Abstract
Background
Symbiotic ant-plant associations, in which ants live on plants, feed on plant-provided food, and protect host trees against threats, are ubiquitous across the tropics, with the Azteca-Cecropia associations being amongst the most widespread interactions in the Neotropics. Upon colonization of Cecropia’s hollow internodes, Azteca queens form small patches with plant parenchyma, which are then used as waste piles when the colony grows. Patches—found in many ant-plant mutualisms—are present throughout the colony life cycle and may supplement larval food. Despite their initial nitrogen (N)-poor substrate, patches in Cecropia accommodate fungi, nematodes, and bacteria. In this study, we investigated the atmospheric N2 fixation as an N source in patches of early and established ant colonies.
Results
Via 15N2 tracer assays, N2 fixation was frequently detected in all investigated patch types formed by three Azteca ant species. Quantified fixation rates were similar in early and established ant colonies and higher than in various tropical habitats. Based on amplicon sequencing, the identified microbial functional guild—the diazotrophs—harboring and transcribing the dinitrogenase reductase (nifH) gene was highly diverse and heterogeneous across Azteca colonies. The community composition differed between early and established ant colonies and partly between the ant species.
Conclusions
Our data show that N2 fixation can result in reasonable amounts of N in ant colonies, which might not only enable bacterial, fungal, and nematode growth in the patch ecosystems but according to our calculations can even support the growth of ant populations. The diverse and heterogeneous diazotrophic community implies a functional redundancy, which could provide the ant-plant-patch system with a higher resilience towards changing environmental conditions. Hence, we propose that N2 fixation represents a previously unknown potential to overcome N limitations in arboreal ant colonies.
Similar content being viewed by others
Background
The parallel rise and diversification of ants and plants have favored intimate ant-plant associations, many of which are mutually beneficial for both partners [1,2,3]. Loose associations between ants and plants, in which plants secrete small volumes of nectar from easily accessible organs outside the flowers (extrafloral nectaries), are widespread [4] and involve many different species. Specialized ant-plant interactions are characterized by modified hollow plant structures for shelter—so called domatia—observed in around 700 vascular plant species [5]. Additionally, the plant is providing food in the form of extrafloral nectar [6] and/or protein and carbohydrate-rich food bodies [7,8,9,10,11,12]. In return, ants protect their host plants against herbivores, pathogens, and encroaching vegetation [13, 14], and the ant’s wastes provide nutrients [15,16,17,18,19]. The extraordinary abundance of ants in the rainforest canopies and their N demand have been extensively discussed in the literature. Arboreal ants are able to adapt to resource imbalance due to N-poor food sources, e.g., through fueling foraging activities [20] or N-recycling and upgrading non-essential to essential amino acids via gut bacteria [21]. Additionally, biological N fixation (BNF), a process performed by certain bacteria and archaea (diazotrophs) that makes atmospheric N2 biologically available, has been hypothesized to supply ants with N and facilitate their high abundance in the rainforest canopies [22]. Although potential diazotrophs were found in the gut of arboreal ants [23,24,25] and Pseudomonas bacteria isolated from an ant-plant association were shown to grow on N-free media [26], BNF activity associated with arboreal ants could not be shown yet [27].
One of the most widespread ant-plant mutualisms in the Neotropics is the Azteca-Cecropia interaction. Arboreal Azteca species (Dolichoderinae) spend their entire life in and on Cecropia (Urticaceae) trees [28, 29] and nest in the internodes of the hollow Cecropia stem while defending their host plant effectively against intruders [30,31,32]. As the ants do not leave the tree, they are solely reliant on host plant-located food sources, such as plant-provided food bodies and hemipteran honeydew. Like in many other ant-plant interactions across the tropics, well-defined, dark-colored “patches” built by the ants can be found in inhabited domatia [33,34,35,36]. The first patch is initialized by the foundress queen right after colonization [37]. As the plant grows and the colony size increases, the worker ants enlarge the nesting space by colonizing additional internodes and form further patches throughout the whole nesting space of an established Azteca ant colony [38]. The patches contain numerous bacteria (pers. observation and Fukuda et al. [26]), nematodes [39, 40], and ascomycete fungi [37, 38], which are likely transmitted vertically from the mother colony [37] and might serve as food [35]. Several observations suggest that these patches are essential to ant colonies in the Azteca-Cecropia association: (i) the colony-founding (“foundress”) queen inoculates a small pile of scratched-off plant tissue (parenchyma) from the domatia walls with a cocktail of introduced patch organisms before laying eggs and (ii) the larval food in the early colony stage seems to be supplemented with patch material [37].
As the initial patch substrate is parenchyma, which is rich in carbon (C) but low in nitrogen (N), additional N might be needed to support the initial growth of the patch organisms. Upon entering, Azteca foundress queens seal the domatium and raise their first brood in isolation (claustral colony founding) by metabolizing their own internal nutritional reserves [41]. Besides defecation onto the patch by the queen, potential N sources are limited; yet, an alternative and potentially unlimited N source could be BNF.
The notable biomass of diverse patch organisms despite low N content of the initial parenchyma substrate leads to the hypothesis that (I) C-dominated patches of foundress queens are supplemented with N by BNF. Due to N-containing organic waste that Azteca worker ants frequently deposit (pers. observation), we hypothesize that (II) BNF is less prevalent in patches of established ant colonies than in more recently founded colonies. As these ant-maintained patches are supposedly essential for colony development and patch particles are transmitted vertically [37], we further hypothesize that (III) the diazotrophic community compositions in early and established ant colonies are tightly associated with the ant species and therefore vary little across colonies of the same ant species. Accordingly, we also expect that (IV) the nifH-transcribing community is dominated by a few specialized taxa.
Our study assesses the significance of BNF in this mutualistic ant-plant association by quantifying N2 fixation activity in patches of different Azteca species and colony ages. Additionally, we identify potential diazotrophs in this system by analyzing the marker gene and transcripts for N2 fixation, the dinitrogenase reductase (nifH) [42, 43].
Results
N2 fixation activity detected in all investigated patch types
In order to evaluate the N2-fixing capability in the ant-plant Cecropia, we collected plant parenchyma, early patches built by foundress queens of the genus Azteca (“initial patch” (IP); Fig. 1a) and patch material from established Azteca colonies formed and maintained by worker ants (“established patch” (EP); Fig. 1b, c). Parenchyma and patches of three different Azteca species were incubated for 72 h in an artificial 15N2:O2 atmosphere, and subsequently, the 15N/14N isotopic ratios were determined. We found that 15N atom percent excess (APE) in patches was significantly higher than that in the parenchyma (Pa) from the walls of uncolonized plant internodes, in which 15N enrichment was not detectable (Pa vs. IP and Pa vs. EP: both P < 0.001; IP vs. EP: P = 0.118; Fig. 1d). In IPs, 0.011 μg N mg−1 dry weight (dw) was fixed on average, with a maximum of 0.068 μg N mg−1 dw. In EPs, 0.012 μg N mg−1 dw was fixed on average with a maximum of 0.189 μg N mg−1 dw. The quantity of fixed N did not vary significantly across ant species and patch types (between patch types: P = 0.28; between ant species within patch types: IP: P = 0.07; EP: P = 0.23; Fig. 1e). The overall mean BNF rate measured in Azteca-made patches was 3.9 μg N g−1 (dw) day−1.
