Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Biomineral armor in leaf-cutter ants

Biomineral armor in leaf-cutter ants ARTICLE https://doi.org/10.1038/s41467-020-19566-3 OPEN 1,2,3 4 5 1 5 6 Hongjie Li , Chang-Yu Sun , Yihang Fang , Caitlin M. Carlson , Huifang Xu , Ana Ješovnik , 6,7 8,9 10 5 8,9 Jeffrey Sosa-Calvo , Robert Zarnowski , Hans A. Bechtel , John H. Fournelle , David R. Andes , ✉ ✉ 6 4,5,11 1,2 Ted R. Schultz , Pupa U. P. A. Gilbert & Cameron R. Currie Although calcareous anatomical structures have evolved in diverse animal groups, such structures have been unknown in insects. Here, we report the discovery of high-magnesium calcite [CaMg(CO ) ] armor overlaying the exoskeletons of major workers of the leaf-cutter 3 2 ant Acromyrmex echinatior. Live-rearing and in vitro synthesis experiments indicate that the biomineral layer accumulates rapidly as ant workers mature, that the layer is continuously distributed, covering nearly the entire integument, and that the ant epicuticle catalyzes biomineral nucleation and growth. In situ nanoindentation demonstrates that the biomineral layer significantly hardens the exoskeleton. Increased survival of ant workers with biomi- neralized exoskeletons during aggressive encounters with other ants and reduced infection by entomopathogenic fungi demonstrate the protective role of the biomineral layer. The dis- covery of biogenic high-magnesium calcite in the relatively well-studied leaf-cutting ants suggests that calcareous biominerals enriched in magnesium may be more common in metazoans than previously recognized. 1 2 Department of Bacteriology, University of Wisconsin-Madison, Madison, WI 53706, USA. Department of Energy Great Lakes Bioenergy Research Center, Wisconsin Energy Institute, University of Wisconsin-Madison, Madison, WI 53726, USA. State Key Laboratory for Managing Biotic and Chemical Threats to the Quality and Safety of Agro-products, Key Laboratory of Biotechnology in Plant Protection of Ministry of Agriculture and Zhejiang Province, Institute of Plant Virology, Ningbo University, 315211 Ningbo, China. Department of Physics, University of Wisconsin-Madison, Madison, WI 53706, USA. 5 6 Department of Geoscience, University of Wisconsin-Madison, Madison, WI 53706, USA. Department of Entomology, National Museum of Natural History, Smithsonian Institution, Washington, DC 20560-0003, USA. School of Life Sciences, Arizona State University, Tempe, AZ 85287, USA. 8 9 Department of Medicine, University of Wisconsin-Madison, Madison, WI 53706, USA. Department of Medical Microbiology and Immunology, University of Wisconsin-Madison, Madison, WI 53706, USA. Advanced Light Source Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA. Departments of Chemistry, Materials Science and Engineering, University of Wisconsin-Madison, Madison, WI 53706, USA. Previously published as Gelsomina De Stasio: Pupa U. P. A. Gilbert. email: pupa@physics.wisc.edu; currie@bact.wisc.edu NATURE COMMUNICATIONS | (2020) 11:5792 | https://doi.org/10.1038/s41467-020-19566-3 | www.nature.com/naturecommunications 1 1234567890():,; ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19566-3 iomineral skeletons first appeared more than 550 million 1–5 years ago , and by the early Cambrian biomineral-based Bdefensive structures had evolved in most extant metazoan phyla, apparently in response to increasing predation pressure . The minerals involved, as well as the biogenic structures they form, are diverse. Calcium carbonate biomineralization is parti- 1,7 8 cularly widespread among metazoans : the hard parts of corals , 9 10 mollusk shells , stomatopod dactyl club , and sea urchin spines contain calcium carbonate, as do the light-focusing eye 12,13 lenses of chitons and brittlestars . Magnesium-enriched calcite (CaCO ) has been discovered in the central part of the sea urchin tooth, where the increased hardness imparted by magnesium is 14–16 thought to aid in the grinding of limestone . Given the 0.5 mm importance of calcareous anatomical structures across metazoan phyla and given that magnesium significantly strengthens such structures , it is surprising that high-magnesium calcite appears to be rare in animals . It is also surprising that despite the near ubiquity of biogenic mineralization across metazoan phyla and the widespread presence of calcium carbonate in the Crustacea, biomineralized calcium carbonate has so far remained unknown in the most diverse group of animals, the insects, which arose from within the Crustacea . Here we report the discovery of a dense layer of biogenic high-magnesium calcite in the leaf-cutter ant Acromyrmex echinatior. Fungus-growing attine ants (tribe Attini, subtribe Attina) engage in an ancient and obligate mutualism with coevolved fungi 1 µm (order Agaricales), which they cultivate for food. Fungus farming, cd which has been described as a major transition in evolution , evolved only once in ants around 60 million years ago . Leaf- Mineral layer cutting ants (genera Acromyrmex and Atta), a phylogenetically derived lineage that arose within the fungus-growing ants around Cuticle 20 million years ago, harvest fresh vegetation as the substrate on Mineral layer which they grow their fungal mutualists. They are ecologically 20,21 dominant herbivores in the New World tropics and serve important roles in carbon and nitrogen cycling . A mature Atta leaf-cutter ant colony comprises a superorganism with >5 million Epicuticlar microstructure workers, a single queen, and complex society with a highly refined division of labor based both on worker size and age. Acromyrmex Cuticle 1 µm leaf-cutter colonies vary in size, generally within the range of 100 µm 23,24 15,000–100,000 workers , and in some species may have more than one queen. Acromyrmex echinatior, the focus of the present Fig. 1 Morphological and structural characterization of minerals on the study, has a mean colony size of 137,500 workers and is facul- cuticle of Ac. echinatior.a Ac. echinatior ant with a whitish cuticular coating 25,26 tatively polygynous . In addition to the leaf-cutters, 17 other (Photo T.R.S.). b SEM image of ant cuticle with crystalline coating. genera of ants occur within the Attina, all of which grow fungus c Backscattered electron (BSE) image of a polished cuticular cross-section gardens, form colonies of hundreds to a few thousand workers, of an ant. This layer is brighter than the cuticle in backscattered electron and use dead vegetative matter or caterpillar frass rather than (BSE) mode scanning electron microscopy (SEM), indicating that it consists fresh leaves and grasses as substrates for their gardens. In addi- of heavier elements and is continuous, covering nearly the entire surface. tion to the symbiotic association with their fungal cultivars, many d BSE image close-up of a polished cuticular cross-section of an ant. fungus-growing ants engage in a second mutualism with Acti- nobacteria (genus Pseudonocardia), which produce antibiotics Based on combined data from in situ X-ray diffraction (XRD), 27–29 that help defend the garden from fungal pathogens . Fungus- electron microscopy, electron backscatter diffraction (EBSD), growing ant colonies, containing both fungal crops and immature quantitative electron probe micro-analysis (EPMA), and Raman ant brood, represent a rich nutritional resource for a wide variety and attenuated total reflectance Fourier-transform infrared of marauding ant species, including army ants and other known (ATR-FTIR) spectroscopy, we report here that this coating is in agro-predatory raiders of ant agriculture . Smaller fungus- fact a mineral layer covering the ant exoskeleton. The layer is growing ant colonies, including those of Acromyrmex, are also composed of euhedral rhombohedral crystals with curved faces, occasionally subject to attack by the large-sized soldier castes of 3–5 μm in size (Fig. 1b). To examine the mechanism of crystal Atta leaf-cutter ants, which use their powerful zinc-enriched growth, we conducted synchrotron X-ray PhotoEmission electron mandibles to defend their colonies’ territories against other, spectro-microscopy (X-PEEM), in vitro synthesis, and in vivo 31–33 encroaching ant species . observation of crystallization and growth in an ant-rearing Many species of fungus-growing ants are variably covered with experiment. We measure the cuticle hardness of Ac. echinatior a whitish granular coating, uniformly distributed on their ants with and without the cuticular layer using in situ nanoin- otherwise dark brown cuticles , including, in addition to Acro- dentation and explore two of several possible benefits associated myrmex echinatior (Fig. 1a), some species of Trachymyrmex and with the biomineral armor in experimental ant battles and Sericomyrmex. infections by entomopathogenic fungi. 2 NATURE COMMUNICATIONS | (2020) 11:5792 | https://doi.org/10.1038/s41467-020-19566-3 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19566-3 ARTICLE Results amino acid profiling of the protein layer of the Ac. echinatior Morphological, structural, and chemical characteristics of epicuticle confirms the presence of phenylalanine in the ant epicuticular minerals. Microscopic imaging of polished cuticular cuticle (Supplementary Table 4). cross-sections of the leaf-cutting ant Acromyrmex echinatior reveals a clear interface between the crystalline layer and the ant In vitro high-Mg calcite synthesis. To assess whether epicuti- cuticle (Fig. 1c). This layer is brighter than the cuticle in back- cular proteins mediate the precipitation of high-magnesium cal- scattered electron (BSE) mode scanning electron microscopy cite in Acromyrmex echinatior, we performed synthetic (SEM) (Fig. 1c, d), indicating that it consists of heavier elements biomineralization experiments in which the cuticle of Ac. echi- and that it is continuously distributed, covering nearly the entire natior was incubated in saturated carbonate solutions with a integumental surface. Energy-dispersive X-ray spectroscopy 2+ 2+ 40 [Mg ]/[Ca ] ratio of 5 at ambient conditions (Fig. 3a and (EDS) characterization of the cuticular coating further indicates Supplementary Fig. 6). In these in vitro experiments, nanocrystal that the crystalline layer contains significant amounts of mag- aggregates precipitated on the epicuticle of Ac. echinatior (Fig. 3b, nesium and calcium (Supplementary Fig. 1a–f), suggestive of an c), and were identified as anhydrous high-magnesium calcite by Mg-bearing calcite biomineral. XRD analysis confirms the high- XRD and EDS analyses (Fig. 3d and Supplementary Fig. 7). As a magnesium calcite composition of the biomineral layer in Ac. negative control, we performed the same in vitro mineralization echinatior, as indicated by the d-spacing of (104) peak at 2.939 Å experiments using the cuticle of the leaf-cutter ant Atta cepha- (Fig. 2a and Supplementary Table 1). Quantitative electron probe lotes, which belongs to the sister genus of Acromyrmex, does not micro-analysis (EPMA) reveals a magnesium concentration of have a biomineral cuticular layer, and has different cuticular 32.9 ± 2.7 mol% (Supplementary Table 2). Using bright-field structures (Supplementary Fig. 8). We found that only aragonite transmission electron microscopy (TEM), selected area electron crystals formed, mainly on the hairs of At. cephalotes and almost diffraction (SAED), TEM-EDS, and Raman and Attenuated Total never on the epicuticle (Supplementary Figures 9 and 10), indi- Reflectance Fourier-Transform InfraRed (ATR-FTIR) spectra, we cating direct precipitation from solution since aragonite is the further confirmed that the biomineral is high-magnesium calcite 41 favorable crystalline precipitate in high-Mg conditions .In with chemically heterogeneous crystals and with no observable control experiments using Ac. echinatior epicuticles either treated Ca–Mg ordering (i.e., no evidence for dolomite, the only Ca–Mg with KOH to hydrolyze surface proteins or coated with a 10 nm carbonate phase with cation ordering) (Supplementary Figs. 2 platinum layer to disable protein function, only aragonite crystals and 3). Extensive XRD analyses of Ac. echinatior, including both formed (Fig. 3d and Supplementary Fig. 7). Interestingly, in lab-reared and field-collected workers from Panama and Brazil, synthetic biomineralization experiments using cuticle from dif- confirms the consistent presence of high-magnesium calcite in ferent developmental stages (pupae to fully mature adult work- quantities of 23–35 mol% MgCO (Supplementary Tables 2 and ers), we found that only mature worker epicuticles catalyze the 3). To investigate the presence of amorphous calcium carbonate precipitation of high-magnesium calcite (Fig. 3e), consistent with within the biomineral cuticular layer, we heat-treated or irra- the presence of a more substantial protein layer in mature diated biomineral-bearing ants, and acquired XRD and TEM- workers indicated by SEM examination (Supplementary Figs. 11 SAED data before and after either treatment. We did not find and 12). These in vitro synthesis results, in particular the syn- evidence of increased crystallinity in either the heat or radiation thetic crystal morphology, are consistent with a solution con- treatments in either X-ray or electron diffraction. taining organic molecules that may have originated in the ant The mineral-cuticle interface of Acromyrmex echinatior was thorax exoskeletons, suggesting that the protein layer in the investigated using X-PEEM at the Advanced Light Source epicuticle of Ac. echinatior and the unusual morphological (Lawrence Berkeley National Laboratory, Berkeley, CA) structures on the cuticles of the ants catalyze the low-temperature (Fig. 2b) . Distinct X-ray absorption near-edge structure nucleation and growth of magnesium-rich calcite on the epicu- (XANES) spectra occur at the carbon K-edge for each of three ticles of mature workers of Ac. echinatior. regions: cuticle, epicuticle, and mineral layer (Fig. 2c; mapped as spectral components in Fig. 2d). The C spectra for the mineral In vivo crystallization and growth of high-Mg calcite.To layer show a strong carbonate peak at 290.3 eV (Fig. 2c). The explore the developmental timing of biomineral formation on the oxygen K-edge spectra extracted from the ant mineral layer epicuticles of Acromyrmex echinatior workers, we conducted indicate that the carbonate crystals are crystalline with a strong rearing experiments. Twenty pupae at the same developmental crystal orientation dependence of peak 1 at 534 eV (Fig. 2e). 36,37 stage were collected, randomly sorted into two groups of ten, and Polarization-dependent imaging contrast (PIC) mapping reared to callow adults (i.e., adults that have just emerged from across the mineral layer, in which color quantitatively displays the pupal stage), then one worker from each of the two groups the orientation of the crystal c-axes, indicates that crystals are was collected every second day and analyzed by XRD and eSEM randomly oriented (Fig. 2f, g; Fig. 2f magnifies images of Fig. 2g). (Fig. 3f, g). No biomineral layer was visible nor detected with The width of peak 2 in all O spectra (Fig. 2e) indicates a mixture 17 XRD on workers 0 to 6 days after eclosion (emergence of the of phases with high- and low-Mg concentrations . This chemical adult stage from the pupal stage). In contrast, 8 days after eclo- heterogeneity is consistent with the XRD data (Fig. 2a), with sion visible and XRD-detectable high-magnesium calcite was electron microprobe analyses (Supplementary Table 2), with present on workers. Magnesium was rapidly integrated into the backscatter diffraction (EBSD) results (Supplementary Fig. 4), calcareous biomineral in these older workers, with XRD mea- and with the magnified PIC map regions (Fig. 2f). surements of mol% MgCO reaching ~35% within 2 days after Unlike the typical chitin spectrum of insect epicuticle, the the initiation of biomineralization on individual worker ants (i.e., Acromyrmex echinatior XANES epicuticle spectrum is consistent from days 6 to 8 after eclosion; Fig. 3g). with a protein-enriched insect epicuticle . Protein hydrolysis of the cuticular layers verifies that the epicuticle is proteinaceous (Supplementary Fig. 5). Furthermore, the epicuticle spectrum Mechanical protection of epicuticular high-Mg calcite.Itis shows a very intense peak at 285.2 eV (Fig. 2c), which, based on plausible that epicuticular high-magnesium calcite enhances the its energy position and its symmetric line shape, suggests that the structural robustness of the ant exoskeleton, providing better epicuticle contains one or more phenylalanine (Phe) enriched defense for ants engaged in ‘wars’ with other ants or under attack proteins . High-performance liquid chromatography (HPLC) from predators or parasites. To test this hypothesis, we first NATURE COMMUNICATIONS | (2020) 11:5792 | https://doi.org/10.1038/s41467-020-19566-3 | www.nature.com/naturecommunications 3 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19566-3 Ac. echinatior Mineral Epicuticle Cuticle Dolomite High-Mg Calcite Calcite 10 15 20 25 30 2 µm 2 theta (°) d Cuticle Epicuticle Mineral Epicuticle C=O CO 2.5 C=C Mineral 2.0 Cuticle 1.5 1.0 285 290 295 300 Photon energy (eV) 2 µm f g Mineral crystal c-axis angle 0 0° ° –30° –30° +30° +30° 2 µm +90° +90° c’ angle 530 535 540 545 550 Photon energy (eV) –90° –60° –30° 0° 30° 60° 90° quantified the increase in hardness conferred by the protective has a greater than two-fold increase in hardness (1.55 ± 0.48 GPa, biomineral layer using in situ nanoindentation in an SEM compared to cuticle alone of 0.73 ± 0.04 GPa) (Fig. 4a and Sup- (Fig. 4a, Supplementary Fig. 13 and Supplementary Movie 1). plementary Figs. 14 and 15). Given that the biomineral layer has Since the surface of the exoskeleton is not flat, conventional an average thickness of 2.3 µm and that it overlays a cuticle with nanoindentation could not be