Serpentine soils affect heavy metal tolerance but not genetic diversity in a common Mediterranean ant

Chemosphere 180 (2017) 326e334

Contents lists available at ScienceDirect

Chemosphere journal homepage: www.elsevier.com/locate/chemosphere

Serpentine soils affect heavy metal tolerance but not genetic diversity in a common Mediterranean ant Filippo Frizzi a, *, Alberto Masoni a, Mine Çelikkol b, Enrico Palchetti c, Claudio Ciofi a, Guido Chelazzi a, Giacomo Santini a
degli Studi di Firenze, Dipartimento di Biologia, Via Madonna del Piano, 6, 50019, Sesto Fiorentino, Florence, Italy Universita _ _ Istanbul University, Institute of Science, Department of Biology, PK 34134, Vezneciler, Fatih, Istanbul, Turkey c
degli Studi di Firenze, Dipartimento di Scienze delle Produzioni Agroalimentari e dell’Ambiente (DISPAA), Piazzale delle Cascine, 18, 50144, Universita Florence, Italy a

b

h i g h l i g h t s
Ants in serpentine soils are affected by the presence of heavy metals.
Here we test the effect of exposure to nickel on the ant Crematogaster scutellaris.
Chronic exposure to intermediate metal levels increases ants’ tolerance to nickel.
Microsatellites analysis reveals no detectable effect on genetic diversity.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 18 January 2017 Received in revised form 3 April 2017 Accepted 6 April 2017 Available online 10 April 2017

Natural habitats with serpentine soils are rich in heavy metal ions, which may significantly affect ecological communities. Exposure to metal pollutants results, for instance, in a reduction of population genetic diversity and a diffused higher tolerance towards heavy metals. In this study, we investigated whether chronic exposure to metals in serpentine soils affect accumulation patterns, tolerance towards metal pollutants, and genetic diversity in ants. In particular, we studied colonies of the common Mediterranean ant, Crematogaster scutellaris, along a contamination gradient consisting of two differently contaminated forests and a reference soil with no geogenic contamination. We first evaluated the metal content in both soil and ants’ body. Then, we tested for tolerance towards metal pollutants by evaluating the mortality of ants fed with nickel (Ni) solutions of increasing concentrations. Finally, differences in genetic diversity among ants from different areas were assessed using eight microsatellite loci. Interestingly, a higher tolerance to nickel solutions was found in ants sampled in sites with intermediate levels of heavy metals. This may occur, because ants inhabiting strongly contaminated areas tend to accumulate higher amounts of contaminants. Additional ingestion of toxicants beyond the saturation threshold would lead to death. There was no difference in the genetic diversity among ant colonies sampled in different sites. This was probably the result of queen mediated gene flow during nuptial flights across uncontaminated and contaminated areas of limited geographical extent. © 2017 Elsevier Ltd. All rights reserved.

Handling Editor: A. Gies Keywords: Serpentine soils Crematogaster scutellaris Heavy metals Genetic diversity Metal tolerance

1. Introduction Metal pollution due to anthropogenic activity is a widely documented phenomenon of environmental and medical relevance €rstner and worldwide (Duruibe et al., 2007; Agarwal, 2009; Fo

* Corresponding author. E-mail address: filippo.frizzi@unifi.it (F. Frizzi). http://dx.doi.org/10.1016/j.chemosphere.2017.04.026 0045-6535/© 2017 Elsevier Ltd. All rights reserved.

Wittmann, 2012). High metal concentrations also occur naturally in areas with specific rock compositions, such as serpentine soils. These terranes originate from ultramafic rocks and are characterized by low nutrient content and high nickel, cobalt, and chrome concentrations (Brooks, 1987; Angelone et al., 1993). Not surprisingly, they harbor particular and often highly specialized biological communities (Prasad, 2001). For example, nickel hyperaccumulator plants and high-nickel herbivorous insects are well-known examples of extreme adaptations to these harsh environments (e.g.

F. Frizzi et al. / Chemosphere 180 (2017) 326e334

Brooks, 1987; Reeves and Baker, 2000; Boyd, 2009). Additionally, arthropod predators such as spiders or mantids that feed in these areas accumulate considerable amounts of nickel in their tissues with different responses to heavy metal poisoning (Boyd and Wall, 2001). In the absence of compensative systems limiting ion absorption, the effects of heavy metal toxicity can be severe and may ultimately have behavioral, functional, or genotoxic consequences (e.g. nchez, 2008; Agarwal, 2009; Boyd, Reinecke and Reinecke, 2004; Sa 2010; Grzes et al., 2015). Avoidance of polluted resources can be an efficient method for preventing poisoning (Tyler et al., 1989; Lefcort et al., 2004). However, when this is not possible other physiological mechanisms may allow for improved tolerance to metal exposure (Morgan et al., 2007; Janssens et al., 2009). These mechanisms may include, for example, overexpression of genes related to metallothionein production (Amiard-Triquet et al., 2011; Isani and
, 2014) or the ability to store metals in a non-soluble Carpene form within intra-cytoplasmic membrane-bound granules (Jeantet et al., 1977; Carneiro et al., 2013). Moreover, prolonged exposure to contaminated soils may work as a selective drive through the survival of tolerant genotypes (e.g. Klerks and Weis, 1987; Shaw, 1989; Maes et al., 2005). It is also known that in many taxa, from plants to crustaceans and fishes, populations living in contaminated areas show a lower genetic diversity than populations from uncontaminated sites (Nadig et al., 1998; Mengoni et al., 2000; Ross et al., 2002; Fratini et al., 2008). Ants are found in virtually every habitat type except permanently frozen soils and some oceanic island’s forests, such as the Azores (Ribeiro et al., 2005; Bolton et al., 2006). Due to their key role in ecosystem functioning, high abundance, ease of identification, and the spatial stability of colonies, ants are widely used as bioindicators to assess biological effects of degraded or polluted habitats (e.g. Majer, 1983; Madden and Fox, 1997; Hoffmann et al., 2000; Ottonetti et al., 2006; Nummelin et al., 2007; Ribas et al., 2012). Some species exhibit considerable resistance to metal pollution and maintain viable populations even in contaminated areas (Migula and Głowacka, 1996; Eeva et al., 2004). However, different species show different responses to metal contamination exposure. For example, in Myrmica rubra, Zinc tolerance is higher for ants living in polluted areas, suggesting an adaptation to heavy metal exposure (Grzes, 2010a). Conversely, in Formica aquilonia there is no evidence for adaptation to metal pollution, implying damage occurs to the immune systems of exposed individuals (Sorvari et al., 2007). Notwithstanding that the accumulation of metals in ants is well documented (e.g. Rabitsch, 1995, 1997; Del Toro et al., 2010; Gramigni et al., 2013), the long-term consequences of prolonged exposure are still poorly known (Grzes, 2010b), and to our knowledge, there is no data on the adaptation of ants to ophiolitic soils. In this study, we analyzed the effects of chronic exposure to geogenic heavy metal contamination due to serpentine soils on populations of the acrobat ant Crematogaster scutellaris, a dominant Myrmicine species found widespread throughout the Mediterranean basin. This species has been previously used to evaluate metal pollution in urban areas (Gramigni et al., 2013), and it is known that it may selectively accumulate some metallic ions in different body parts and tissues (Gramigni et al., 2011). This species is a top-ranked competitor found in both natural, semi-natural, and urban areas (Morris et al., 1998; Schatz and Hossaert-Mckey, 2003; Santini et al., 2011). Similarly to many other ants, it has a variety of feeding strategies including preying, scavenging, and homopteran tending (Ottonetti et al., 2008). The wide resource spectrum used by C. scutellaris may facilitate the uptake of metal ions (Gramigni et al., 2013). We sampled ants from three sites across a gradient of