Visual appearance of patch types maintained by Azteca ants and quantified N2 fixation activity. a An initial patch (IP), here inside a cross-sectioned internode, is built up by a foundress queen scratching off parenchyma. Patch material of established b A. alfari and c A. constructor colonies (established patches (EP)) is consisting of plant material, dead nestmates, fungi, and nematodes. Patch samples and parenchyma from the inner walls of uninhabited plant internodes were incubated for 72 h in an artificial 15N2:O2 atmosphere. d The 15N atom percent excess (APE) [%] depicted on the y-axis shows significant N2 fixation in patches, but not in the plant parenchyma (Pa). e The amount of fixed N per mg sample (dry weight) depicted on the y-axis shows variations, but insignificant differences between Azteca species (alf, A. alfari; con, A. constructor; xan, A. xanthochroa) within patch types (IP, EP). Sample sizes are given in brackets, and significant differences are indicated by asterisks (***P < 0.001). Scale bars, 1 cm (a–c)
Azteca patches contain highly diverse diazotrophic communities
Aiming to assess the abundance of potential diazotrophs in patches of established colonies, we quantified the copy numbers of the functional marker gene for N2 fixation (nifH) and the 16S rRNA gene by quantitative PCR (qPCR) assays. The average proportion of diazotrophic to total microbial copy numbers per ng DNA of EPs was 3.2% and reached up to 14.7%. The ratios and abundance measures varied insignificantly between ant species (nifH: P = 0.786; 16S rRNA: P = 0.120; ratio: P = 0.349; Table 1).
To examine the diazotrophic communities in both ant-built patch types (IP and EP), we sequenced the nifH gene using Illumina MiSeq and processed the raw amplicon reads using the bioinformatic pipeline NifMAP [44]. Our data revealed a high diversity of diazotrophs, with a total of 1147 operational taxonomic units (OTUs) spanning eleven different phyla, based on the classification of nifH amino acid sequences. Across all patch samples, most reads were assigned to Proteobacteria (on average 81.3% of reads per sample), followed by Firmicutes (7.2%), Cyanobacteria (6.0%), Verrucomicrobia (3.0%), and Bacteroidetes (2.1%). The most abundant orders were Rhizobiales (on average 30.8% of reads per sample, Alphaproteobacteria), Enterobacterales (17.8%, Gammaproteobacteria), and Rhodocyclales (14.7%, Betaproteobacteria), though each ranged from 0% to more than 90% of relative read abundance amongst patches of individual ant colonies (Fig. 2). In accordance with this revealed heterogeneity in relative read abundances of bacterial orders, only 16 of 1147 nifH OTUs could be detected in 75% of the samples and only 6 of 74 genera were ubiquitous, namely Bradyrhizobium, Azorhizobium and Rhodopseudomonas (all Rhizobiales), Azospirillum (Rhodospirillales), Paraburkholderia (Burkholderiales), and Klebsiella (Enterobacterales). Neither, a single OTU nor genus, accounted constantly for more than 0.5% of reads per sample.
Diazotrophic community composition in Azteca ant-associated patch samples. Taxonomic assignments of OTUs are summarized on the order level based on BLASTP analysis. On the x-axis, every bar represents the diazotrophic community per ant colony, grouped by patch type (initial patch, established patch) and ant species (Azteca alfari, A. constructor, A. xanthochroa), whereas the y-axis shows the relative read abundance of taxonomic orders
Despite the variations in relative read abundances of bacterial orders within IPs and EPs (Fig. 2, Additional file 1: Table S1), we observed distinct patterns in the diazotrophic community composition. IPs were mostly dominated by one of the abovementioned abundant orders, or by Nostocales (Cyanobacteria), while in EPs, other additional orders were highly abundant in single samples, e.g., Rhodospirillales (Alphaproteobacteria) and Clostridiales (Firmicutes) (Additional file 1: Table S1). IPs were overall significantly less diverse on the OTU level than EPs (Shannon index; on average: A. alfari, IP = 1.3, EP = 2.1, P = 0.021; A. constr., IP = 1.9, EP = 2.9, P = 0.012; A. xanthochroa, IP = 1.2, Additional file 2: Fig. S1). The diazotrophic community composition of IPs and EPs displayed significant differences (P < 0.001, R2=0.083), visible in the constrained ordination (Fig. 3a). This was also reflected within ant species (IP vs. EP: A. alfari, P < 0.001, R2 = 0.102; A. constr., P < 0.001, R2 = 0.079). Of several established ant colonies, patch material was sampled from up to three different locations throughout the nesting space representing different patch ages. Young patches were taken from recently colonized new domatia that develop apically during plant growth, while older patch material was sampled from lower parts of the ant’s nesting space. The position of the collected EPs within the plant, representing different patch age, did not affect the diazotrophic community composition (P = 0.994). Consistently, the EPs of the same ant colony harbored similar diazotrophic communities; thus, the most significant predictor for diazotroph community composition was the individual ant colony that inhabited one Cecropia tree (P < 0.001, R2 = 0.749, Fig. 3b). The individual Cecropia tree played an insignificant role, as in Cecropia saplings with multiple independent colonization events, the IP community compositions did not correlate with the plant ID (P = 0.084). Within patch types, the ant species significantly correlated with the community composition of EPs (P = 0.0029; R2 = 0.049; Fig. 3c), but not of IPs (P = 0.316).
Distance-based redundancy analysis showing the variation of diazotrophic communities on OTU level a due to patch types, b within established patches due to individual ant colonies, and c within established patches due to ant species. Patch types are represented by shapes (IPs as triangles, EPs as circles). In b, multiple circles of the same color represent patch samples taken from different positions within the same plant
Ecophysiological traits of diazotrophs were inferred from nifH amino acid sequences, by assigning OTUs to one of four phylogenetic clusters based on regression trees (Fig. 4). In both patch types, most reads belonged to the phylogenetic cluster I (IP: on average 71% of reads per sample, EP: on average 72%), harboring mainly the canonical Mo-Fe nitrogenase of aerobic Proteobacteria and Cyanobacteria. A considerable percentage of reads was assigned to cluster II, which represents the alternative nitrogenase anfH using only Fe as a cofactor (IP 26%, EP 14%). Reads of cluster III consisting mainly of obligate anaerobic diazotrophs harboring the canonical Mo-Fe nitrogenase were almost exclusively present in patches of established ant colonies and more abundant in EPs of A. constructor than A. alfari (IP 1%, EP 12%; alf 5%, con 15%). The potentially inactive nitrogenases of cluster IV were hardly present in both sample types (IP 0.7%, EP 0.7%).