used, whereas in situ nanoinden- an average thickness of 33.5 µm, this more than two-fold increase tation with real-time microscopic imaging allowed near- in hardness is conferred by only a 7% increase in cuticle thickness perpendicular contact of the probe tip with the surface (Supple- (Supplementary Fig. 16). These cuticle thickness values agree with mentary Fig. 13). Typical non-biomineralized ant cuticle, made those reported by Peeters et al. . Additional in situ nano- primarily of chitin, has a hardness of H ~0.73 ± 0.04 GPa (Fig. 4a mechanical testing of the cuticles of Atta cephalotes ants, which and Supplementary Fig. 14). In contrast, when high-magnesium do not have a biomineral layer, as well as of other common calcite and cuticular layers are combined, the composite structure insects, including a beetle (Xylotrechus colonus) and a honeybee 4 NATURE COMMUNICATIONS | (2020) 11:5792 | https://doi.org/10.1038/s41467-020-19566-3 | www.nature.com/naturecommunications Intensity (a.u.) Intensity (a.u.) Intensity (a.u.) (012) (104) d = 2.939 Å (006) (110) (113) d = 2.237 Å (202) (024) (108) (116) d = 1.554 Å d = 1.392 Å NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19566-3 ARTICLE Fig. 2 Chemical characterization of minerals on the cuticle of Ac. echinatior.a In situ XRD analysis identifying the cuticular crystalline layer as high-Mg calcite. b–g XANES spectroscopy and mapping with PEEM of a cuticular cross-section. b Average of PEEM images acquired across the C K-edge, showing crystalline layer tightly attached to the cuticle. Three distinct component spectra were identified in the regions labeled cuticle, epicuticle, and mineral, from the most internal part of the ant to the outer surface. c Normalized component spectra extracted from the corresponding labeled regions. Characteristic peaks are marked, including the 285.2 eV (C=C), 288.2 eV (C=O) and 290.3 eV (carbonate) peaks. d Component map where each pixel is colored according to the chemical components it contains. Black pixels are masked areas containing epoxy or gaps. Faint carbonate components within the cuticle and epicuticle were emphasized by enhancing the blue channel 5×, thus this is a semi-quantitative map. A fully quantitative RGB component map is presented in Supplementary Fig. 21. Individual maps of each component are presented in Supplementary Fig. 22, clearly showing an increasing gradientof carbonates towards the surface in the cuticle. e O K-edge spectra extracted from the mineral crystals correspondingly colored in the Polarization- dependent Imaging Contrast (PIC) maps in f and g. f Magnified PIC maps for the regions represented by boxes in the complete PIC map in g. g PIC map quantitatively displaying the orientations of the mineral crystals’ c-axes in colors. This map was acquired from the same area shown in b and d at precisely the same magnification. These are interspersed high- and low-Mg calcite, and heterogenous at the nanoscale. Biomineral crystals do not show preferred orientations but are randomly oriented. High-magnesium calcite in carbon spectra is identified by the carbonate peak at 290.3 eV, which occurs in all carbonates, amorphous, or crystalline. The O spectra in d clearly indicate crystallinity, and their line shape indicates a mixture of high-magnesium calcite and low Mg-bearing calcite. (Apis mellifera), produced similar hardness values in the range of and 6 days, respectively (Fig. 4e) (P = 0.05; two-sample t-test). 0.4–0.7 GPa (Fig. 4a and Supplementary Fig. 14) as they are all On day 6, all ants had succumbed to infection, and examination mainly made of chitin. The nano-mechanical measurements of workers without biominerals exposed to M. anisopliae revealed indicate that the biomineralized layer substantially hardens the substantial fungal growth and emergence (Fig. 4e, inset). In this exoskeleton of Ac. echinatior, consistent with the hypothesis that experiment, a reduced abundance of the antibiotic-producing the biomineral layer functions as protective armor. bacterial symbiont Pseudonocardia associated with the To further test the role of the biomineral as protective armor, biomineral-free workers could have contributed to the reduction we exposed Acromyrmex echinatior major workers with and in survival. without biomineral armor to Atta cephalotes soldiers in ant aggression experiments designed to mimic territorial ‘ant wars’ Discussion 31,43,44 that are a relatively common occurrence in nature .In Three independent lines of evidence indicate that epicuticular direct combat with the substantially larger and stronger At. biomineral crystals are ant-generated rather than adventitiously cephalotes soldier workers (average body length of 10.4 mm and a precipitated from the environment or generated by bacteria. First, head capsule width of 6.1 mm, compared to major Ac. echinatior in both C component maps (e.g., Fig. 2d) and PIC maps (e.g., body length of 6.4 mm and head capsule width of 2.9 mm) Fig. 2f) the magnesium-rich calcite crystals outside the epicuticle (Fig. 4b), ants with biomineralized cuticles lost significantly fewer are space-filling, a characteristic of biominerals formed by body parts (Fig. 4c and Supplementary Fig. 17) and had eukaryotes . Second, magnesium-rich calcite biominerals are significantly higher survival rates compared to biomineral-free spatially co-localized with epicuticle protein(s), which are likely ants (Fig. 4d, Supplementary Movies 2 and 3). Further, in direct involved in biomineral formation, consistent with the absence of aggression experiments in which biomineral-armored Ac. echi- biomineral formation in in vitro synthesis experiments in which natior workers were pitted against At. cephalotes soldiers, all of ant epicuticles were either coated with platinum or hydrolyzed. the At. cephalotes soldiers died, whereas only a few such deaths Third, the ant-rearing experiments were carried out in sterile, occurred when Atta soldiers were pitted against biomineral-free clean Petri dishes, eliminating the possibility of biominerals ants. SEM examination of biomineral-armored Ac. echinatior ants acquired from external sources. after combat with Atta cephalotes soldiers showed significantly The biota of the Ediacaran period (635–541 million years ago) less damage to their exoskeletons (Supplementary Fig. 18). included organisms of known and unknown phylogenetic affi- Notably, biomineral armor is present in mature major workers, nities that lived in oceans with a high ratio of magnesium to which forage outside of the nest, further indicating that calcium. Most were soft-bodied, but some possessed rudimentary epicuticular high-magnesium calcite is critical in a highly skeletons composed either of aragonite (a form of calcium car- competitive environment (Supplementary Figs. 19 and 20). These bonate) or, notably, of high-magnesium calcite . Around 550 results, taken together, are consistent with a role for epicuticular million years ago, coinciding with a shift in the Earth’s oceans to high-magnesium calcite as armor that defends workers from significantly lower magnesium-to-calcium ratios, metazoans with aggressive interactions with other ants, even though more ant strongly calcified internal and external skeletons appeared, species need to be further investigated. including the most familiar modern phyla. In spite of its Biomineral armor could also help protect ants from pathogens. strengthening properties, the enrichment of calcareous structures In a series of experiments, we focused on entomopathogenic with high concentrations of magnesium in Cambrian and modern fungi, which establish infection by penetrating the insect metazoans has until now remained only known from a very small exoskeleton and have significant impacts on survival. We exposed plate within the tooth of sea urchins. The ability of fungus- the propleural plates of Ac. echinatior major worker ants with and growing ants to facilitate the formation of magnesium-rich bio- without biomineralized exoskeletons to the spores of the minerals on their epicuticles is thus surprising. Further, given that entomopathogenic fungus Metarhizium anisopliae (Ascomycota, fungus-growing ants are among the most extensively studied Hypocreales). Compared to biomineral-free workers, major tropical insects, our finding raises the intriguing possibility that workers with biomineralized exoskeletons were significantly more high-magnesium calcite biomineralization may be more wide- resistant to infection. Specifically, we found that a majority of ants spread in insects than previously suspected, suggesting a pro- without biominerals died from infection within 4 days (1.0 ± 0.4 mising avenue for future research. and 0 ± 0 ants survived to 4 and 6 days, respectively), whereas an Fungus farming in ants originated ~60 million years ago in average of 2.2 ± 0.4 and 1.4 ± 0.5 (out of 3 individuals per sub- South America when a hunter-gatherer ancestor irreversibly colony over 5 sub-colonies) ants with biominerals survived to 4 committed to subsistence-scale cultivation of fungal crops for NATURE COMMUNICATIONS | (2020) 11:5792 | https://doi.org/10.1038/s41467-020-19566-3 | www.nature.com/naturecommunications 5 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19566-3 bc + 2+ - Mg2 / Ca / Cl + - / Na / HCO solution 19 °C, 7 days 1 µm 1 µm Ac. echinatior In vitro precipitated ant original cuticle carbonates d e Cuticle Original cuticle Old worker Precipitation on Young worker original cuticle A A A A A A Newly eclosed Precipitation on Pt worker Pt coated cuticle Pupa Precipitation on cuticle after hydrolysis A A New pupa A A A A A High-Mg High-Mg calcite calcite Calcite Calcite Aragonite Aragonite 6 10 15 20 25 6 10 15 20 25 2 theta (°) 2 theta (°) 1 µ µm 1 µm 1 µm 0 <3 h 2 d 4 d 6 d 8 d 10 d 12 d >30 d Post-eclosion 2 d 6 d 10 d Fig. 3 Mineral precipitation on the cuticle in both in vitro cuticle synthetic studies and ant-rearing experiments. a Scheme of in vitro mineralization experiment using Acromyrmex echinatior leaf-cutting ant cuticles as templates for biomineralization (Photo C.M.C.). b, c Pre- and post-incubation SEM 2+ 2+ − images showing the original, uncoated cuticle (b) and the cuticle covered by a layer of precipitated carbonate (c) after incubation in Mg /Ca /Cl / + − 2+ 2+ − Na /HCO solution for 7 days at 19 °C. d, XRD patterns of, from top to bottom, an uncoated ant cuticle, a cuticle after incubation in Mg /Ca /Cl / + − 2+ 2+ − + − Na /HCO solution, a platinum-coated cuticle incubated in Mg /Ca /Cl /Na /HCO solution, and a cuticle after KOH protein hydrolysis incubated 3 3 2+ 2+ − + − in Mg /Ca /Cl /Na /HCO solution. H: high-magnesium-calcite, A: aragonite, Pt: platinum. e XRD patterns of cuticles of ants representing different 2+ 2+ − + − developmental stages, ranging from (from bottom to top), a newly formed pupa to an older worker, after incubation in Mg /Ca /Cl /Na /HCO solution. f Environmental scanning electron micrographs (eSEM) of ant epicuticles taken over a 10-day time series, from immediately after eclosion from pupa to adult (left), to 10 days post-eclosion (right), showing the formation of the biomineral layer over time (Photo H.L.). g Estimated magnesium concentration of the biomineral layer during 30 days of ant development based on the XRD d value according to Graf and Goldsmith , showing the (104) rapid integration of magnesium from days 6 to 8 and the continued presence of high-magnesium content for up to 30 days (n = 2 per treatment and the corresponding standard error are shown). food . The transition to industrial-scale agriculture occurred ~20 importance of agriculture in driving the expansion of human million years ago with the origin of the ecologically dominant populations and the elaboration of human social systems . leaf-cutting ants, in which colony populations are orders of Further paralleling human agriculture, the fungal cultivars of the magnitude greater in size and in which physically distinct worker ants are highly susceptible to pathogens and the ants have castes enable the complex division of labor, paralleling the similar responded, in part, by evolving associations with antibiotic- 6 NATURE COMMUNICATIONS | (2020) 11:5792 | https://doi.org/10.1038/s41467-020-19566-3 | www.nature.com/naturecommunications (2.980 Å) (2.244 Å) Intensity (a.u.) (012) (012) (111) (021) (104) (104) H(104) (002) (006) (2.984 Å) (121) (006) (012) (102) (110) (110) (113) H(113) (211) (113) (2.253 Å) (220) (202) (202) (221) (024) (041) (108) (024) (202) (116) (108) (132) (116) (113) Ac. echinatior original cuticle Fomation Intensity (a.u.) stage Mol% MgCO (012) (012) (111) (021) (104) (104) H(104) (002) (006) (121) (2.975 Å) (006) In vitro precipitated (012) carbonates (102) (110) (110) (113) H(113) (211) (113) (2.259 Å) (220) (202) (202) (221) (024) (041) (108) (024) (202) (116) (108) (132) (116) (113) NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19566-3 ARTICLE 02 0.5 1 1.5 2.5 Hardness (GPa) 2 mm c d 100 ** ** At. cephalotes soldier At. cephalotes soldier 06 10 20 30 40 500 70 vs. mineral-free ants vs. mineral-present ants Time (min) +mineral 2 mm 2 mm - mineral 0246 135 Day producing bacteria to protect their crops . Early sedentary parallel with agriculture-driven human cultural evolution, human agricultural settlements represented rich resources that fungus-growing ants have evolved biomineralized armor that were highly susceptible to marauding bands of human raiders, serves, at least in part, both to protect them from other ants, leading to the development of multiple modes of defense, including other fungus-growing ants, in disputes over territory including specialized warrior castes, fortified cities, weapons, and and agro-predatory ants that are known to raid their colonies and protective armor . Here we show that, in another striking to consume their gardens and brood, and protect them from NATURE COMMUNICATIONS | (2020) 11:5792 | https://doi.org/10.1038/s41467-020-19566-3 | www.nature.com/naturecommunications 7 Ac. echinatior mineral+cuticle Ac. echinatior cuticle At. cephalotes soldier At. cephalotes worker Beetle (X. colonus) Honey bee (A. mellifera) Ant survival (%) Body parts lost (No.) Ant survival (%) ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19566-3 Fig. 4 Mechanical protection afforded by the epicuticular mineral layer. a Quantitative nano-mechanical properties of insect cuticles, including honeybee (Apis mellifera), beetle (Xylotrechus colonus), leaf-cutting ants [Atta cephalotes worker, Atta cephalotes soldier (purple), and Acromyrmex echinatior worker without biomineral (green, with green minus circle beside the ant image)] and Ac. echinatior ant worker with biomineral epicuticular layer (orange, with orange plus circle beside the ant image), measured by an in situ nanoindenters with a cube-corner probe (n = 12, 15, 13, 15, 12, and 13 for insect measured above, respectively; center, median; box, upper and lower quantiles; whisker, 1.5× interquartile range; points, outlier). Atta ants images, Xylotrechus beetle image, and Apis bee image provided with permission from the copyright holder, Alexander L. Wild, Jon Rapp, and Don Farrall, respectively. b–d Aggressive interaction between three Ac. echinatior workers (with/without biomineral, respectively) and Atta cephalotes soldier (Photo C.M.C.). b Ac. echinatior worker (left) aggressively interacts with Atta cephalotes soldier (right). c In aggressive encounters with Atta cephalotes soldiers, Ac. echinatior workers with biomineral armor (orange) lose substantially fewer body parts (i.e., legs, antennae, abdomen, and head) compared to Ac. echinatior worker without biomineral (green). d Survivorship of Ac. echinatior workers without (green) and with (orange) biomineral armor in aggressive encounters with Atta cephalotes soldiers (purple). Asterisks indicate significant differences via a two-sample t-test (*P < 0.05, **P < 0.001; P-value = 0.0184, 0.0001, and 0.0006 from left to right, respectively; n = 5 per treatment and the corresponding standard error are shown.). e Survivorship curves of Ac. echinatior worker with (orange) and without (green) an epicuticular biomineral layer exposed to the entomopathogenic fungus Metarhizium. The inset images show more substantial fungal growth and emergence from biomineral-free workers (Photo H.L.). c, e n = 3 per treatment, and the corresponding standard error are shown. disease organisms that might otherwise spread rapidly in their averaged spectrum was then normalized to the beamline I curve, acquired with precisely the same energy steps. densely populated colonies. The three spectra were then aligned between 280.0 and 283.7 eV. The cuticle and epicuticle spectra were shifted in energy so that the first peak was at 285.2 eV for Methods chitin and proteins, following Cody et al. , whereas the mineral spectra were shifted Ants. The fungus-farming ants used in this study were Acromyrmex echinatior and in energy so that the last peak, characteristic of carbonates, was at 290.3 eV, following 56,57 Atta cephalotes, originally collected in Costa Rica and Panama, and subsequently Madix and Stöhr . The cuticle spectrum is identical to that published by Cody maintained in the lab. These ant species co-occur in the same nesting areas in et al. 2011 obtained from scorpion cuticle and interpreted as chitin. The spectrum has Panama and other regions. a peak at 285.2 corresponding to C=C in aromatic carbon, a shoulder at ~287 eV, and a peak at 288.2 eV corresponding to C=O in chitin. The epicuticle spectrum shows the characteristics C=C of aromatic amino acids (tyrosine, tryptophan, and PhotoEmission electron microscopy (PEEM). Acromyrmex echinatior ants were phenylalanine) , a shoulder at 287.6 eV corresponding to C–H aliphatic carbon, and freeze-dried prior to PEEM sample preparation. The heads of the ants were then a sharp peak at 288.2 eV corresponding to the carboxyl