Symbiotic Nitrogen Fixation in the Fungus Gardens of Leaf-Cutter Ants

See allHide authors and affiliations

Science  20 Nov 2009:
Vol. 326, Issue 5956, pp. 1120-1123
DOI: 10.1126/science.1173036


Bacteria-mediated acquisition of atmospheric N2 serves as a critical source of nitrogen in terrestrial ecosystems. Here we reveal that symbiotic nitrogen fixation facilitates the cultivation of specialized fungal crops by leaf-cutter ants. By using acetylene reduction and stable isotope experiments, we demonstrated that N2 fixation occurred in the fungus gardens of eight leaf-cutter ant species and, further, that this fixed nitrogen was incorporated into ant biomass. Symbiotic N2-fixing bacteria were consistently isolated from the fungus gardens of 80 leaf-cutter ant colonies collected in Argentina, Costa Rica, and Panama. The discovery of N2 fixation within the leaf-cutter ant−microbe symbiosis reveals a previously unrecognized nitrogen source in neotropical ecosystems.

Ants play a critical role in shaping terrestrial ecosystems. They make up at least one-third of the global insect fauna biomass and 86% of the arthropod biomass in tropical forest canopies, and, in the Amazon forest, they represent four times more biomass than do all land vertebrates combined (13). Among ants, the leaf cutters (tribe Attini: genera Atta and Acromyrmex) play an important role as one of the most dominant herbivores in New World tropical ecosystems, stimulating new plant growth and facilitating nutrient cycling (4). Mature Atta colonies are among the largest of any social insect, consisting of up to 8 million workers and occupying an underground volume of more than 20 m3 (Fig. 1, A and B) (5). These “superorganisms” harvest more than 240 kg dry weight of leaf material per year (4), which they use to cultivate a fungus for food (6). This ability to grow a specialized fungal crop using freshly cut plant material is a key factor in the ecological success of leaf-cutter ants (7). In addition to their relationship with fungal mutualists (family Lepiotaceae), the ants engage in a second mutualism with Actinobacteria (genus Pseudonocardia), which produce antibiotics to help defend the fungus garden from parasites (8, 9). We explored the possibility that leaf-cutter ants engage in mutualistic associations with N2-fixing symbionts to supplement the nitrogen budget of their fungus gardens.

Fig. 1

Evidence for N2 fixation in the fungus gardens of leaf-cutter ants. (A and B) The agricultural system of leaf-cutter ants is extremely efficient, allowing colonies to grow from a single fungus garden chamber [(A) incipient At. cephalotes colony with queen (black arrow) on top of fungus garden; scale bar = 1 cm] to a massive underground operation with hundreds of chambers, intricate tunnel systems, and millions of workers [(B) partially excavated nest of a mature Atta colony]. (C) Nitrogen content of the different components of five At. cephalotes colonies. (D) N2-fixation activity measured by acetylene reduction for different components of 10 Atta spp. colonies. All results are shown as means ± SEM. Means labeled with different letters (a to e) are statistically different (P < 0.05). [Photo credits: (A) Graham D. Anderson, (B) M. Moffett/Minden Pictures]

Nitrogen is expected to be a growth-limiting resource in leaf-cutter ant agriculture: The primary nutrient input into their colonies is fresh leaves, which have a much lower nitrogen-to-carbon (N:C) ratio than is required by insects (10, 11). In contrast to this expectation, several field studies have shown that the exhausted leaf substrate removed from the bottom of fungus gardens by ant workers contains higher proportions of N than either freshly harvested leaf material or surrounding leaf litter does, indicating that N enrichment occurs as the plant substrate is processed by the colony (4, 12, 13). Although these findings suggest the presence of N2-fixing symbionts, potential additional sources are mineralized N from the soil and compensatory feeding by the ants (13, 14). We analyzed the N content of laboratory-maintained colonies of Atta cephalotes in which we prevented N input from these alternate sources (15). We found an increase in N content as leaf substrate passes through the system: N content was lowest in fresh leaf cuttings, significantly higher in the fungus garden, and even higher in the ants’ refuse dump, where the exhausted leaf substrate is placed by workers (Fig. 1C) (F(5,24) = 458.34, P < 0.0001, n = 5 colonies). Our results confirm that N is enriched as substrate passes through leaf-cutter ant colonies, and because other sources of N input were prevented, it suggests that, like other insects with low-N diets (11, 16, 17), leaf-cutter ants acquire additional N from symbiotic N2 fixers [see also supporting online material (SOM) text].