Relative abundance of phylogenetic nifH clusters based on classification and regression trees (CART) analysis. On the x-axis, every bar represents the diazotrophic community of an individual ant colony, grouped by patch type (initial patch, established patch) and ant species (Azteca alfari, A. constructor, A. xanthochroa), whereas the y-axis shows the relative read abundance
High diversity of diazotrophs actively transcribing nifH genes
To identify potential key diazotrophs that were actively expressing the nifH gene at the time of sampling, we sequenced reversely transcribed RNA (cDNA) of EPs. For both ant species, the actively nifH-transcribing communities were significantly less diverse than the nifH gene-harboring communities (Additional file 2: Fig. S2; mean Shannon indices; A. alfari: DNA 1.8, RNA 0.8, P = 0.036; A. constr.: DNA 2.9, RNA 2.1, P < 0.001). In Fig. 5, data of the nifH gene-harboring communities (based on DNA) are depicted together with the nifH-transcribing communities (based on RNA) derived from the same patch samples, along with their corresponding 15N APE after incubations. We recovered nifH transcripts from samples with varying N2 fixation activity, which was not associated with a particular taxonomic group. In general, reads assigned to Alpha-, Beta-, and Gammaproteobacteria were particularly abundant in actively nifH-transcribing communities of both ant species. However, no specific taxonomic group was ubiquitous and dominating these communities; instead, they were composed of diverse orders. Interestingly, in nearly all cases, the most abundant nifH-transcribing taxa at the time of sampling did not reflect the most abundant nifH gene-harboring taxa within the same samples (Fig. 5, denoted by triangles). This discrepancy was particularly apparent for A. alfari. In patches showing higher 15N enrichments than the median APE, reads assigned to the genera Azospirillum (Rhodospirillales), Frankia (Frankiales), Kosakonia (Enterobacterales), Methyloferula (Rhizobiales), and Pseudomonas (Pseudomonadales) were dominating the nifH transcripts of A. alfari EPs, whereas reads assigned to Lachnoclostridium (Clostridiales), Rhodopseudomonas, Rhodomicrobium (both Rhizobiales), Magnetospirillum (Rhodospirillales), Sphingomonas (Sphingomonadales), Dactylopiibacterium (Rhodocyclales), and an unclassified Opitutaceae (Opitutales) were most abundant amongst the nifH transcripts of A. constructor EPs (Additional file 2: Fig. S3).
High heterogeneity of the nifH-transcribing community in patches of established Azteca alfari and A. constructor colonies at the time of sampling. Taxonomic assignments of OTUs are summarized on the order level based on BLASTP analysis. On the x-axis, every bar represents the diazotrophic community (DNA, bottom) and the actively transcribing community (RNA, middle) of one EP sample, grouped by ant species. The y-axis shows the relative read abundance of taxonomic orders. Triangles denote samples in which the most abundant transcribing order is not the most abundant one on the diazotrophic community level. The top graph shows the corresponding 15N atom percent excess (APE) after samples were incubated with 15N2. The dashed line denotes the median APE of all measured EP samples
Discussion
N2 fixation is a prevalent process in the Azteca-Cecropia symbiosis
In the domatia of their Cecropia host plant, Azteca ants create and maintain patches housing an intricate ecosystem, which might be crucial for colony survival. The observed abundance of patch organisms (fungi, nematodes, and bacteria) in C-dominated initial patches indicated the presence of additional N sources. Our study now provided the first evidence for atmospheric N2 fixation in ant-made patches, which can serve as an N source in arboreal Azteca ant colonies (Fig. 1). Via 15N2 tracer assays, we detected significant amounts of fixed N in initial patches (IPs) formed by foundress queens, thereby supporting our hypothesis that these patches obtain N via biological nitrogen fixation (BNF). This additional N presumably supports the growth of patch organisms in the C-dominated IP which, in turn, could help inhabiting ants raising their first larvae as their diet is supplemented with particles from IPs [37].
Contrary to our hypothesis, BNF was frequently detected and had similar rates in patches of established colonies (EPs) formed and maintained by worker ants as in foundress queen-initiated IPs. This is surprising, as these patches were frequently enriched by N-containing debris such as dead nestmates and occasionally food bodies (pers. observation), thus potentially decreasing the demand for BNF. In EPs of both ant species, the estimated proportion of diazotrophs to total bacteria was up to 14%, with an average of 3%, which is considerably higher than in various other terrestrial environments. For instance, in grassland or agricultural soil, as well as on plant surfaces and roots, diazotrophs have been estimated to account for less than 1% of the total microbial community [45,46,47]. This indicates selection in favor of BNF in patch ecosystems, potentially driven by the increasing demand for N due to increasing bacterial, fungal, and nematode biomass in later-stage patches.
N2 fixation can support ant colony growth
The only previous evidence for ant-associated BNF activity was found in fungal gardens of ground-dwelling leaf cutter ants, which was estimated to account for around 50% of the garden’s N-demand based on natural isotope fractionation analysis [48]. While it is not possible to similarly model the N-input by BNF in Azteca colonies, BNF rates in fungal gardens of leaf cutter ants could be assessed based on measurements conducted by Pinto-Tomás and colleagues [48] (for detailed calculations, see Additional file 3: supplementary methods, Additional file 1: Table S2). The mean BNF rates we detected in the Azteca ant-built patches were around 200 times higher than those in the abovementioned fungal gardens, though the 15N2 incubation setup used for fungal garden pieces is believed to underestimate BNF activity (Pinto-Tomás, pers. communication). The mean BNF rate measured in Azteca-made patches (3.9 μg N g−1 (dw) day−1) also exceeded rates of other tropical habitats, such as canopy leaves, mosses, litter, and soil samples that range from appr. 0.001 to 2.8 μg N g−1 (dw) day−1 [49,50,51,52]. Only BNF rates of cyanolichens on volcanic soil in young forests and the highest detected rates in litter of a tropical lowland rainforest (up to 6 μg N g−1 (dw) day−1 [53, 54]) surpassed the mean BNF rate in ant-built patches but were still below the maximum rate (63.2 μg N g−1 (dw) day−1). Although the information on the life cycle and nutrition of these Azteca ant species is scarce, our approximation indicates that N provided by BNF can play a considerable role in ant population growth (for calculation details, see Additional file 3: supplementary methods). In established A. alfari and A. constructor colonies, up to 9.32 and 139.59 μg N, respectively, can be fixed daily, which correspond to approximately 66% and 900% of a worker ant’s N content. On average, it takes around 2 months from an egg to an adult ant [1, 55]. Under the assumption that all freshly fixed N ends up in Azteca larvae, the maximum N input extrapolated for a 2-month period corresponds to an “instant production” of up to 40 A. alfari and 598 A. constructor workers. Due to the small patch size of IPs, the amount of fixed N during the colony founding stage would be only sufficient for up to 0.5 workers, according to this calculation. Yet, its impact might be crucial for this most precarious colonization phase [56]. The first worker generation in claustral colony founding is typically miniature workers developed with the restricted nutrients provided solely by the queen [1]. Even putatively small amounts of additional N might lead to more workers or shorter development times and therefore could increase the chance of survival. In any case, BNF might enable the growth of patch organisms and thus facilitate the development of the diverse patch community including fungi, nematodes, and bacteria.
