Caste specialization in behavioral defenses against fungus garden parasites in <Emphasis Type="Italic">Acromyrmex octospinosus</Emphasis> leaf-cutting ants

Division of labor and caste specialization plays an important role in many aspects of social insect colony organization, including parasite defense. Within leaf-cutting ant colonies, worker caste specialization permeates colony tasks ranging from foraging, substrate incorporation, brood care, and chemical defenses via glandular secretions and mutualistic bacteria. Leaf-cutting ants rely on protecting a mutualistic fungus they grow for food from microfungi in the genus Escovopsis that parasitizes the ant–fungus relationship. Here, we examine whether Acromyrmex octospinosus leaf-cutter ant castes (minors and majors) display task specialization in two behavioral defenses against Escovopsis: fungus grooming (the removal of Escovopsis spores) and weeding (the removal of large pieces of Escovopsis-infected fungus garden). Using behavioral observations, we show that minors are the primary caste that performs fungus grooming, while weeding is almost exclusively performed by majors. In addition, using a sub-colony infection experimental setup, we show that at the early stages of infection, minors more efficiently remove Escovopsis spores from the fungus garden, thereby restricting Escovopsis spore germination and growth. At later stages of infection, after Escovopsis spore germination, we find that major workers are as efficient as minors in defending the fungus garden, likely due to the increased importance of weeding. Finally, we show, using SEM imaging, that the number of sensory structures is similar between minor and major workers. If these structures are invoked in recognition of the parasites, this finding suggests that both castes are able to sense Escovopsis. Our findings support that leaf-cutter ant behavioral defense tasks against Escovopsis are subject to caste specialization, likely facilitated by worker sizes being optimal for grooming and weeding by minors and majors, respectively, with important consequences for cultivar defense.     

Experimental evidence of a tripartite mutualism: bacteria protect ant fungus gardens from specialized parasites

Symbioses shape all levels of biological organization. Although symbiotic interactions are typically viewed as bipartite associations, with two organisms interacting largely in isolation from other organisms, the presence and importance of additional symbionts is becoming increasingly more apparent. This study examines the importance of a third mutualist within the ancient symbiosis between leaf-cutting ants and their fungal cultivars. Specifically, we experimentally examine the role of a filamentous bacterium (actinomycete), which is typically carried on the cuticle of fungus-growing ants, in suppressing the growth of a specialized microfungal parasite ( Escovopsis ) of the fungus garden. We conducted two-by-two factorial design experiments crossing the presence/absence of actinomycete with the presence/absence of Escovopsis within small sub-colonies of Acromyrmex octospinosus . In these experiments, infection by Escovopsis became much more extensive within fungus gardens and had a greater impact on the health of gardens in those sub-colonies with the bacterium removed from workers as compared to gardens with the bacterium still present on the ants. We establish that the actinomycete bacterium is most abundant on those major workers tending the garden, providing further support that the bacterium is involved in garden hygiene. We also found a significantly higher abundance of actinomycete on workers in colonies experimentally infected with Escovopsis as compared to uninfected control colonies. We suggest that mutualisms between antibiotic-producing microbes and higher organisms may be common associations that are mostly overlooked and that the role of symbionts in reducing the impact of parasites is likely an important aspect in the cost-benefit assessment of mutualisms.     

Humidity preference for fungus culturing by workers of the leaf-cutting ant Atta sexdens rubropilosa

Summary: The hygropreference of gardening workers of the leaf-cutting ant Atta sexdens rubropilosa was investigated in the laboratory using a gradient of relative humidity. Gardening workers were placed, together with pieces of fungus garden, in small, interconnected nest chambers offering four different relative humidities: 33 %, 75 %, 84 % and 98 % RH. Workers were allowed to move freely between them and to relocate the fungus following their humidity preference. While workers distributed themselves randomly in the nest chambers, they located the fungus gardens in the chamber with the highest humidity. These results indicate that gardening workers are able to sense differences in relative humidity, and that this ability is shown when they are engaged in fungus culturing. Humidity is discussed as one of the relevant variables that probably underlay the evolution of regulatory responses for the control of fungus growth in leaf-cutting ants.     