327

contaminated soils: an uncontaminated oak-pine forest used as a reference site, an oak-pine forest located at the margin of an ophiolite outcrop, and a pine forest in the middle of another ophiolite outcrop. We first quantified the metal content both in the soil and in the ants in order to assess which elements were accumulated or metabolically regulated by ants. Second, we performed mortality experiments to test for evidence of some form of adaptation to chronic contamination. This was conducted by feeding ants with nickel-contaminated food. Finally, we genotyped ants using nuclear DNA microsatellite markers to assess whether longterm exposure to metals resulted in a reduced genetic diversity. In particular, we expected that ants from contaminated sites showed a lower mortality and genetic diversity with respect to ants from uncontaminated areas. 2. Materials and methods The study was carried out between May 2013 and September 2014 in late spring and summer. Ants were sampled in three sites in Tuscany, Italy: the Casaglia oak-pine forest near the town of Florence (CS, the reference site), an oak-pine forest located at the margins of the Monterufoli Nature Reserve ophiolite outcrop near Pomarance (PO, the intermediate contaminated area), and a pine forest located over the ophiolitic Monteferrato Protected Area outcrop, near Prato (MF, the most contaminated area). The reference site is far from other sources of human induced pollution (urban habitats, heavy traffic roads) and has no geogenic contamination. None of the three areas is actively managed for economic purposes. Boundaries for the ophiolite outcrops were defined from the Geological Map of Tuscany (http://www502.regione.toscana.it/ geoscopio/cartoteca.html, accessed on March 29, 2017). For each site, ants were sampled from six C. scutellaris nests located at least 50 m apart. Each sample to be used for subsequent metal content analysis was formed by a total of 100 ants from six out of the ten nests. Ants were immediately taken to the laboratory and kept without food for three days to allow gastric emptying. Water was available to ants during this time. Ants were then stored at 80  C. Soil samples were randomly sampled within a 5 m radius from each nest. Approximately 3 cm of soil layer were first removed in order to discard temporary atmospheric depositions (Carter, 1993) and 100 cm3 of soil were then sampled and stored in a plastic container. In summary, we collected a total of 18 ant (six replicate nests x three sites) and 18 soil samples. The metal content in ants and soil was assessed in mg/kg by an Inductively Coupled Plasma Optical Emission Spectrometer (ICPOES) analyzer (IRIS Intrepid II XSP Radial, Thermo Fisher Scientific, USA). Analyses were conducted on six of the most important elements found in ophiolitic soils: Cd, Co, Cr, Cu, Ni, and Zn. Ants were first cleaned with double distilled water and then dried at 60  C in an oven for 48 h. Soil samples were purified from debris and visible organic material, such as roots or invertebrates, dried at 60  C in an oven for 48 h, and then ground using a powder mill (Pulverisette 6, Fritsch GmbH). Approximately 0.5 g of pulverized soil and 0.1 g of dried ants were diluted in 10 ml and 5 ml nitric acid, respectively. Each sample was mineralized using a MARS microwave reaction system (CEM Corporation, Matthews, NC) according to the protocol provided by the manufacturer. The pH of soil was also measured. For each metal and site, we computed a biota-to-soil accumulation factor (BSAF, Cortet et al., 1999), which is the ratio of metal in ant tissues to that in the soil. Metal concentrations assessed by the ICPOES and BSAF values were log-transformed prior to analysis due to the wide difference in the relative abundance of metals. Differences in the metal contents (ants and soil) and BSAF among sites were assessed using ANOVA and the Tukey post-hoc test for multiple comparisons. We employed non-metric multidimensional scaling