Initial diazotrophic community is not shaped by the ant species
By scratching off plant parenchyma and forming the initial substrate for fungal cultivation [37], the queen is assumed to inoculate the C-dominated substrate with vertically transmitted bacteria from its mother colony, though very heterogeneous diazotrophic communities of various taxonomic orders dominated the nifH libraries of individual IP samples (Fig. 2). Contrary to our expectations, the diazotroph community composition in IPs was not shaped by the ant species. This does not confirm vertical transmission of diazotrophs from mother to daughter colonies; though, this also does not necessarily rule it out. As ascomycete fungi and nematodes were not only found in patches but also in the infrabuccal pockets of foundress queens when entering the plant internode [37], bacteria are likely transmitted vertically as well. The discovered diazotrophic heterogeneity within an ant species may be explained by two scenarios: the IP communities did not originate from the same mother colony, or varying environmental conditions and stochastic effects favor certain diazotrophs which further dominate patches in the early colony stage. Interestingly, despite detected taxa, e.g., Rhizobiales and Enterobacterales which are regularly reported as plant-associated diazotrophs, the host plants did not determine the diazotrophic community. The correlation between Cecropia saplings harboring multiple initial ant colonies and the diazotrophs was non-significant but had a low p-value. While this could indicate that some diazotrophs in patches are plant associated and were potentially introduced from the plant surface when the internode was entered, it might rather mirror that some initial ant colonies in one Cecropia plant originated from the same mother colony. Various environmental conditions can be expected to vary significantly between the plant surface and the patch ecosystem inside the ant nesting space (e.g., humidity, light intensity, temperature, nutrient availability, activity, and biomass of fungi, nematodes, and ants) implying a challenging environmental shift. Studying an increased number of Cecropia plants colonized by different Azteca species and investigating bacteria inside the infrabuccal pockets of foundress queens during their nuptial flight would be needed to finally clarify the role of the host plant and the mother colonies. Rhizobiales, Enterobacterales, and Burkholderiales have previously been hypothesized to be involved in recycling, upgrading, or fixing N2 in the gastrointestinal tract of arboreal ants [21, 23, 24, 57, 58], though most genera found in ant guts were not present in studied patches. Only Klebsiella and Pantoea (both Enterobacterales) were widely distributed and have been found in ant guts [23, 57], fungal gardens of leaf cutter ants [48], and now also in patches of Azteca colonies.
Mature diazotrophic community differs from initial patches and is associated with the ant species
Interestingly, the diazotrophic community composition changed during colony development and differed significantly between IPs and EPs (Fig. 3a), even though both patch types displayed similar BNF rates. The change in community composition when transitioning from IPs to EPs could be explained by multiple factors: First, by the colonization history, as it is common for Azteca-Cecropia and various other ant-plant associations [37, 59,60,61,62] that one plant houses several colony foundress queens in spatially separated domatia. As only one young ant colony takes over the Cecropia plant, several initial patch communities are finally fused and expected to increase microbial diversity. Second, new diazotrophs could be introduced from outside when ant workers in established Azteca colonies are patrolling on the surface of the Cecropia tree [63]. Third, the difference between diazotrophic communities in IPs and EPs could be due to the patch transformation during ant colony growth. The patch texture changes from brittle IPs (Fig. 1a) to extensive lawn-like EPs in A. alfari colonies (Fig. 1b) and dense voluminous EPs in A. constructor (Fig. 1c). Different textures will create variable environmental conditions, which could lead to changes in the diazotrophic community composition.
The diazotrophic community composition in EPs from different internodes of the same ant colony did not vary significantly. Our data imply that newly built EPs of subsequently colonized apical plant internodes are not randomly colonized by diazotrophs, but that worker ants must transfer EP material from existing patches as an inoculum onto new patches. The variation in diazotroph community composition correlated significantly with ant species, thereby supporting our hypothesis that the diazotrophs in established ant colonies are associated with the ant species (Fig. 3c). Differences in the community composition between EPs of both ant species could be caused by differing environmental conditions in these EPs. For instance, we identified a strong O2 gradient towards anoxia in EPs of A. constructor, whereas in A. alfari EPs, only atmospheric O2 concentrations were measured (Additional file 2: Fig. S4, Additional file 3: supplementary methods). Reduced O2 concentrations could offer niches for potentially anaerobic diazotrophs (e.g., cluster III, Fig. 4) and are reflected in the higher relative abundance of mainly anaerobic bacterial groups such as Bacteroidales, Clostridiales, and Opitutales in A. constructor compared to A. alfari EPs (Fig. 2, established patches).
Heterotrophic nifH-transcribing community implies resilience towards variable conditions
As BNF is a highly regulated process [64], sequencing of nifH genes merely identifies the community harboring these genes, while sequencing the nifH-transcripts reflects the diazotrophic community that actively expressed the genes and thus was likely involved in fixing atmospheric N2. Interestingly, in the majority of samples, the most abundant taxonomic order in DNA-based libraries did not coincide with the most abundant order in RNA-based libraries at the time of sampling (Fig. 5, denoted by triangles). This underlines the importance in investigating functional guilds beyond the DNA level, as actively transcribing community members cannot be inferred based only on the nifH-gene harboring community composition. We hypothesized that nifH-transcribing communities are dominated by few specialized taxa actively expressing the nifH gene, and indeed, the detected nifH-transcribing communities were less diverse than the nifH gene-harboring communities (Additional file 2: Fig. S2). However, there was no ubiquitous taxonomic group dominating all samples. Instead, in patches showing higher 15N enrichments than the median APE, we observed an unspecific active diazotrophic community, whose transcripts belonged mainly to twelve different taxonomic genera of nine orders (Additional file 2: Fig. S3). Amongst others, we found in situ nifH transcription of the genus Pseudomonas in investigated patches, thus supporting a recent study on the BNF capability of Pseudomonas isolates originating from an Azteca-Cecropia association in Brazil [26]. The genera Klebsiella and Pantoea, which are supposed to contribute to BNF in fungal gardens of leaf cutter ants [48] could be detected in DNA-based libraries, but not within the nifH transcripts.
Conclusions
Our study provides the first evidence for atmospheric N2 fixation in ant-made patches located in the nesting space of the host trees—a so far overlooked potential N source for arboreal ants to overcome N limitations. Future research will show if this phenomenon is exclusive to the Azteca-Cecropia or can be found in other ant-plant associations. Although the sources of the diazotrophs serving as inoculum for freshly formed IPs need further investigation, the diverse and heterogeneous diazotrophic communities across ant colonies imply functional redundancy in ant-built patches, which could provide the ant-plant-patch system with a higher resilience towards changing environmental conditions like nutrient shortage, but also towards changes in substrate or O2 concentrations. Our data indicate that BNF can result in reasonable amounts of N especially in established ant colonies. This might not only enable bacterial, fungal, and nematode growth as well as community development in the patch ecosystems but can even support the growth of ant populations. Even though larval food seems to be supplemented with patch material during the early stage of a colony’s life cycle, more information on the nutrient fluxes inside the domatia and ants’ population dynamics are needed to confirm our indications that BNF is of nutritional relevance for Azteca population growth during early foundation and in established ant colonies.
Methods
Study site characteristics and sample collection
Azteca-colonized Cecropia trees were investigated between 2015 and 2018 in lowland areas of the Golfo Dulce region in SW Costa Rica close to the Tropical Research Facility La Gamba (www.lagamba.at; N08° 42′ 03″, W083° 12′ 06″, 70 m asl). Sampled Cecropia trees were collected from disturbed habitats at the margins of primary and secondary rain forests, alongside rivers (Rio Bonito, Rio Sardinal) and roads. Azteca-colonized Cecropia trees were cut in the field, wrapped in plastic bags, and transported to the research station. Colonies were categorized into two colonization stages, an early founding stage limited to one inhabited plant internode and an established ant colony stage with at least two internodes being inhabited. In the founding stage, solely the ant queen, possible brood, and an initial pile of scratched-off plant parenchyma (“initial patch” (IP)) were present. The entrance hole created by the founding queen to enter the hollow stem internode was always sealed with chewed plant tissue [37]. Established ant colonies, in contrast, reopened the entrance and connected additional internodes, in many of which they also formed clearly defined dark-colored patches (“established patches” (EPs)). After opening Cecropia stems, we collected patch material using sterile tweezers. Queens and worker ants (if present) were preserved in ethanol and classified using an Azteca-specific key [65]. We investigated IPs of A. alfari [Emery], A. constructor [Emery], and A. xanthochroa [Roger] and EPs of A. alfari and A. constructor, as no established A. xanthochroa ant colony inhabited one of our sampled Cecropia trees.