group (C=O) in the peptide detached and embedded in Epofix epoxy (EMS, Hatfield, PA), ground with SiC bonds of all proteins. Compared to the spectra in tyrosine and tryptophan, the sandpapers, polished with Al O suspensions of 300 nm (MicroPolish II, Buehler, 2 3 phenylalanine spectrum has a more symmetric peak at 285.2 eV, allowing us to assign 8,49 Lake Bluff, IL) and 50 nm (Masterprep, Buehler, Lake Bluff, IL) particle sizes . this peak to phenylalanine in the epicuticle .The C=O occurs at the expected 288.2 22 g/L Na CO saturated solution was added regularly onto the pad during 2 3 eV . These normalized and averaged spectra were then adopted as component grinding and polishing to prevent carbonate dissolution, and the Al O suspen- 2 3 spectra, displayed in Fig. 2c, and used to obtain a component map in Fig. 2d. sions were also dialyzed against 22 g/L Na CO saturated solution . The samples 2 3 were re-embedded to fill as much as possible the interior of the ants and the gap between mineral and epoxy, and then the polishing procedures were repeated. Component mapping. The extracted, averaged, normalized, and aligned compo- After final polishing, the samples were rinsed with ethanol and gently wiped with nent spectra were made references by multiplying the I . Spectrum in each pixel of TexWipe Cotton (Texwipe, Kernersville, NC), air-dried, and coated with 1 nm Pt the stack was analyzed and best-fitted to a linear combination of the component on the areas to be analyzed and 40 nm Pt around it . references: cuticle, epicuticle, and mineral. The resulting component proportion For C K-edge spectra, PEEM stacks were acquired by scanning across 280–320 maps were exported as a gray level image and combined by the Merge Channel eV range with 0.1-eV step between 284 and 292 eV, and 0.5-eV step elsewhere, function in Adobe Photoshop, which became a fully quantitative RGB image resulting in 145 images per stack . For O K-edge spectra, PEEM stacks were (Supplementary Fig. 21). Individual component distribution maps were presented acquired by scanning across 525–555 eV range with 0.1-eV step between 530 and in Supplementary Fig. 22. For Fig. 2d, we enhanced the blue channel by adjusting 52,53 545 eV, and 0.5-eV step elsewhere, resulting in 181 images per stack . The the midtone value in levels five times greater than the other two channels, to images were stacked and processed with GG Macros in Igor Pro 6.37 . emphasize the presence of a mineral in the cuticle and epicuticle. For PIC mapping, a stack of 19 images was acquired by fixing the photon energy at the O K-edge π* peak (534 eV) and changing the X-ray polarization from 49,52,55 In vitro synthesis. All synthesis experiments were carried out in sealed plastic horizontal to vertical with a 5° step . Colored PIC maps were then produced 54 bottles at 19 °C for 7 d. Solutions were prepared by dissolving 50 mM MgCl ·6H O, 2 2 using Igor Pro 6.37 with GG Macros . 10 mM CaCl ·2H O, and 50 mM NaHCO with pH buffer to ~8.0 with NaOH to 2 2 3 simulate modern seawater chemistry. The solution was mixed for 20 min and then Masking the component map. The component map in Fig. 2d was masked using divided into 100 mL bottles with ant exoskeleton (Supplementary Fig. 6). Peptide an image of the same region acquired in SEM in backscattered electron (BSE) synthesis experiments with 1 mM, 5 mM, and 10 mM of phenylalanine peptide (H- mode. Unfortunately, in both the PEEM average image in Fig. 2b and in the BSE Phe-Phe-Phe-OH) (Bachem, CA) were mixed into solutions without ants. All image, the gray levels in the embedding epoxy and those in the cuticle are similar. vessels during the experiments have been washed with deionized water and pre- Therefore, there is no rigorous and quantitative method to select one but not the treated with 6 M hydrochloric acid to prevent carbonate contamination. Filters and other. We used Adobe Photoshop and the Magic Wand tool with a tolerance of 30 ants were air-dried for XRD and SEM characterization. to select all of the cuticles and deleted all those pixels from a black mask. The brighter mineral and all of the mineral debris deposited on the epoxy were then Scanning electron microscopy (SEM) and electron backscatter diffraction selected using the Magic Wand and a tolerance of 50 on the BSE image. These were (EBSD). Scanning electron microscopy (SEM) was done using a Hitachi S3400 at also deleted from the same black mask. The black pixels in the BSE image corre- 15 kV. Images were obtained in both variable pressure and vacuum mode. Energy- spond to gaps between the cuticle and the epoxy, or holes between mineral crystals, dispersive X-ray spectroscopy (EDS) and electron backscatter diffraction (EBSD) those black pixels were remained black in the black mask. The bright mineral was carried out using an AZtecOne system with silicon-drift detector from Oxford debris is presumably an artifact of polishing, as they appear both in PEEM and instruments. Samples were coated with 5 nm Pt coating. Phases used in EBSD are SEM images and are spectroscopically identified without a doubt as a mineral. constructed based on Mg-poor (a = 4.990 Å, c = 17.062 Å), Mg-medium (a = These were also removed from the mask and therefore displayed in Fig. 2d, as 4.920 Å, c = 16.656 Å; calculated), Mg-rich (a = 4.850 Å, c = 16.250 Å) regions. removing them would have been an artifact. The BSE image was warped to cor- However, given the EBSD is not particularly sensitive to unit-cell parameter dif- respond correctly to the PEEM image using Adobe Photoshop and specifically the ferences, chemical heterogeneity from EBSD is only qualitative. Puppet Warp tool. Obtaining component spectra. We extracted single-pixel spectra from the cuticle, Transmission electron microscopy (TEM). TEM measurements were carried out the epicuticle, and the mineral regions. These were identified as the only three using a Philips CM200UT TEM instrument operating at 200 kV acceleration vol- reliable components that were spectroscopically distinct from one another and not tage with 0.5 mm spherical aberration (Cs) and a point resolution of 0.19 nm. linear combinations of other components. The single-pixel spectra from the same Images and electron diffraction were collected with a CCD camera and analyzed material were extracted from each stack, aligned in energy, and averaged. The with Gatan DigitalMicrograph software. Samples that have been previously 8 NATURE COMMUNICATIONS | (2020) 11:5792 | https://doi.org/10.1038/s41467-020-19566-3 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19566-3 ARTICLE examined by XRD and SEM were rinsed with ethanol and DI water to remove glue (30–100% [vol/vol]), followed by critical point drying and coating with 1-nm residue. Samples were then crushed in an agar mortar, suspend in acetone, and platinum. Scanning electron microscopy (SEM) of samples was performed using a drop onto Lacey/ carbon 200 mesh copper grid. The composition of phases was LEO 1530 microscope to investigate the cuticular structure. confirmed with energy-dispersive X-ray spectroscopy (EDS) and analyzed with Thermo Noran software. High-performance liquid chromatography (HPLC) analysis. Dissected ant cuticle samples were placed in 100 µL 25% TFA containing 10 mM DTT and were In situ X-ray diffraction (XRD) analyzes. In situ X-ray diffraction (XRD) was hydrolyzed at 110 °C for 24 h. Hydrolyzed samples were then dried at 45 °C under a performed using a Rigaku Rapid II X-ray diffraction system with Mo Kα radiation. stream of nitrogen and resuspended in 50 mM HCl. Amino acids were then con- This XRD instrument uses a 2-D image-plate detector for signal collection and verted into respective fluorescent derivatives using o-phthalaldehyde (OPA) integrated using Rigaku’s 2DP software. XRD was run at 50 kV and a 100-µm (Agilent #5061-3335). Briefly, 5 µL sample aliquot was added to 20 µL of 40 mM diameter collimator. Whole fresh ant samples were glued onto an American potassium tetraborate buffer (pH 9.8) followed by the addition of 5 µL OPA, mixed Durafilm Kapton® tube with vacuum grease. Ant samples were then spin around gently and another 40 µL water was added. The mixture was filtered through a 0.45 phi and oscillate on omega. Synthesized powder samples were sealed in Kapton µm cellulose acetate 4-mm syringe filter (Nalgene #171-0045). Freshly prepared tube and run with fixed omega and phi spin. Refinements for phase percentage and samples were immediately subjected to HPLC analysis. unit-cell parameters were run using Jade 9.0 software with American Mineralogist Amino acids were analyzed using a modified method . In brief, the apparatus Crystal Structure Database (AMCSD) and the PDF-4+ database from the Inter- used was a custom-built dual analytical/semi-preparative Shimadzu system national Centre for Diffraction Data (ICDD). Disordered dolomite reference was consisting of a SIL-20AC autosampler, a CBM-20A system controller, two LC- constructed based on the unit-cell parameter of disordered dolomite with 50 mol.% 20AR pumps, a C50-20AC oven, a PDA S10-M20A detector, and a CPP-10Avp MgCO3 and powder diffraction pattern calculated by CrystalMaker built-in detector. Chromatographic separation of OPA-derivatized amino acids was CrystalDiffract software. performed using an Agilent ZORBAX Eclipse AAA column (4.6 mm × 150 mm × 3.5 µm; Agilent #963400-902) coupled with a ZORBAX Eclipse AAA Analytical Guard Column (4.6 mm × 12.5 mm × 5 µm; Agilent #820950-931) heated at 40 °C. Sectioning and transmission electron microscopy (TEM). Ants for sectioning The gradient elution was applied using 40 mM sodium phosphate dibasic buffer and transmission electron microscopy were fixed in cold 2% glutaraldehyde in Na- (pH 7.8) as solvent A and a mixture of acetonitrile, methanol, and water (45:45:10, cacodylate buffer. Postfixation was done in 2% osmium tetroxide and specimens v/v/v) as solvent B. HPLC-grade acetonitrile and methanol were supplied from were subsequently dehydrated in a graded acetone series. Specimens were Fisher Scientific and were used without further purification. The optimum embedded in Araldite and sectioned with a Reichert Ultracut E microtome. separation of amino acids was obtained using the following gradient program: 0% B Semithin 1-µm sections for light microscopy were stained with methylene blue and for 1.9 min, then increase to 57% B up to 28.10 min followed by an increase to thionin. Double-stained 70-nm thin sections were examined in a Zeiss EM900 100% B up to 38.60 min, then hold at 100% B till 47.30 min, and decrease to 22.3% electron microscope. B up to 48.20 min and down to 0% B till 60 min. The flow rate was 1 mL/min. Aliquots of 10 μL standards/samples were injected at 0.5 min and amino acids were Quantitative electron probe micro-analysis (EPMA). The carbonate EPMA data detected at the maximum wavelength of 338 nm with a 4 nm bandwidth. Retention were acquired with a CAMECA SXFive FE electron probe in the Cameca Electron times, as well as spectral information, was given by the PDA detector was used for Probe Lab in the Department of Geoscience at the University of Wisconsin-Madison. peak identification. Calibration curves of individual and mixed amino acids were Operating conditions were 7 kv and 10 nA (Faraday cup), using a focused beam. A prepared using either 250 pmol stocks of corresponding individual amino acids or low accelerating voltage was used to shrink the analytical volume to less than 300 nm. a 250 pmol amino acid standard mix (Agilent #5061-3331). Quantification was Peak counting time was 10 s, with background acquired for 10 s. Mg Ka was acquired performed using the calibration curves of the respective amino acid standards. with a TAP crystal and Ca Ka with an LPET crystal. The standard used as Delight Dolomite. Automation and data reduction utilized Probe for EPMA (Probe Software), Biomineral-free Ac. echinatior ant generation. Biomineral-free Ac. echinatior Carbon and oxygen were accounted for in a robust procedure in the Probe for EPMA ants were generated using a sub-colony setup. Sub-colonies were set up in small software: oxygen was calculated based upon stoichiometry to the measured Mg and (diameter 6 cm, height 4 cm) clear plastic containers. After sterilizing containers for Ca, with carbon calculated relative to that oxygen value (1:3), with this being iterated several times within the Armstrong/Love Scott matrix correction. The resulting values at least 20 min using UV light, cotton moistened with distilled water was placed at the bottom to help provide humidity. A small (width 4.12 cm, length 4.12 cm, and were then evaluated for actual accuracy, based upon two criteria: a non-normalized analytical total close to 100 wt% (~98–102 wt%), and for a formula basis of 3 oxygens, height 0.79 cm) weigh boat (Fisher catalog #08-732-112) was placed on top of the wet cotton, and then 0.1 g of the fungus garden, 2 minor workers, and a major the carbon formula value being close to 1.00 (~0.99–1.01). With these conditions met, worker pupa being reared to derive a biomineral-free adult (n = 10 sub-colonies). the determined compositions were deemed acceptable. A ∼1cm leaf fragment of pin oak (Quercus palustris) was added 24 h or more after pupa eclosion for the ants to cut and incorporate into the fungus garden. We Raman and attenuated total reflectance Fourier-transform infrared (ATR- monitored sub-colonies daily to record the eclosion date for the major worker pupa FTIR) spectroscopy. Ant mineral samples for Raman experiments were prepared until 14–21 days after eclosion. Then we performed environmental Scanning by bleaching the freeze-dried ant samples in 8.25% NaClO commercial bleach for Electron Microscopy (eSEM) and XRD on a subset of the ants to confirm the 24 h at room temperature to remove the exposed organic materials . Raman absence of the biomineral. Meanwhile, we established that Ac. echinatior ants could spectra were collected using a LabRam Raman microprobe (JY Horiba, Inc.) grow a biomineral layer normally in sub-colonies with the addition of two major equipped with a Microscope (Olympus DX41, ×50 and ×100 objectives) and a 633 worker adults (n = 10 sub-colonies) and other colony components were main- nm laser. Spectra were acquired with a CCD camera behind a spectrometer (the tained as above. accumulations and integration time varied). The ant carbonate powders were dropped on a microscope slide just before individual measurement. For ATR-FTIR, freeze-dried ant samples were used directly. ATR-FTIR data Nano-mechanical testing. Nanoindentation tests were carried out using a Bruker collection was conducted on a Perkin-Elmer 1720x spectrometer according to the Hysitron PI-85 SEM Picoindenter in a Zeiss Leo 1550VP SEM at the Wisconsin manufacturer’s instructions. Centers for Nanoscale Technology, UW-Madison. The samples were tested using a cube-corner probe with a basic quasistatic trapezoid load-controlled function, where the maximum load was 500 µN, the hold time was 2 s, and the loading/ Rearing experiments. In a total of 20 worker pupae of Ac. echinatior at the same unloading rate was 100 µN/s . SEM imaging was done in a high vacuum using an developmental stage and its ~10 g fungus garden were collected and randomly accelerating voltage of 3 kV with secondary electron mode. In order to simulate the sorted into two groups of ten and maintain them in chambers (diameter 6 cm, defense mechanism of the actual ant exoskeleton, we tested the combination of ant height 4 cm) with wet cotton. Reared to callow workers (around 3 h) and followed mineral and ant cuticle by indenting from the outside in, as illustrated in Sup- by one worker from each of two groups was collected every second day. The fresh plementary Fig. 13. The samples for testing the combination of ant mineral and ant ant samples were subjected to XRD analyses immediately and followed by an cuticle were prepared by slicing through the transverse plane of the head of the ant Environmental scanning electron microscope examination (eSEM). For eSEM, an to allow probing only on the flatter top part of the head. All samples were then FEI QUANTA 200 eSEM (FEI Company) was used. Ants were placed directly onto attached firmly to carbon tape on an SEM stub. The combination sample was the eSEM stub and examined without any preparation (i.e., samples were not fixed additionally pressed carefully with tweezers to ensure good attachment and flat- or coated for this analysis). All samples were analyzed at 5.0 torr, 3.0 spot size, ness. The indentation data and corresponding SEM video were analyzed not only and 4 °C. to ensure that the correct contact point of the probe with the surface was chosen, but also that there was no movement of the sample during indentation, as well as Ant sample preparation for SEM analyses of the cuticular structure. Worker that the load-displacement curve was smooth and there was no abrupt pop-out or ant cuticles from Ac. echinatior at different developmental stages (pupae to fully discontinuity. Based on these criteria, 13, 13, 10, and 12 valid data points were mature adult workers) and At. cephalotes were immediately fixed with 4% (vol/vol) selected for Ac. echinatior ant mineral plus cuticle and Ac. echinatior ant cuticle formaldehyde and 1% glutaraldehyde at 22 °C RT overnight. Samples were then cross-section, respectively. Measurements on other insect cuticles were also done washed with PBS and treated with 1% osmium tetroxide for 30 min at 22 °C. by indenting from the outside in following the same analyzing criteria, with which Samples were subsequently washed with a series of increasing ethanol dilutions 15, 13, 15, and 12 valid data points were selected for Atta soldier ant cuticle, Atta NATURE COMMUNICATIONS | (2020) 11:5792 | https://doi.org/10.1038/s41467-020-19566-3 | www.nature.com/naturecommunications 9 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19566-3 worker ant cuticle, beetle elytra, and honeybee cuticle, respectively. All the valid References points were then used to obtain the hardness values presented in Fig. 4a. 1. Knoll, A. H. Biomineralization and evolutionary history. Rev. Mineral. Additional hardness comparison of the mineral phases alone was done on polished Geochem. 