Behavioural Evidence and Chemical Identification of a Female Sex Pheromone in Anagrus atomus (Hymenoptera: Mymaridae)

Anagrus atomus (L.) is an egg parasitoid involved in the biological control of Empoasca vitis (Göthe) in vineyards. Sex pheromones play a crucial role in mate finding for several parasitoid species and could be used for monitoring under field conditions. We carried out laboratory and field studies aimed at assessing the existence and identity of a possible A. atomus sex pheromone. We found that males were significantly attracted by virgin females independent of age. Males were not attracted to individuals of the same sex, but they were attracted by a crude extract from an unmated female and its polar fraction. Eugenol (4-allyl-2-methoxyphenol) was identified as the attractive substance and proved to be attractive not only in the olfactometer but also in another laboratory bioassay and under field conditions. Attraction of males, but not females, confirms that this is not an aggregation pheromone. This is the first sex-pheromone component identified in Mymaridae, however more compounds could be involved in the mating behaviour of A. atomus. The utility of a sex pheromone in A. atomus is discussed in the context of fitness returns.


Introduction
Mate nding is a crucial step in the mating system of insects and the use of reliable information may increase both the probability of nding a mate and mating success. Several stimuli (e.g. visual and tactile) can be used for this purpose but chemicals produced by members of both sexes are mainly involved (Symonds et al. 2009). Depending on the function, the range of activity of pheromones can vary: highly volatile compounds released by females are used by males for long-range orientation during mate nding (> 2 cm) (Steiner et al. 2006), whereas chemicals of relative low volatility mediate male courtship behaviour at close range (Steiner et al. 2006). Sex pheromones play a crucial role in mate nding in most Hymenopterous parasitoids (Kainoh 1999;Ruther 2013); until now, the sex-pheromone components of more than 20 species, belonging to 10 families, have been identi ed (Quicke 1997;Kainoh 1999;Eller et al. 1984;Pompanon et al. 1997;Keeling et al. 2004;Nichols et al. 2010;Ruther et al. 2011;Salerno et al. 2012;Ruther 2013;Hrabar et al. 2015). In the family Mymaridae, evidence of a sex-pheromone has been reported for one species only (Cormier et al. 1998); however, no sex-pheromone components have been identi ed so far. Parasitoids' sex pheromones are more often produced by females but there is increasing evidence of sex pheromones production by males as well (Steiner et al. 2006;Steiner e Ruther 2009;Ruther 2013).
Practical interest in sex pheromones is normally limited to insect pests' attractants for their possible use in integrated pest management (Witzgall et al. 2010;Mainers and Peri 2013). Among the semiochemicals associated with natural enemies, mostly kairomones and synomones are considered in the pest control strategies (James 2003;James and Price 2004;James 2005;Rodriguez-Saona et al. 2012). However, the possibility to capture natural enemies with traps baited with alive females (Jewett and Carpenter 2001) or with lures containing their sex pheromones (Eller et al. 1984;Gabrýs et al. 1997;De Lury et al. 1999;Suckling et al. 2002;Itadani and Ueno 2014) has been exploited for research purposes.
In spring, the parasitoid moves into the vineyards where it completes several generations per year in E. vitis eggs (Cerutti et al. 1991). In autumn, A. atomus migrates back to the surrounding vegetation, where overwintering eggs are laid in eggs of leafhoppers other than E. vitis Pavan 2011, 2013).
As many other parasitoids, the mating system of A. atomus is characterized by protandry with males emerging rst and followed by females (Zanolli and Pavan 2013); the sex ratio is generally 1:1 (Zanolli and Pavan 2011). A. atomus females begin to lay eggs as soon as they emerge and have a life expectancy of 9 days at 24 °C (Agboka et al. 2004). In A. atomus, as in most in hymenopteran parasitoids, male offspring emerges from unfertilized haploid eggs (arrhenotokous parthenogenesis) and female offspring from fertilized diploid eggs (Choudhury and Copland 2003;Heimpel and De Boer 2008).
In male parasitoids, the perception of the female sex pheromone is often followed by wing fanning (Vinson 1972, Böttinger et al. 2020; this behaviour has also been exploited for the purpose of pheromone isolation . When males of A. incarnatosimilis Soyka and A. breviphragma Soyka, that are closely related to A. atomus, recognize a virgin female, they raise their wings perpendicular to the dorsum and quickly move towards the female (Moratorio and Chiappini 1995). In A. breviphragma, mated females can still attract males such that, occasionally, a second insemination can occur (Moratorio and Chiappini 1995).
This study was carried out to ascertain the occurrence and identity of a female sex pheromone in the parasitoid A. atomus through behavioural bioassays and chemical analyses. The research work included three consecutive steps: (1) veri cation of the existence of a female sex pheromone through behavioural bioassays, (2) identi cation of the chemical eliciting males' response, and (3) laboratory and eld tests of the attractiveness of the identi ed compound towards males.
The gathered information on the female sex pheromone of A. atomus could be useful to monitor wasp populations in the eld, to delineate their spatial distributions and to determine potential sources of parasitoids in the context of habitat management strategies.
Parasitized eggs inside the leaf portions were individually isolated in vials (1.5 × 10 cm) and kept in a climatic chamber (Sanyo Versatile Environmental Test Chamber) at 60 ± 5% RH and 24 ± 1 °C with a daily light/dark cycle of 16:8 hours. The vials were checked daily for adult emergence. Unmated males and females to be used in the bioassays were maintained in the same vial where development took place until use. Mated females were obtained by putting couples of newly emerged females and males in a vial until mating was observed and then placed back to their respective vial. After emergence parasitoid wasps were maintained under the same conditions described above.