Quantifying N2 fixation activity
For proof of N2 fixation activity in ant-built patches, a 15N2 tracer assay was used [66]. Fifty-six Cecropia plants were investigated, and in total, 122 samples were incubated in 2-mL glass vials containing an artificial 15N2:O2 (80:20) atmosphere. Early colony stage IPs (total n = 23; per species: A. alfari (10), A. constructor (2), A. xanthochroa (11)) were transferred directly from the internodes into the glass vial. EPs per established ant colonies (total n = 99; per species: A. alfari (36), A. constructor (59)) were homogenized first on a weighing paper before transferring parts into the glass vial for incubations and storing parts for molecular work. In addition to patch material, plant parenchyma of uncolonized internodes (n = 6), typically serving as initial substrate for patches, was sampled for incubations as well. The head space in the glass vials was exchanged by first inducing vacuum using a syringe and then injecting 2 mL of the artificial atmosphere. 15N2 gas (98 at%) used in incubations was purchased from CAMPRO Scientific (Berlin, Germany, lot # MBBB0968V) and Cambridge Isotope Laboratories (Tewksbury, MA, USA, lot # I-21117A) and tested negative for 15N-labelled ammonia gas contaminations using the hypobromite oxidation method [67, 68]. After incubation for 72 h, the vials were opened and patches were dried at 60 °C. The 15N/14N isotopic ratios were determined using an elemental analyzer (EA 110; CE Instruments, Milan, Italy) coupled to an isotope ratio mass spectrometer (DELTA Plus; Finnigan MAT, Bremen, Germany) in Vienna, Austria. Non-incubated parenchyma (n = 4), IPs (n = 33), and EPs (n = 84) served as natural abundance controls. The measured 15N enrichment per sample was calculated to atom percent excess \(\left(\mathrm{APE}=\left[\mathrm{at}{\%}_{\mathrm{inc}.}-\mathrm{at}{\%}_{\mathrm{nat}.\mathrm{ab}.}\right]\times \left[\frac{100}{\mathrm{at}{\%}_{{}^{15}{\mathrm{N}}_2\mathrm{gas}}-\mathrm{at}{\%}_{\mathrm{nat}.\mathrm{ab}.}}\right]\right)\). The amount of fixed N was calculated for each sample as N per mg patch dry weight [μg N mg−1 dw] \(\left(\mathrm{N}=\left[\frac{{\mathrm{APE}}_{\mathrm{sample}}}{100}\right]\times \left[\frac{{\left(\mathrm{Amt}\%\mathrm{N}\right)}_{\mathrm{sample}}}{100}\right]\times 1000\right)\).
Investigating the diazotrophic community composition
In total, the diazotrophic communities of 33 early colony stages (A. alfari (19), A. constructor (6), A. xanthochroa (8)) and 36 established ant colonies (A. alfari (12), A. constructor (24)) were investigated (detailed overview in Additional file 1: Table S3). Some Cecropia plants were colonized by several foundress queens, each in a spatially segregated domatium. From the early ant colony stage, the single, initial patch was sampled. From 15 established ant colonies, patch material was sampled up to three times throughout the nesting space. As new domatia are formed apically during plant growth, recently created patches were from the top, while older patch material was sampled from lower parts of the plants. Collected patch samples were preserved in RNAlater (Thermo Fisher Scientific, Waltham, MA, USA) and stored at − 20 °C in the field and at − 80 °C after returning to Vienna. Details of the molecular analysis to investigate the diazotrophic community composition can be found in the Additional file 3: supplementary methods. In short, nucleic acid extractions of RNAlater-removed samples were performed according to a bead-beating phenol/chloroform-based protocol [69]. To examine the diazotrophic communities, we amplified the dinitrogenase reductase-encoding gene (nifH) [42, 43] using Ueda19F as forward and R6 as reverse primer [70, 71], as these showed the least amplification bias regarding nifH homologous genes (bchX, chlL, bchL, parA) [44]. To uncover the diazotrophic communities transcribing the nifH gene and therefore potentially actively fixing N2 at the time of sampling, purified RNA was reverse transcribed into cDNA. As extractions of IPs did not yield sufficient nucleic acids for RNA purification, only EPs were used for investigating the transcribing diazotrophic communities. NifH genes and transcripts of the template DNA and cDNA were amplified in a two-step barcoding approach for subsequent Illumina MiSeq sequencing according to Angel et al. [44]. Raw MiSeq amplicon reads were deposited in the NCBI Short Read Archive under the BioProject accession number PRJNA777006 [72] and processed using the pipeline NifMAP [44]. The clustered OTUs (3% radius) were taxonomically classified using BLASTP [73] against the RefSeq database [74]. OTUs were assigned to phylogenetic clusters based on Classification And Regression Trees (CART) [75]. Samples resulting in less than 300 nifH reads were discarded. The remaining samples harbored on average around 6000 nifH reads amplified from DNA and 1900 nifH reads amplified from cDNA.
Quantitative PCR assays
The abundances of potential diazotrophs in patches of established ant colonies were assessed by quantitative PCR assays (qPCR) targeting the nifH gene using the same primer pair as for amplicon sequencing. The total microbial abundances were investigated using primers 515F-mod and 806R-mod [76] targeting the 16S rRNA gene. Ten EPs of both Azteca species were analyzed. For quantification of both genes, serially diluted standards (ranging between 101 and 106 copies) were added per reaction plate. Details on the used standards and the performed qPCR assay according to Angel et al. [44] are listed in the Additional file 3: supplementary methods. The nifH gene copy number did not have to be corrected, as homologous non-nifH reads were not detected in amplicon libraries of the same samples.
The abundance of diazotrophs per EP sample was calculated as the proportion of diazotrophic bacteria (nifH copy number) to total bacteria (16S rRNA gene copy number). As copy numbers of both genes might be different in the genomes, this abundance is an estimate. While the 16S rRNA gene copy numbers are known to vary in genomes [77, 78], also additional nifH copies can exist in the chromosome, e.g., in Clostridium pasteurianum [79], or on a symbiotic plasmid, e.g., in Bradyrhizobium sp. DOA9 [80]. Furthermore, every known diazotroph which encodes an alternative V or Fe-only nitrogenase also possesses the canonical Mo nitrogenase [81].
Statistical analyses
All analyses were performed using R (version 3.5.2 [82];) and were plotted using the package ggplot2 (version 3.3.3 [83];). All functions originated from the package vegan (version 2.5-6 [84];), unless otherwise mentioned.
A Wilcoxon rank sum test (function wilcox.test, package stats, version 3.5.2 [82];) was used to test whether APE differs significantly across patch types and uncolonized parenchyma (pairwise, p-value corrected), and whether BNF rates correlate with patch types or ant species within EPs. Kruskal-Wallis rank sum test (function kruskal.test of the package stats) was used to test if BNF rates of IPs correlate with the ant species. The Wilcoxon rank sum test was also used to test for correlations between quantitative abundance data (gene copy numbers and their ratio) and ant species.