The epizootiology of a Metarhizium infection in mini-nests of the leaf-cutting ant Atta sexdens rubropilosa

There is growing interest in the use of entomopathogenic organisms to control leaf-cutting ants (Hymenoptera: Formicidae: Attini). However, the way leaf-cutting ants react as a colony to biohazards is poorly understood. We investigated the effects of Metarhizium anisopliae (Metschnikoff) (Deuteromycotina: Hyphomycetes) applied to the foraging arenas of mini-nests (queenless sub-colonies) of the leaf-cutting ant Atta sexdens rubropilosa (Forel). Dry spores were applied either alone or mixed with citrus powder, at 0.5 g or 0.05 g per mini-nest. The spores were removed four days after application, and all dead ants removed every three days. Ant numbers near the Metarhizium increased as the ants attempted to clean up the biohazard. The ants attempted to place the spores in piles, which they then covered over with other material. They were able to deal with the low doses in this way, but the high doses overwhelmed them. All treated mini-nests suffered increased ant mortality during the first ten days after application. This mortality was particularly high in the media worker caste which had played the major role in attempting to clean up the spores. Foraging activity decreased, as did the health of the fungus gardens. The mini-nests exposed to the low dose of spores mixed with citrus powder then recovered fully. The health of the other treated mini-nests declined gradually until around 26 days after application, when they began deteriorating sharply. However, the decline of these mini-nests after day 26 was not due directly to the pathogenic action of the Metarhizium, nor to the initial ant mortality it had caused. The results suggest that the social stress caused by even such a short-lived Metarhizium epizootic was sufficient to cause the decline and ultimate death of the mini-nests. This has important implications for the control of leaf-cutting ants. It also demonstrates how important the social homeostasis of the colony is to leaf-cutting ants.     

Limited impacts of the fungus Syncephalastrum on nests of leaf-cutting ants

Leaf-cutting ants interact naturally with a range of antagonistic microorganisms, among them the soil-borne fungus Syncephalastrum. The antagonism of this fungus to the leaf-cutting ants’ fungal cultivar has been shown in studies without the ant queens. So far, the impacts of this fungus on whole colonies (queenright) of leaf-cutting ants are unknown. We assessed the impacts of Syncephalastrum on queenless and queenright colonies of Acromyrmex subterraneus subterraneus. In general, Syncephalastrum negatively impacted leaf cutting but not midden production or colony weight. This impact was greater in queenless colonies. Nevertheless, it did not compromise the survival of any colony. This indicates that the virulence of this fungus to leaf-cutting ant colonies may be limited in a more realistic set-up than previously reported. We propose that future laboratory studies also use queenright colonies where possible, and that the diverse species of leaf-cutting ants also be considered.     

The dynamics of plant cell-wall polysaccharide decomposition in leaf-cutting ant fungus gardens

The degradation of live plant biomass in fungus gardens of leaf-cutting ants is poorly characterised but fundamental for understanding the mutual advantages and efficiency of this obligate nutritional symbiosis. Controversies about the extent to which the garden-symbiont Leucocoprinus gongylophorus degrades cellulose have hampered our understanding of the selection forces that induced large scale herbivory and of the ensuing ecological footprint of these ants. Here we use a recently established technique, based on polysaccharide microarrays probed with antibodies and carbohydrate binding modules, to map the occurrence of cell wall polymers in consecutive sections of the fungus garden of the leaf-cutting ant Acromyrmex echinatior. We show that pectin, xyloglucan and some xylan epitopes are degraded, whereas more highly substituted xylan and cellulose epitopes remain as residuals in the waste material that the ants remove from their fungus garden. These results demonstrate that biomass entering leaf-cutting ant fungus gardens is only partially utilized and explain why disproportionally large amounts of plant material are needed to sustain colony growth. They also explain why substantial communities of microbial and invertebrate symbionts have evolved associations with the dump material from leaf-cutting ant nests, to exploit decomposition niches that the ant garden-fungus does not utilize. Our approach thus provides detailed insight into the nutritional benefits and shortcomings associated with fungus-farming in ants.     