328

F. Frizzi et al. / Chemosphere 180 (2017) 326e334

(nMDS) using a Bray-Curtis dissimilarity matrix (Clarke and Warwick, 2001) as an input to investigate multivariate differences among sites. We also used non-parametric Multivariate Analysis of Variance (npMANOVA) to test differences in metal content among sites. Both nMDS and npMANOVA analyses were performed using the ‘vegan’ R 3.3.1 package (Oksanen et al., 2013). Finally, we ran a generalized linear mixed model (GLMM), using sampling site as a random factor to assess differences in the amounts of each metal in the soil and in the ants, respectively. Models were fitted using ‘lme4’ R 3.3.1 package (Bates et al., 2015). 2.1. Mortality tests Five out of the ten nests were randomly chosen in each of the three study areas, and at least 50 ants were sampled from each nest. Ants were stored in 50 ml tubes with water available to prevent dehydration and immediately taken to the laboratory, where they were separated into groups of ten ants each. Twenty groups per site were selected for subsequent tests. Ants were left to acclimatize to experimental conditions (25  C, 75% relative humidity) for 24 h. Four nickel (Ni) concentrations were used in a 10% sucrose solution: control (sucrose solution without pollutant), 2, 4, and 10 mM Ni. Each group was placed in a plastic Petri dish with Fluon® coated walls to prevent ants from escaping. Petri dishes were prepared each with a 1.5 ml tube containing a different sucrose solution. Tubes were refilled when necessary to ensure a continuous solution supply and to prevent mildew formation. The dishes were kept in a thermostatic chamber at 25  C and 75% relative humidity. The number of ants that were alive was recorded each day at 10:00 am. Observations continued until all ants had died out or up to a maximum of 35 days. In order to ensure blindness in the experiment, each dish was labeled with a numerical code and the Ni solution concentration was not known to the observer (van Wilgenburg and Elgar, 2013). The number of survivors was fitted using Kaplan-Meier estimator curves for survival data (Bland and Altman, 1998). Differences in survival rate associated with different test solutions were assessed using Cox proportional hazards regression models with Efron approximation in order to handle time ties (Andersen and Gill, 1982; Hertz-Picciotto and Rockhill, 1997). Complete models included the site of origin as the main (fixed) factor and nest identity as a random frailty term to account for non-independence of nest-mates (Therneau et al., 2003). The full model was compared to a null model using AIC values. Finally, the differences between all possible site pairs were assessed by performing multiple comparisons (Hsu, 1996). We used the ‘survival’ (Therneau and Grambsch, 2000) and ‘multcomp’ (Hothorn et al., 2008) R 3.3.1 packages to estimate models and perform multiple comparisons, respectively. 2.2. Genetic analyses Six ants were sampled per nest in the three study areas for a total number of 180 individuals. Ants were placed in 90% ethanol and subsequently stored at 80  C. A single leg was excised from each ant and ground in a microcentrifuge tube containing 300 ml 10% CHELEX 100 resin (BIO-RAD) and incubated at 95  C for 20 min. Samples were vortexed thoroughly, centrifuged for 15 s at 13,000 rpm, and 1 ml of supernatant was used for subsequent polymerase chain reaction (PCR) amplification. Allelic variation was assessed at eight microsatellite loci described in Frizzi et al. (2009). PCR products were resolved in an Applied Biosystems 3130xl genetic analyzer, and allele sizes were scored against a GeneScan 500 ROX size standard using GENEMAPPER 5.0. For each sampling area, deviation from Hardy Weinberg (HW) equilibrium was assessed for each locus using 1999 Monte Carlo simulations implemented in

‘adegenet’ R 3.3.1 package (Jombart, 2008), by randomly selecting one individual per colony to account for the high level of relationship among nest mates. Statistical significance was corrected using a sequential Bonferroni method (Rice, 1989). The HW test was repeated 1000 times per locus and per area. We considered a locus to be at HW equilibrium if the number of tests that significantly deviated from HW proportions was less than 50 (5% of the number of reiterations). A parentage analysis was first performed among workers to determine if the colonies we sampled were monogynous or polygynous. We compared colonies characterized by the same arrangement of the colony, given that a different number of queens may result in a significantly different number of alleles (Bourke and Franks, 1995; Crozier and Pamilo, 1996). The number of queens in each nest was estimated from microsatellite data by a maximum likelihood method implemented in COLONY 2.0.5.0 (Jones and Wang, 2010). For each site, kinship among colonies was evaluated by randomly choosing one ant per colony and then estimating pairwise relatedness using KINGROUP v2 (Konovalov et al., 2004). Differences in relatedness among sites were tested using paired ttests with the Holm-Bonferroni correction for multiple tests. Genetic diversity at each site was assessed by estimating the mean number of alleles per locus (allelic diversity), mean observed heterozygosity, and gene diversity using ‘hierfstat’ R 3.3.1 package (Goudet, 2013). Differences in genetic diversity among sites were tested using a one-way ANOVA for repeated measures. 3. Results The pH value of soils ranged from 6.50 to 7.50 at CS, from 7.10 to 7.61 at PO and from 6.80 to 7.50 at MF. These values were not statistically different (df ¼ 2, F ¼ 1.43, P ¼ 0.27). Average amounts of metals in soils and ants from the three study sites are reported in Table 1. Low concentrations of Cd, Co, Cr, and Ni were recorded in the CS oak-pine forest. Intermediate values were recovered in the PO forest located at the margin of the Monterufoli ophiolite outcrop, while the highest metal concentration was found in the MF pine forest located over the Monteferrato ophiolite outcrop. A reverse trend was recorded for Cu, while Zn concentrations were highly variable. The soil content of Cd, Co, Cr, and Ni usually differed between sites with the exception of Cd, which did not differ between CS and PO. Cu content was significantly higher at CS than PO and MF, while Zn only differed between CS and PO. In ants, trends of Cd, Co, Cr, and Ni broadly paralleled soils. Conversely, Cu showed no significant trend, while Zn was high in all three sites. Significant differences emerged between CS and MF and between PO and MF for the majority of metals. Only Zn showed no significant differences between sites. Mean values for the biota-to-soil accumulation factor (BSAF) are shown in Table 2. Values were always lower than 1 with the only exception of Zn, which ranged from 1.714 (±0.288) in CS to 3.586 (±0.456) in PO. MF exhibited the lowest BSAF values for most metals, with the exception of Cd and Cu. Results of nMDS analysis showed a clear distinction between soil samples from CS and MF. On the other hand, PO samples were more scattered and positioned between CS and MF (Fig. 1A). In particular, the main separation was along the first axis, which is correlated with the metals typical of serpentine soils. Differences in metal concentrations among the three sites were statistically significant (npMANOVA: df ¼ 2,15; F ¼ 15.20; P < 0.001). Similar results were obtained for ant samples (npMANOVA: df ¼ 2,15; F ¼ 26.49; P < 0.001; Fig. 1B). Ant samples were positioned along the gradient with CS and MF reporting the lowest and highest values, respectively, and PO was intermediate. Concentrations of Cd, Co, Cr, and Ni in ants were significantly related to those

F. Frizzi et al. / Chemosphere 180 (2017) 326e334

329

Table 1 Mean (±SE) metal content in each area for both soil and ants and results of multiple comparisons (MC) between sites in pair. Concentration values are expressed in mg/g. Significance: e not significant; *P < 0.05; **P < 0.01; ***P < 0.001. Metal

Soils

CS

3.1 ± 0.3 17.7 ± 1.4 58.6 ± 5.5 64.2 ± 4.9 57.3 ± 5.8 103.5 ± 8.3 0.25 ± 0.1 0.30 ± 0.0 0.50 ± 0.1 10.8 ± 0.7 0.8 ± 0.1 169.6 ± 13.6