Olfactometer Bioassays
All behavioural experiments were performed using a four-arms olfactometer (diameter 10 cm) (Pettersson 1970;Vet et al. 1983); the stimuli to be tested (i.e. alive insects, extracts and pure substances on a strip of lter paper) were placed in a syringe connected with an arm upwind the exposure chamber and closed at the opposite end with a mesh, so that one eld of the olfactometer was treated with the odour to be tested and three others arms served as odourless control elds. The olfactometer was placed in a dark cardboard box to exclude the interference of possible external visual cues. All bioassays were carried out between 10:00 am and 6:00 pm at 24 ± 1.5 °C, in a darkened room. The olfactometer was illuminated from the top under a 7.500 lux light source. The air stream was maintained at 400 ml/min. One hour before the beginning of an experimental session, parasitoid wasps in vials were placed in the test room to get acclimatized to experimental conditions. To avoid any asymmetrical bias, the olfactometer was rotated 90° clockwise after each bioassay. Parasitoid wasps were introduced individually through a hole in the centre of the olfactometer's ceiling. The test started as soon as the insect entered the olfactometer and lasted for 10 min. Each wasp was tested only once. The rst choice made by the insects and the time spent in each odour eld were recorded with the computer program "OLFA 1.0" (Nazzi 1995). After each test, the olfactometer was dismantled, scrupulously washed with 70% ethanol and then rinsed with water.
If not otherwise speci ed, in the different experiments 12 parasitoid wasps per treatment were tested.
During the olfactometer essays, the male behaviour (i.e., general behaviour, wing fanning and locomotory activity) was also recorded.

Behavioural Evidence of a Female Attractive Pheromone
To ascertain the existence of a female sex pheromone in A. atomus, four experiments were conducted. Experiments 1 and 2 were carried out to test the hypothesis that A. atomus females release a pheromone to attract males and whether male responsiveness is in uenced by the age and mating status of female; these experiments were performed using the four-arms olfactometer. In Experiment 1 and 2, the response of males towards 0 to 4 day-old unmated females and towards 0 to 3 day-old mated females, was evaluated, respectively. 0 day-old females were tested within 24 hours from the emergence.
Experiment 3 was carried out to test the possible attraction of males to conspeci c males, using 0-1 dayold unmated males both as a bait and test insects.
Experiment 4 aimed at testing the response of males to the crude extract of 0-1 day-old unmated female.
For this purpose, twelve 0-1 day-old unmated females were killed by freezing (30 min, -20 °C) and then extracted µl at room temperature. The resulting extracts were stored at -20 °C until use. For each of the 12 replicates of the bioassay, one female equivalent of the extract (10 µl) was applied to a strip of lter paper (absorbent paper 500 µm, 0.5 × 3.0 cm), and the solvent was allowed to evaporate. Then the strip was inserted into the syringe in the treated arm of the four-arm olfactometer. For each replicate, a new strip of paper was used.

