposed plant materials (Fig. 1e) 48 . Between the dark-colored fresh substrates and the whitish old substrates is the active zone in which the Termitomyces conidiophores are mainly produced. The upper and lower parts were sampled separately by dividing the active zone at the middle, and therefore the relatively fresh and aged active zone will be sampled and included in the upper and lower parts, respectively.
Termites and inquilines in each nest were sought and collected exhaustively. To investigate whether the foraging termites carry nematodes, foraging termites on feeding sites (logs, leaves, tree bark) found within 20 m of nests were collected. Feeding sites of O. formosanus were distinguished by the presence of soil-sheeting, which is specifically built by foragers of O. formosanus. Inquilines in O. formosanus nests were identified based on the morphological descriptions of termitophilous insects in Taiwan 49 (Fig. 2). To examine whether inquilines carry nematodes in the dispersal stage, dispersing adults of a termitophilous beetle, Ziaelas formosanus Hozawa (Coleoptera: Tenebrionidae) (Fig. 2d) were collected using a light trap at Xiaping Tropical Botanical Garden, the Experimental Forest, National Taiwan University, Nantou, Taiwan (23.77°N, 120.67°E). Ziaelas formosanus is commonly observed in fungus gardens of O. formosanus. Dispersing termite alates (Fig. 1a) were collected from four localities (Supplementary Information, Table S3) to examine whether nematodes were transferred from parental colonies. Voucher specimens of nematodes, termites, and inquilines were deposited at National Chung Hsing University (NCHU).
To clearly observe nematodes on agar plates, termite eggs were evenly spread on agar plates. Head capsules and digestive systems of termite larvae, workers, and alates were further dissected. To avoid high densities of body parts interfering with the observation of nematodes, the number of larvae and major and minor workers on each agar plate was limited to 50 individuals, and that of termite alates was limited to 40 individuals (20 males and 20 females). Nematodes of inquilines were examined using the same methodology applied to termites. The number of inquilines on each agar plate was approximately 30 individuals. All inquilines were examined on a single agar plate if the nest contained less than 30 individuals. The fungus gardens, eggs, dissected bodies of termites, and inquilines were kept on agar plates for one month at room temperature to allow nematodes to propagate for inspection. The percentage of nematodes present in fungus gardens, termite castes, and inquilines was calculated by dividing the number of plates with nematodes with the total number of plates examined and multiplying by 100.
Morphotyping and molecular characterization of nematodes. When nematode propagation was confirmed on agar plates, the nematodes were removed manually, their morphologies assessed, and were transferred to nematode digestion buffer 50,51 for molecular identification. However, if the number of individuals was www.nature.com/scientificreports www.nature.com/scientificreports/ sufficient to establish cultures, they were transferred to artificial media (i.e., nematode growth medium inoculated with OP50 Escherichia coli strain for bacteria feeders or 2% malt extract agar inoculated with Botrytis cinerea Pers. for fungal feeders). Cultured nematodes were kept as laboratory strains and the detailed taxonomic studies are presented elsewhere.
Since all nematodes collected during dissection were dauer (dispersal) juveniles that do not have genus/ species-specific morphological characters, they were transferred to nematode digestion buffer and molecularly characterized based on ribosomal RNA sequences, that is, near full-length 18S (SSU) and/or D2-D3 expansion segments of 28S (D2-D3 LSU) regions. Molecular sequences were determined through PCR-based direct sequencing according to Kanzaki and Futai 52 and Ye, et al. 53 . Generated sequences were compared with those deposited in the GenBank database (http://www.genome.jp/dbget-bin/www_bfind?genbank-today) using the BLAST homology search program (http://blast.ddbj.nig.ac.jp/blastn?lang=en).  Table S4) were isolated and cultured on Potato Dextrose Agar (PDA) to examine their nematocidal activity against four nematode genera isolated from termites, fungus gardens, or inquilines. Termitomyces spp. strains and their identification based on the internal transcribed spacer (ITS) region were provided by Mycology Laboratory of NCHU (Supplementary Information). Fungus cultures were inoculated with five individuals of one species from each nematode genus. To test whether nematocidal activity was present on fresh and aged substrates of fungus gardens, three fungus gardens were collected from NCHU and separated into fresh and aged parts. Six grams of fresh or aged substrates were vortexed for homogenization and each of the 0.3-g substrates placed on different 2% agar plates (ø = 50 mm). Each plate www.nature.com/scientificreports www.nature.com/scientificreports/ was inoculated with 10 individuals of a single nematode species. Locomotion and pharyngeal pumping in the nematodes were examined at 1, 24, or 48 hours after exposure to fungal hyphae to observe if the fungal culture would exhibit any toxicity to the nematodes.