For all OTU-based distance matrices, Bray-Curtis dissimilarity measures were used. Permutational multivariate ANOVA (PERMANOVA) was used for variance partitioning analyses (adonis function) [85]. For research question-based data subsetting and computing Shannon alpha diversity indices (function estimate_richness), the phyloseq package was used (version 1.26.1 [86];). For testing correlations of alpha diversity differences with categories (patch type and ant species), the Wilcoxon and Kruskal-Wallis rank sum tests were used. The package microbiome was used for core microbiome analyses (core function, version 1.5.28 [87];). Read counts were converted into relative abundance for community composition analyses. For comparisons between communities of IPs and EPs, as well as EP communities between ant species, biological replicates of the same ant colony (Additional file 1: Table S3) were merged by calculating the average community composition. A PERMANOVA was used for testing correlations between categories (patch type, ant species within patch types, assumed patch age and ant colony within multiple-sampled established ant colonies, and tree specimens within multiple queen-colonized Cecropia trees) and variation in community composition. Significant factors were displayed as constrained distance-based redundancy analysis (dbRDA) via the function capscale. To evaluate per ant species if alpha diversity measures of the general diazotrophic and nifH-transcribing communities in EPs are significantly varying, the Wilcoxon rank sum tests were used. Analysis scripts are available at https://github.com/mnepel/nitrogen_fixation_arboreal_ant_nests [88].
Availability of data and materials
The raw sequencing reads were deposited in the NCBI Short Read Archive under the BioProject accession number PRJNA777006 [72].
Abbreviations
- APE:
-
Atom percent excess
- BNF:
-
Biological N fixation
- C:
-
Carbon
- dw:
-
Dry weight
- EP:
-
Established patch
- IP:
-
Initial patch
- N:
-
Nitrogen
- OTU:
-
Operational taxonomic unit
- Pa:
-
Parenchyma
- qPCR:
-
Quantitative PCR
References
Hölldobler B, Wilson EO. The ants. Cambridge: Harvard University Press; 1990.
Nelsen MP, Ree RH, Moreau CS. Ant-plant interactions evolved through increasing interdependence. Proc Natl Acad Sci U S A. 2018;115:12253–8.
Moreau CS, Bell CD, Vila R, Archibald SB, Pierce NE. Phylogeny of the ants: diversification in the age of angiosperms. Science. 2006;312:101–4.
Weber MG, Keeler KH. The phylogenetic distribution of extrafloral nectaries in plants. Ann Bot. 2013;111:1251–61.
Chomicki G, Renner SS. Phylogenetics and molecular clocks reveal the repeated evolution of ant-plants after the late Miocene in Africa and the early Miocene in Australasia and the Neotropics. New Phytol. 2015;207:411–24.
Heil M. Extrafloral nectar at the plant-insect interface: a spotlight on chemical ecology, phenotypic plasticity, and food webs. Annu Rev Entomol. 2015;60:213–32.
Rickson FR. The ultrastructure of Acacia cornigera L. Beltian body tissue. Am J Bot. 1975;62:913–22.
Heil M, Baumann B, Kruger R, Linsenmair KE. Main nutrient compounds in food bodies of Mexican Acacia ant-plants. Chemoecology. 2004;14:45–52.
Heil M, Fiala B, Kaiser W, Linsenmair KE. Chemical contents of Macaranga food bodies: adaptations to their role in ant attraction and nutrition. Funct Ecol. 1998;12:117–22.
Bischof S, Umhang M, Eicke S, Streb S, Qi W, Zeeman SC. Cecropia peltata accumulates starch or soluble glycogen by differentially regulating starch biosynthetic genes. Plant Cell. 2013;25:1400–15.
Rickson FR. Anatomical development of leaf trichilium and Müllerian bodies of Cecropia peltata L. Am J Bot. 1976;63:1266–71.
Fischer RC, Richter A, Wanek W, Mayer V. Plants feed ants: food bodies of myrmecophytic Piper and their significance for the interaction with Pheidole bicornis ants. Oecologia. 2002;133:186–92.
Heil M, McKey D. Protective ant-plant interactions as model systems in ecological and evolutionary research. Annu Rev Ecol Evol Syst. 2003;34:425–53.
Davidson DW, Longino JT, Snelling RR. Pruning of host plant neighbors by ants: an experimental approach. Ecology. 1988;69:801–8.
Treseder KK, Davidson DW, Ehleringer JR. Absorption of ant-provided carbon dioxide and nitrogen by a tropical epiphyte. Nature. 1995;375:137–9.
Fischer RC, Wanek W, Richter A, Mayer V. Do ants feed plants? A 15N labelling study of nitrogen fluxes from ants to plants in the mutualism of Pheidole and Piper. J Ecol. 2003;91:126–34.
Gegenbauer C, Mayer VE, Zotz G, Richter A. Uptake of ant-derived nitrogen in the myrmecophytic orchid Caularthron bilamellatum. Ann Bot. 2012;110:757–65.
Chomicki G, Renner SS. Farming by ants remodels nutrient uptake in epiphytes. New Phytol. 2019;223:2011–23.
Dejean A, Petitclerc F, Roux O, Orivel J, Leroy C. Does exogenic food benefit both partners in an ant-plant mutualism? The case of Cecropia obtusa and its guest Azteca plant-ants. C R Biol. 2012;335:214–9.
Davidson DW. The role of resource imbalances in the evolutionary ecology of tropical arboreal ants. Biol J Linn Soc. 1997;61:153–81.
Hu Y, Sanders JG, Łukasik P, D’Amelio CL, Millar JS, Vann DR, et al. Herbivorous turtle ants obtain essential nutrients from a conserved nitrogen-recycling gut microbiome. Nat Commun. 2018;9:964.
Davidson DW, Cook SC, Snelling RR, Chua TH. Explaining the abundance of ants in lowland tropical rainforest canopies. Science. 2003;300:969–72 isi:000182719800054.
Eilmus S, Heil M. Bacterial associates of arboreal ants and their putative functions in an obligate ant-plant mutualism. Appl Environ Microbiol. 2009;75:4324–32.
Russell JA, Moreau CS, Goldman-Huertas B, Fujiwara M, Lohman DJ, Pierce NE. Bacterial gut symbionts are tightly linked with the evolution of herbivory in ants. Proc Natl Acad Sci U S A. 2009;106:21236–41.
van Borm S, Buschinger A, Boomsma JJ, Billen J. Tetraponera ants have gut symbionts related to nitrogen-fixing root-nodule bacteria. Proc R Soc B. 2002;269:2023–7.
Fukuda TTH, Pereira CF, Melo WGP, Menegatti C, Andrade PHM, Groppo M, et al. Insights into the ecological role of Pseudomonas spp. in an ant-plant symbiosis. Front Microbiol. 2021;12:621274.
Bar-Shmuel N, Behar A, Segoli M. What do we know about biological nitrogen fixation in insects? Evidence and implications for the insect and the ecosystem. Insect Sci. 2020;27:392–403.
Berg CC, Rosselli PF. Cecropia. Bronx: The New York Botanical Garden Press; 2005.
Longino JT. Taxonomy of the Cecropia-inhabiting Azteca ants. J Nat Hist. 1991;25:1571–602.
Vasconcelos HL, Casimiri AB. Influence of Azteca alfari ants on the exploitation of Cecropia trees by a leaf-cutting ant. Biotropica. 1997;29:84–92.