Identi cation of a Candidate Pheromone
To isolate and identify the female sex pheromone demonstrated as above, we carried out experiment 5. This involved testing the biological activity of three fractions of the female extract fractions using 0-1 day-old unmated males as test insects. For this purpose, ether extracts of A. atomus females were fractionated by liquid chromatography, and the resulting fractions bioassayed as usual. Batches of 15 freshly emerged females were extracted with 150 µl at room temperature as described above, then the extract was loaded on a column packed with 100 mg silica gel (200-400 mesh, pore size 60 Å. Sigma Aldrich US) and eluted sequentially with 1 ml each of hexane and ether. One female equivalent of each fraction in 10 µl of the solvent was tested in the olfactometer against sixteen individual A. atomus males. Ether and hexane alone were also tested.

Chemical Analysis of the Crude Extract
Ether extracts of males and females (N = 10 for each sex) were prepared as described above and analysed by coupled gas chromatography-mass spectrometry (GC-MS) on a Varian 3400 gas chromatograph coupled with a Varian Saturn 2000 mass spectrum detector, equipped with CIP-SIL 8 capillary column (30 m × 0.25 mm I.D.: 0.25 µm thickness). The oven temperature was 50 °C for 1 min, followed by a ramp to 320 °C at 10 °C/min; nal temperature was hold for 2 min. Carrier gas was He, maintained at a coµl °C. Interface temperature was maintained at 250 °C. Mass Spectrometry (MS) detection was performed with electron impact (EI) mode at 70 eV by operating in the l-mul delay in the 40-650 amu range. The identi cation of the volatile compounds emitted by females but not by males was performed by comparison of the females and males crude extract; unknown spectra were identi ed using the NIST Library (National Institute of Standards and Technologies, US) as a reference for spectra and retention indexes.

Coinjection of the Female Crude Extract and Synthetic Eugenol
An ether extract of females (N = 50) was prepared as described above, reduced under nitrogen to 5 µl, and 2 µl of this extract analysed by GC-MS. Then 1 µl of a solution of synthetic eugenol in ether (0.01 µg/µl, Sigma Aldrich US) was added to the remaining extract and 1 µl of this mixture analysed by GC-MS. This analysis served also for the purpose of roughly quantifying the amount of eugenol associated to the major peak found in the female extract but not in the male one.

Testing of the Candidate Sex Pheromone Component
In Experiment 6, the activity of synthetic eugenol on males and females was evaluated in the olfactometer bioassays. Responses of A. atomus males (N = 16) to eugenol were assessed at different doses (0.1-0.5-1-5-10-100 ng). Ether was used as control as the treatment stimulus was dissolved in that solvent.
The response to eugenol was also tested with 16 individual A. atomus unmated females (N = 16) in order to exclude that the substance is an aggregation rather than a sex pheromone. Ether was used as control as above.

Cage Bioassays
To con rm the attractiveness of eugenol in the lab, the response of 0-1 day-old unmated males was tested in a Plexiglas's cage with ventilation holes covered with mesh (30 × 60 × 50 cm) and exposed to natural light. 50 µl eugenol, corresponding to 50 ng of pure compound were pipetted onto a strip of lter paper (absorbent paper, 0.5 × 3.0 cm) and, after evaporation of the solvent, the strip was xed in the middle part of a yellow sticky trap (15 × 5 cm). Another trap baited with 50 µl of ether was used as control. Traps were hanging from the top of the cage and males were released inside once a day. The bioassay lasted ve days. Every 24 hours, a batch of ve males was released and the traps were replaced. The number of males captured on traps was counted.