Comparative analysis of fungal communities in colonies of two leaf-cutting ant species with different substratum preferences

Fungus gardens of leaf-cutting ants harbor diverse alien fungi in addition to their fungal cultivar. Previous work suggested that alien microorganisms are likely derived from the substrata foraged by ant workers and incorporated into the fungus gardens. To test this hypothesis, we sampled 1014 garden fragments from 16 field colonies of Atta sexdens rubropilosa (a dicot-cutting ant) and Atta capiguara (a grass-cutting ant) in Brazil. From a total of 615 fungal isolates recovered, we observed similar diversity of fungi between colonies of both ant species. However, fungal communities differed in composition of taxa between ant colonies. Trichoderma spirale, Trichosporon chiarellii and Penicillium citrinum were prevalent accounting for 18.5%, 12.2% and 11.7% of the total isolates, respectively. As expected, fungal communities clustered in two major groups supporting the hypothesis that plant substratum has an impact on the composition of the alien fungi found in leaf-cutting ant gardens.     

Individual complexity and self-organization in foraging by leaf-cutting ants

Leaf-cutting ants cut vegetation into small fragments that they transport to the nest, where a symbiotic fungus cultivated by the ants processes the material. Since the harvested leaf fragments are incorporated into the fungus garden and not directly consumed by the workers, it is expected that foraging workers select plants by responding to those physical or chemical traits that promote maximal fungal growth, irrespective of the potential direct effects of these leaf features on them. In this paper I summarize experimental work focusing on the decision-making processes that occur at the individual level, and discuss to what extent individual complexity contributes to the emergence of collective foraging patterns. Although some basic features of self-organizing systems, such as the existence of regulatory positive and negative feedback loops, are expected to be involved in the collective organization of leaf-cutting ant foraging, I contend that they are combined with complex individual responses that may result from the integration of local information during food collection with an assessment of colony conditions.     

Hydrolytic enzymes of leaf-cutting ant fungi

The production of enzymes and the colonization of leaves by Leucoagaricus gongylophorus were investigated to further understand the digestive interactions of leaf-cutting ant colonies. The enzymes detected were indicative of a saprophytic origin of this fungus, producing all the enzymes necessary for plant tissue breakdown. Enhanced activities of certain enzymes in the fungus garden extracts may be due to the particular behaviour of the adult worker ants that concentrate fungal acquired enzymes in the rectal fluid and subsequently defaecate these enzymes onto the leaves. The production of chitinases by the fungus may be an ancestral vestige of lower attines, and may have a role as agonists of invading microbes. Growth of the fungus on plant cell wall medium resulted in highest enzyme activity against pectin, reflecting the fact that polygalacturonans comprise the main matrix of the primary plant cell wall. SEM shows that L. gongylophorus does not form specialized structures for cell wall penetration, but gains access to the inner plant tissue at the cut edges of the leaf fragments. Enzymes secreted by the fungus were compared to those seen in larval and adult leaf-cutting ants, demonstrating the inter-dependence of the symbiotic relationship between the ants and their fungi.     

Rapid shifts in <Emphasis Type="Italic">Atta cephalotes</Emphasis> fungus-garden enzyme activity after a change in fungal substrate (Attini,Formicidae)

Fungus gardens of the basidiomycete Leucocoprinus gongylophorus sustain large colonies of leaf-cutting ants by degrading the plant material collected by the ants. Recent studies have shown that enzyme activity in these gardens is primarily targeted toward starch, proteins and the pectin matrix associated with cell walls, rather than toward structural cell wall components such as cellulose and hemicelluloses. Substrate constituents are also known to be sequentially degraded in different sections of the fungus garden. To test the plasticity in the extracellular expression of fungus-garden enzymes, we measured the changes in enzyme activity after a controlled shift in fungal substrate offered to six laboratory colonies of Atta cephalotes. An ant diet consisting exclusively of grains of parboiled rice rapidly increased the activity of endo-proteinases and some of the pectinases attacking the backbone structure of pectin molecules, relative to a pure diet of bramble leaves, and this happened predominantly in the most recently established top sections of fungus gardens. However, fungus-garden amylase activity did not significantly increase despite the substantial increase in starch availability from the rice diet, relative to the leaf diet controls. Enzyme activity in the older, bottom sections of fungus gardens decreased, indicating a faster processing of the rice substrate compared to the leaf diet. These results suggest that leaf-cutting ant fungus gardens can rapidly adjust enzyme activity to provide a better match with substrate availability and that excess starch that is not protected by cell walls may be digested by the ants rather than by the fungus-garden symbiont.     