Cd Co Cr Cu Ni Zn Cd Co Cr Cu Ni Zn

Ants

PO

MF

3.2 ± 0.1 38.1 ± 4.8 382.1 ± 125.6 39.7 ± 5.7 447.0 ± 163.6 62.8 ± 9.0 0.59 ± 0.1 0.34 ± 0.1 0.71 ± 0.1 18.1 ± 1.9 1.6 ± 0.4 206.1 ± 14.1

Table 2 Mean BSAF values (±SE) in each area and results of multiple comparisons (MC) between sites in pair. Significance: e not significant; *P < 0.05; **P < 0.01; ***P < 0.001. Metal CS

PO

MF

MC significance CS-PO CS-MF PO-MF

Cd Co Cr Cu Ni Zn

0.081 0.017 0.009 0.176 0.015 1.714

± ± ± ± ± ±

0.013 0.002 0.002 0.020 0.003 0.228

0.183 0.011 0.005 0.509 0.010 3.586

± ± ± ± ± ±

0.038 0.004 0.003 0.102 0.006 0.456

0.233 0.006 0.001 0.623 0.003 2.096

± ± ± ± ± ±

0.035 0.002 0.000 0.074 0.001 0.301

e e e ** e **

** * * *** e e

e e e e e *

recorded in the soil. No significant relationship was found for Cu and Zn (Table 3 and Fig. 2). Differences among sites from the Kaplan-Meier analysis were not significant for the ants fed with unpolluted sucrose solution (Table 4), since a small proportion died during the last ten days (0.06 from CS, 0.15 from PO and 0.2 from MF, Fig. 3A). Nonetheless, the models that included the Site factor were always better than the corresponding null model (DAIC > 15) for all groups tested with different Ni concentrations indicating a difference among sites. Ants fed with 2% Ni solution (Fig. 3B) started to die ten days after the beginning of the test. Survival rate at the end of the test period was lower than 0.2 in all tested locations. Slightly higher survival was observed in ants from PO, which differed significantly from CS. Ants fed with 4% Ni solution suffered a much greater mortality, which was clearly site-related. Ants from CS started to die almost immediately after the beginning of the test. Mortality rapidly increased after the 7th day and all ants were dead after 20 days. Ants from PO and MF began to die after the 12th and 16th days, respectively, and they were all dead after 30 days. Individuals fed with 10% Ni solution suffered the highest mortality (Fig. 3D). Multiple comparisons revealed a statistically significant difference between PO and CS, the latter characterized by ants showing a quicker mortality rate. Parentage analysis showed that two colonies sampled at PO, one sampled at CS, and one from MF were polygynous. Considering that polygynous colonies have an inherently higher genetic variability than monogynous colonies, they were excluded from subsequent analysis of allelic diversity. We randomly removed one colony from both CS and MF in order to balance our data set, and performed genetic analyses using eight colonies per site. All loci were in Hardy Weinberg equilibrium. The number of tests that deviated from equilibrium never exceeded 5% of the 1000 reiterated HW tests. Genetic diversity for each locus and for each site is summarized in

7.5 ± 1.0 122.9 ± 21.8 2576.1 ± 704.1 30.2 ± 3.4 1849.3 ± 321.3 90.6 ± 15.0 1.6 ± 0.1 0.6 ± 0.1 4.3 ± 2.2 18.9 ± 3.1 4.7 ± 1.8 170.5 ± 9.4

MC significance CS-PO

CS-MF

PO-MF

e * ** * ** * e e e * e e

*** *** *** *** *** e *** * * * ** e

** ** *** e ** e *** e * e * e

Table 5. Allelic diversity was 4.6 ± 0.47SE, 4.9 ± 0.42SE, and 5.7 ± 0.48SE for CS, PO, and MF, respectively. No differences among sites were detected (F ¼ 1.74, df ¼ 2, P ¼ 0.21). Similarly, gene diversity (mean values: CS, 0.49 ± 0.03SE; PO, 0.51 ± 0.02SE; MF, 0.52 ± 0.01SE) did not differ among sites (F ¼ 0.43, df ¼ 2, P ¼ 0.65). Mean relatedness was 0.0324, 0.0246, and 0.0351 for CS, PO, and MF, respectively. The t-tests showed no significant pairwise differences between sites (P z 1 for each test). 4. Discussion The observed trends in metal contents in both the soil and ant body mirrored the expected pattern whereby an increase in the amounts of the four metals known to be abundant in serpentine soils (nickel, cadmium, cobalt, and chromium) increased from the uncontaminated area of Casaglia to the outcrop of Monteferrato. The pH value of the soils, a factor that can affect metal availability (Kashem and Singh, 2001), did not vary between sites. The amount of metals found in ants broadly paralleled the pattern recorded in the soil with some notable exceptions. The amount of Ni, Cr, Co, and Cd found in ants increased from Casaglia to Monteferrato, although quantities in ant bodies were remarkably lower than values recorded in the soil. On the other hand, Zn was significantly more abundant in ants than in the substrate, while Cu showed no significant trend. These findings were not unexpected, since C. scutellaris is known to accumulate considerable amounts of Zn in body tissues (Gramigni et al., 2013). However, this species seems to be able to regulate the uptake and/or the accumulation of Zn to some extent. The mean BSAF value for this element was, in fact, higher (3.6) in Pomarance, where the element was less abundant, and lower in the other two sites (1.7 and 2.1 for Casaglia and Monteferrato, respectively) that were characterized by higher Zn abundance. High body concentrations of Zn have been documented in several species of Formicinae and Myrmicinae, while the Myrmicine Myrmica rubra showed an ability to efficiently regulate Zn uptake and maintain constant metal body contents with changing environmental metal availability (Grzes, 2010a,b). For all the other metals, BSAF values were always lower than 1, suggesting that C. scutellaris could be an excluder species for these elements. However, it has to be emphasized that we only assessed metals in soil and ants in this study and did not consider their main feeding sources (arthropod prey and homopteran honeydew). Therefore, we cannot be sure whether metal exclusion is due to ants or if it is already a feature of lower trophic levels. The analysis of the metal content for the intermediate levels of the food chain could also be an interesting future development of this study. Despite the low BSAF values, however, some significant differences among sites