Field Experiments
This study was carried out in an insecticide unsprayed vineyard (cultivar Tocai Friulano) located in a hilly grape-growing area (locality Buttrio, 46° 01' latitude N, 13° 21' longitude E, 90 m a.s.l.). Yellow sticky traps (11.5 × 24 cm) smeared with Temoocid® (Kollant S.r.l., Vigonovo, Venice, Italy) were used for A. atomus monitoring Zanolli and Pavan 2011). Six traps baited with eugenol were compared with six control traps. Lures were prepared with 9 µg (1 × 3 cm) with perforated plastic capsUnder laboratory conditions, vials used as bait released about 20 ng of eugenol per day, but under eld conditions greater release is likely. On each baited trap a vial with eugenol was hanged in the upper part while in the control traps empty vials were applied. Traps were distributed in the vineyard following a randomized blocks scheme with six replicates (rows). Traps were replaced twice a week and baits were changed at each replacement. Anagrus atomus male captures at each sampling interval were counted in the laboratory under a dissection microscope. Within A. atomus species, the intraspeci c taxa A. atomus and A. parvus Soyka were also distinguished (Zanolli et al. 2016).

Statistical Analysis
Wasp residence time in the treated and control arms of the olfactometer was compared using a paired one-tailed t-test.
To compare the male response to females of different ages or physiological status (unmated/mated), the proportions of residence time in the treated arm with respect to control arms were calculated and before statistical analysis data were arcsine transformed. To compare the male response to females of different age, ANOVA and Tukey test were performed, whereas to compare the male response to unmated/mated females an unpaired t-test was performed.
Number of males captured on yellow sticky traps, in cage bioassays, was compared with Binomial test.
To compare the total eld captures of A. atomus males on the eugenol baited and control traps over all the monitoring periods, Wilcoxon matched-pairs signed-ranks test was applied considering sampling dates as replications.
Statistical analyses were performed with GraphPad Instat 3.0 for Macintosh.
Moreover, in both Experiments 1 and 2, males exposed to the odour released by a virgin female showed wing fanning and increased locomotory activity. In particular, most males fan their wings in presence of a virgin female, whereas, only a small proportion of those exposed to the odour of a mated female did so.
The comparison between the proportion of time that males spent in the arm of the olfactometer treated with the odour of a virgin or a mated female revealed that one-and two-days old males were signi cantly more attracted to unmated females than to mated ones (unpaired t test, N = 12; 0 days: t = 2.04, P = 0.054, 1 day: t = 3.10, P = 0.005; 2 days: t = 3.04, P = 0.006; 3 days: t = 0.47, P = 0.64; Fig. 3).
In Experiment 3, males did not stay longer in the arm of the olfactometer treated with the odour of a young male (paired t test, N = 12, t = 1.49, P = 0.08; Fig. 1).
In Experiment 4, males showed a strong attraction towards one female equivalent of the crude extract of unmated females (paired t test, N = 16, t = 7.43, P < 0.001. Figure 4). The male response to the extract was similar to that displayed to a living unmated female, although no wing fanning was noted in this case.

Chemical Analysis of A. atomus Crude whole Extract
The comparison of GC-MS analyses of virgin males' and females' crude extracts revealed only minor differences; the most notable difference regarded one peak that was found only in the female extracts (Fig. 5). The peak was identi ed as eugenol by comparison of the mass-spectrum and retention time with of an authentic standard and con rmed by coinjection. The amount of eugenol calculated per female was less than 1 ng.

Cage bioassays
Under lab conditions yellow sticky traps baited with eugenol captured more caged males than control traps (binomial test, P = 0.02. Figure 7).

