Contamination and mortality of leaf-cutting ant workers by the quinone inside inhibitor fungicide after social interactions

Leaf-cutting ants of the genera Atta and Acromyrmex (Hymenoptera: Formicidae) are the most important pests in forest and agricultural plantations and livestock. Toxic baits are the main method to manage these insects. The objective was to determine whether the behavior of allogrooming, touch, and self-grooming among Atta sexdens rubropilosa Forel, 1908 (Hymenoptera: Formicidae) workers disperse the fungicide quinone inside inhibitor and whether this product is toxic to them. This fungicide was applied, topically, in groups of workers and the social interactions between them and their mortality with and without the fungicide were evaluated. The interactions and the quinone inside inhibitor fungicide contamination increased with the number of leaf-cutting ant workers per group. Excessive touches, with subsequent allogrooming, and self-grooming among the ant workers dispersed the quinone inside the inhibitor fungicide causing 100% mortality and indicating its toxicity to this insect. The hypothesis that social interactions contaminated ant colony mates and the toxicity of the fungicide quinone inside inhibitor to workers of the leaf-cutting ant A. sexdens rubropilosa was proven.

www.nature.com/scientificreports/complex 11 .The crop of ants, which collect fluids (nectar) during foraging, such as the genus Camponotus, is more developed and adapted to this foraging strategy 11 .
Low rates of fluid absorption by leaf-cutting ants affect other foraging strategies.Atta sexdens Forel 1908 (Hymenoptera: Formicidae: Myrmicinae) workers ingested liquids during cutting and processing leaves collected to cultivate their symbiotic fungus and rarely visited nectar sources 12 .Furthermore, leaf-cutting ants ingest liquids during gongylid harvesting in the fungus garden, but their crops are not adapted to store large volumes of liquid 13 .The rate of fluid intake by C. rufipes workers was 6.7 µl/min, higher than that of A. sexdens, at most, 0.6 µl/min 11 .This indicates that trophallaxy may be infrequent or absent in leaf-cutting ants 14 .
Oral trophallaxy may not disperse insecticides in colonies of leaf-cutting ants and these insects should be intoxicated by direct contact during processing and incorporation of toxic baits into the fungus garden.This can also occur through indirect contact with the active ingredient, during social interactions between workers in the colony 15 .These interactions include hygienic behaviors such as allogrooming, self-grooming and touching, besides contact between contaminated workers or not dispersing the insecticide among colony members 14,16,17 .This social interaction can be the route of contamination with other substances, such as fungicides?An experiment was carried out to test this hypothesis with a fungicide (toxic to insects).The quinone inside inhibitor fungicide was applied topically on Atta sexdens rubropilosa Forel, 1908 (Hymenoptera: Formicidae) workers and the social interactions between them, with this fungicide or not, and the mortality of these insect were evaluated.

Material and methods
Colonies studied.Sixty colonies of A. sexdens rubropilosa with approximately six months old were collected in March 2020 in the municipality of Botucatu, São Paulo state, Brazil, maintained at the Laboratory of Social Insect Pests of UNESP in Botucatu and used in the experiment.Each colony was kept in a container (length: 15 cm, width: 15 cm and height: 15 cm) with a fungus garden and received Acalypha spp.(leaves and stems) at a temperature of 24 ± 2 °C, relative humidity of 80% and a photoperiod of 12 h of light.The handling of plant (Acalypha spp.leaves and stems) were carried out in accordance with relevant guidelines and regulations.
The A. sexdens rubropilosa workers were removed from their colonies and separated by size, based on head width, from 1.2 to 2.2 mm.The pronotum of these workers, which would be contaminated, was marked with a small dot of ink made using a white colored pen (Edding®) with excellent adhesion, fast drying and good visibility 18 .After marking, these workers remained for two hours in 350 ml plastic containers with the edges smeared with Fluon (fluoroethylene resin) to prevent them from escaping.Then, topically, 1.0 μl of the fungicide was applied on each worker's pronotum using a Hamilton microsyringe (5.0 μl) and placed in their respective groups.Self-cleaning, mutual-cleaning and touching frequencies were recorded during 24 h of video recording.