Results
Nematodes isolated. Seven genotypes of nematodes were recognized from the culture plates, namely, a fungal feeder, Aphelenchoides sp. and 6 bacteria feeders, comprising three types of Acrostichus sp. (Type A, B, and C), 2 types of Halicephalobus spp. (Type A, B), and a Diplogastrellus sp. The molecular sequences, BLAST results, and notes on each genotype determined in the present study are presented in Supplementary Information.
No nematodes were observed when dissecting 21 fresh and 21 aged substrates of fungus gardens (Table S2). A few Halicephalobus sp. (type B) were observed in only one plate 3 weeks after inoculation of the fresh substrate from Xiaping (1 plate detected/21 plates cultured, ~4% occurrence) (Table S2). No nematode was observed in www.nature.com/scientificreports www.nature.com/scientificreports/ the plates hosting the aged substrates (0/21 plates, 0%) ( Table S2). The results indicate that termite fungiculture systems were almost nematode free.
No nematodes were isolated from the 160 swarming alates (80 males and 80 females) of O. formosanus from the four localities that were examined (Tables S3 and S5) (0/4 plates, 0%); the results indicated that nematodes were not transferred from parental to incipient colonies. Therefore, within-colony transmission of nematodes via alates was unlikely.
Nematodes were not found when dissecting the termite workers collected in or outside the nests. After culturing on plates, four nematode genotypes were identified from foraging major workers (Supplementary Information, Table S5): an Aphelenchoides sp., a Diplogastrellus sp., and two genotypes of Halicephalobus spp. (type A and B) (3/6 plates, 50%) ( Table S5). The major workers collected in the nest hosted three nematode genotypes, namely Diplogastrellus sp., Halicephalobus spp. (type A and B) (4/7 plates, 57%) (Table S5). Minor workers in nests hosted two nematode genotypes (Aphelenchoides sp. and Diplogastrellus sp.) (1/5 plates, 25%) (Table S5). No nematodes were detected in termite eggs (0/6 plates, 0%) or larvae (0/13 plates, 0%) ( Table S5). The results suggest that nematodes are largely transferred to nests by major workers, which forage and bring plant materials to nests. Therefore, it is likely that horizontal transmission of nematodes via foragers occurred frequently.
General information on presence of nematodes on termites, inquilines, and fungus gardens is summarized in Fig. 4. The results revealed that although nematodes were transferred to fungus gardens, the populations were suppressed before or after they entered fungus gardens.

Discussion transmission and management of nematodes.
The results of the present study revealed that nematodes commonly enter termite nests via two sources: (1) termite foraging workers and (2) coleopteran inquilines. However, being a potential pest, nematodes were not transferred to other castes after entering nests, considering the absence of nematodes on swarming alates, larvae, and eggs. In addition, although nematodes were not www.nature.com/scientificreports www.nature.com/scientificreports/ repelled by fungus and fungus gardens, they rarely occurred on fungus gardens, which indicated that nematodes were likely suppressed before entering fungus gardens. We suggest that the low quantities of nematodes in agricultural systems of fungus-growing termites were due to integrated management by termites through three strategies: (1) prevention of transmission of nematodes by partitioning of tasks and diets; (2) suppression of nematodes in fungus gardens by passing fresh substrates through their guts; and (3) decontaminating the vectored nematodes through allogrooming behavior.