Schupp EW. Azteca protection of Cecropia: ant occupation benefits juveniles trees. Oecologia. 1986;70:379–85.
Davidson DW. Cecropia and its biotic defenses. In: Berg CC, Rosselli PF, editors. Cecropia. Bronx: The New York Botanical Garden Press; 2005. p. 214–26.
Defossez E, Selosse MA, Dubois MP, Mondolot L, Faccio A, Djieto-Lordon C, et al. Ant-plants and fungi: a new threeway symbiosis. New Phytol. 2009;182:942–9.
Voglmayr H, Mayer V, Maschwitz U, Moog J, Djieto-Lordon C, Blatrix R. The diversity of ant-associated black yeasts: insights into a newly discovered world of symbiotic interactions. Fungal Biol. 2011;115:1077–91.
Blatrix R, Djiéto-Lordon C, Mondolot L, La Fisca P, Voglmayr H, McKey D. Plant-ants use symbiotic fungi as a food source: new insight into the nutritional ecology of ant-plant interactions. Proc R Soc B. 2012;279:3940–7.
Vasse M, Voglmayr H, Mayer V, Gueidan C, Nepel M, Moreno L, et al. A phylogenetic perspective on the association between ants (Hymenoptera: Formicidae) and black yeasts (Ascomycota: Chaetothyriales). Proc R Soc B. 2017;284:20162519.
Mayer VE, Nepel M, Blatrix R, Oberhauser FB, Fiedler K, Schönenberger J, et al. Transmission of fungal partners to incipient Cecropia-tree ant colonies. PLoS One. 2018;13:e0192207.
Nepel M, Voglmayr H, Blatrix R, Longino JT, Fiedler K, Schönenberger J, et al. Ant-cultivated Chaetothyriales in hollow stems of myrmecophytic Cecropia sp. trees - diversity and patterns. Fungal Ecol. 2016;23:131–40.
Esquivel A, Abolafia J, Hanson P, Pinto-Tomás AA. A new species of nematode, Sclerorhabditis neotropicalis sp.n. (Rhabditida), associated with Azteca ants in Cecropia obtusifolia. Nematropica. 2012;42:163–9.
Morera J, Mora-pineda G, Esquivel A, Hanson P, Pinto-Tomás AA. Detection, ultrastructure and phylogeny of Sclerorhabditis neotropicalis (Nematoda: Rhabditidae) nematodes associated with the Azteca ant-Cecropia tree symbiosis. Rev Biol Trop. 2018;66:368–80.
Brown MJF, Bonhoeffer S. On the evolution of claustral colony founding in ants. Evol Ecol Res. 2003;5:305–13.
Zehr J, Jenkins B, Short S, Steward G. Nitrogenase gene diversity and microbial community structure: a cross-system comparison. Environ Microbiol. 2003;5:539–54.
Gaby JC, Buckley DH. A global census of nitrogenase diversity. Environ Microbiol. 2011;13:1790–9.
Angel R, Nepel M, Panhölzl C, Schmidt H, Herbold CW, Eichorst SA, et al. Evaluation of primers targeting the diazotroph functional gene and development of NifMAP – a bioinformatics pipeline for analyzing nifH amplicon data. Front Microbiol. 2018;9:703.
Gupta VVSR, Zhang B, Penton CR, Yu J, Tiedje JM. Diazotroph diversity and nitrogen fixation in summer active perennial grasses in a mediterranean region agricultural soil. Front Mol Biosci. 2019;6:115.
Pereira e Silva MC, Schloter-Hai B, Schloter M, van Elsas JD, Salles JF. Temporal dynamics of abundance and composition of nitrogen-fixing communities across agricultural soils. PLoS One. 2013;8:e74500.
Regan K, Stempfhuber B, Schloter M, Rasche F, Prati D, Philippot L, et al. Spatial and temporal dynamics of nitrogen fixing, nitrifying and denitrifying microbes in an unfertilized grassland soil. Soil Biol Biochem. 2017;109:214–26.
Pinto-Tomás AA, Anderson MA, Suen G, Stevenson DM, Chu FST, Cleland WW, et al. Symbiotic nitrogen fixation in the fungus gardens of leaf-cutter ants. Science. 2009;326:1120–3.
Reed SC, Cleveland CC, Townsend AR. Controls over leaf litter and soil nitrogen fixation in two lowland tropical rain forests. Biotropica. 2007;39:585–92.
Reed SC, Cleveland CC, Townsend AR. Tree species control rates of free-living nitrogen fixation in a tropical rain forest. Ecology. 2008;89:2924–34.
Cusack DF, Silver W, McDowell WH. Biological nitrogen fixation in two tropical forests: ecosystem-level patterns and effects of nitrogen fertilization. Ecosystems. 2009;12:1299–315.
Barron AR, Wurzburger N, Bellenger JP, Wright SJ, Kraepiel AML, Hedin LO. Molybdenum limitation of asymbiotic nitrogen fixation in tropical forest soils. Nat Geosci. 2009;2:42–5.
Reed SC, Townsend AR, Cleveland CC, Nemergut DR. Microbial community shifts influence patterns in tropical forest nitrogen fixation. Oecologia. 2010;164:521–31.
Benner JW, Conroy S, Lunch CK, Toyoda N, Vitousek PM. Phosphorus fertilization increases the abundance and nitrogenase activity of the cyanolichen Pseudocyphellaria crocata in Hawaiian montane forests. Biotropica. 2007;39:400–5.
Tschinkel WR. The natural history of the arboreal ant, Crematogaster ashmeadi. J Insect Sci. 2002;2:12.
Nishi AH, Romero GQ. Colonization pattern of Cecropia by Azteca ants: influence of plant ontogeny, environment and host plant choice by queens. Sociobiology. 2008;52:367–76.
Stoll S, Gadau J, Gross R, Feldhaar H. Bacterial microbiota associated with ants of the genus Tetraponera. Biol J Linn Soc. 2007;90:399–412.
Anderson KE, Russell JA, Moreau CS, Kautz S, Sullam KE, Hu Y, et al. Highly similar microbial communities are shared among related and trophically similar ant species. Mol Ecol. 2012;21:2282–96.
Kuhn KM. Colony founding by the ant Myrmelachista flavocotea. Insect Soc. 2014;61:239–45.
Sanchez A. Establishing an ant-plant mutualism: foundress queen mortality and acquiring the third partner. Insect Soc. 2016;63:155–62.
Fiala B, Maschwitz U. Studies on the south east Asian ant-plant association Crematogaster borneensis/Macaranga: adaptations of the ant partner. Insect Soc. 1990;37:212–31.
Longino JT. Geographic variation and community structure in an ant-plant mutualism: Azteca and Cecropia in Costa Rica. Biotropica. 1989;21:126–32.
Gianoli E, Sendoya S, Vargas F, Mejía P, Jaffé R, Rodríguez M, et al. Patterns of Azteca ants’ defence of Cecropia trees in a tropical rainforest: support for optimal defence theory. Ecol Res. 2008;23:905–8.
Burris RH, Roberts GP. Biological nitrogen fixation. Annu Rev Nutr. 1993;13:317–35.
Longino JT. A taxonomic review of the genus Azteca (Hymenoptera: Formicidae) in Costa Rica and a global revision of the aurita group. Zootaxa. 2007;1491:1–63.