Experiment 2.
The mortality per group of contaminated A. sexdens workers was evaluated in an experiment similar to the first.A control treatment, consisting of vegetable oil at a concentration of 1.0%, and two others with the fungicide at 0.1% and 1.0% (mass/mass) were evaluated.After marking and contaminating the worker, each group was placed in a transparent plastic container with a diameter of 7.5 cm and 5.5 cm height, with hermetic lids and, at its base, a 1.0 cm of plaster to maintain moisture.The contaminated worker and its group of uncontaminated ants were kept for 24 h without food.After this period, they received approximately 3.0 g of the symbiont fungus and about 20 minimal workers with a head width of approximately 0.8 to 1.0 mm, as they efficiently cultivate the fungus 19 .Ant mortality was evaluated in the 1st, 2nd, 3rd, 5th, 7th, 9th, 11th, 14th, 17th and 21st days after their contamination 20 .

Data analysis.
A regression model, considering the frequency of each behavior with different fungicide concentrations and A. sexdens rubropilosa worker proportions, was applied using the Generalized Linear Model (GLM) with negative binomial variance and logarithmic linkage function according to overdispersion data 21 .A residual deviation (36,887.41,989and 39,656 with 30 degrees of freedom) were non-significant (P-value = 0.18, 0.07 and 0.11) indicating a lack of evidence against this model for the frequency of allogrooming, self-cleaning and touching behaviors.Multiple comparison tests, between group means and fungicide concentrations, were performed using the log odds ratio.The p value was adjusted using Tukey's method to compare a family with four and three estimates, respectively 22 .
Multiple comparisons of survival curves were performed using the Log-Rank test 23 and the False Discovery Rate 24 to adjust their P-value correction.The MASS, multicomp, emmeans, survival, survminer, ggplot2 and RColorBrewer packages of the R statistical and graphics programming environment version 4.1.2used the R Core Software 25 .

Results
Experiment 1.The frequency of allogrooming behavior was higher in the 01:04, 01:09 and 01:19 groups than in the 01:01 and similar between the concentrations of quinone inside inhibitor fungicide and the control (Table 1).
The frequency of self-grooming behavior, 01:01, 01:04, 01:09 and 01:19 groups, did not differ between the concentrations of quinone inside inhibitor fungicide and the control (Table 1).The frequency of self-grooming behavior increased, reaching a higher value in the 01:04, 01:09 and 01:19 groups than in the 01:01 (Table 1, Fig. 1), but it did not differ between the concentrations of 0.1%, 1% of quinone inside inhibitors fungicide and the control (Table 1).Table 1.Coefficients (C), intercept (Int.) and estimated values (Est.)(mean + standard error-SE), z.value (Zv) and Pr > z (Pv) of the generalized linear model of the touching, self-grooming and allogrooming frequencies between Atta sexdens rubropilosa (Hymenoptera: Formicidae) workers contaminated with quinone inside inhibitor fungicide at 0.1% (0.1%) and 1.0% (1.0%) with the proportions of 04, 09 and 19.Superscripted values are the standard error of the mean for groups.The frequency of touching behavior was higher in the 01:04, 01:09 and 01:19 groups than in the 01:01 (Table 1, Fig. 1).This frequency was higher with the concentrations of quinone inside inhibitors fungicide at 0.1% and 1% than in the control (Fig. 1).Touching behavior increased with the number of workers of the leaf-cutting ant A. sexdens rubropilosa per group (Table 2), but it did not differ between the concentrations of quinone inside inhibitor fungicide 0.1% and 1% and the control (Fig. 2).
Experiment 2. The survival curves of A. sexdens rubropilosa workers, with the concentrations of the fungicide quinone inside inhibitor at 0.1% p < 2e-16 and 1.0% p < 2e-16) and the control differed from each other (p = 0.019) with the death of 100% of the workers with the two concentrations of this fungicide (Fig. 3).