330

F. Frizzi et al. / Chemosphere 180 (2017) 326e334

Fig. 1. nMDS ordination plot with metal contents as variables: A) soils; B) ants. Red squares, Casaglia; cyan triangles, Pomarance; green circles, Monteferrato. Arrows are the correlation vectors for each metal. In parentheses the significance level (- not significant; *P < 0.05; **P < 0.01; ***P < 0.001). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Table 3 Summary of GLMM results to quantify the relationship between metal amounts in the soils and in ants. In bold significant P values (< 0.05). Metal

df

F

P

Cd Co Cr Cu Ni Zn

15.89 16.00 16.00 15.45 16.00 14.25

7.51 6.31 14.82 0.12 15.60 0.02

0.0146 0.0232 0.0014 0.7289 0.0011 0.9002

emerged, especially between CS, the reference site, and MF, the most contaminated one. These differences changed with the metal considered. Cadmium (Cd) showed higher BSAF in MF than in the two other sites. This result broadly confirms the findings reported in Gramigni et al. (2013), although in that study the Cd concentrations within the ant body was more than ten times higher than in the soil. A reverse pattern was observed for Cu. For this metal, in

fact, the BSAF value decreased as the amount in the soil increased, as has already been reported by Gramigni et al. (2013). Some differences were also observed for Co and Cr. Although no statistically significant variation was observed for Ni, the BSAF value was the lowest in MF (see Table 2), suggesting a local adaptation to this metal. Our hypothesis that survival ability directly relates to previous long-term exposure was only partly confirmed. Ants from Casaglia were always the first to die under all Ni concentrations tested. Besides, ants from Pomarance had the highest resistance to heavy metal exposure. This result is quite unexpected, because the amount of soil metals at Pomarance were between levels recorded at Casaglia and Monteferrato. Contrary to our expectations, ants from Monteferrato did not have the highest resistance to metal intoxication. A possible explanation for this result is that the high metal concentrations in the body of ants from Monteferrato are close to a physiologically acceptable threshold, so additional exposure to toxicants would cause faster death for the ants. In fact, even if chronic exposure to sub-lethal doses of metal toxicants may

F. Frizzi et al. / Chemosphere 180 (2017) 326e334

331

Fig. 2. Correlation between metal amounts (log transformed values) measured in soil samples and ants. Values on axes are log(mg/Kg). All the samples (letters) and the regression line are shown (dashed line). c, Casaglia; p, Pomarance; m, Monteferrato.

Table 4 Results of multiple comparisons of survival rate between each pair of areas: CR, Casaglia; PO, Pomarance; MF, Monteferrato. z values and associated probability values (P) are shown (- not significant; *P < 0.05; **P < 0.01; ***P < 0.001). Pair

MF - CS PO - CS PO - MF

Table 5 Genetic diversity for each locus and for each site. He, effective alleles; Hs, gene diversity. CS, Casaglia; PO, Pomarance; MF, Monteferrato. Locus

Concentration

z P z P z P

0

2

4

10

2.20 e 1.03 e 1.29 e

1.66 e 3.19 ** 1.56 e

4.34 *** 7.64 *** 3.49 **

0.47 e 2.77 * 2.23 e

reduce the effects of subsequent acute exposure to the same or different metals, it is also possible that very high concentrations in body tissues hamper the functionality of regulatory systems, such as excretion or metabolic binders to eventually lead to an early death (Rainbow and Dallinger, 1993; Jensen and Trumble, 2003). Taylor et al. (1995), for example, described a similar process of early saturation in the freshwater crayfish Cambarus robustus. In this species, animals from strongly metal-polluted freshwaters showed minor behavioral impairments after Cu exposure, although they

Crem12 Crem16 Crem21 Crem22 Crem23 Crem24 Crem45 Crem46

He

Hs

CS

PO

MF

CS

PO

MF

3.802 5.811 4.264 3.614 5.097 4.805 6.902 2.674

5.009 5.592 2.569 6.373 5.855 5.242 5.026 3.942

5.009 7.373 7.945 4.669 5.137 5.738 5.923 3.955

0.495 0.548 0.540 0.485 0.523 0.578 0.588 0.350

0.535 0.660 0.583 0.457 0.513 0.518 0.593 0.538

0.577 0.523 0.487 0.515 0.550 0.558 0.557 0.497

were unable to regulate additional amounts of pollutant intake. In fact, when regulatory systems start to fail, for instance when metallothinein molecules are saturated, the toxicant can flow within the cellular compartments, causing adverse effects on the vitality of the individual. This phenomenon is also termed as ‘spillover’ (e.g. Brown et al., 1977; Hamilton and Mehrle, 1986; Geffard et al., 2007). Tolerance to toxic stressors can result from a combination of behavioral strategies, such as active avoidance of toxicants and

332

F. Frizzi et al. / Chemosphere 180 (2017) 326e334

Fig. 3. Kaplan-Meier survival curves. 0, 2, 4 and 10 indicate the millimolar concentration of Ni test solutions (A, B, C, D in the figure, respectively). Continuous line e Casaglia; dotdash line e Pomarance; dashed line e Monteferrato.

metabolic adaptations (Janssens et al., 2009). In our study, ants were fed solely with polluted sucrose solutions. That should have excluded behavioral adaptations to reduce the intake of toxicants. It is therefore likely that the observed variation in heavy metal tolerance among colonies has a metabolic rather than behavioral basis. There are several examples of these mechanisms in invertebrates (see Janssens et al., 2009; for a review). Genetic analysis showed no differences in either allelic diversity, gene diversity, or relatedness among ant populations from different ophiolitic outcrops. This may suggest that different environmental features had no major effect on C. scutellaris genetic diversity and/or that sufficient gene flow occurs between contaminated and uncontaminated areas. The latter would reduce differences in genetic variation among sites and possible, detrimental effects of genetic erosion (van Straalen and Timmermans, 2002; Lopes et al., 2004; Nowak et al., 2009; Ribeiro and Lopes, 2013). Finally, we would emphasize that although our analysis did not reveal a clear segregation, genetic diversity may still be occurring. This point will have to be further investigated possibly using genome-wide association studies. The ophiolitic outcrops considered in our study were part of a relatively small area, probably not large enough to be considered an independent site from the rest of the environmental matrix (Latter, 1973; Allendorf and Luikart, 2009). The Monteferrato outcrop has the shape of an ellipse where the maximum distance between the center and the edge does not exceed 1.5 km. Additionally in the absence of human intervention, the amount of metals in the soil rapidly decreases with the distance from the source of contamination (Rodríguez et al., 2009). Queens of the ant C. scutellaris adopt an independent colony foundation strategy and seem to travel

long-distance dispersal routes during nuptial flights (Frizzi et al., 2015). This is a behavior observed in other ant species, which are capable of flying for over 10 km from the nest (Markin et al., 1971). The relatively small area of the study sites and high dispersal ability may enable a continuous flow of queens from and to the outcrops. This, in turn, may prevent genetic segregation and inbreeding (Morgan et al., 2007; Janssens et al., 2009).