Montoya JP, Voss M, Kähler P, Capone DG. A simple, high-precision, high-sensitivity tracer assay for N2 fixation. Appl Environ Microbiol. 1996;62:986–93.
Preisler A, De Beer D, Lichtschlag A, Lavik G, Boetius A, Jørgensen BB. Biological and chemical sulfide oxidation in a Beggiatoa inhabited marine sediment. ISME J. 2007;1:341–53.
Warembourg FR. Nitrogen fixation in soil and plant systems. In: Knowles R, Blackburn TH, editors. Nitrogen Isotope Techniques. Amsterdam: Elsevier; 1993. p. 127–56.
Angel R, Claus P, Conrad R. Methanogenic archaea are globally ubiquitous in aerated soils and become active under wet anoxic conditions. ISME J. 2012;6:847–62.
Ueda T, Suga Y, Yahiro N, Matsuguchi T. Remarkable N2-fixing bacterial diversity detected in rice roots by molecular evolutionary analysis of nifH gene sequences. J Bacteriol. 1995;177:1414–7.
Marusina AI, Boulygina ES, Kuznetsov BB, Tourova TP, Kravchenko IK, Gal’chenko VF. A system of oligonucleotide primers for the amplification of nifH genes of different taxonomic groups of prokaryotes. Microbiology. 2001;70:73–8.
Sequencing microbial communities in the ant-plant Azteca-Cecropia association. NCBI BioProject PRJNA777006. https://www.ncbi.nlm.nih.gov/bioproject/PRJNA777006.
Altschul SF, Madden TL, Schäffer AA, Zhang J, Zhang Z, Miller W, et al. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 1997;25:3389–402.
Pruitt KD, Tatusova T, Maglott DR. NCBI reference sequences (RefSeq): a curated non-redundant sequence database of genomes, transcripts and proteins. Nucleic Acids Res. 2007;35:501–4.
Frank IE, Turk-Kubo KA, Zehr JP. Rapid annotation of nifH gene sequences using classification and regression trees facilitates environmental functional gene analysis. Environ Microbiol Rep. 2016;8:905–16.
Walters W, Hyde ER, Berg-lyons D, Ackermann G, Humphrey G, Parada A, et al. Improved bacterial 16S rRNA gene (V4 and V4-5) and fungal internal transcribed spacer marker gene primers for microbial community surveys. mSystems. 2015;1:e0009–15.
Větrovský T, Baldrian P. The variability of the 16S rRNA gene in bacterial genomes and its consequences for bacterial community analyses. PLoS One. 2013;8:e57923.
Sun DL, Jiang X, Wu QL, Zhou NY. Intragenomic heterogeneity of 16S rRNA genes causes overestimation of prokaryotic diversity. Appl Environ Microbiol. 2013;79:5962–9.
Wang S-Z, Chen J-S, Johnson JL. The presence of five nifH-like sequences in Clostridium pasteurianum: sequence divergence and transcription properties. Nucleic Acids Res. 1988;16:439–54.
Okazaki S, Noisangiam R, Okubo T, Kaneko T, Oshima K, Hattori M, et al. Genome analysis of a novel Bradyrhizobium sp. DOA9 carrying a symbiotic plasmid. PLoS One. 2015;10:e0117392.
McRose DL, Zhang X, Kraepiel AML, Morel FMM. Diversity and activity of alternative nitrogenases in sequenced genomes and coastal environments. Front Microbiol. 2017;8:267.
R Core Team. R. A language and environment for statistical computing; 2018. https://www.r-project.org/
Wickham H. ggplot2: elegant graphics for data analysis. New York: Springer-Verlag; 2016. https://ggplot2.tidyverse.org
Oksanen J, Blanchet FG, Friendly M, Kindt R, Legendre P, McGlinn D, et al. vegan: community ecology package; 2019. https://cran.r-project.org/package=vegan.
McArdle BH, Anderson MJ, Ecology S, Jan N. Fitting multivariate models to community data: a comment on distance-based redundancy analysis. Ecology. 2001;82:290–7.
McMurdie PJ, Holmes S. phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data. PLoS One. 2013;8:e61217.
Lahti L, Shetty S. Tools for microbiome analysis in R. 2017. https://microbiome.github.io/tutorials/.
Community analysis scripts in mnepel/nitrogen_fixation_arboreal_ant_nests. GitHub. 2022. https://github.com/mnepel/nitrogen_fixation_arboreal_ant_nests.
Acknowledgements
We thank the staff at the Estación Tropical La Gamba for great working conditions and help, wonderful food, and a lot of fun beyond work. We thank Gaute Lavik from the Max Planck Institute for Marine Microbiology, Bremen, Germany, for performing the ammonia measurements of 15N2 gas samples. We thank Costa Rican authorities for their permission to conduct this research (R-046-2015-OT-CONAGEBIO, SINAC INV-ACOSA-001-16, INV ACOSA-013-18).
Funding
This work was funded by a project grant (P-31990-B to V.E.M.) from the Austrian Science Fund (FWF). M.N. was supported by a DOC fellowship (24388) from the Austrian Academy of Sciences (ÖAW) and by the Austrian Science Fund (FWF) DK+ program ‘Microbial Nitrogen Cycling’ [W1257-B20] and a fellowship by the Society for the promotion of the La Gamba Field Station. F.B.O. was funded by a short-term grant abroad (KWA) from the University of Vienna and the DFG under Germany’s Excellence Strategy - EXC 2117 – 422037984.
Author information
Authors and Affiliations
Contributions
V.E.M. and A.R. designed the stable isotope and D.W. and M.N. the molecular part of this study. V.E.M., M.N., F.B.O., and D.W. conducted the fieldwork and collected the samples. M.N. and J.P. performed the lab work. M.N. analyzed the data and wrote the original draft of the manuscript. All authors contributed substantially to the revisions. All authors read and approved the final manuscript.
Corresponding authors
Ethics declarations
Ethics approval and consent to participate
Not applicable
Consent for publication
Not applicable
Competing interests
The authors declare that they have no competing interests.
Additional information
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Additional file 1: Table S1.
Relative read abundances of the most abundant taxonomic orders per patch type. Table S2. Calculating BNF rates in fungal gardens of leaf cutter ants. Table S3. Overview on the number of sampled Cecropia trees, early stage and established ant colonies.
Additional file 2: Figure S1.
Alpha diversities in samples due to different patch types of individual Azteca species. Figure S2. Alpha diversities in patch samples of established ant colonies on nifH gene and transcript level. Figure S3. NifH transcribing community on genus level in selected patch samples of established Azteca alfari and A. constructor colonies with measured APE higher than the overall median. Figure S4. Oxygen gradients through patches of established Azteca colonies.
Additional file 3: Supplementary methods.
Detailed information on molecular analysis and performed qPCR assays, calculating BNF rates in fungal gardens of leaf cutter ants, evaluating the potential ecological role of BNF for ant colony growth, and oxygen profiling through established patches.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
About this article
Cite this article
Nepel, M., Pfeifer, J., Oberhauser, F.B. et al. Nitrogen fixation by diverse diazotrophic communities can support population growth of arboreal ants. BMC Biol 20, 135 (2022). https://doi.org/10.1186/s12915-022-01289-0
Received:
Accepted:
Published:
DOI: https://doi.org/10.1186/s12915-022-01289-0