Discussion
The hypothesis was confirmed that social interactions between A. sexdens rubropilosa workers increase the dispersion of the fungicide quinone inside inhibitor and contaminates others through allogrooming, self-cleaning, and touch behaviors.These results corroborate the higher mortality of ant workers due to the dispersion of fat-soluble substances, including insecticides, in social interactions 15,26 .This mode of action should be similar to that of the sulfluramid insecticide, widely used to manage leaf-cutting ants in Brazil 3 , but that of the latter compound is slower and only by contact 20 .The higher frequency of allogrooming, self-grooming and touching behaviors among A. sexdens rubropilosa workers, after contamination with quinone inside inhibitor fungicide at 0.1% and 1.0%, proves its dispersion by this ant.This dispersion type has been reported for insecticides on different social and semi-social insects such as the Argentine ant, Linepithema humile (Mayr) (Hymenoptera: Formicidae) [27][28][29] and black carpenter ants, Camponotus pennsylvanicus De Geer, 1773 (Hymenoptera, Formicidae) transferring the insecticide fipronil 29 and the German cockroach Blattella germanica (L.) (Blattodea: Blattellidae) the indoxacarb 30 to nestmates.This occurs mainly when those contaminated at the application place transfer the compound to others in the population 29 by tarsal or antennal contact in groups or, randomly, by touch between an alive and a dead worker 27 .The frequency of touching behavior by A. sexdens with workers exposed to the concentrations of quinone inside inhibitor fungicide at 0.1% and 1% was higher than in the control." The higher frequency of allogrooming in the 01:04, 01:09 and 01:19 groups than in the 01:01 group.This behavior increases the dispersion of this fungicide due to excessive touching among workers.This is similar to that reported for social interactions among those of A. sexdens with allogrooming, self-cleaning, and touching behaviors between contaminated and non-contaminated individuals dispersing the insecticides fipronil and sulfluramid to colony mates 15 .Social interactions disperse insecticidal substances and fungicides among A. sexdens rubropilosa workers.
The higher frequency of self-grooming behavior in the groups 01:04, 01:09 and 01:19 is important to remove potentially pathogenic organisms on the surface of the bodies of social insects and from their colony mates 15 .Self-grooming effectively removed parasites such as Metarhizium from the ant cuticle [31][32][33] .The higher frequency of self-grooming behavior in the 01:04, 01:09 and 01:19 groups than in the 01:01 shows an increase in social interactions with the group size and, consequently, the dispersion of insecticidal substances among the colony nests 16,26,34 .Active ingredients, used to manage leaf-cutting ants, act by ingestion and contact and, therefore, hygiene behaviors increase contact with the insecticide and, consequently, ant contamination 3 .On the other hand, the similar frequency of self-cleaning, between the concentrations of quinone inside inhibitors fungicide and the control, demonstrates that the A. sexdens rubropilosa workers did not detect this fungicide.This is desirable, as ants, during self-grooming and collective cleaning, ingest particles and substances collected during cleaning, including insecticides, which are absorbed by the post-pharyngeal gland and, consequently, intoxicating all or most of the colony companions 26 .The similarity between the self-grooming behavior with the concentrations of 0.1% and 1% of quinone inside inhibitor fungicide and in the control.Also, this demonstrates that this fungicide not modified the ant workers' self-cleaning behavior, increasing the insecticide dispersion and contamination".
The touch behavior as the number of ant workers per group increased, but without differences between the concentrations of quinone inside inhibitor fungicide at 0.1%, 1% and the control, demonstrates that the fungicide did not modify this behavior of the ant workers A. rubropilosa sexdens.The non-detection of insecticides or www.nature.com/scientificreports/fungicides is important to circulate the active ingredient, contaminating and killing as many workers as possible and, consequently, causing the collapse of the leaf-cutting ant colony 3 .The death of 100% of A. sexdens rubropilosa workers in 21 days with the two concentrations of the quinone inside inhibitor fungicide is unprecedented for a fungicide, probably toxic by contact and ingestion, acting like the insecticide fipronil 15 with two modes of action increasing the contamination and death of workers after its dispersion by social interactions.
The interactions between A. sexdens rubropilosa workers dispersed the fungicide quinone inside inhibitor among them.The mode and mechanism of action of this fungicide on leaf-cutting ant workers is still unknown.

Figure 3 .
Figure 3. Survival probability curves for Atta sexdens rubropilosa (Hymenoptera: Formicidae) workers in groups with the quinone inside inhibitors fungicide 0.1% and 1.0% and in the control.