5. Conclusions This study provides insights into the effects of long-term metal exposure on Crematogaster scutellaris populations living in ophiolitic outcrops. We found that the amount of a number of metals in the body increased with increasing concentrations in the soil. Mortality tests revealed an unexpected pattern with ants from intermediate contaminated sites being more tolerant to acute exposure to nickel. This finding opens new questions on the development of tolerance mechanisms in chronically stressed organisms. Finally, no evidence of significant variation in genetic diversity was observed among study sites. This pattern can probably be explained by long-range dispersion flights of queens, which may allow for a continuous flow towards and from contaminated areas.

Acknowledgements Many thanks are due to two anonymous referees for their comments on an earlier version of the manuscript.

F. Frizzi et al. / Chemosphere 180 (2017) 326e334

References Agarwal, S.K., 2009. Heavy Metal Pollution. APH Publishing, Delhi. Allendorf, F.W., Luikart, G., 2009. Conservation and the Genetics of Populations. John Wiley & Sons, New York. o, M., 2011. Tolerance to Environmental Amiard-Triquet, C., Rainbow, P.S., Rome Contaminants. CRC Press, Boca Raton. Andersen, P.K., Gill, R.D., 1982. Cox’s regression model for counting processes: a large sample study. Ann. Stat. 10, 1100e1120. Angelone, M., Vaselli, O., Bini, C., Coradossi, N., 1993. Pedogeochemical evolution and trace elements availability to plants in ophiolitic soils. Sci. Total Environ. 129, 291e309. Bates, D., Maechler, M., Bolker, B., Walker, S., 2015. Fitting linear mixed-effects models using lme4. J. Stat. Soft 67, 1e48. Bland, J.M., Altman, D.G., 1998. Survival probabilities (the Kaplan-Meier method). BMJ 317, 1572e1580. Bolton, B., Alpert, G., Ward, P.S., Naskrecki, P., 2006. Bolton’s Catalogue of Ants of the World: 1758e2005. Harvard University Press, Cambridge MA. Bourke, A.F., Franks, N.R., 1995. Social Evolution in Ants. Princeton University Press, Princeton. Boyd, R.S., 2009. High-nickel insects and nickel hyperaccumulator plants: a review. Insect Sci. 16, 19e31. Boyd, R.S., 2010. Heavy metal pollutants and chemical ecology: exploring new frontiers. J. Chem. Ecol. 36, 46e58. Boyd, R.S., Wall, M.A., 2001. Responses of generalist predators fed high-Ni Melanotrichus boydi (Heteroptera: Miridae): elemental defense against the third trophic level. Am. Midl. Nat. 146, 186e198. Brooks, R.R., 1987. Serpentine and its Vegetation. A Multidisciplinary Approach. Croom Helm Ltd, London. Brown, D.A., Bawden, C.A., Chatel, K.W., Parsons, T.R., 1977. The wildlife community of Iona Island jetty, Vancouver, BC, and heavy-metal pollution effects. Environ. Conserv. 4, 213e216. Carneiro, F.G., Keim, C.N., Acosta-Avalos, D., Farina, M., 2013. Elemental composition of biomineralized amorphous mineral granules isolated from ants: correlation with ingested mineral particles from the soil. Micron 44, 120e124. Carter, M.R., 1993. Soil Sampling and Methods of Analysis. CRC Press, Boca Raton. Clarke, K.R., Warwick, R.M., 2001. Change in Marine Communities: an Approach to Statistical Analysis and Interpretation, second ed. Primer-e Ltd, Plymouth Marine Laboratory, Plymouth. Cortet, J., Gomot-De Vauflery, A., Poinsot-Balaguer, N., Gomot, L., Texier, C., Cluzeau, D., 1999. The use of invertebrate soil fauna in monitoring pollutant effects. Eur. J. Soil Biol. 35, 115e134. Crozier, R.H., Pamilo, P., 1996. Evolution of Social Insect Colonies - Sex Allocation and Kin Selection. Oxford Series in Ecology and Evolution, Oxford. Del Toro, I., Floyd, K., Gardea-Torresdey, J., Borrok, D., 2010. Heavy metal distribution and bioaccumulation in Chihuahuan Desert Rough Harvester ant (Pogonomyrmex rugosus) populations. Environ. Pollut. 158, 1281e1287. Duruibe, J.O., Ogwuegbu, M.O.C., Egwurugwu, J.N., 2007. Heavy metal pollution and human biotoxic effects. Int. J. Phys. Sci. 2, 112e118. Eeva, T., Sorvari, J., Koivunen, V., 2004. Effects of heavy metal pollution on red wood ant (Formica s. str.) populations. Environ. Pollut. 132, 533e539. €rstner, U., Wittmann, G.T., 2012. Metal Pollution in the Aquatic Environment. Fo Springer Science & Business Media, New York. Fratini, S., Zane, L., Ragionieri, L., Vannini, M., Cannicci, S., 2008. Relationship between heavy metal accumulation and genetic variability decrease in the intertidal crab Pachygrapsus marmoratus (Decapoda; Grapsidae). Estuar. Coast. Shelf S 79, 679e686. Frizzi, F., Ciofi, C., Dapporto, L., Natali, C., Chelazzi, G., Turillazzi, S., Santini, G., 2015. The rules of aggression: how genetic, chemical and spatial factors affect intercolony fights in a dominant species, the mediterranean acrobat ant Crematogaster scutellaris. PloS One 10, e0137919. Frizzi, F., Santini, G., Natali, C., Chelazzi, G., Ciofi, C., 2009. Characterization of polymorphic microsatellite loci in the ant Crematogaster scutellaris. Conserv. Genet. Resour. 1, 425e428. Geffard, A., Geffard, O., Amiard, J.C., His, E., Amiard-Triquet, C., 2007. Bioaccumulation of metals in sediment elutriates and their effects on growth, condition index, and metallothionein contents in oyster larvae. Arch. Environ. Contam. Toxicol. 53, 57e65. Goudet, J., 2013. Hierfstat: Estimation and Tests of Hierarchical F-statistics. R package version 0.04-10. https://CRAN.R-project.org/package¼hierfstat. Gramigni, E., Calusi, S., Chelazzi, G., Del Greco, F., Delfino, G., Gelli, N., Giuntini, L., Massi, M., Santini, G., 2011. Analysis of metal deposit distribution in ants (Crematogaster scutellaris) at the Florence external scanning microbeam. X-Ray Spectrom. 40, 186e190. Gramigni, E., Calusi, S., Gelli, N., Giuntini, L., Massi, M., Delfino, G., Chelazzi, G., Baracchi, D., Frizzi, F., Santini, G., 2013. Ants as bioaccumulators of metals from soils: body content and tissue-specific distribution of metals in the ant Crematogaster scutellaris. Eur. J. Soil Biol. 58, 24e31. Grzes, I.M., 2010a. Zinc and cadmium regulation efficiency in three ant species originating from a metal pollution gradient. Bull. Environ. Contam. Toxicol. 84, 61e65. Grzes, I.M., 2010b. Ants and heavy metal pollutionea review. Eur. J. Soil Biol. 46, 350e355. Grzes, I.M., Okrutniak, M., Woch, M.W., 2015. Monomorphic ants undergo within-