Discussion
The attraction of A. atomus male wasps to both alive virgin females and their crude extract support the hypothesis that a volatile sex pheromone is released by females similarly to what previously found in other hymenopterous parasitoids (Simser and Coppel 1980;Eller et al. 1984;Nazzi et al. 1995;De Lury et al. 1999;Ruther et al. 2000;Salerno et al. 2012;). Attraction of males, but not females, con rms that this is not an aggregation pheromone.
Olfactometer and cage bioassays suggest that the attractant is a long-range pheromone, since the distance at which it is perceived by A. atomus males is greater than 2 cm (Keeling et al. 2004). The females of other Anagrus species, such as A. delicatus (Cronin and Strong 1990), A. incarnatosimilis and A. breviphragma (Moratorio and Chiappini 1995), mate at the emergence site. For this reason, it is likely they do not produce a long-range attractant in analogy with other hymenopterous parasitoids showing the same reproductive behaviour (Godfray 1994). In fact, in this case, males emerge rst and then wait for the female's emergence to mate on site. In contrast, in the case of A. atomus, that does not mate at emergence site because it parasitizes solitary eggs (Agboka et al. 2004), and where only males are normally produced when females do not meet a male before egg laying, a female produced long range attractant facilitates male search. The natural selection may favour sex pheromone production if the population is not at sex ratio equilibrium or if there are advantages in having the capability of producing both sons and daughters (Godfray 1994). Metabolic costs are involved in the production of long-range sex pheromones and so it is likely that A. atomus females gain advantage from mating.
Although some difference between unmated and mated females was noted, they both appeared to be attractive to males. Therefore, the production/release of the pheromone by A. atomus females seems not to cease after mating as would be expected if multiple matings occurred. Polyandry is known in other Mymarids such as, for example Anaphes nitens (Santolamazza-Carbone and Pestaña 2010) and a second mating was occasionally observed also in Anagrus breviphragma (Moratorio and Chiappini 1995). The tness returns from multiple matings of old females, are not clear since egg supply and quality usually decrease with mother age (Giron and Casas 2003); nevertheless, a high tendency to remate of old individuals was observed in other wasps such as Nasonia vitripennis (Burton-Chellew et al. 2007) and could be related to the limited storage capacity of the spermatheca (Santolamazza-Carbone and Pestaña 2010). In hymenopteran haplodiploid species, sperm depletion followed by the oviposition of unfertilized male eggs is possible (Godfray 1990). In this case, since the evolutionary stable strategy sex ratio in panmitic populations is 1:1, as reported for A. atomus in previous studies (Zanolli and Pavan, 2011), sperm-depleted females are likely to re-mate thus explaining the persistence of female produced attractant after mating. The release of sex pheromone by mated females could be convenient if multiple matings increase lifespan of females, as observed in Trichogramma evanescens (Jacob and Boivin 2005), or if sex pheromones have multiple functions as in Venturia canescens (Metzger et al. 2010).
The compound triggering male attraction was identi ed as eugenol (4-allyl-2-methoxyphenol), a sex pheromone component previously detected in male butter ies of genus Amauris spp. (Lepidoptera, Danaidae) (Schulz et al. 1993) and in Bactrocera spp. (Symonds et al. 2009). No more compounds were identi ed in the active fraction of the crude extract but the limited quantity of material extractable from females may have precluded the identi cation of other substances. In fact, the magnitude of the biological effect elicited by several dosages of the compound appeared to be smaller than that observed against alive females. The presence of eugenol only in the female extract was con rmed also by analysing males and females of A. atomus emerged from grapevine leaves collected in the eld (data not reported).
Furthermore, wing fanning has never been observed when extracts or pure compounds were tested in the bioassay. Since many females Hymenopterous parasitoids produce sex pheromones that are constituted by a blend of two or more compounds, one attracting the males and others mediating the subsequent phases of the courtship behavior (Quicke 1997;Ruther 2013), the hypothesis that more compounds involved in A. atomus reproduction are still to be identi ed is very likely. Nevertheless, this is the rst study in which a sex-pheromone component was identi ed in Mymaridae.
Under open eld conditions, eugenol signi cantly increased the already elevated captures of A. atomus males in yellow sticky traps, con rming the biological activity detected under lab conditions. Unbaited yellow sticky traps have already proved to be a valid tool to monitor eld populations of A. atomus wasps Zanolli and Pavan 2011); a complete identi cation of the pheromone blend of the insect could signi cantly increase the e ciency of such traps and their potential for monitoring of this bene cial insect as bioindicator in the contest of studies on the effects of chemical control and cultural practices on natural enemies (Gabrýs et al. 1997;Suckling et al. 2002).