Fitness consequences of nest infiltration by the mutualist‐exploiter Megalomyrmex adamsae

1. Fungus‐growing ants are obligate mutualists. Their nutrient‐rich fungus garden provides a valuable food store that sustains the ant hosts, but can also attract social parasites. 2. The ‘guest ant' Megalomyrmex adamsae Longino parasitises the fungus‐growing Trachymyrmex zeteki Weber queen just after nest founding. The parasitic queen infiltrates the incipient nest, builds a cavity in the fungal garden, and lays eggs that develop into workers and reproductive males and females. 3. This study compared young parasitised and non‐parasitised laboratory colonies by measuring garden growth and biomass, and the number of host workers and reproductives. Host queen survival and parasite colony growth were also monitored. 4. Parasitised Trachymyrmex colonies had reduced host worker and alate numbers, as well as lower garden biomass, compared with non‐parasitised control colonies, confirming that M. adamsae is a xenobiotic social parasite. Host queen survival was not significantly different between parasitised and control colonies. 5. This is the first study that experimentally infects host colonies with a xenobiotic social parasite to measure fitness cost to the host. The natural history of M. adamsae and the fungus‐growing ant mutualism are evaluated in the context of three general predictions of (Bronstein, Ecology Letters, 4 , 277–287, 2001a) regarding the cost of mutualism exploiters.     

Instability of novel ant-fungal associations constrains horizontal exchange of fungal symbionts

One of the more fascinating features of fungus-gardening ants (Attini: Formicidae) is their fidelity to their lineage-specific fungal symbionts. Among the derived higher-attine ants (leafcutter ants and close relatives), it is thought that most leaf-cutting ants grow Attamyces fungus whereas most Trachymyrmex ants grow ‘Trachymyces’ fungus, but there exist exceptions to this clade-to-clade correspondence between ants and fungi. The exceptions are inconsistent with strict one-to-one coevolution, which suggests that ants sometimes are able to switch to novel fungi. Such switches appear to be largely constrained and ants are generally faithful to their species-specific fungi. Prior experiments demonstrated no clear fitness consequences of growing novel fungi over the short-term when the ant Trachymyrmex septentrionalis was symbiont-switched by forcing it to grow Attamyces leaf-cutter fungus. We hypothesized that long-term ant-fungal fidelity is constrained either by physiological differences among fungal species or by garden diseases that symbiont-switched ants cannot control. Repeat experiments in a different location show that T. septentrionalis colonies switched to grow Attamyces exhibit sudden declines in garden biomass and consequent fitness reductions due to garden destruction by pathogens, whereas control colonies (Trachymyrmex ants cultivating Trachymyces fungus) do not show parallel garden declines. These patterns are mirrored in symbiont-switch experiments conducted on colonies in Trachymyrmex turrifex. Disease microbes selecting on ant-cultivar combinations therefore can constrain switches to novel cultivars and maintain combinations that are more resistant to disease.     

Prevalence and impact of a virulent parasite on a tripartite mutualism

The prevalence and impact of a specialized microfungal parasite (Escovopsis) that infects the fungus gardens of leaf-cutting ants was examined in the laboratory and in the field in Panama. Escovopsis is a common parasite of leaf-cutting ant colonies and is apparently more frequent in Acromyrmex spp. gardens than in gardens of the more phylogenetically derived genus Atta spp. In addition, larger colonies of Atta spp. appear to be less frequently infected with the parasite. In this study, the parasite Escovopsis had a major impact on the success of this mutualism among ants, fungi, and bacteria. Infected colonies had a significantly lower rate of fungus garden accumulation and produced substantially fewer workers. In addition, the extent of the reduction in colony growth rate depended on the isolate, with one isolate having a significantly larger impact than two others, suggesting that Escovopsis has different levels of virulence. Escovopsis is also spatially concentrated within parts of ant fungus gardens, with the younger regions having significantly lower rates of infection as compared to the older regions. The discovery that gardens of fungus-growing ants are host to a virulent pathogen that is not related to any of the three mutualists suggests that unrelated organisms may be important but primarily overlooked components of other mutualistic associations.     