54, 329–356 (2003). cross-sections of geologic dolomite and Ac. echinatior mineral. Cross-section 2. Eder, M., Amini, S. & Fratzl, P. Biological composites-complex structures for samples were prepared using the same embedding and polishing procedures as functional diversity. Science 362, 543–547 (2018). described for the PEEM samples. Representative load-depth curves in Supple- 3. Sperling, E. A. et al. Oxygen, ecology, and the Cambrian radiation of animals. mentary Fig. 14 were selected from the data point closest to the averaged hardness Proc. Natl Acad. Sci. USA 110, 13446–13451 (2013). value in each sample. 4. Gilbert, P. U. P. A. et al. Biomineralization by particle attachment in early animals. Proc. Natl Acad. Sci. USA 116, 17659–17665 (2019). 5. Gilbert, P. U. P. A., Abrecht, M. & Frazer, B. H. The organic-mineral interface Aggressive experiments between At. cephalotes soldier and Ac. echinatior in biominerals. Mol. Geomicrobiol. 59, 157–185 (2018). workers. We confronted one major At. cephalotes soldier and three mineral-pre- 6. Bengtson, S. Early Life on Earth. Nobel Symposium No. 84. (ed. Bengtson, S.) sent/mineral-free workers of Ac. echinatior . The experiment was replicated 5 412–425 (Columbia University Press, New York, 1994). times under reduced light in 9 cm Petri dishes. The survival of ants was counted 7. Sanderson, K. Artificial armour. Nature 519, S14–S15 (2015). after 3 h of confrontation. 8. Mass, T. et al. Amorphous calcium carbonate particles form coral skeletons. Using the time-lapse setting of an iPad Pro 2018, we recorded aggressive encounters between At. cephalotes soldier and Ac. echinatior workers Proc. Natl Acad. Sci. USA 114, E7670–E7678 (2017). (Supplementary Movies S2 and S3). For the aggressive experiments with mineral- 9. DeVol, R. T. et al. Nanoscale transforming mineral phases in fresh nacre. J. present ants, the video was started when At. cephalotes soldier was placed in the Am. Chem. Soc. 137, 13325–13333 (2015). petri dish with the Ac. echinatior ants. The video was stopped when the soldier ant 10. Weaver, J. C. et al. The stomatopod dactyl club: a formidable damage-tolerant was killed and the Ac. echinatior worker was able to separate from the soldier ant. biological hammer. Science 336, 1275–1280 (2012). The video is a total of 42 min filmed in time-lapse at 120 times its speed reducing 11. Politi, Y., Arad, T., Klein, E. & Addadi, L. Sea urchin spine calcite forms via a the video to 21 s in length, followed by editing it down to 10% of its speed using transient amorphous calcium carbonate phase. Science 306, 1161–1165 (2004). Premiere Pro resulting in a 3:27 min video. For the aggressive experiments with 12. Li, L. et al. Multifunctionality of chiton biomineralized armor with an mineral-free ants, the video was started when the soldier ant was placed into the integrated visual system. Science 350, 952–956 (2015). petri dish, video was stopped after all three Ac. echinatior ants were killed by the 13. Aizenberg, J., Tkachenko, A., Weiner, S., Addadi, L. & Hendler, G. Calcitic soldier ant. The video was recorded for 1 h in time-lapse at 120 times its speed microlenses as part of the photoreceptor system in brittlestars. Nature 412, reducing the video to 30 s in length, followed by editing down to 10% of its original 819 (2001). speed resulting in a video that is 5:00 min. The counting of lost body parts was 14. Killian, C. E. et al. Self-sharpening mechanism of the sea urchin tooth. Adv. based on the video record. Funct. Mater. 21, 682–690 (2011). 15. Ma, Y. et al. The grinding tip of the sea urchin tooth exhibits exquisite control Entomopathogenic fungi infection. Using two groups of Ac. echinatior major over calcite crystal orientation and Mg distribution. Proc. Natl Acad. Sci. USA worker ants in which the mineral was either present or absent, we conducted 106, 6048–6053 (2009). infection experiments using the entomopathogenic fungus Metarhizium aniso- 16. Wang, R. Z., Addadi, L. & Weiner, S. Design strategies of sea urchin teeth: pliae (Ascomycota, Hypocreales) . In brief, groups of three ant individuals were Structure, composition and micromechanical relations to function. Philos. placed into an individual Petri dish with a ring of moist cotton. Then, each ant’s Trans. R. Soc. B Biol. Sci. 352, 469–480 (1997). propleural plate was inoculated with 0.5 μLof Metarhizium spores of ca. 1.00 × 17. Polishchuk, I. et al. Coherently aligned nanoparticles within a biogenic single 7 −1 10 conidiospores mL suspension + 0.01% Tween 20 by using a micropipette crystal: a biological prestressing strategy. Science 358, 1294–1298 (2017). under a dissecting microscope. The experiment was replicated five times. The 18. Morse, J. W., Wang, Q. & Tsio, M. Y. Influences of temperature and Mg:Ca control for each replicate consisted of inoculating with a control solution of ratio on CaCO precipitates from seawater. Geology 25,85–87 (1997). sterile, deionized water + 0.01% Tween 20 . The survival of ants (three indi- 19. Oakley, T. H., Wolfe, J. M., Lindgren, A. R. & Zaharoff, A. K. viduals per sub-colony over five sub-colonies) was monitored every 24 h post- Phylotranscriptomics to bring the understudied into the fold: monophyletic treatment for 6 days. Ostracoda, fossil placement, and pancrustacean phylogeny. Mol. Biol. Evol. 30, The Metarhizium anisopliae var. anisopliae strain was isolated from dead fungus- 215–233 (2013). growing ants in Gamboa, Panama, and grown on a pure medium of potato dextrose 20. Schultz, T. R. & Brady, S. G. Major evolutionary transitions in ant agriculture. agar at room temperature until a full plate of spores was observed after around 7 days. Proc. Natl Acad. Sci. USA 105, 5435–5440 (2008). The spores (conidia) were centrifuged, washed three times with sterile 0.05% Triton-X 21. Li, H. et al. Convergent evolution of complex structures for ant-bacterial solution, and then harvested fresh for each infection experiment. For the infection defensive symbiosis in fungus-farming ants. Proc. Natl Acad. Sci. USA 115, experiments, spore suspensions were set up from recently sporulating cultures in a 10720–10725 (2018). solution of sterile-deionized water containing 0.01% Tween 20. Spore concentration 22. Pinto-Tomás, A. A. et al. Symbiotic nitrogen fixation in the fungus gardens of was quantified using a hemocytometer and diluted to achieve a concentration of ca. leaf-cutter ants. Science 326, 1120–1123 (2009). 7 −1 1.00 × 10 conidiospores mL . 23. Weber, N. A. Gardening Ants: The Attines (American Philosophical Society, Philadelphia, 1972). Reporting summary. Further information on research design is available in the Nature 24. Wetterer, J. K. Forager size and ecology of Acromyrmex coronatus and other Research Reporting Summary linked to this article. leaf-cutting ants in Costa Rica. Oecologia 104, 409–415 (1995). 25. Bekkevold, D., Frydenberg, J. & Boomsma, J. Multiple mating and facultative polygyny in the Panamanian leafcutter ant Acromyrmex echinatior. Behav. Data availability Ecol. Sociobiol. 46, 103–109 (1999). Source data files are provided on https://figshare.com/s/29b90a59a699bb33bf32. Each 26. Ferguson-Gow, H., Sumner, S., Bourke, A. F. G. & Jones, K. E. Colony data set is provided as an Igor Pro experiment with file extension.pxp. All the relevant size predicts division of labour in attine ants. Proc. R. Soc. B 281, 20141411 files for C and O spectra are zipped separately. Permits for collections and accessing (2014). genetic resources in Brazil were issued by SISBIO #46555–5 and CNPq #010936/2014–9. 27. Currie, C. R., Poulsen, M., Mendenhall, J., Boomsma, J. J. & Billen, J. Costa Rican collecting permits were issued by the Comisión Institucional de Coevolved crypts and exocrine glands support mutualistic bacteria in fungus- Biodiversidad (Institutional Biodiversity Committee, University of Costa Rica; growing ants. Science 311,81–83 (2006). Resolutions # 012 and 020; Material Transfer Agreement MTA VI-4307–2013) and 28. Currie, C. R., Scott, J. A., Summerbell, R. C. & Malloch, D. Fungus-growing authorized by La Selva Biological Station. Photos of Atta ants image and Apis bee image ants use antibiotic-producing bacteria to control garden parasites. Nature 398, are used under a perpetual commercial license from Alexander L. Wild and Getty 701–704 (1999). Images, respectively (Fig. 4). Xylotrechus beetle image provided with permission from the 29. Currie, C. R., Mueller, U. G. & Malloch, D. The agricultural pathology of ant copyright holder, Jon Rapp (Fig. 4). fungus gardens. Proc. Natl Acad. Sci. USA 96, 7998–8002 (1999). 30. Adams, R. M. M., Mueller, U. G., Schultz, T. R. & Norden, B. Agro-predation: Code availability usurpation of attine fungus gardens by Megalomyrmex ants. The Igor Pro macros, called GG Macros, used to produce PIC maps are available free of Naturwissenschaften 87, 549–554 (2000). charge on https://home.physics.wisc.edu/gilbert/software/. The code to measure the 31. Fowler, H. G. Field response of Acromyrmex crassispinus (forel) to aggression angular distances of c-axes in Fig. 2 is available on https://home.physics.wisc.edu/gilbert/ by Atta sexdens (linn.) and predation by Labidus praedator (fr. smith) software/. (Hymenoptera: Formicidae). Aggress. Behav. 3, 385–391 (1977). 32. Vilela, E. F. & Howse, P. E. Fire Ants and Leaf-Cutting Ants. Biology and Management (eds Lofgren, C. S. & Vander Meer, R. K.) 159–171 (Westview Received: 16 March 2020; Accepted: 15 October 2020; Press. Boulder, CO, 1986). 10 NATURE COMMUNICATIONS | (2020) 11:5792 | https://doi.org/10.1038/s41467-020-19566-3 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19566-3 ARTICLE 33. Schofield, R. M., Nesson, M. H. & Richardson, K. A. Tooth hardness increases 62. Wang, C. & Wang, S. Insect pathogenic fungi: genomics, molecular with zinc-content in mandibles of young adult leaf-cutter ants. interactions, and genetic improvements. Annu. Rev. Entomol. 62,73–90 Naturwissenschaften 89, 579–583 (2002). (2017). 34. Weber, N. A. Gardening Ants: The Attines (American Philosophical Society, 63. Wang, B., Kang, Q., Lu, Y., Bai, L. & Wang, C. Unveiling the biosynthetic Philadelphia, 1972). puzzle of destruxins in Metarhizium species. Proc. Natl Acad. Sci. USA 109, 35. Myers, C. E. et al. Exceptional preservation of organic matrix and shell 1287–1292 (2012). microstructure in a Late Cretaceous Pinna fossil revealed by photoemission 64. Graf, D. L. & Goldsmith, J. R. Some hydrothermal syntheses of dolomite and electron spectromicroscopy. Geology 46, 711–714 (2018). protodolomite. J. Geol. 64, 173–186 (1956). 36. Gilbert, P. U. P. A., Young, A. & Coppersmith, S. N. Measurement of c-axis angular orientation in calcite (CaCO ) nanocrystals using X-ray absorption Acknowledgements spectroscopy. Proc. Natl Acad. Sci. USA 108, 11350–11355 (2011). We thank U. P. Agarwal and S. A. Ralph from U.S. Forest Products Laboratory for the 37. Metzler, R. A. et al. Polarization-dependent imaging contrast in abalone shells. Raman spectroscopy measurements; B. Schneider and R. Noll for expert help conducting Phys. Rev. B 77, 064110-1/9 (2008). SEM work; J. Morasch for assistance with nanoindentation measurements; E. Okonski 38. Cody, G. D. et al. Molecular signature of chitin-protein complex in Paleozoic for ant imaging and laboratory assistance; and R. J. Massey for assistance with the arthropods. Geology 39, 255–258 (2011). microtome. This work was primarily supported by the National Institutes of Health 39. Kaznacheyev, K. et al. Innershell absorption spectroscopy of amino acids. J. (NIH) Grant U19 TW009872-05, NIH Grant U19 AI109673 and the Department of Phys. Chem. A 106, 3153–3168 (2002). Energy Great Lakes Bioenergy Research Center Office of Science Grant DE-FC02- 40. Zhang, F., Xu, H., Konishi, H., Shelobolina, E. S. & Roden, E. E. 07ER64494 to C.R.C. P.U.P.A.G. acknowledges support from the U.S. Department of Polysaccharide-catalyzed nucleation and growth of disordered dolomite: a Energy, Office of Science, Office of Basic Energy Sciences, Chemical Sciences, Geos- potential precursor of sedimentary dolomite. Am. Mineral. 97, 556–567 ciences, and Biosciences Division, under Award DE-FG02-07ER15899, and NSF grant (2012). DMR-1603192. PEEM experiments were done at the Advanced Light Source, which is a 41. Lippmann, F. Sedimentary Carbonate Minerals (Springer, New York, 1973). DOE Office of Science User Facility supported by grant DE-AC02-05CH11231. T.R.S. is 42. Peeters, C., Molet, M., Lin, C. C. & Billen, J. Evolution of cheaper workers in supported by the National Science Foundation award DEB 1654829 and DEB 1927224. ants: a comparative study of exoskeleton thickness. Biol. J. Linn. Soc. 121, H.X. and Y.F. is supported by the NASA Astrobiology Institute (NNA13AA94A) and 556–563 (2017). S. W. Bailey Scholarship of the Department of Geoscience. 43. Hölldobler, B. & Wilson, E. O. The Superorganism: The Beauty, Elegance, and Atrangeness of Insect Societies (WW Norton, New York, 2008). 44. Dijkstra, M. B. & Boomsma, J. J. Gnamptogenys hartmani Wheeler Author contributions (Ponerinae: Ectatommini): an agro-predator of Trachymyrmex and Study design: H.L., C.-Y.S., P.U.P.A.G., and C.R.C. Experimental design and supervision: Sericomyrmex fungus-growing ants. Naturwissenschaften 90, 568–571 (2003). H.L., C.-Y.S., T.R.S., P.U.P.A.G., and C.R.C. Data collection and analysis: H.L., C.-Y.S., 45. Yang, L., Killian, C. E., Kunz, M., Tamura, N. & Gilbert, P. U. P. A. Biomineral Y.F., C.M.C., H.X., A.J., J.S.-C., R.Z., H.A.B., J.H.F., D.R.A., T.R.S., P.U.P.A.G., and C.R.C. nanoparticles are space-filling. Nanoscale 3, 603–609 (2011). Initial draft: H.L., C.R.C., T.R.S. C.-Y.S., and P.U.P.A.G. Final version: all authors. 46. Wood, R., Ivantsov, A. Y. & Zhuravlev, A. Y. First macrobiota biomineralization was environmentally triggered. Proc. R. Soc. B 284, 20170059 (2017). Competing interests The authors declare no competing interests. 47. Nygaard, S. et al. Reciprocal genomic evolution in the ant-fungus agricultural symbiosis. Nat. Commun. 7,1–9 (2016). 48. Diamond, J. M. & Ordunio, D. Guns, Germs, and Steel (Books on Tape, 1999). Additional information 49. Sun, C. Y. et al. Spherulitic growth of coral skeletons and synthetic Supplementary information is available for this paper at https://doi.org/10.1038/s41467- aragonite: nature’s three-dimensional printing. ACS Nano 11, 6612–6622 020-19566-3. (2017). 50. Gong, Y. U. et al. Phase transitions in biogenic amorphous calcium carbonate. Correspondence and requests for materials should be addressed to P.U.P.A.G. or C.R.C. Proc. Natl Acad. Sci. USA 109, 6088–6093 (2012). 51. De Stasio, G., Frazer, B. H., Gilbert, B., Richter, K. L. & Valley, J. W. Peer review information Nature Communications thanks Charissa de Bekker and the Compensation of charging in X-PEEM: a successful test on mineral inclusions other, anonymous, reviewers for their contribution to the peer review of this work. Peer in 4.4Ga old zircon. Ultramicroscopy 98,57–62 (2003). reviewer reports are available. 52. DeVol, R. T. et al. Oxygen spectroscopy and polarization-dependent imaging contrast (PIC)-mapping of calcium carbonate minerals and biominerals. J. Reprints and permission information is available at http://www.nature.com/reprints Phys. Chem. B 118, 8449–8457 (2014). 53. Zou, Z. et al. A hydrated crystalline calcium carbonate phase: Calcium Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in carbonate hemihydrate. Science 363, 396–400 (2019). published maps and institutional affiliations. 54. Macros, G. G. http://home.physics.wisc.edu/gilbert/software.htm (2019). 55. Pokroy, B. et al. Narrowly distributed crystal orientation in biomineral vaterite. Chem. Mater. 27, 6516–6523 (2015). 56. Stohr, J. NEXAFS Spectroscopy (Springer-Verlag, Berlin, 1992). Open Access This article is licensed under a Creative Commons 57. Madix, R. J., Solomon, J. L. & Stohr, J. The orientation of the carbonate anion Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give on Ag(110). Surf. Sci. 197, L253–L259 (1988). 58. Henderson, J. W. Jr., Ricker, R. D., Bidlingmeyer, B. A. & Woodward, C. appropriate credit to the original author(s) and the source, provide a link to the Creative Rapid, Accurate, Sensitive and Reproducible HPLC Analysis of Amino Acids Commons license, and indicate if changes were made. The images or other third party (Agilent Pub.# 5980–1193E, 2000). material in this article are included in the article’s Creative Commons license, unless 59. Amini, S., Tadayon, M., Idapalapati, S. & Miserez, A. The role of quasi- indicated otherwise in a credit line to the material. If material is not included in the plasticity in the extreme contact damage tolerance of the stomatopod dactyl article’s Creative Commons license and your intended use is not permitted by statutory club. Nat. Mater. 14, 943–950 (2015). regulation or exceeds the permitted use, you will need to obtain permission directly from 60. Oliver, W. C. & Pharr, G. M. An improved technique for determining the copyright holder. To view a copy of this license, visit http://creativecommons.org/ hardness and elastic modulus using load and displacement sensing licenses/by/4.0/. indentation experiments. J. Mater. Res. 7, 1564–1583 (1992). 61. Hölldobler, B. & Wilson, E. O. The Ants (Harvard Univ. Press, Cambridge, © The Author(s) 2020 MA, 1990). NATURE COMMUNICATIONS | (2020) 11:5792 | https://doi.org/10.1038/s41467-020-19566-3 | www.nature.com/naturecommunications 11 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nature Communications Springer Journals

Loading next page...
