Leaf-cutting ants and their fungus
The leaf-cutting ants
Ants culturing fungi as their primary food source belong to the tribe Attini (Hymenoptera, Formicidae), which is a monophyletic group of more than 210 species of fungus growers, distributed in 12 genera (Chapela et al., 1994; Wetterer et al., 1998; Mueller et al., 2001). The group of phylogenetically so-called "lower" attines consists of 10 attine genera that are characterized by their use of detritus of plants or insect faeces as substrate for fungal cultivation (Weber, 1966; 1972; Hölldobler and Wilson, 1990). The remaining two "higher" attine genera, Atta and Acromyrmex, use fresh plant material as the substrate for fungus cultivation, hence their common name of "leaf-cutting ants". Leaf-cutting ants harvest plant material in the area surrounding their nest and bring it to the underground chambers where the mutualistic fungus is cultivated. Here, the plant material is licked and chewed into a pulp and distributed as a substrate for the fungus (e.g. Quinlan and Cherrett, 1977).The exploitation of fresh plant material has contributed significantly to the success of leaf-cutting ants, and has made them one of the most dominant herbivores in Central and South America, where they are highly abundant and have become a severe agricultural pest (Weber, 1972; Hölldobler and Wilson, 1990).
A leaf-cutting ant colony is established by one (or several) queen(s). Leaf-cutting ant gynes (winged prospective queens) have been shown to mate with several males (Fjerdingstad and Boomsma, 1998; Boomsma et al., 1999; Bekkevold et al., 1999). However, the exact benefits of exhibiting multiple mating are yet to be fully clarified. The leaf-cutting ant queen is long lived (at least 5-10 years for Atta; Weber, 1972) and her primary role is to produce workers and colony reproductives. Mature colonies of the genus Acromyrmex typically contain in the order of 50-100.000 workers, which is few compared to Atta colonies that can contain several million workers (Weber, 1966; Hölldobler and Wilson, 1990; Wetterer, 1999). Within colony division of labour and task specialization is apparent and mirrored in worker polymorphism (Wilson, 1980a,b; Hölldobler and Wilson, 1990; Wetterer, 1999). The worker caste system in Acromyrmex was previously believed to be trimodal (a minor, a media and a major caste) (Weber, 1972). However, Wetterer (1999) has recently found that workers of intermediate sizes (media) cannot be considered to belong to a distinct caste, rather is a group of workers of sizes intermediary to minors and majors. In Atta the trait of worker polymorphism is most pronounced, in that four distinct castes is observed, with the largest one (the soldier caste) being specialized in mechanically defending the colony from vertebrate intruders. However, as Wetterer (1999) has pointed out, the exact size distribution within colonies is variable and may most probably represent an optimization of the working force in relation to given environmental factors.
Within colonies the smallest sized workers have been found to primarily be involved in fungus garden maintenance and brood care (Weber, 1972; Wilson, 1980a), presumably due their size being optimal for doing this task (Bass and Cherrett, 1996). Young workers of the major caste may likewise be involved in garden tending, whereas older workers of the major caste engage in foraging and waste management outside the fungus garden (Weber, 1972; Wilson, 1980a; Schmid-Hempel, 1998). In this way, the most valuable workers (i.e. young workers with most of their lives ahead) are protected, whereas older workers of less value to the colony are subject to tasks that incur larger risks, due to increased predator and parasite exposure (e.g. Schmid-Hempel, 1998). When the colony reaches maturity, sexuals (winged males and females) are produced. They leave the colony, mate and can subsequently initiate new leaf-cutting ant colonies (e.g. Weber, 1972; Rissing et al., 1989; Hölldobler and Wilson, 1990)
The fungi cultivated by leaf-cutting ants
The fungi cultivated by attine ants can be phylogenetically divided into three groups called G1, G2 and G3 by Chapela et al. (1994). The groups G1 and G3 belong to the Leucocoprineae (Lepiotaceae; Agaricales), whereas G2 is a group consisting of more distantly related fungi (Chapela et al., 1994). The "lower" attines cultivate fungal clones that belong to the G2 or G3 groups, whereas the symbionts associated with the leaf-cutting ants are basidiomycetous fungi (Leucocoprineae; Basidiomycotina) of the G1 group (Chapela et al., 1994). The assocation with the ants is likely to go as far as 65 million years back (Mueller et al., 2001) and has developed into an association where both partners are mutually interdependent (e.g. Weber, 1972), a feature that is expected to strongly align interests and reduce the risk of conflicts between partners (Herre et al., 1999).