Communication between the fungus garden and workers of the leaf-cutting ant, Atta sexdens rubropilosa, regarding choice of substrate for the fungus

Bait made from orange peel, containing the fungicide cycloheximide, was initially harvested by workers of Atta sexdens rubropilosa (Forel) and incorporated into the fungus garden as substrate for the fungus. The bait was subsequently rejected by the worker ants days later. Exposure of the fungus to cycloheximide, in laboratory sub-colonies, resulted in the fungus being ‘stressed’. By interchanging normal fungus garden with ‘stressed’ fungus garden, a change in the foraging behaviour of the workers was evident. –Two hypotheses to explain the behavioural changes were tested: a volatile semiochemical is produced by the fungus which affects the foragers directly, or contact between workers (and fungus garden) is necessary for information regarding fungal substrate to be transmitted through the worker force. When pairs of sub-colonies were connected (one colony of each pair exposed to cycloheximide in the bait) and workers were initially prevented from passing from one colony to the other, one colony continued to forage on orange bait while the other did not. When both colonies were allowed to make full contact then both colonies failed to accept orange bait. This discounted the first hypothesis, but supported the second, as a highly volatile chemical should be able to diffuse between colonies. When large foragers were prevented from making contact with the second colony, the information may be communicated by smaller workers.     

Toxic effects of Canavalia ensiformis L. (Leguminosae) on laboratory colonies of Atta sexdens L. (Hym., Formicidae)

Laboratory colonies of the leaf-cutting ants Atta sexdens that were fed daily with leaves of Canavalia ensiformis showed a high ant mortality, and a significant decrease in fungal garden volume, with complete depletion of nests after 11 weeks of treatment.     

Nest Enlargement in Leaf-Cutting Ants: Relocated Brood and Fungus Trigger the Excavation of New Chambers

During colony growth, leaf-cutting ants enlarge their nests by excavating tunnels and chambers housing their fungus gardens and brood. Workers are expected to excavate new nest chambers at locations across the soil profile that offer suitable environmental conditions for brood and fungus rearing. It is an open question whether new chambers are excavated in advance, or will emerge around brood or fungus initially relocated to a suitable site in a previously-excavated tunnel. In the laboratory, we investigated the mechanisms underlying the excavation of new nest chambers in the leaf-cutting ant Acromyrmex lundi. Specifically, we asked whether workers relocate brood and fungus to suitable nest locations, and to what extent the relocated items trigger the excavation of a nest chamber and influence its shape. When brood and fungus were exposed to unfavorable environmental conditions, either low temperatures or low humidity, both were relocated, but ants clearly preferred to relocate the brood first. Workers relocated fungus to places containing brood, demonstrating that subsequent fungus relocation spatially follows the brood deposition. In addition, more ants aggregated at sites containing brood. When presented with a choice between two otherwise identical digging sites, but one containing brood, ants'' excavation activity was higher at this site, and the shape of the excavated cavity was more rounded and chamber-like. The presence of fungus also led to the excavation of rounder shapes, with higher excavation activity at the site that also contained brood. We argue that during colony growth, workers preferentially relocate brood to suitable locations along a tunnel, and that relocated brood spatially guides fungus relocation and leads to increased digging activity around them. We suggest that nest chambers are not excavated in advance, but emerge through a self-organized process resulting from the aggregation of workers and their density-dependent digging behavior around the relocated brood and fungus.     

Nest-founding in <Emphasis Type="Italic">Acromyrmex octospinosus</Emphasis> (Hymenoptera,Formicidae, Attini): demography and putative prophylactic behaviors

Summary Foundresses of the leaf-cutting ant Acromyrmex octospinosus in central Panamá forage for leaves as garden substrate (semi-claustral foundation). The fungal pellet and substrate usually are attached to rootlets, which are used as a platform for the garden. This arrangement keeps the garden suspended away from the earthen chamber of the underground nest during early colony growth, and we hypothesize that it serves to minimize contact between the garden and contaminants. A. octospinosus foundresses produce from 3 to 7 workers in 2.7 months after founding the nest, but workers do not forage for substrate at this time. Incipient nests died or were abandoned at a monthly rate of ca. 50%. We show that ants routinely clean their legs before manipulating the garden substrate. We also describe how foundresses use their fore-legs to rub the surface of the metapleural gland (MPG), and they then use typical grooming behaviors to pass the forelegs through the mouthparts, after which the ant then licks the garden substrate. Similarly, ants apparently use their mouths to transfer fecal droplets to their legs. We briefly discuss the functional significance of these grooming behaviors, and hypothesize that they are prophylactic behaviors that may help the foundress maintain a hygienic garden.     