We conducted acetylene reduction (AR) assays to determine whether N enrichment within leaf-cutter ant colonies is occurring, at least in part, through N2 fixation. AR is a functional test that demonstrates the presence of an active nitrogenase enzyme complex and has been widely used to provide evidence for biological N2 fixation (1618). We detected positive AR activity in the fungus gardens of all leaf-cutter ant colonies evaluated [mean = 1.03 ± 0.06 SEM nmol of ethylene per hour per g dry weight (dw)], including five species of Acromyrmex (n = 14 colonies) and three Atta species (n = 21 colonies; table S1). Overall, mean AR activity in the fungus garden was higher in Atta colonies (1.16 ± 0.07 SEM nmol of ethylene per hour per g dw) than in Acromyrmex colonies (0.81 ± 0.06 SEM nmol of ethylene per hour per g dw). Long-term monitoring of three Atta spp. colonies showed that N2 fixation was both consistent and continuous over a 2-year period (SOM text and fig. S1). N2-fixation rates were significantly higher in the middle of fungus gardens, where most feeding on the cultivated fungus occurs (fig. S2). In contrast to the fungus garden, only marginal AR activity (<0.1 nmol of ethylene per hour per g dw) was detected in ant workers, ant brood, and leaf material used by the ants to grow their fungus (Fig. 1D) (F(3,36) = 210.40, P < 0.0001, n = 10 colonies; see also figs. S3 and S4). These results showed that N2 fixation occurs in leaf-cutter ant fungus gardens and indicated the presence of N2-fixing bacteria within the garden matrix.

To locate N2 fixers, we used a culture-independent approach to detect nifH, a gene that encodes a subunit of the evolutionarily conserved nitrogenase enzyme complex, which is used by most N2-fixing bacteria (19). nifH was detected in 50% of the fungus garden samples tested from seven different leaf-cutter ant species in the genera Atta and Acromyrmex (table S2). The presence of nifH was found in only 25% of the ant workers, and because the ants showed insignificant N2-fixation activity as measured by AR, it is likely that these amplicons correspond to undigested fungus garden in their digestive systems. nifH was not detected in pure cultures of either Pseudonocardia or the fungal cultivar, the latter of which was tested because of the possible presence of endosymbiotic bacteria, because N2 fixation is not known in any fungi. These findings, coupled with the results from the AR assays, reveal that N2 fixation occurs in vivo within the fungus gardens of leaf-cutter ant colonies.

To determine whether fixed N2 directly benefits leaf-cutter ants, we conducted enrichment experiments using the natural stable isotope of nitrogen 15N2. Paired leaf-cutter ant subcolonies were placed in airtight containers (Fig. 2A), whose internal atmosphere was exchanged with either a mixture of 80% 15N2 and 20% O2 (treatment) or regular atmospheric air. Under these conditions, eight At. cephalotes subcolonies were assayed for 2 weeks (Fig. 2, B and C, and tables S3 and S4). After 1 week, significant 15N enrichment was found in the fungus garden (paired t test; t = 7.45, P = 0.0001), worker ants (t = 5.28, P = 0.001), and ant brood (t = 2.47, P = 0.04) (Fig. 2B). At the end of 2 weeks, significant enrichment was detected in the fungus garden (t = 4.87, P = 0.002) and in ant workers (t = 3.08, P = 0.02), but not in ant brood (t = 1.28, P = 0.247), because most immature individuals had developed into pupae, which do not feed (Fig. 2C). Using the 15N natural abundance measurements for fungus gardens and leaves used to feed the leaf-cutter ant colonies in the laboratory, we estimate that At. cephalotes fungus gardens can obtain between 45 and 61% of their N supply through their symbiotic association with N2-fixing bacteria (SOM text and figs. S5 and S6). These findings show that leaf-cutter ant workers obtain nitrogen from N2-fixing bacteria, which appear to serve as a substantial source of N for the ants’ fungal crop.