 
/lp/springer-journals/biomineral-armor-in-leaf-cutter-ants-IU2MTy3six

References (103)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2020
eISSN
2041-1723
DOI
10.1038/s41467-020-19566-3
Publisher site
See Article on Publisher Site

Abstract

ARTICLE https://doi.org/10.1038/s41467-020-19566-3 OPEN 1,2,3 4 5 1 5 6 Hongjie Li , Chang-Yu Sun , Yihang Fang , Caitlin M. Carlson , Huifang Xu , Ana Ješovnik , 6,7 8,9 10 5 8,9 Jeffrey Sosa-Calvo , Robert Zarnowski , Hans A. Bechtel , John H. Fournelle , David R. Andes , ✉ ✉ 6 4,5,11 1,2 Ted R. Schultz , Pupa U. P. A. Gilbert & Cameron R. Currie Although calcareous anatomical structures have evolved in diverse animal groups, such structures have been unknown in insects. Here, we report the discovery of high-magnesium calcite [CaMg(CO ) ] armor overlaying the exoskeletons of major workers of the leaf-cutter 3 2 ant Acromyrmex echinatior. Live-rearing and in vitro synthesis experiments indicate that the biomineral layer accumulates rapidly as ant workers mature, that the layer is continuously distributed, covering nearly the entire integument, and that the ant epicuticle catalyzes biomineral nucleation and growth. In situ nanoindentation demonstrates that the biomineral layer significantly hardens the exoskeleton. Increased survival of ant workers with biomi- neralized exoskeletons during aggressive encounters with other ants and reduced infection by entomopathogenic fungi demonstrate the protective role of the biomineral layer. The dis- covery of biogenic high-magnesium calcite in the relatively well-studied leaf-cutting ants suggests that calcareous biominerals enriched in magnesium may be more common in metazoans than previously recognized. 1 2 Department of Bacteriology, University of Wisconsin-Madison, Madison, WI 53706, USA. Department of Energy Great Lakes Bioenergy Research Center, Wisconsin Energy Institute, University of Wisconsin-Madison, Madison, WI 53726, USA. State Key Laboratory for Managing Biotic and Chemical Threats to the Quality and Safety of Agro-products, Key Laboratory of Biotechnology in Plant Protection of Ministry of Agriculture and Zhejiang Province, Institute of Plant Virology, Ningbo University, 315211 Ningbo, China. Department of Physics, University of Wisconsin-Madison, Madison, WI 53706, USA. 5 6 Department of Geoscience, University of Wisconsin-Madison, Madison, WI 53706, USA. Department of Entomology, National Museum of Natural History, Smithsonian Institution, Washington, DC 20560-0003, USA. School of Life Sciences, Arizona State University, Tempe, AZ 85287, USA. 8 9 Department of Medicine, University of Wisconsin-Madison, Madison, WI 53706, USA. Department of Medical Microbiology and Immunology, University of Wisconsin-Madison, Madison, WI 53706, USA. Advanced Light Source Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA. Departments of Chemistry, Materials Science and Engineering, University of Wisconsin-Madison, Madison, WI 53706, USA. Previously published as Gelsomina De Stasio: Pupa U. P. A. Gilbert. email: pupa@physics.wisc.edu; currie@bact.wisc.edu NATURE COMMUNICATIONS | (2020) 11:5792 | https://doi.org/10.1038/s41467-020-19566-3 | www.nature.com/naturecommunications 1 1234567890():,; ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19566-3 iomineral skeletons first appeared more than 550 million 1–5 years ago , and by the early Cambrian biomineral-based Bdefensive structures had evolved in most extant metazoan phyla, apparently in response to increasing predation pressure . The minerals involved, as well as the biogenic structures they form, are diverse. Calcium carbonate biomineralization is parti- 1,7 8 cularly widespread among metazoans : the hard parts of corals , 9 10 mollusk shells , stomatopod dactyl club , and sea urchin spines contain calcium carbonate, as do the light-focusing eye 12,13 lenses of chitons and brittlestars . Magnesium-enriched calcite (CaCO ) has been discovered in the central part of the sea urchin tooth, where the increased hardness imparted by magnesium is 14–16 thought to aid in the grinding of limestone . Given the 0.5 mm importance of calcareous anatomical structures across metazoan phyla and given that magnesium significantly strengthens such structures , it is surprising that high-magnesium calcite appears to be rare in animals . It is also surprising that despite the near ubiquity of biogenic mineralization across metazoan phyla and the widespread presence of calcium carbonate in the Crustacea, biomineralized calcium carbonate has so far remained unknown in the most diverse group of animals, the insects, which arose from within the Crustacea . Here we report the discovery of a dense layer of biogenic high-magnesium calcite in the leaf-cutter ant Acromyrmex echinatior. Fungus-growing attine ants (tribe Attini, subtribe Attina) engage in an ancient and obligate mutualism with coevolved fungi 1 µm (order Agaricales), which they cultivate for food. Fungus farming, cd which has been described as a major transition in evolution , evolved only once in ants around 60 million years ago . Leaf- Mineral layer cutting ants (genera Acromyrmex and Atta), a phylogenetically derived lineage that arose within the fungus-growing ants around Cuticle 20 million years ago, harvest fresh vegetation as the substrate on Mineral layer which they grow their fungal mutualists. They are ecologically 20,21 dominant herbivores in the New World tropics and serve important roles in carbon and nitrogen cycling . A mature Atta leaf-cutter ant colony comprises a superorganism with >5 million Epicuticlar microstructure workers, a single queen, and complex society with a highly refined division of labor based both on worker size and age. Acromyrmex Cuticle 1 µm leaf-cutter colonies vary in size, generally within the range of 100 µm 23,24 15,000–100,000 workers , and in some species may have more than one queen. Acromyrmex echinatior, the focus of the present Fig. 1 Morphological and structural characterization of minerals on the study, has a mean colony size of 137,500 workers and is facul- cuticle of Ac. echinatior.a Ac. echinatior ant with a whitish cuticular coating 25,26 tatively polygynous . In addition to the leaf-cutters, 17 other (Photo T.R.S.). b SEM image of ant cuticle with crystalline coating. genera of ants occur within the Attina, all of which grow fungus c Backscattered electron (BSE) image of a polished cuticular cross-section gardens, form colonies of hundreds to a few thousand workers, of an ant. This layer is brighter than the cuticle in backscattered electron and use dead vegetative matter or caterpillar frass rather than (BSE) mode scanning electron microscopy (SEM), indicating that it consists fresh leaves and grasses as substrates for their gardens. In addi- of heavier elements and is continuous, covering nearly the entire surface. tion to the symbiotic association with their fungal cultivars, many d BSE image close-up of a polished cuticular cross-section of an ant. fungus-growing ants engage in a second mutualism with Acti- nobacteria (genus Pseudonocardia), which produce antibiotics Based on combined data from in situ X-ray diffraction (XRD), 27–29 that help defend the garden from fungal pathogens . Fungus- electron microscopy, electron backscatter diffraction (EBSD), growing ant colonies, containing both fungal crops and immature quantitative electron probe micro-analysis (EPMA), and Raman ant brood, represent a rich nutritional resource for a wide variety and attenuated total reflectance Fourier-transform infrared of marauding ant species, including army ants and other known (ATR-FTIR) spectroscopy, we report here that this coating is in agro-predatory raiders of ant agriculture . Smaller fungus- fact a mineral layer covering the ant exoskeleton. The layer is growing ant colonies, including those of Acromyrmex, are also composed of euhedral rhombohedral crystals with curved faces, occasionally subject to attack by the large-sized soldier castes of 3–5 μm in size (Fig. 1b). To examine the mechanism of crystal Atta leaf-cutter ants, which use their powerful zinc-enriched growth, we conducted synchrotron X-ray PhotoEmission electron mandibles to defend their colonies’ territories against other, spectro-microscopy (X-PEEM), in vitro synthesis, and in vivo 31–33 encroaching ant species . observation of crystallization and growth in an ant-rearing Many species of fungus-growing ants are variably covered with experiment. We measure the cuticle hardness of Ac. echinatior a whitish granular coating, uniformly distributed on their ants with and without the cuticular layer using in situ nanoin- otherwise dark brown cuticles , including, in addition to Acro- dentation and explore two of several possible benefits associated myrmex echinatior (Fig. 1a), some species of Trachymyrmex and with the biomineral armor in experimental ant battles and Sericomyrmex. infections by entomopathogenic fungi. 2 NATURE COMMUNICATIONS | (2020) 11:5792 | https://doi.org/10.1038/s41467-020-19566-3 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19566-3 ARTICLE Results amino acid profiling of the protein layer of the Ac. echinatior Morphological, structural, and chemical characteristics of epicuticle confirms the presence of phenylalanine in the ant epicuticular minerals. Microscopic imaging of polished cuticular cuticle (Supplementary Table 4). cross-sections of the leaf-cutting ant Acromyrmex echinatior reveals a clear interface between the crystalline layer and the ant In vitro high-Mg calcite synthesis. To assess whether epicuti- cuticle (Fig. 1c). This layer is brighter than the cuticle in back- cular proteins mediate the precipitation of high-magnesium cal- scattered electron (BSE) mode scanning electron microscopy cite in Acromyrmex echinatior, we performed synthetic (SEM) (Fig. 1c, d), indicating that it consists of heavier elements biomineralization experiments in which the cuticle of Ac. echi- and that it is continuously distributed, covering nearly the entire natior was incubated in saturated carbonate solutions with a integumental surface. Energy-dispersive X-ray spectroscopy 2+ 2+ 40 [Mg ]/[Ca ] ratio of 5 at ambient conditions (Fig. 3a and (EDS) characterization of the cuticular coating further indicates Supplementary Fig. 6). In these in vitro experiments, nanocrystal that the crystalline layer contains significant amounts of mag- aggregates precipitated on the epicuticle of Ac. echinatior (Fig. 3b, nesium and calcium (Supplementary Fig. 1a–f), suggestive of an c), and were identified as anhydrous high-magnesium calcite by Mg-bearing calcite biomineral. XRD analysis confirms the high- XRD and EDS analyses (Fig. 3d and Supplementary Fig. 7). As a magnesium calcite composition of the biomineral layer in Ac. negative control, we performed the same in vitro mineralization echinatior, as indicated by the d-spacing of (104) peak at 2.939 Å experiments using the cuticle of the leaf-cutter ant Atta cepha- (Fig. 2a and Supplementary Table 1). Quantitative electron probe lotes, which belongs to the sister genus of Acromyrmex, does not micro-analysis (EPMA) reveals a magnesium concentration of have a biomineral cuticular layer, and has different cuticular 32.9 ± 2.7 mol% (Supplementary Table 2). Using bright-field structures (Supplementary Fig. 8). We found that only aragonite transmission electron microscopy (TEM), selected area electron crystals formed, mainly on the hairs of At. cephalotes and almost diffraction (SAED), TEM-EDS, and Raman and Attenuated Total never on the epicuticle (Supplementary Figures 9 and 10), indi- Reflectance Fourier-Transform InfraRed (ATR-FTIR) spectra, we cating direct precipitation from solution since aragonite is the further confirmed that the biomineral is high-magnesium calcite 41 favorable crystalline precipitate in high-Mg conditions .In with chemically heterogeneous crystals and with no observable control experiments using Ac. echinatior epicuticles either treated Ca–Mg ordering (i.e., no evidence for dolomite, the only Ca–Mg with KOH to hydrolyze surface proteins or coated with a 10 nm carbonate phase with cation ordering) (Supplementary Figs. 2 platinum layer to disable protein function, only aragonite crystals and 3). Extensive XRD analyses of Ac. echinatior, including both formed (Fig. 3d and Supplementary Fig. 7). Interestingly, in lab-reared and field-collected workers from Panama and Brazil, synthetic biomineralization experiments using cuticle from dif- confirms the consistent presence of high-magnesium calcite in ferent developmental stages (pupae to fully mature adult work- quantities of 23–35 mol% MgCO (Supplementary Tables 2 and ers), we found that only mature worker epicuticles catalyze the 3). To investigate the presence of amorphous calcium carbonate precipitation of high-magnesium calcite (Fig. 3e), consistent with within the biomineral cuticular layer, we heat-treated or irra- the presence of a more substantial protein layer in mature diated biomineral-bearing ants, and acquired XRD and TEM- workers indicated by SEM examination (Supplementary Figs. 11 SAED data before and after either treatment. We did not find and 12). These in vitro synthesis results, in particular the syn- evidence of increased crystallinity in either the heat or radiation thetic crystal morphology, are consistent with a solution con- treatments in either X-ray or electron diffraction. taining organic molecules that may have originated in the ant The mineral-cuticle interface of Acromyrmex echinatior was thorax exoskeletons, suggesting that the protein layer in the investigated using X-PEEM at the Advanced Light Source epicuticle of Ac. echinatior and the unusual morphological (Lawrence Berkeley National Laboratory, Berkeley, CA) structures on the cuticles of the ants catalyze the low-temperature (Fig. 2b) . Distinct X-ray absorption near-edge structure nucleation and growth of magnesium-rich calcite on the epicu- (XANES) spectra occur at the carbon K-edge for each of three ticles of mature workers of Ac. echinatior. regions: cuticle, epicuticle, and mineral layer (Fig. 2c; mapped as spectral components in Fig. 2d). The C spectra for the mineral In vivo crystallization and growth of high-Mg calcite.To layer show a strong carbonate peak at 290.3 eV (Fig. 2c). The explore the developmental timing of biomineral formation on the oxygen K-edge spectra extracted from the ant mineral layer epicuticles of Acromyrmex echinatior workers, we conducted indicate that the carbonate crystals are crystalline with a strong rearing experiments. Twenty pupae at the same developmental crystal orientation dependence of peak 1 at 534 eV (Fig. 2e). 36,37 stage were collected, randomly sorted into two groups of ten, and Polarization-dependent imaging contrast (PIC) mapping reared to callow adults (i.e., adults that have just emerged from across the mineral layer, in which color quantitatively displays the pupal stage), then one worker from each of the two groups the orientation of the crystal c-axes, indicates that crystals are was collected every second day and analyzed by XRD and eSEM randomly oriented (Fig. 2f, g; Fig. 2f magnifies images of Fig. 2g). (Fig. 3f, g). No biomineral layer was visible nor detected with The width of peak 2 in all O spectra (Fig. 2e) indicates a mixture 17 XRD on workers 0 to 6 days after eclosion (emergence of the of phases with high- and low-Mg concentrations . This chemical adult stage from the pupal stage). In contrast, 8 days after eclo- heterogeneity is consistent with the XRD data (Fig. 2a), with sion visible and XRD-detectable high-magnesium calcite was electron microprobe analyses (Supplementary Table 2), with present on workers. Magnesium was rapidly integrated into the backscatter diffraction (EBSD) results (Supplementary Fig. 4), calcareous biomineral in these older workers, with XRD mea- and with the magnified PIC map regions (Fig. 2f). surements of mol% MgCO reaching ~35% within 2 days after Unlike the typical chitin spectrum of insect epicuticle, the the initiation of biomineralization on individual worker ants (i.e., Acromyrmex echinatior XANES epicuticle spectrum is consistent from days 6 to 8 after eclosion; Fig. 3g). with a protein-enriched insect epicuticle . Protein hydrolysis of the cuticular layers verifies that the epicuticle is proteinaceous (Supplementary Fig. 5). Furthermore, the epicuticle spectrum Mechanical protection of epicuticular high-Mg calcite.Itis shows a very intense peak at 285.2 eV (Fig. 2c), which, based on plausible that epicuticular high-magnesium calcite enhances the its energy position and its symmetric line shape, suggests that the structural robustness of the ant exoskeleton, providing better epicuticle contains one or more phenylalanine (Phe) enriched defense for ants engaged in ‘wars’ with other ants or under attack proteins . High-performance liquid chromatography (HPLC) from predators or parasites. To test this hypothesis, we first NATURE COMMUNICATIONS | (2020) 11:5792 | https://doi.org/10.1038/s41467-020-19566-3 | www.nature.com/naturecommunications 3 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19566-3 Ac. echinatior Mineral Epicuticle Cuticle Dolomite High-Mg Calcite Calcite 10 15 20 25 30 2 µm 2 theta (°) d Cuticle Epicuticle Mineral Epicuticle C=O CO 2.5 C=C Mineral 2.0 Cuticle 1.5 1.0 285 290 295 300 Photon energy (eV) 2 µm f g Mineral crystal c-axis angle 0 0° ° –30° –30° +30° +30° 2 µm +90° +90° c’ angle 530 535 540 545 550 Photon energy (eV) –90° –60° –30° 0° 30° 60° 90° quantified the increase in hardness conferred by the protective has a greater than two-fold increase in hardness (1.55 ± 0.48 GPa, biomineral layer using in situ nanoindentation in an SEM compared to cuticle alone of 0.73 ± 0.04 GPa) (Fig. 4a and Sup- (Fig. 4a, Supplementary Fig. 13 and Supplementary Movie 1). plementary Figs. 14 and 15). Given that the biomineral layer has Since the surface of the exoskeleton is not flat, conventional an