In fungus-growing termites, major workers are generally foragers while minor workers are mostly nest-keepers. For example, in a fungus-growing termite, O. distans Holmgren and Holmgren, 94.8% of the individuals collected from feeding sites were major workers, and 95.1% of the individuals collected from queen chambers were minor workers 54 . Similarly, in another fungus-growing termite, Macrotermes subhyalinus Rambur, 88.8% of foragers were major workers, and 56.1% of individuals in the nest were minor workers 55 . Such partitioning of tasks predictably decreases the level of interactions between major workers and queens, larvae, or eggs, and www.nature.com/scientificreports www.nature.com/scientificreports/ may lower the probability of transferring nematodes. In addition to task partitioning, age polyethism on diets was also observed in M. subhyalinus. Young major workers that molted less than 30 days before stay in the nest, consume the plant material collected by foragers, and construct fungus gardens, while the older major workers that molted more than 30 days before are more likely to forage for plant materials in the field and largely consume aged substrates of fungus gardens in their diets 55 . The diet partitioning in fungus-growing termites separates the constructors of fungus gardens and the foragers, which likely decreases the probability of foragers contaminating fresh substrates. Thomas 47 reported that an average of 229.8 and 8.6 fungal isolates were found in a gram of plant material collected and in fresh substrate of fungus gardens constructed by the fungus-growing termite, M. bellicosus (Smeathman), respectively, which supported the claim that microbial populations were suppressed after passing the gut of major workers. We propose that O. formosanus controlled the populations of nematodes inhabiting the plant materials collected by integrating the task and diet partitioning, and suppressing the microbial populations in fresh substrates of fungus gardens.
Allogrooming behavior, which cleans microbes growing on body surfaces [56][57][58][59] , was reported in a fungus-growing termite, Macrotermes michaelseni (Sjöstedt) 60 . Suppression of microbial populations by allogrooming behavior has been confirmed in multiple groups of termites, such as genus Reticulitermes (Blattodea: Rhinotermitidae) 61 and genus Zootermopsis (Blattodea: Archotermopsidae) 59 . In the present study, no nematodes were observed on the body surfaces of alates, workers, larvae, eggs, and most of the inquilines. Even in the millipedes, which commonly harbor internal parasites [62][63][64] and external phoretic nematodes 65 , no nematodes were isolated. For inquilines, nematodes were only found beneath the elytron of coleopteran inquilines, which is a site that is not likely to be cleaned by allogrooming behavior in termites. In the laboratory, we observed that O. formosanus performed allogrooming behavior on all inquilines (unpublished data). We propose that allogrooming behavior in the fungus-growing termites not only managed the populations of nematodes carried by termites, but also the populations carried by inquilines.  [66][67][68][69] . The genus Diplogastrellus is commonly found in rotten plants 70 and insects associated with rotten plants 71 , as well as in humid soils that are rich in organic matter 72 . Similarly, the genus Aphelenchoides sp. is associated with humid soils that are rich in organic matter. The nematode species composition in O. formosanus is likely associated to soil environment, similar to in termites foraging or nesting in soil. For example, Halicephalobus was isolated from the subterranean termite Reticulitermes lucifagus in Corsica 23 , from dry/dampwood termites in both Florida and Taiwan 27,73 , from Coptotermes formosanus in Florida 74 , and from many different termite species in Central America 30 . Aphelenchoides was isolated from two termite species: Reticulitermes lucifagus in Corsica 23 and Hodotermopsis sjostedti in Japan 75 . We suggest that the nematodes isolated from termite workers were from their foraging environments, e.g., soils and/or substrates that the termites foraged.

Diversity of nematodes of termites and inquilines.
Only a single nematode genus, Acrostichus, was isolated from inquilines, which was not found in termites. The result indicated that nematodes were not transferred between termites and inquilines, and the sources of nematodes in inquilines were different from the sources of nematodes in termites. Nevertheless, the life cycle and ecology of Acrostichus spp. is unknown.
Conclusion. Vertical transmission of nematodes through termite dispersal alates to eggs and larvae was not observed in the fungus-growing termite, O. formosanus. Nematodes were transferred into termite nests through two horizontal transmission pathways, including via termite foragers and inquilines. Nematode species observed on termite foragers are associated with the soil environment, but inquilines carried specific nematode species, which indicates the two horizontal transmission pathways are exclusive. Laboratory experiments revealed neither fungus garden substrates nor Termitomyces fungi were toxic to nematodes, but nematodes were barely present in fungus gardens in the field, which indicates termite hygienic behavior plays a vital role in nematode management in the nests.

Data Availability
The sequences of nematodes and GenBank accession number of Termitomyces spp. are available as Supplementary Information to this paper.