333

colony morphological changes along the metal-pollution gradient. Environ. Sci. Pollut. R. 22, 6126e6134. Hamilton, S.J., Mehrle, P.M., 1986. Metallothionein in fish: review of its importance in assessing stress from metal contaminants. Trans. Am. Fish. Soc. 115, 596e609. Hertz-Picciotto, I., Rockhill, B., 1997. Validity and efficiency of approximation methods for tied survival times in Cox regression. Biometrics 53, 1151e1156. Hoffmann, B.D., Griffiths, A.D., Andersen, A.N., 2000. Responses of ant communities to dry sulfur deposition from mining emissions in semi-arid tropical Australia, with implications for the use of functional groups. Austral Ecol. 25, 653e663. Hothorn, T., Bretz, F., Westfall, P., 2008. Simultaneous inference in general parametric models. Biom. J. 50, 346e363. Hsu, J., 1996. Multiple Comparisons: Theory and Methods. CRC Press, Boca Raton.
, E., 2014. Metallothioneins, unconventional proteins from unIsani, G., Carpene conventional animals: a long journey from nematodes to mammals. Biomolecules 4, 435e457. Janssens, T.S., Roelofs, D., vanStraalen, N., 2009. Molecular mechanisms of heavy metal tolerance and evolution in invertebrates. Insect Sci. 16, 3e18. Jeantet, A.Y., Ballan-Dufrancais, C., Martoja, R., 1977. Insects resistance to mineral pollution. Importance of spherocrystal in ionic regulation. Rev. Ecol. Biol. Sol. 14, 563e582. Jensen, P., Trumble, J.T., 2003. Ecological consequences of bioavailability of metals and metalloids in insects. Recent Res. Dev. Entomol. 42, 1e17. Jombart, T., 2008. adegenet: a R package for the multivariate analysis of genetic markers. Bioinformatics 24, 1403e1405. Jones, O.R., Wang, J., 2010. COLONY: a program for parentage and sibship inference from multilocus genotype data. Mol. Ecol. Resour. 10, 551e555. Kashem, M.A., Singh, B.R., 2001. Metal availability in contaminated soils: I. Effects of flooding and organic matter on changes in Eh, pH and solubility of Cd, Ni and Zn. Nutr. Cycl. Agroecosys. 61, 247e255. Klerks, P.L., Weis, J.S., 1987. Genetic adaptation to heavy metals in aquatic organisms: a review. Environ. Pollut. 45, 173e205. Konovalov, D.A., Manning, C., Henshaw, M.T., 2004. KINGROUP: a program for pedigree relationship reconstruction and kin group assignments using genetic markers. Mol. Ecol. Notes 4, 779e782. Latter, B.D.H., 1973. The island model of population differentiation: a general solution. Genetics 73, 147e157. Lefcort, H., Abbott, D.P., Cleary, D.A., Howell, E., Keller, N.C., Smith, M.M., 2004. Aquatic snails from mining sites have evolved to detect and avoid heavy metals. Arch. Environ. Con. Tox. 46, 478e484. Lopes, I., Baird, D.J., Ribeiro, R., 2004. Genetic determination of tolerance to lethal and sublethal copper concentrations in field populations of Daphnia longispina. Arch. Environ. Con. Tox. 46, 43e51. Madden, K.E., Fox, B.J., 1997. Arthropods as indicators of the effects of fluoride pollution on the succession following sand mining. J. Appl. Ecol. 1239e1256. Maes, G.E., Raeymaekers, J.A.M., Pampoulie, C., Seynaeve, A., Goemans, G., Belpaireb, C., Volckaerta, F.A.M., 2005. The catadromous European eel Anguilla anguilla (L.) as a model for freshwater evolutionary ecotoxicology: relationship between heavy metal bioaccumulation, condition and genetic variability. Aquat. Toxicol. 73, 99e114. Majer, J.D., 1983. Ants: bio-indicators of minesite rehabilitation, land-use, and land conservation. Environ. Manag. 7, 375e383. Markin, G.P., Dillier, J.H., Hill, S.O., Blum, M.S., Hermann, H.R., 1971. Nuptial flight and flight ranges of the imported fire ant, Solenopsis saevissima richteri (Hymenoptera: formicidae). J. Ga. Entomol. So 6, 145e156. Mengoni, A., Gonnelli, C., Galardi, F., Gabbrielli, R., Bazzicalupo, M., 2000. Genetic diversity and heavy metal tolerance in populations of Silene paradoxa L. (Caryophyllaceae): a random amplified polymorphic DNA analysis. Mol. Ecol. 9, 1319e1324. Migula, P., Głowacka, E., 1996. Heavy metals as stressing factors in the red wood ants (Formica polyctena) from industrially polluted forests. Fresen. J. Anal. Chem. 354, 653e659. Morgan, A.J., Kille, P., Stürzenbaum, S.R., 2007. Microevolution and ecotoxicology of metals in invertebrates. Environ. Sci. Technol. 41, 1085e1096. Morris, T.I., Campos, M., Jervis, M.A., McEwen, P.K., Kidd, N.A.C., 1998. Potential effects of various ant species on green lacewing, Chrysoperla carnea (Stephens)(Neuropt., Chrysopidae) egg numbers. J. Appl. Entomol. 122, 401e403. Nadig, S.G., Lee, K.L., Adams, S.M., 1998. Evaluating alterations of genetic diversity in sunfish populations exposed to contaminants using RAPD assay. Aquat. Toxicol. 43, 163e178. Nowak, C., Vogt, C., Pfenninger, M., Schwenk, K., Oehlmann, J., Streit, B., Oetken, M., 2009. Rapid genetic erosion in pollutant-exposed experimental chironomid populations. Environ. Pollut. 157, 881e886. Nummelin, M., Lodenius, M., Tulisalo, E., Hirvonen, H., Alanko, T., 2007. Predatory insects as bioindicators of heavy metal pollution. Environ. Pollut. 145, 339e347. Oksanen, J., Blanchet, F.G., Kindt, R., Legendre, P., Minchin, P.R., O’Hara, R.B., Simpson, G.L., Solymos, P., Henry, M., Stevens, H., Wagner, H., 2013. Vegan: Community Ecology Package. R package version 2.0-10. https://CRAN.R-project. org/package¼vegan. Ottonetti, L., Tucci, L., Santini, G., 2006. Recolonization patterns of ants in a rehabilitated lignite mine in central Italy: potential for the use of Mediterranean ants as indicators of restoration processes. Restor. Ecol. 14, 60e66. Ottonetti, L., Tucci, L., Chelazzi, G., Santini, G., 2008. Stable isotopes analysis to assess the trophic role of ants in a Mediterranean agroecosystem. Agric. For. Entomol. 10, 29e36. Prasad, M.N.V., 2001. Metals in the Environment: Analysis by Biodiversity. CRC