The ants provide the fungus with a suitable environment, in that it is cultured in nearly axenic (pure) gardens (Craven et al., 1970; Currie et al. 1999a; Currie et al., 1999b). The fungus is continuously manured with fresh plant material that is prepared by the ants (Quinlan and Cherrett, 1977) and is protected against adverse abiotic conditions, parasites, competitors and predators (Bass and Cherrett, 1994; North et al., 1997; Currie et al., 1999b; Currie and Stuart, 2001). In return, the fungus serves as the exclusive food source for the ant brood (Quinlan and Cherrett, 1979; Hölldobler and Wilson, 1990), which are fed with fungal hyphae and gongylidia (specialized fungal cells rich in amino acids and carbohydrates: Quinlan and Cherrett, 1979). Adult ants also feed on the fungus, they do however rely on plant sap from the chewed leaf fragments to obtain most of their energetic requirements (Littledyke and Cherrett, 1976; Quinlan and Cherrett, 1979).
The mutualistic fungus is vertically transmitted by the colony founding gyne (the prospective queen), who carries a bit of her natal fungus garden in a cavity in the mouth, the infrabuccal pocket, during her mating flight (Hölldobler and Wilson, 1990 after Von Ihering, 1898 and Autuori, 1956). After mating, the queen selects and excavates a suitable nest site in which she expels the contents of her infrabuccal pocket and uses the fungus from the colony she left as an inoculum to start a new fungus garden (Rissing et al., 1989; Hölldobler and Wilson, 1990; Mueller et al., 2001). Thus, vertical transmission between generations appears to be the primary transmission mode, leading to the expectations of ancient, clonally propagated fungal lineages that evolve in parallel with their lineages of ant hosts (Mueller et al., 2001). This strict vertical transmission produces a bottleneck and may through this constrain the ability of the fungus to evolve in response to e.g. changes in parasite selection pressure. It is, however, consistent with expectations of the importance of genotypic uniformity in the alignment of interests between the partner species in mutualistic relationships (Herre et al., 1999), and with the proposal put forth by Law and Lewis (1983) that symbionts in close mutualistic relationships should exhibit reduced sexual reproduction and thus exhibit lower genetic exchange compared to free-living taxa.
Phylogenetic studies have revealed that horizontal transmissions have occurred several times on an evolutionary time scale and that the phylogenetically lower attines occasionally re-acquire fungal cultivars from free-living stock (Mueller et al., 1998). Horizontal transmission may happen when a colony is founded by multiple gynes (pleiometrosis), which appears obligate in some species (Rissing et al., 1989; Mintzer; 1990), and facultative in others (Bekkevold et al., 1999). Alternatively, horizontal fungus exchange may occur due to several colonies being founded in close proximity (Rissing et al., 1989) and/or due to garden stealing by colonies having experienced fungus garden loss (Adams et al., 2000, a study on Cyphomyrmex). Horizontal exchange and mixing of genetically different fungal clones can lead to competion between clones (Bot et al., 2001). Mechanisms of mycelial incompatibility based on the recognition of genetic similarity between fungal clones exist which allow genetically different fungi to defend themselves so that one clone usually prevails (Hansen et al., 1993; Bot et al., submitted manuscript). Despite the usual asexual reproduction of fungi associated with leaf-cutting ants, horizontal mixing may initiate the occasional production of fruiting bodies, which is often associated with severe worker mortality and collapse of the fungus garden (e.g. Fisher et al., 1994). A possible explanation for this may be that the fungal symbiont has been selected to keep competitive (virulent) traits, which enhances its fitness through transmission and sexual reproduction under unfavourable conditions, e. g. when contamination with foreign fungi occurs (e.g. Fisher et al., 1994; Bot et al., 2001). The costs of keeping this trait have apparently been outweighed by the benefits of being able to compete effectively on rare occasions.