Fig. 2

15N2-enrichment experiments demonstrate that leaf-cutter ants obtain N from N2-fixing symbionts. (A) The airtight apparatus used in the enrichment experiments consists of a fungus garden chamber (GC) and a feeding chamber (FC), joined by a foraging tube. It included injection ports for gas exchange (white arrow), a CO2 trap (white arrowhead), and a moisture device (black arrow). (B and C) Results for 15N2-enrichment experiments with eight At. cephalotes subcolonies, 1 week [(B) n = 8 subcolonies] and 2 weeks [(C) n = 7] after the initial treatment. The mean δ15N value [N isotopic ratio, per mil (‰)] is provided for the different components of leaf-cutter ant subcolonies maintained in a control (white bars) and a 15N2-enriched (gray bars) atmosphere. The asterisks indicate significant 15N enrichment (two-tailed paired t test, P < 0.05). Error bars represent SEM.

To further support our findings of N2 fixation in the fungus gardens of leaf-cutter ants and to begin identifying the microbes responsible, we conducted extensive enrichment isolations using N-free media. Bacteria in the genus Klebsiella were isolated from the fungus gardens of 69% of leaf-cutter colonies from four different ant species collected across different ecosystems in Costa Rica and Panama (n = 58 colonies). These isolates fix N2 at rates comparable to rates of known free-living diazotrophs, including Azospirillum sp. (Fig. 3A). We also re-isolated Klebsiella from leaf-cutter ant fungus gardens several months after field collection, which suggests that these N2 fixers proliferate within the fungus garden. Phylogenetic analyses revealed that leaf-cutter–associated Klebsiella form a specific clade that includes the type strain for K. variicola (Fig. 3B). Bacteria in the genus Pantoea were also consistently isolated, but these isolates fix N2 at a significantly lower rate than the Klebsiella isolates do and were obtained less frequently in the field (in 30% of leaf-cutter ant colonies collected in Costa Rica). Leaf-cutter–associated Pantoea also form a well-supported clade in our multilocus phylogeny (Fig. 3B). We sequenced the full nif cluster from leaf-cutter–associated Klebsiella and Pantoea, and we found that both contain intact clusters that use the canonical molybdenum-iron nitrogen-fixing pathway (20), and both are identical with respect to gene content and synteny (SOM text and table S6). The presence of these clusters in the fungus garden was further confirmed through real-time polymerase chain reaction (PCR); we detected leaf-cutter–associated Klebsiella and Pantoea nifH within the fungus gardens from all eight Atta colonies tested (SOM text and fig. S9). Given the high N2-fixation rates measured for these Klebsiella isolates, the prevalence and apparent persistence in leaf-cutter fungus gardens, and the fact that they form a specialized monophyletic clade in our phylogenetic analyses, it is likely that Klebsiella is an important N2-fixing symbiont of leaf-cutter ants (SOM text).

Fig. 3

N2-fixing symbionts associated with leaf-cutter ant colonies. (A) N2-fixing activity measured by acetylene reduction for bacterial isolates obtained from multiple leaf-cutter ant colonies. Azospirillum sp., a known free-living diazotroph, was used as a positive control. Results are shown as means ± SEM [Pantoea, n = 23; Klebsiella, n = 44; Azospirillum, n = 4; isolates from other genera were obtained in N-free media (n = 27)]. (B) Bayesian phylogeny of leaf-cutter ant–associated N2-fixing symbionts. The phylogenetic tree was constructed using concatenated gene fragments from three housekeeping genes (gapA, icdA, and mtlD) of each isolate. Isolates of leaf-cutter ant–associated N2 fixers are labeled with the host ant species and color-coded to indicate whether they belong in the genus Klebsiella (blue) or Pantoea (green). The numbers above the branches represent their Bayesian-calculated posterior probabilities.