334

F. Frizzi et al. / Chemosphere 180 (2017) 326e334

Press, Boca Raton. Rabitsch, W.B., 1995. Metal accumulation in arthropods near a lead/zinc smelter in Arnoldstein, Austria. II. Formicidae. Environ. Pollut. 90, 239e247. Rabitsch, W.B., 1997. Seasonal metal accumulation patterns in the red wood ant Formica pratensis (Hymenoptera) at contaminated and reference sites. J. Appl. Ecol. 34, 1455e1461. Rainbow, P.S., Dallinger, R., 1993. Ecotoxicology of Metals in Invertebrates. CRC Press, Boca Raton. Reeves, R.D., Baker, A.J., 2000. Metal-accumulating plants. In: Raskin, I., Ensley, B.D. (Eds.), Phytoremediation of Toxic Metals: Using Plants to Clean up the Environment. Wiley, New York, pp. 193e229. Reinecke, S.A., Reinecke, A.J., 2004. The comet assay as biomarker of heavy metal genotoxicity in earthworms. Arch. Environ. Con. Tox. 46, 208e215. Ribas, C.R., Solar, R.R., Campos, R.B., Schmidt, F.A., Valentim, C.L., Schoereder, J.H., 2012. Can ants be used as indicators of environmental impacts caused by arsenic? J. Insect Conserv. 16, 413e421. Ribeiro, S.P., Borges, P.A., Gaspar, C., Melo, C., Serrano, A.R., Amaral, J., Aguiar, C., , G., Quartau, J.A., 2005. Canopy insect herbivores in the Azorean LaurAndre isilva forests: key host plant species in a highly generalist insect community. Ecography 28, 315e330. Ribeiro, R., Lopes, I., 2013. Contaminant driven genetic erosion and associated hypotheses on alleles loss, reduced population growth rate and increased susceptibility to future stressors: an essay. Ecotoxicology 22, 889e899. Rice, W.R., 1989. The sequential Bonferroni test. Evolution 43, 223e225. rate, J., Rinco n, J., 2009. Heavy metal distribution Rodríguez, L., Ruiz, E., Alonso-Azca and chemical speciation in tailings and soils around a PbeZn mine in Spain. J. Environ. Manag. 90, 1106e1116. Ross, K., Cooper, N., Bidwell, J.R., Elder, J., 2002. Genetic diversity and metal tolerance of two marine species: a comparison between populations from

contaminated and reference sites. Mar. Pollut. Bull. 44, 671e679. nchez, M.L., 2008. Causes and Effects of Heavy Metal Pollution. Nova Science Sa Publishers, New York. Santini, G., Ramsay, P.M., Tucci, L., Ottonetti, L., Frizzi, F., 2011. Spatial patterns of the ant Crematogaster scutellaris in a model ecosystem. Ecol. Entomol. 36, 625e634. Schatz, B., Hossaert-Mckey, M., 2003. Interactions of the ant Crematogaster scutellaris with the fig/fig wasp mutualism. Ecol. Entomol. 28, 359e368. Shaw, J., 1989. Heavy Metal Tolerance in Plants: Evolutionary Aspects. CRC Press, Boca Raton. Sorvari, J., Rantala, L.M., Rantala, M.J., Hakkarainen, H., Eeva, T., 2007. Heavy metal pollution disturbs immune response in wild ant populations. Environ. Pollut. 145, 324e328. Taylor, R.M., Watson, G.D., Alikhan, M.A., 1995. Comparative sub-lethal and lethal acute toxicity of copper to the freshwater crayfish, Cambarus robustus (Cambaridae, Decapoda, Crustacea) from an acidic metal-contaminated lake and a circumneutral uncontaminated stream. Water Res. 29, 401e408. Therneau, T.M., Grambsch, P.M., 2000. Modeling Survival Data: Extending the Cox Model. Springer, New York. Therneau, T.M., Grambsch, P.M., Pankratz, V.S., 2003. Penalized survival models and frailty. J. Comput. Graph. Stat. 12, 156e175. Tyler, G., Påhlsson, A.M.B., Bengtsson, G.E., Bååth, E., Tranvik, L., 1989. Heavy-metal ecology of terrestrial plants, microorganisms and invertebrates. Water Air Soil Poll. 47, 189e215. van Straalen, N.M., Timmermans, M.J., 2002. Genetic variation in toxicant-stressed populations: an evaluation of the “genetic erosion” hypothesis. Hum. Ecol. Risk Assess. 8, 983e1002. van Wilgenburg, E., Elgar, M.A., 2013. Confirmation bias in studies of nestmate recognition: a cautionary note for research into the behaviour of animals. PloS One 8, e53548.