average thickness of 2.3 µm and that it overlays a cuticle with nanoindentation could not be used, whereas in situ nanoinden- an average thickness of 33.5 µm, this more than two-fold increase tation with real-time microscopic imaging allowed near- in hardness is conferred by only a 7% increase in cuticle thickness perpendicular contact of the probe tip with the surface (Supple- (Supplementary Fig. 16). These cuticle thickness values agree with mentary Fig. 13). Typical non-biomineralized ant cuticle, made those reported by Peeters et al. . Additional in situ nano- primarily of chitin, has a hardness of H ~0.73 ± 0.04 GPa (Fig. 4a mechanical testing of the cuticles of Atta cephalotes ants, which and Supplementary Fig. 14). In contrast, when high-magnesium do not have a biomineral layer, as well as of other common calcite and cuticular layers are combined, the composite structure insects, including a beetle (Xylotrechus colonus) and a honeybee 4 NATURE COMMUNICATIONS | (2020) 11:5792 | https://doi.org/10.1038/s41467-020-19566-3 | www.nature.com/naturecommunications Intensity (a.u.) Intensity (a.u.) Intensity (a.u.) (012) (104) d = 2.939 Å (006) (110) (113) d = 2.237 Å (202) (024) (108) (116) d = 1.554 Å d = 1.392 Å NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19566-3 ARTICLE Fig. 2 Chemical characterization of minerals on the cuticle of Ac. echinatior.a In situ XRD analysis identifying the cuticular crystalline layer as high-Mg calcite. b–g XANES spectroscopy and mapping with PEEM of a cuticular cross-section. b Average of PEEM images acquired across the C K-edge, showing crystalline layer tightly attached to the cuticle. Three distinct component spectra were identified in the regions labeled cuticle, epicuticle, and mineral, from the most internal part of the ant to the outer surface. c Normalized component spectra extracted from the corresponding labeled regions. Characteristic peaks are marked, including the 285.2 eV (C=C), 288.2 eV (C=O) and 290.3 eV (carbonate) peaks. d Component map where each pixel is colored according to the chemical components it contains. Black pixels are masked areas containing epoxy or gaps. Faint carbonate components within the cuticle and epicuticle were emphasized by enhancing the blue channel 5×, thus this is a semi-quantitative map. A fully quantitative RGB component map is presented in Supplementary Fig. 21. Individual maps of each component are presented in Supplementary Fig. 22, clearly showing an increasing gradientof carbonates towards the surface in the cuticle. e O K-edge spectra extracted from the mineral crystals correspondingly colored in the Polarization- dependent Imaging Contrast (PIC) maps in f and g. f Magnified PIC maps for the regions represented by boxes in the complete PIC map in g. g PIC map quantitatively displaying the orientations of the mineral crystals’ c-axes in colors. This map was acquired from the same area shown in b and d at precisely the same magnification. These are interspersed high- and low-Mg calcite, and heterogenous at the nanoscale. Biomineral crystals do not show preferred orientations but are randomly oriented. High-magnesium calcite in carbon spectra is identified by the carbonate peak at 290.3 eV, which occurs in all carbonates, amorphous, or crystalline. The O spectra in d clearly indicate crystallinity, and their line shape indicates a mixture of high-magnesium calcite and low Mg-bearing calcite. (Apis mellifera), produced similar hardness values in the range of and 6 days, respectively (Fig. 4e) (P = 0.05; two-sample t-test). 0.4–0.7 GPa (Fig. 4a and Supplementary Fig. 14) as they are all On day 6, all ants had succumbed to infection, and examination mainly made of chitin. The nano-mechanical measurements of workers without biominerals exposed to M. anisopliae revealed indicate that the biomineralized layer substantially hardens the substantial fungal growth and emergence (Fig. 4e, inset). In this exoskeleton of Ac. echinatior, consistent with the hypothesis that experiment, a reduced abundance of the antibiotic-producing the biomineral layer functions as protective armor. bacterial symbiont Pseudonocardia associated with the To further test the role of the biomineral as protective armor, biomineral-free workers could have contributed to the reduction we exposed Acromyrmex echinatior major workers with and in survival. without biomineral armor to Atta cephalotes soldiers in ant aggression experiments designed to mimic territorial ‘ant wars’ Discussion 31,43,44 that are a relatively common occurrence in nature .In Three independent lines of evidence indicate that epicuticular direct combat with the substantially larger and stronger At. biomineral crystals are ant-generated rather than adventitiously cephalotes soldier workers (average body length of 10.4 mm and a precipitated from the environment or generated by bacteria. First, head capsule width of 6.1 mm, compared to major Ac. echinatior in both C component maps (e.g., Fig. 2d) and PIC maps (e.g., body length of 6.4 mm and head capsule width of 2.9 mm) Fig. 2f) the magnesium-rich calcite crystals outside the epicuticle (Fig. 4b), ants with biomineralized cuticles lost significantly fewer are space-filling, a characteristic of biominerals formed by body parts (Fig. 4c and Supplementary Fig. 17) and had eukaryotes . Second, magnesium-rich calcite biominerals are significantly higher survival rates compared to biomineral-free spatially co-localized with epicuticle protein(s), which are likely ants (Fig. 4d, Supplementary Movies 2 and 3). Further, in direct involved in biomineral formation, consistent with the absence of aggression experiments in which biomineral-armored Ac. echi- biomineral formation in in vitro synthesis experiments in which natior workers were pitted against At. cephalotes soldiers, all of ant epicuticles were either coated with platinum or hydrolyzed. the At. cephalotes soldiers died, whereas only a few such deaths Third, the ant-rearing experiments were carried out in sterile, occurred when Atta soldiers were pitted against biomineral-free clean Petri dishes, eliminating the possibility of biominerals ants. SEM examination of biomineral-armored Ac. echinatior ants acquired from external sources. after combat with Atta cephalotes soldiers showed significantly The biota of the Ediacaran period (635–541 million years ago) less damage to their exoskeletons (Supplementary Fig. 18). included organisms of known and unknown phylogenetic affi- Notably, biomineral armor is present in mature major workers, nities that lived in oceans with a high ratio of magnesium to which forage outside of the nest, further indicating that calcium. Most were soft-bodied, but some possessed rudimentary epicuticular high-magnesium calcite is critical in a highly skeletons composed either of aragonite (a form of calcium car- competitive environment (Supplementary Figs. 19 and 20). These bonate) or, notably, of high-magnesium calcite . Around 550 results, taken together, are consistent with a role for epicuticular million years ago, coinciding with a shift in the Earth’s oceans to high-magnesium calcite as armor that defends workers from significantly lower magnesium-to-calcium ratios, metazoans with aggressive interactions with other ants, even though more ant strongly calcified internal and external skeletons appeared, species need to be further investigated. including the most familiar modern phyla. In spite of its Biomineral armor could also help protect ants from pathogens. strengthening properties, the enrichment of calcareous structures In a series of experiments, we focused on entomopathogenic with high concentrations of magnesium in Cambrian and modern fungi, which establish infection by penetrating the insect metazoans has until now remained only known from a very small exoskeleton and have significant impacts on survival. We exposed plate within the tooth of sea urchins. The ability of fungus- the propleural plates of Ac. echinatior major worker ants with and growing ants to facilitate the formation of magnesium-rich bio- without biomineralized exoskeletons to the spores of the minerals on their epicuticles is thus surprising. Further, given that entomopathogenic fungus Metarhizium anisopliae (Ascomycota, fungus-growing ants are among the most extensively studied Hypocreales). Compared to biomineral-free workers, major tropical insects, our finding raises the intriguing possibility that workers with biomineralized exoskeletons were significantly more high-magnesium calcite biomineralization may be more wide- resistant to infection. Specifically, we found that a majority of ants spread in insects than previously suspected, suggesting a pro- without biominerals died from infection within 4 days (1.0 ± 0.4 mising avenue for future research. and 0 ± 0 ants survived to 4 and 6 days, respectively), whereas an Fungus farming in ants originated ~60 million years ago in average of 2.2 ± 0.4 and 1.4 ± 0.5 (out of 3 individuals per sub- South America when a hunter-gatherer ancestor irreversibly colony over 5 sub-colonies) ants with biominerals survived to 4 committed to subsistence-scale cultivation of fungal crops for NATURE COMMUNICATIONS | (2020) 11:5792 | https://doi.org/10.1038/s41467-020-19566-3 | www.nature.com/naturecommunications 5 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19566-3 bc + 2+ - Mg2 / Ca / Cl + - / Na / HCO solution 19 °C, 7 days 1 µm 1 µm Ac. echinatior In vitro precipitated ant original cuticle carbonates d e Cuticle Original cuticle Old worker Precipitation on Young worker original cuticle A A A A A A Newly eclosed Precipitation on Pt worker Pt coated cuticle Pupa Precipitation on cuticle after hydrolysis A A New pupa A A A A A High-Mg High-Mg calcite calcite Calcite Calcite Aragonite Aragonite 6 10 15 20 25 6 10 15 20 25 2 theta (°) 2 theta (°) 1 µ µm 1 µm 1 µm 0 <3 h 2 d 4 d 6 d 8 d 10 d 12 d >30 d Post-eclosion 2 d 6 d 10 d Fig. 3 Mineral precipitation on the cuticle in both in vitro cuticle synthetic studies and ant-rearing experiments. a Scheme of in vitro mineralization experiment using Acromyrmex echinatior leaf-cutting ant cuticles as templates for biomineralization (Photo C.M.C.). b, c Pre- and post-incubation SEM 2+ 2+ − images showing the original, uncoated cuticle (b) and the cuticle covered by a layer of precipitated carbonate (c) after incubation in Mg /Ca /Cl / + − 2+ 2+ − Na /HCO solution for 7 days at 19 °C. d, XRD patterns of, from top to bottom, an uncoated ant cuticle, a cuticle after incubation in Mg /Ca /Cl / + − 2+ 2+ − + − Na /HCO solution, a platinum-coated cuticle incubated in Mg /Ca /Cl /Na /HCO solution, and a cuticle after KOH protein hydrolysis incubated 3 3 2+ 2+ − + − in Mg /Ca /Cl /Na /HCO solution. H: high-magnesium-calcite, A: aragonite, Pt: platinum. e XRD patterns of cuticles of ants representing different 2+ 2+ − + − developmental stages, ranging from (from bottom to top), a newly formed pupa to an older worker, after incubation in Mg /Ca /Cl /Na /HCO solution. f Environmental scanning electron micrographs (eSEM) of ant epicuticles taken over a 10-day time series, from immediately after eclosion from pupa to adult (left), to 10 days post-eclosion (right), showing the formation of the biomineral layer over time (Photo H.L.). g Estimated magnesium concentration of the biomineral layer during 30 days of ant development based on the XRD d value according to Graf and Goldsmith , showing the (104) rapid integration of magnesium from days 6 to 8 and the continued presence of high-magnesium content for up to 30 days (n = 2 per treatment and the corresponding standard error are shown). food . The transition to industrial-scale agriculture occurred ~20 importance of agriculture in driving the expansion of human million years ago with the origin of the ecologically dominant populations and the elaboration of human social systems . leaf-cutting ants, in which colony populations are orders of Further paralleling human agriculture, the fungal cultivars of the magnitude greater in size and in which physically distinct worker ants are highly susceptible to pathogens and the ants have castes enable the complex division of labor, paralleling the similar responded, in part, by evolving associations with antibiotic- 6 NATURE COMMUNICATIONS | (2020) 11:5792 | https://doi.org/10.1038/s41467-020-19566-3 | www.nature.com/naturecommunications (2.980 Å) (2.244 Å) Intensity (a.u.) (012) (012) (111) (021) (104) (104) H(104) (002) (006) (2.984 Å) (121) (006) (012) (102) (110) (110) (113) H(113) (211) (113) (2.253 Å) (220) (202) (202) (221) (024) (041) (108) (024) (202) (116) (108) (132) (116) (113) Ac. echinatior original cuticle Fomation Intensity (a.u.) stage Mol% MgCO (012) (012) (111) (021) (104) (104) H(104) (002) (006) (121) (2.975 Å) (006) In vitro precipitated (012) carbonates (102) (110) (110) (113) H(113) (211) (113) (2.259 Å) (220) (202) (202) (221) (024) (041) (108) (024) (202) (116) (108) (132) (116) (113) NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19566-3 ARTICLE 02 0.5 1 1.5 2.5 Hardness (GPa) 2 mm c d 100 ** ** At. cephalotes soldier At. cephalotes soldier 06 10 20 30 40 500 70 vs. mineral-free ants vs. mineral-present ants Time (min) +mineral 2 mm 2 mm - mineral 0246 135 Day producing bacteria to protect their crops . Early sedentary parallel with agriculture-driven human cultural evolution, human agricultural settlements represented rich resources that fungus-growing ants have evolved biomineralized armor that were highly susceptible to marauding bands of human raiders, serves, at least in part, both to protect them from other ants, leading to the development of multiple modes of defense, including other fungus-growing ants, in disputes over territory including specialized warrior castes, fortified cities, weapons, and and agro-predatory ants that are known to raid their colonies and protective armor . Here we show that, in another striking to consume their gardens and brood, and protect them from NATURE COMMUNICATIONS | (2020) 11:5792 | https://doi.org/10.1038/s41467-020-19566-3 | www.nature.com/naturecommunications 7 Ac. echinatior mineral+cuticle Ac. echinatior cuticle At. cephalotes soldier At. cephalotes worker Beetle (X. colonus) Honey bee (A. mellifera) Ant survival (%) Body parts lost (No.) Ant survival (%) ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19566-3 Fig. 4 Mechanical protection afforded by the epicuticular mineral layer. a Quantitative nano-mechanical properties of insect cuticles, including honeybee (Apis mellifera), beetle (Xylotrechus colonus), leaf-cutting ants [Atta cephalotes worker, Atta cephalotes soldier (purple), and Acromyrmex echinatior worker without biomineral (green, with green minus circle beside the ant image)] and Ac. echinatior ant worker with biomineral epicuticular layer (orange, with orange plus circle beside the ant image), measured by an in situ nanoindenters with a cube-corner probe (n = 12, 15, 13, 15, 12, and 13 for insect measured above, respectively; center, median; box, upper and lower quantiles; whisker, 1.5× interquartile range; points, outlier). Atta ants images, Xylotrechus beetle image, and Apis bee image provided with permission from the copyright holder, Alexander L. Wild, Jon Rapp, and Don Farrall, respectively. b–d Aggressive interaction between three Ac. echinatior workers (with/without biomineral, respectively) and Atta cephalotes soldier (Photo C.M.C.). b Ac. echinatior worker (left) aggressively interacts with Atta cephalotes soldier (right). c In aggressive encounters with Atta cephalotes soldiers, Ac. echinatior workers with biomineral armor (orange) lose substantially fewer body parts (i.e., legs, antennae, abdomen, and head) compared to Ac. echinatior worker without biomineral (green). d Survivorship of Ac. echinatior workers without (green) and with (orange) biomineral armor in aggressive encounters with Atta cephalotes soldiers (purple). Asterisks indicate significant differences via a two-sample t-test (*P < 0.05, **P < 0.001; P-value = 0.0184, 0.0001, and 0.0006 from left to right, respectively; n = 5 per treatment and the corresponding standard error are shown.). e Survivorship curves of Ac. echinatior worker with (orange) and without (green) an epicuticular biomineral layer exposed to the entomopathogenic fungus Metarhizium. The inset images show more substantial fungal growth and emergence from biomineral-free workers (Photo H.L.). c, e n = 3 per treatment, and the corresponding standard error are shown. disease organisms that might otherwise spread rapidly in their averaged spectrum was then normalized to the beamline I curve, acquired with precisely the same energy steps. densely populated colonies. The three spectra were then aligned between 280.0 and 283.7 eV. The cuticle and epicuticle spectra were shifted in energy so that the first peak was at 285.2 eV for Methods chitin and proteins, following Cody et al. , whereas the mineral spectra were shifted Ants. The fungus-farming ants used in this study were Acromyrmex echinatior and in energy so that the last peak, characteristic of carbonates, was at 290.3 eV, following 56,57 Atta cephalotes, originally collected in Costa Rica and Panama, and subsequently Madix and Stöhr . The cuticle spectrum is identical to that published by Cody maintained in the lab. These ant species co-occur in the same nesting areas in et al. 2011 obtained from scorpion cuticle and interpreted as chitin. The spectrum has Panama and other regions. a peak at 285.2 corresponding to C=C in aromatic carbon, a shoulder at ~287 eV, and a peak at 288.2 eV corresponding to C=O in chitin. The epicuticle spectrum shows the characteristics C=C of aromatic amino acids (tyrosine, tryptophan, and PhotoEmission electron microscopy (PEEM). Acromyrmex echinatior ants were phenylalanine) , a shoulder at 287.6 eV corresponding to C–H aliphatic carbon, and freeze-dried prior to PEEM sample preparation. The heads of the ants