Leucoagaricus gongylophorus uses leaf-cutting ants to vector proteolytic enzymes towards new plant substrate

The mutualism between leaf-cutting ants and their fungal symbionts revolves around processing and inoculation of fresh leaf pulp in underground fungus gardens, mediated by ant fecal fluid deposited on the newly added plant substrate. As herbivorous feeding often implies that growth is nitrogen limited, we cloned and sequenced six fungal proteases found in the fecal fluid of the leaf-cutting ant Acromyrmex echinatior and identified them as two metalloendoproteases, two serine proteases and two aspartic proteases. The metalloendoproteases and serine proteases showed significant activity in fecal fluid at pH values of 5–7, but the aspartic proteases were inactive across a pH range of 3–10. Protease activity disappeared when the ants were kept on a sugar water diet without fungus. Relative to normal mycelium, both metalloendoproteases, both serine proteases and one aspartic protease were upregulated in the gongylidia, specialized hyphal tips whose only known function is to provide food to the ants. These combined results indicate that the enzymes are derived from the ingested fungal tissues. We infer that the five proteases are likely to accelerate protein extraction from plant cells in the leaf pulp that the ants add to the fungus garden, but regulatory functions such as activation of proenzymes are also possible, particularly for the aspartic proteases that were present but without showing activity. The proteases had high sequence similarities to proteolytic enzymes of phytopathogenic fungi, consistent with previous indications of convergent evolution of decomposition enzymes in attine ant fungal symbionts and phytopathogenic fungi.     

Effects of leaves of Ipomoea batatas (Convolvulaceae) on nest development and on respiratory metabolism of leaf-cutting ants Atta sexdens L. (Hym., Formicidae)

Abstract: Laboratory colonies of the leaf-cutting ants Atta sexdens feed daily with leaves of Ipomoea batatas s howed ant mortality and a significant decrease in the size of the fungal garden after the second week, with complete depletion of nests after 5 weeks of treatment. The mean oxygen consumption rate of these ants was higher than the control (ants collected from nests feed with leaves of Eucalyptus alba ), suggesting a physiological action of the leaves of I. batatas on the ants in addition to the effect of inhibiting the growth of the fungal garden.     

Leaf-cutting ant fungi produce cell wall degrading pectinase complexes reminiscent of phytopathogenic fungi

Background

Leaf-cutting (attine) ants use their own fecal material to manure fungus gardens, which consist of leaf material overgrown by hyphal threads of the basidiomycete fungus Leucocoprinus gongylophorus that lives in symbiosis with the ants. Previous studies have suggested that the fecal droplets contain proteins that are produced by the fungal symbiont to pass unharmed through the digestive system of the ants, so they can enhance new fungus garden growth.

Results

We tested this hypothesis by using proteomics methods to determine the gene sequences of fecal proteins in Acromyrmex echinatior leaf-cutting ants. Seven (21%) of the 33 identified proteins were pectinolytic enzymes that originated from the fungal symbiont and which were still active in the fecal droplets produced by the ants. We show that these enzymes are found in the fecal material only when the ants had access to fungus garden food, and we used quantitative polymerase chain reaction analysis to show that the expression of six of these enzyme genes was substantially upregulated in the fungal gongylidia. These unique structures serve as food for the ants and are produced only by the evolutionarily advanced garden symbionts of higher attine ants, but not by the fungi reared by the basal lineages of this ant clade.

Conclusions

Pectinolytic enzymes produced in the gongylidia of the fungal symbiont are ingested but not digested by Acromyrmex leaf-cutting ants so that they end up in the fecal fluid and become mixed with new garden substrate. Substantial quantities of pectinolytic enzymes are typically found in pathogenic fungi that attack live plant tissue, where they are known to breach the cell walls to allow the fungal mycelium access to the cell contents. As the leaf-cutting ant symbionts are derived from fungal clades that decompose dead plant material, our results suggest that their pectinolytic enzymes represent secondarily evolved adaptations that are convergent to those normally found in phytopathogens.