The ecological success of leaf-cutter ants is derived, in large part, from the combined ability of the ants to break down plant antifungal barriers and of the fungus garden to neutralize plant anti-insect toxins (21). Consequently, the leaf-cutters are able to use a high diversity of plant families (22), in contrast to most herbivorous insects, which specialize in a few plant species (14). However, the reliance on leaf material means that leaf-cutter ant colonies are potentially N-limited. Our work shows that these ants can potentially overcome such limitation through symbiotic associations with N2-fixing bacteria. Indeed, our estimates, using the natural abundance of the 15N isotope, suggest that leaf-cutter colonies obtain a substantial proportion of their N requirements from symbiotic N2 fixers, roughly equivalent to estimates calculated for termite species that are known to heavily rely on N2-fixing symbionts for N acquisition (22). Klebsiella spp., the main leaf-cutter ant–associated N2 fixer we identified in this study, is also known to engage in symbiotic associations with termites (23) and fruit flies (17). In these other insect/N2-fixing bacteria mutualisms, symbiotic N2 fixation occurs in the host’s gut, whereas in leaf-cutter ants, N2 fixation takes place in the fungus garden. This is consistent with the role of the fungus garden as the external digestive system of the leaf-cutter ant superorganism.

Herbivory by leaf-cutter ants has a major impact on the structure of neotropical plant communities (5), and through their collection of vast quantities of leaf biomass, these ants also play a critical role in nutrient cycling in the ecosystems in which they live (24). Nutrient enrichment by leaf-cutter ant colonies results in significant enhancement of fine root production and higher overall plant diversity near leaf-cutter refuse dumps (4, 25, 26). Nitrogen, an element that typically limits primary production across terrestrial ecosystems (27), is 26-fold more abundant in At. colombica refuse dumps than in the surrounding leaf litter (4). It is assumed that this N enrichment is achieved through the concentration of nutrients by the ants’ extensive foraging activity, but here we show that additional N is acquired through symbiotic N2-fixing bacteria that are present in the fungus gardens of leaf-cutter ant colonies. We estimate that a single mature leaf-cutter ant colony may contribute as much as 1.8 kg of fixed N per year into neotropical ecosystems (see SOM text for details). Associations with symbiotic N2 fixers may be a widespread phenomenon in tropical ants: Davidson and collaborators (28) reported that many arboreal ant species obtain little N through predation and scavenging, indicating that these insects must obtain additional N through other means, such as microbial symbionts. Given the dominance of ants in general, and leaf-cutter species in particular, these discoveries identify a potential major source of fixed N in neotropical ecosystems.

Supporting Online Material

Materials and Methods

SOM Text

Figs. S1 to S10

Tables S1 to S6

References and Notes

References and Notes

  1. Materials and methods are available as supporting material on Science Online.
  2. We thank R. Steffensen, L. Schwab, L. Uribe, M. Mora, B. Matarrita, D. Brenes, R. Araya, H. Read, J. Mentzer, D. Maly, and G. Pine for technical assistance; Y. Zhang, E. Pohlmann, and G. Roberts for assistance with acetylene reduction assays; A. Little, S. Price, and U. Mueller for leaf-cutter ant colony collection; M. Rogel-Hernández and E. Martínez-Romero for providing isolate K. variicola F2R9; B. Ma, A. Charkowski, and N. Perna for assistance with phylogenetic analyses; E. Sánchez, R. Moreira, and T. Escalante for assistance with microscopic analyses; N. Keuler for statistical advice; the sequencing and production teams at the Joint Genome Institute; and S. Adams, F. Aylward, E. Caldera, N. Gerardo, H. Goodrich-Blair, K. Grubbs, S. Marsh, M. Poulsen, K. Raffa, G. Roberts, E. Ruby, T. Schultz, and J. Scott for comments on the manuscript. We acknowledge the Organization for Tropical Studies (OTS) and the Ministerio de Ambiente y Energía in Costa Rica, the Autoridad Nacional del Ambiente in Panama, and the Government of Argentina for facilitating the research and granting collecting permits. This work was funded by NSF grants MCB-0731822, MCB-0702025, and DEB-0747002 to C.R.C.; NIH grant GM 18938 to W.W.C., and an OTS research fellowship to A.A.P.-T. G.S. and C.R.C. were supported by the U.S. Department of Energy’s Great Lakes Bioenergy Research Center under contract DE-FC02-07ER64494; D.M.S. and P.J.W. were supported by U.S. Department of Agriculture–Agricultural Research Service Current Research Information System project 3655-41000-005-00D. DNA sequence data were deposited in GenBank under accession numbers FJ593730 to FJ593840 and GQ342603 to GQ342604.
View Abstract

Stay Connected to Science

Navigate This Article