were then a sharp peak at 288.2 eV corresponding to the carboxyl group (C=O) in the peptide detached and embedded in Epofix epoxy (EMS, Hatfield, PA), ground with SiC bonds of all proteins. Compared to the spectra in tyrosine and tryptophan, the sandpapers, polished with Al O suspensions of 300 nm (MicroPolish II, Buehler, 2 3 phenylalanine spectrum has a more symmetric peak at 285.2 eV, allowing us to assign 8,49 Lake Bluff, IL) and 50 nm (Masterprep, Buehler, Lake Bluff, IL) particle sizes . this peak to phenylalanine in the epicuticle .The C=O occurs at the expected 288.2 22 g/L Na CO saturated solution was added regularly onto the pad during 2 3 eV . These normalized and averaged spectra were then adopted as component grinding and polishing to prevent carbonate dissolution, and the Al O suspen- 2 3 spectra, displayed in Fig. 2c, and used to obtain a component map in Fig. 2d. sions were also dialyzed against 22 g/L Na CO saturated solution . The samples 2 3 were re-embedded to fill as much as possible the interior of the ants and the gap between mineral and epoxy, and then the polishing procedures were repeated. Component mapping. The extracted, averaged, normalized, and aligned compo- After final polishing, the samples were rinsed with ethanol and gently wiped with nent spectra were made references by multiplying the I . Spectrum in each pixel of TexWipe Cotton (Texwipe, Kernersville, NC), air-dried, and coated with 1 nm Pt the stack was analyzed and best-fitted to a linear combination of the component on the areas to be analyzed and 40 nm Pt around it . references: cuticle, epicuticle, and mineral. The resulting component proportion For C K-edge spectra, PEEM stacks were acquired by scanning across 280–320 maps were exported as a gray level image and combined by the Merge Channel eV range with 0.1-eV step between 284 and 292 eV, and 0.5-eV step elsewhere, function in Adobe Photoshop, which became a fully quantitative RGB image resulting in 145 images per stack . For O K-edge spectra, PEEM stacks were (Supplementary Fig. 21). Individual component distribution maps were presented acquired by scanning across 525–555 eV range with 0.1-eV step between 530 and in Supplementary Fig. 22. For Fig. 2d, we enhanced the blue channel by adjusting 52,53 545 eV, and 0.5-eV step elsewhere, resulting in 181 images per stack . The the midtone value in levels five times greater than the other two channels, to images were stacked and processed with GG Macros in Igor Pro 6.37 . emphasize the presence of a mineral in the cuticle and epicuticle. For PIC mapping, a stack of 19 images was acquired by fixing the photon energy at the O K-edge π* peak (534 eV) and changing the X-ray polarization from 49,52,55 In vitro synthesis. All synthesis experiments were carried out in sealed plastic horizontal to vertical with a 5° step . Colored PIC maps were then produced 54 bottles at 19 °C for 7 d. Solutions were prepared by dissolving 50 mM MgCl ·6H O, 2 2 using Igor Pro 6.37 with GG Macros . 10 mM CaCl ·2H O, and 50 mM NaHCO with pH buffer to ~8.0 with NaOH to 2 2 3 simulate modern seawater chemistry. The solution was mixed for 20 min and then Masking the component map. The component map in Fig. 2d was masked using divided into 100 mL bottles with ant exoskeleton (Supplementary Fig. 6). Peptide an image of the same region acquired in SEM in backscattered electron (BSE) synthesis experiments with 1 mM, 5 mM, and 10 mM of phenylalanine peptide (H- mode. Unfortunately, in both the PEEM average image in Fig. 2b and in the BSE Phe-Phe-Phe-OH) (Bachem, CA) were mixed into solutions without ants. All image, the gray levels in the embedding epoxy and those in the cuticle are similar. vessels during the experiments have been washed with deionized water and pre- Therefore, there is no rigorous and quantitative method to select one but not the treated with 6 M hydrochloric acid to prevent carbonate contamination. Filters and other. We used Adobe Photoshop and the Magic Wand tool with a tolerance of 30 ants were air-dried for XRD and SEM characterization. to select all of the cuticles and deleted all those pixels from a black mask. The brighter mineral and all of the mineral debris deposited on the epoxy were then Scanning electron microscopy (SEM) and electron backscatter diffraction selected using the Magic Wand and a tolerance of 50 on the BSE image. These were (EBSD). Scanning electron microscopy (SEM) was done using a Hitachi S3400 at also deleted from the same black mask. The black pixels in the BSE image corre- 15 kV. Images were obtained in both variable pressure and vacuum mode. Energy- spond to gaps between the cuticle and the epoxy, or holes between mineral crystals, dispersive X-ray spectroscopy (EDS) and electron backscatter diffraction (EBSD) those black pixels were remained black in the black mask. The bright mineral was carried out using an AZtecOne system with silicon-drift detector from Oxford debris is presumably an artifact of polishing, as they appear both in PEEM and instruments. Samples were coated with 5 nm Pt coating. Phases used in EBSD are SEM images and are spectroscopically identified without a doubt as a mineral. constructed based on Mg-poor (a = 4.990 Å, c = 17.062 Å), Mg-medium (a = These were also removed from the mask and therefore displayed in Fig. 2d, as 4.920 Å, c = 16.656 Å; calculated), Mg-rich (a = 4.850 Å, c = 16.250 Å) regions. removing them would have been an artifact. The BSE image was warped to cor- However, given the EBSD is not particularly sensitive to unit-cell parameter dif- respond correctly to the PEEM image using Adobe Photoshop and specifically the ferences, chemical heterogeneity from EBSD is only qualitative. Puppet Warp tool. Obtaining component spectra. We extracted single-pixel spectra from the cuticle, Transmission electron microscopy (TEM). TEM measurements were carried out the epicuticle, and the mineral regions. These were identified as the only three using a Philips CM200UT TEM instrument operating at 200 kV acceleration vol- reliable components that were spectroscopically distinct from one another and not tage with 0.5 mm spherical aberration (Cs) and a point resolution of 0.19 nm. linear combinations of other components. The single-pixel spectra from the same Images and electron diffraction were collected with a CCD camera and analyzed material were extracted from each stack, aligned in energy, and averaged. The with Gatan DigitalMicrograph software. Samples that have been previously 8 NATURE COMMUNICATIONS | (2020) 11:5792 | https://doi.org/10.1038/s41467-020-19566-3 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19566-3 ARTICLE examined by XRD and SEM were rinsed with ethanol and DI water to remove glue (30–100% [vol/vol]), followed by critical point drying and coating with 1-nm residue. Samples were then crushed in an agar mortar, suspend in acetone, and platinum. Scanning electron microscopy (SEM) of samples was performed using a drop onto Lacey/ carbon 200 mesh copper grid. The composition of phases was LEO 1530 microscope to investigate the cuticular structure. confirmed with energy-dispersive X-ray spectroscopy (EDS) and analyzed with Thermo Noran software. High-performance liquid chromatography (HPLC) analysis. Dissected ant cuticle samples were placed in 100 µL 25% TFA containing 10 mM DTT and were In situ X-ray diffraction (XRD) analyzes. In situ X-ray diffraction (XRD) was hydrolyzed at 110 °C for 24 h. Hydrolyzed samples were then dried at 45 °C under a performed using a Rigaku Rapid II X-ray diffraction system with Mo Kα radiation. stream of nitrogen and resuspended in 50 mM HCl. Amino acids were then con- This XRD instrument uses a 2-D image-plate detector for signal collection and verted into respective fluorescent derivatives using o-phthalaldehyde (OPA) integrated using Rigaku’s 2DP software. XRD was run at 50 kV and a 100-µm (Agilent #5061-3335). Briefly, 5 µL sample aliquot was added to 20 µL of 40 mM diameter collimator. Whole fresh ant samples were glued onto an American potassium tetraborate buffer (pH 9.8) followed by the addition of 5 µL OPA, mixed Durafilm Kapton® tube with vacuum grease. Ant samples were then spin around gently and another 40 µL water was added. The mixture was filtered through a 0.45 phi and oscillate on omega. Synthesized powder samples were sealed in Kapton µm cellulose acetate 4-mm syringe filter (Nalgene #171-0045). Freshly prepared tube and run with fixed omega and phi spin. Refinements for phase percentage and samples were immediately subjected to HPLC analysis. unit-cell parameters were run using Jade 9.0 software with American Mineralogist Amino acids were analyzed using a modified method . In brief, the apparatus Crystal Structure Database (AMCSD) and the PDF-4+ database from the Inter- used was a custom-built dual analytical/semi-preparative Shimadzu system national Centre for Diffraction Data (ICDD). Disordered dolomite reference was consisting of a SIL-20AC autosampler, a CBM-20A system controller, two LC- constructed based on the unit-cell parameter of disordered dolomite with 50 mol.% 20AR pumps, a C50-20AC oven, a PDA S10-M20A detector, and a CPP-10Avp MgCO3 and powder diffraction pattern calculated by CrystalMaker built-in detector. Chromatographic separation of OPA-derivatized amino acids was CrystalDiffract software. performed using an Agilent ZORBAX Eclipse AAA column (4.6 mm × 150 mm × 3.5 µm; Agilent #963400-902) coupled with a ZORBAX Eclipse AAA Analytical Guard Column (4.6 mm × 12.5 mm × 5 µm; Agilent #820950-931) heated at 40 °C. Sectioning and transmission electron microscopy (TEM). Ants for sectioning The gradient elution was applied using 40 mM sodium phosphate dibasic buffer and transmission electron microscopy were fixed in cold 2% glutaraldehyde in Na- (pH 7.8) as solvent A and a mixture of acetonitrile, methanol, and water (45:45:10, cacodylate buffer. Postfixation was done in 2% osmium tetroxide and specimens v/v/v) as solvent B. HPLC-grade acetonitrile and methanol were supplied from were subsequently dehydrated in a graded acetone series. Specimens were Fisher Scientific and were used without further purification. The optimum embedded in Araldite and sectioned with a Reichert Ultracut E microtome. separation of amino acids was obtained using the following gradient program: 0% B Semithin 1-µm sections for light microscopy were stained with methylene blue and for 1.9 min, then increase to 57% B up to 28.10 min followed by an increase to thionin. Double-stained 70-nm thin sections were examined in a Zeiss EM900 100% B up to 38.60 min, then hold at 100% B till 47.30 min, and decrease to 22.3% electron microscope. B up to 48.20 min and down to 0% B till 60 min. The flow rate was 1 mL/min. Aliquots of 10 μL standards/samples were injected at 0.5 min and amino acids were Quantitative electron probe micro-analysis (EPMA). The carbonate EPMA data detected at the maximum wavelength of 338 nm with a 4 nm bandwidth. Retention were acquired with a CAMECA SXFive FE electron probe in the Cameca Electron times, as well as spectral information, was given by the PDA detector was used for Probe Lab in the Department of Geoscience at the University of Wisconsin-Madison. peak identification. Calibration curves of individual and mixed amino acids were Operating conditions were 7 kv and 10 nA (Faraday cup), using a focused beam. A prepared using either 250 pmol stocks of corresponding individual amino acids or low accelerating voltage was used to shrink the analytical volume to less than 300 nm. a 250 pmol amino acid standard mix (Agilent #5061-3331). Quantification was Peak counting time was 10 s, with background acquired for 10 s. Mg Ka was acquired performed using the calibration curves of the respective amino acid standards. with a TAP crystal and Ca Ka with an LPET crystal. The standard used as Delight Dolomite. Automation and data reduction utilized Probe for EPMA (Probe Software), Biomineral-free Ac. echinatior ant generation. Biomineral-free Ac. echinatior Carbon and oxygen were accounted for in a robust procedure in the Probe for EPMA ants were generated using a sub-colony setup. Sub-colonies were set up in small software: oxygen was calculated based upon stoichiometry to the measured Mg and (diameter 6 cm, height 4 cm) clear plastic containers. After sterilizing containers for Ca, with carbon calculated relative to that oxygen value (1:3), with this being iterated several times within the Armstrong/Love Scott matrix correction. The resulting values at least 20 min using UV light, cotton moistened with distilled water was placed at the bottom to help provide humidity. A small (width 4.12 cm, length 4.12 cm, and were then evaluated for actual accuracy, based upon two criteria: a non-normalized analytical total close to 100 wt% (~98–102 wt%), and for a formula basis of 3 oxygens, height 0.79 cm) weigh boat (Fisher catalog #08-732-112) was placed on top of the wet cotton, and then 0.1 g of the fungus garden, 2 minor workers, and a major the carbon formula value being close to 1.00 (~0.99–1.01). With these conditions met, worker pupa being reared to derive a biomineral-free adult (n = 10 sub-colonies). the determined compositions were deemed acceptable. A ∼1cm leaf fragment of pin oak (Quercus palustris) was added 24 h or more after pupa eclosion for the ants to cut and incorporate into the fungus garden. We Raman and attenuated total reflectance Fourier-transform infrared (ATR- monitored sub-colonies daily to record the eclosion date for the major worker pupa FTIR) spectroscopy. Ant mineral samples for Raman experiments were prepared until 14–21 days after eclosion. Then we performed environmental Scanning by bleaching the freeze-dried ant samples in 8.25% NaClO commercial bleach for Electron Microscopy (eSEM) and XRD on a subset of the ants to confirm the 24 h at room temperature to remove the exposed organic materials . Raman absence of the biomineral. Meanwhile, we established that Ac. echinatior ants could spectra were collected using a LabRam Raman microprobe (JY Horiba, Inc.) grow a biomineral layer normally in sub-colonies with the addition of two major equipped with a Microscope (Olympus DX41, ×50 and ×100 objectives) and a 633 worker adults (n = 10 sub-colonies) and other colony components were main- nm laser. Spectra were acquired with a CCD camera behind a spectrometer (the tained as above. accumulations and integration time varied). The ant carbonate powders were dropped on a microscope slide just before individual measurement. For ATR-FTIR, freeze-dried ant samples were used directly. ATR-FTIR data Nano-mechanical testing. Nanoindentation tests were carried out using a Bruker collection was conducted on a Perkin-Elmer 1720x spectrometer according to the Hysitron PI-85 SEM Picoindenter in a Zeiss Leo 1550VP SEM at the Wisconsin manufacturer’s instructions. Centers for Nanoscale Technology, UW-Madison. The samples were tested using a cube-corner probe with a basic quasistatic trapezoid load-controlled function, where the maximum load was 500 µN, the hold time was 2 s, and the loading/ Rearing experiments. In a total of 20 worker pupae of Ac. echinatior at the same unloading rate was 100 µN/s . SEM imaging was done in a high vacuum using an developmental stage and its ~10 g fungus garden were collected and randomly accelerating voltage of 3 kV with secondary electron mode. In order to simulate the sorted into two groups of ten and maintain them in chambers (diameter 6 cm, defense mechanism of the actual ant exoskeleton, we tested the combination of ant height 4 cm) with wet cotton. Reared to callow workers (around 3 h) and followed mineral and ant cuticle by indenting from the outside in, as illustrated in Sup- by one worker from each of two groups was collected every second day. The fresh plementary Fig. 13. The samples for testing the combination of ant mineral and ant ant samples were subjected to XRD analyses immediately and followed by an cuticle were prepared by slicing through the transverse plane of the head of the ant Environmental scanning electron microscope examination (eSEM). For eSEM, an to allow probing only on the flatter top part of the head. All samples were then FEI QUANTA 200 eSEM (FEI Company) was used. Ants were placed directly onto attached firmly to carbon tape on an SEM stub. The combination sample was the eSEM stub and examined without any preparation (i.e., samples were not fixed additionally pressed carefully with tweezers to ensure good attachment and flat- or coated for this analysis). All samples were analyzed at 5.0 torr, 3.0 spot size, ness. The indentation data and corresponding SEM video were analyzed not only and 4 °C. to ensure that the correct contact point of the probe with the surface was chosen, but also that there was no movement of the sample during indentation, as well as Ant sample preparation for SEM analyses of the cuticular structure. Worker that the load-displacement curve was smooth and there was no abrupt pop-out or ant cuticles from Ac. echinatior at different developmental stages (pupae to fully discontinuity. Based on these criteria, 13, 13, 10, and 12 valid data points were mature adult workers) and At. cephalotes were immediately fixed with 4% (vol/vol) selected for Ac. echinatior ant mineral plus cuticle and Ac. echinatior ant cuticle formaldehyde and 1% glutaraldehyde at 22 °C RT overnight. Samples were then cross-section, respectively. Measurements on other insect cuticles were also done washed with PBS and treated with 1% osmium tetroxide for 30 min at 22 °C. by indenting from the outside in following the same analyzing criteria, with which Samples were subsequently washed with a series of increasing ethanol dilutions 15, 13, 15, and 12 valid data points were selected for Atta soldier ant cuticle, Atta NATURE COMMUNICATIONS | (2020) 11:5792 | https://doi.org/10.1038/s41467-020-19566-3 | www.nature.com/naturecommunications 9 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19566-3 worker ant cuticle, beetle elytra, and honeybee cuticle, respectively. All the valid References points were then used to obtain the hardness values presented in Fig. 4a. 1. Knoll, A. H. Biomineralization and evolutionary history. Rev. Mineral. Additional hardness comparison of the mineral phases alone was done on polished Geochem. 54, 329–356 (2003). cross-sections of geologic dolomite and Ac. echinatior mineral. Cross-section 2. Eder, M., Amini, S. & Fratzl, P. Biological composites-complex structures for samples were prepared using the same embedding and polishing procedures as functional diversity. Science 362, 543–547 (2018). described for the PEEM samples. Representative load-depth curves in Supple- 3. Sperling, E. A. et al. Oxygen, ecology, and the Cambrian radiation of animals. mentary Fig. 14 were selected from the data point closest to the averaged hardness Proc. Natl Acad. Sci. USA 110, 13446–13451 (2013). value in each sample. 4. Gilbert, P. U. P. A. et al. Biomineralization by particle attachment in early animals. Proc. Natl Acad. Sci. USA 116, 17659–17665 (2019). 5. Gilbert, P. U. P. A., Abrecht, M. & Frazer, B. H. The organic-mineral interface Aggressive experiments between At. cephalotes soldier and Ac. echinatior in biominerals. Mol. Geomicrobiol. 59, 157–185 (2018). workers. We confronted one major At. cephalotes soldier and three mineral-pre- 6. Bengtson, S. Early Life on Earth. Nobel Symposium No. 84. (ed. Bengtson, S.) sent/mineral-free workers of Ac. echinatior . The experiment was replicated 5 412–425 (Columbia University Press, New York, 1994). times under reduced light in 9 cm Petri dishes. The survival of ants was counted 7. Sanderson, K. Artificial armour. Nature 519, S14–S15 (2015). after 3 h of confrontation. 8. Mass, T. et al. Amorphous calcium carbonate particles form coral skeletons. Using the time-lapse setting of an iPad Pro 2018, we recorded aggressive encounters between At. cephalotes soldier and Ac. echinatior workers Proc. Natl Acad. Sci. USA 114, E7670–E7678 (2017). (Supplementary Movies S2 and S3). For the aggressive experiments with mineral- 9. DeVol, R. T. et al. Nanoscale transforming mineral phases in fresh nacre. J. present ants, the video was started when At. cephalotes soldier was placed in the Am. Chem. Soc. 137, 13325–13333 (2015). petri dish with the Ac. echinatior ants. The video was stopped when the soldier ant 10. Weaver, J. C. et al. The stomatopod dactyl club: a formidable damage-tolerant was killed and the Ac. echinatior worker was able to separate from the soldier ant. biological hammer. Science 336, 1275–1280 (2012). The video is a total of 42 min filmed in time-lapse at 120 times its speed reducing 11. Politi, Y., Arad, T., Klein, E. & Addadi, L. Sea urchin spine calcite forms via a the video to 21 s in length, followed by editing it down to 10% of its speed using transient amorphous calcium carbonate phase. Science 306, 1161–1165 (2004). Premiere Pro resulting in a 3:27 min video. For the aggressive experiments with 12. Li, L. et al. Multifunctionality of chiton biomineralized armor with an mineral-free ants, the video was started when the soldier ant was placed into the integrated visual system. Science 350, 952–956 (2015). petri dish, video was stopped after all three Ac. echinatior ants were killed by the 13. Aizenberg, J., Tkachenko, A., Weiner, S., Addadi, L. & Hendler, G. Calcitic soldier ant. The video was recorded for 1 h in time-lapse at 120 times its speed microlenses as part of the photoreceptor system in brittlestars. Nature 412, reducing the video to 30 s in length, followed by editing down to 10% of its original 819 (2001). speed resulting in a video that is 5:00 min. The counting of lost body parts was 14. Killian, C. E. et al. Self-sharpening mechanism of the sea urchin tooth. Adv. based on the video record. Funct. Mater. 21, 682–690 (2011). 15. Ma, Y. et al. The grinding tip of the sea urchin tooth exhibits exquisite control Entomopathogenic fungi infection. Using two groups of Ac. echinatior major over calcite crystal orientation and Mg distribution. Proc. Natl Acad. Sci. USA worker ants in which the mineral was either present or absent, we conducted 106, 6048–6053 (2009). infection experiments using the entomopathogenic fungus Metarhizium aniso- 16. Wang, R. Z., Addadi, L. & Weiner, S. Design strategies of sea urchin teeth: pliae (Ascomycota, Hypocreales) . In brief, groups of three ant individuals were Structure, composition and micromechanical relations to function. Philos. placed into an individual Petri dish with a ring of moist cotton. Then, each ant’s Trans. R. Soc. B Biol. Sci. 352, 469–480 (1997). propleural plate was inoculated with 0.5 μLof Metarhizium spores of ca. 1.00 × 17. Polishchuk, I. et al. Coherently aligned nanoparticles within a biogenic single 7 −1 10 conidiospores mL suspension + 0.01% Tween 20 by using a micropipette crystal: a biological prestressing strategy. Science 358, 1294–1298 (2017). under a dissecting microscope. The experiment was replicated five times. The 18. Morse, J. W., Wang, Q. & Tsio, M. Y. Influences of temperature and Mg:Ca control for each replicate consisted of inoculating with a control solution of ratio on CaCO precipitates from seawater. Geology 25,85–87 (1997). sterile, deionized water + 0.01% Tween 20 . The survival of ants (three indi- 19. Oakley, T. H., Wolfe, J. M., Lindgren, A. R. & Zaharoff, A. K. viduals per sub-colony over five sub-colonies) was monitored every 24 h post- Phylotranscriptomics to bring the understudied into the fold: monophyletic treatment for 6 days. Ostracoda, fossil placement, and pancrustacean phylogeny. Mol. Biol. Evol. 30, The Metarhizium anisopliae var. anisopliae strain was isolated from dead fungus- 215–233 (2013). growing ants in Gamboa, Panama, and grown on a pure medium of potato dextrose 20. Schultz, T. R. & Brady, S. G. Major evolutionary transitions in ant agriculture. agar at room temperature until a full plate of spores was observed after around 7 days. Proc. Natl Acad. Sci. USA 105, 5435–5440 (2008). The spores (conidia) were centrifuged, washed three times with sterile 0.05% Triton-X 21. Li, H. et al. Convergent evolution of complex structures for ant-bacterial solution, and then harvested fresh for each infection experiment. For the infection defensive symbiosis in fungus-farming ants. Proc. Natl Acad. Sci. USA 115, experiments, spore suspensions were set up from recently sporulating cultures in a 10720–10725 (2018). solution of sterile-deionized water containing 0.01% Tween 20. Spore concentration 22. Pinto-Tomás, A. A. et al. Symbiotic nitrogen fixation in the fungus gardens of was quantified using a hemocytometer and diluted to achieve a concentration of ca. leaf-cutter ants. Science 326, 1120–1123 (2009). 7 −1 1.00 × 10 conidiospores mL . 23. Weber, N. A. Gardening Ants: The Attines (American Philosophical Society, Philadelphia, 1972). Reporting summary. Further information on research design is available in the Nature 24. Wetterer, J. K. Forager size and ecology of Acromyrmex coronatus and other Research Reporting Summary linked to this article. leaf-cutting ants in Costa Rica. Oecologia 104, 409–415 (1995). 25. Bekkevold, D., Frydenberg, J. & Boomsma, J. Multiple mating and facultative polygyny in the Panamanian leafcutter ant Acromyrmex echinatior. Behav. Data availability Ecol. Sociobiol. 46, 103–109 (1999). Source data files are provided on https://figshare.com/s/29b90a59a699bb33bf32. Each 26. Ferguson-Gow, H., Sumner, S., Bourke, A. F. G. & Jones, K. E. Colony data set is provided as an Igor Pro experiment with file extension.pxp. All the relevant size predicts division of labour in attine ants. Proc. R. Soc. B 281, 20141411 files for C and O spectra are zipped separately. Permits for collections and accessing (2014). genetic resources in Brazil were issued by SISBIO #46555–5 and CNPq #010936/2014–9. 27. Currie, C. R., Poulsen, M., Mendenhall, J., Boomsma, J. J. & Billen, J. Costa Rican collecting permits were issued by the Comisión Institucional de Coevolved crypts and exocrine glands support mutualistic bacteria in fungus- Biodiversidad (Institutional Biodiversity Committee, University of Costa Rica; growing ants. Science 311,81–83 (2006). Resolutions # 012 and 020; Material Transfer Agreement MTA VI-4307–2013) and 28. Currie, C. R., Scott, J. A., Summerbell, R. C. & Malloch, D. Fungus-growing authorized by La Selva Biological Station. Photos of Atta ants image and Apis bee image ants use antibiotic-producing bacteria to control garden parasites. Nature 398, are used under a perpetual commercial license from Alexander L. Wild and Getty 701–704 (1999). Images, respectively (Fig. 4). Xylotrechus beetle image provided with permission from the 29. Currie, C. R., Mueller, U. G. & Malloch, D. The agricultural pathology of ant copyright holder, Jon Rapp (Fig. 4). fungus gardens. Proc. Natl Acad. Sci. USA 96, 7998–8002 (1999). 30. Adams, R. M. M., Mueller, U. G., Schultz, T. R. & Norden, B. Agro-predation: Code availability usurpation of attine fungus gardens by Megalomyrmex ants. The Igor Pro macros, called GG Macros, used to produce PIC maps are available free of Naturwissenschaften 87, 549–554 (2000). charge on https://home.physics.wisc.edu/gilbert/software/. The code to measure the 31. Fowler, H. G. Field response of Acromyrmex crassispinus (forel) to aggression angular distances of c-axes in Fig. 2 is available on https://home.physics.wisc.edu/gilbert/ by Atta sexdens (linn.) and predation by Labidus praedator (fr. smith) software/. (Hymenoptera: Formicidae). Aggress. Behav. 3, 385–391 (1977). 32. Vilela, E. F. & Howse, P. E. Fire Ants and Leaf-Cutting Ants. Biology and Management (eds Lofgren, C. S. & Vander Meer, R. K.) 159–171 (Westview Received: 16 March 2020; Accepted: 15 October 2020; Press. Boulder, CO, 1986). 10 NATURE COMMUNICATIONS | (2020) 11:5792 | https://doi.org/10.1038/s41467-020-19566-3 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-19566-3 ARTICLE 33. Schofield, R. M., Nesson, M. H. & Richardson, K. A. Tooth hardness increases 62. Wang, C. & Wang, S. Insect pathogenic fungi: genomics, molecular with zinc-content in mandibles of young adult leaf-cutter ants. interactions, and genetic improvements. Annu. Rev. Entomol. 62,73–90 Naturwissenschaften 89, 579–583 (2002). (2017). 34. Weber, N. A. Gardening Ants: The Attines (American Philosophical Society, 63. Wang, B., Kang, Q., Lu, Y., Bai, L. & Wang, C. Unveiling the biosynthetic Philadelphia, 1972). puzzle of destruxins in Metarhizium species. Proc. Natl Acad. Sci. USA 109, 35. Myers, C. E. et al. Exceptional preservation of organic matrix and shell 1287–1292 (2012). microstructure in a Late Cretaceous Pinna fossil revealed by photoemission 64. Graf, D. L. & Goldsmith, J. R. Some hydrothermal syntheses of dolomite and electron spectromicroscopy. Geology 46, 711–714 (2018). protodolomite. J. Geol. 64, 173–186 (1956). 36. Gilbert, P. U. P. A., Young, A. & Coppersmith, S. N. Measurement of c-axis angular orientation in calcite (CaCO ) nanocrystals using X-ray absorption Acknowledgements spectroscopy. Proc. Natl Acad. Sci. USA 108, 11350–11355 (2011). We thank U. P. Agarwal and S. A. Ralph from U.S. Forest Products Laboratory for the 37. Metzler, R. A. et al. Polarization-dependent imaging contrast in abalone shells. Raman spectroscopy measurements; B. Schneider and R. Noll for expert help conducting Phys. Rev. B 77, 064110-1/9 (2008). SEM work; J. Morasch for assistance with nanoindentation measurements; E. Okonski 38. Cody, G. D. et al. Molecular signature of chitin-protein complex in Paleozoic for ant imaging and laboratory assistance; and R. J. Massey for assistance with the arthropods. Geology 39, 255–258 (2011). microtome. This work was primarily supported by the National Institutes of Health 39. Kaznacheyev, K. et al. Innershell absorption spectroscopy of amino acids. J. (NIH) Grant U19 TW009872-05, NIH Grant U19 AI109673 and the Department of Phys. Chem. A 106, 3153–3168 (2002). Energy Great Lakes Bioenergy Research Center Office of Science Grant DE-FC02- 40. Zhang, F., Xu, H., Konishi, H., Shelobolina, E. S. & Roden, E. E. 07ER64494 to C.R.C. P.U.P.A.G. acknowledges support from the U.S. Department of Polysaccharide-catalyzed nucleation and growth of disordered dolomite: a Energy, Office of Science, Office of Basic Energy Sciences, Chemical Sciences, Geos- potential precursor of sedimentary dolomite. Am. Mineral. 97, 556–567 ciences, and Biosciences Division, under Award DE-FG02-07ER15899, and NSF grant (2012). DMR-1603192. PEEM experiments were done at the Advanced Light Source, which is a 41. Lippmann, F. Sedimentary Carbonate Minerals (Springer, New York, 1973). DOE Office of Science User Facility supported by grant DE-AC02-05CH11231. T.R.S. is 42. Peeters, C., Molet, M., Lin, C. C. & Billen, J. Evolution of cheaper workers in supported by the National Science Foundation award DEB 1654829 and DEB 1927224. ants: a comparative study of exoskeleton thickness. Biol. J. Linn. Soc. 121, H.X. and Y.F. is supported by the NASA Astrobiology Institute (NNA13AA94A) and 556–563 (2017). S. W. Bailey Scholarship of the Department of Geoscience. 43. Hölldobler, B. & Wilson, E. O. The Superorganism: The Beauty, Elegance, and Atrangeness of Insect Societies (WW Norton, New York, 2008). 44. Dijkstra, M. B. & Boomsma, J. J. Gnamptogenys hartmani Wheeler Author contributions (Ponerinae: Ectatommini): an agro-predator of Trachymyrmex and Study design: H.L., C.-Y.S., P.U.P.A.G., and C.R.C. Experimental design and supervision: Sericomyrmex fungus-growing ants. Naturwissenschaften 90, 568–571 (2003). H.L., C.-Y.S., T.R.S., P.U.P.A.G., and C.R.C. Data collection and analysis: H.L., C.-Y.S., 45. Yang, L., Killian, C. E., Kunz, M., Tamura, N. & Gilbert, P. U. P. A. Biomineral Y.F., C.M.C., H.X., A.J., J.S.-C., R.Z., H.A.B., J.H.F., D.R.A., T.R.S., P.U.P.A.G., and C.R.C. nanoparticles are space-filling. Nanoscale 3, 603–609 (2011). Initial draft: H.L., C.R.C., T.R.S. C.-Y.S., and P.U.P.A.G. Final version: all authors. 46. Wood, R., Ivantsov, A. Y. & Zhuravlev, A. Y. First macrobiota biomineralization was environmentally triggered. Proc. R. Soc. B 284, 20170059 (2017). Competing interests The authors declare no competing interests. 47. Nygaard, S. et al. Reciprocal genomic evolution in the ant-fungus agricultural symbiosis. Nat. Commun. 7,1–9 (2016). 48. Diamond, J. M. & Ordunio, D. Guns, Germs, and Steel (Books on Tape, 1999). Additional information 49. Sun, C. Y. et al. Spherulitic growth of coral skeletons and synthetic Supplementary information is available for this paper at https://doi.org/10.1038/s41467- aragonite: nature’s three-dimensional printing. ACS Nano 11, 6612–6622 020-19566-3. (2017). 50. Gong, Y. U. et al. Phase transitions in biogenic amorphous calcium carbonate. Correspondence and requests for materials should be addressed to P.U.P.A.G. or C.R.C. Proc. Natl Acad. Sci. USA 109, 6088–6093 (2012). 51. De Stasio, G., Frazer, B. H., Gilbert, B., Richter, K. L. & Valley, J. W. Peer review information Nature Communications thanks Charissa de Bekker and the Compensation of charging in X-PEEM: a successful test on mineral inclusions other, anonymous, reviewers for their contribution to the peer review of this work. Peer in 4.4Ga old zircon. Ultramicroscopy 98,57–62 (2003). reviewer reports are available. 52. DeVol, R. T. et al. Oxygen spectroscopy and polarization-dependent imaging contrast (PIC)-mapping of calcium carbonate minerals and biominerals. J. Reprints and permission information is available at http://www.nature.com/reprints Phys. Chem. B 118, 8449–8457 (2014). 53. Zou, Z. et al. A hydrated crystalline calcium carbonate phase: Calcium Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in carbonate hemihydrate. Science 363, 396–400 (2019). published maps and institutional affiliations. 54. Macros, G. G. http://home.physics.wisc.edu/gilbert/software.htm (2019). 55. Pokroy, B. et al. Narrowly distributed crystal orientation in biomineral vaterite. Chem. Mater. 27, 6516–6523 (2015). 56. Stohr, J. NEXAFS Spectroscopy (Springer-Verlag, Berlin, 1992). Open Access This article is licensed under a Creative Commons 57. Madix, R. J., Solomon, J. L. & Stohr, J. The orientation of the carbonate anion Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give on Ag(110). Surf. Sci. 197, L253–L259 (1988). 58. Henderson, J. W. Jr., Ricker, R. D., Bidlingmeyer, B. A. & Woodward, C. appropriate credit to the original author(s) and the source, provide a link to the Creative Rapid, Accurate, Sensitive and Reproducible HPLC Analysis of Amino Acids Commons license, and indicate if changes were made. The images or other third party (Agilent Pub.# 5980–1193E, 2000). material in this article are included in the article’s Creative Commons license, unless 59. Amini, S., Tadayon, M., Idapalapati, S. & Miserez, A. The role of quasi- indicated otherwise in a credit line to the material. If material is not included in the plasticity in the extreme contact damage tolerance of the stomatopod dactyl article’s Creative Commons license and your intended use is not permitted by statutory club. Nat. Mater. 14, 943–950 (2015). regulation or exceeds the permitted use, you will need to obtain permission directly from 60. Oliver, W. C. & Pharr, G. M. An improved technique for determining the copyright holder. To view a copy of this license, visit http://creativecommons.org/ hardness and elastic modulus using load and displacement sensing licenses/by/4.0/. indentation experiments. J. Mater. Res. 7, 1564–1583 (1992). 61. Hölldobler, B. & Wilson, E. O. The Ants (Harvard Univ. Press, Cambridge, © The Author(s) 2020 MA, 1990). NATURE COMMUNICATIONS | (2020) 11:5792 | https://doi.org/10.1038/s41467-020-19566-3 | www.nature.com/naturecommunications 11

Journal

Nature CommunicationsSpringer Journals

Published: Dec 1, 2020

There are no references for this article.