Composition of Strawberry Floral Volatiles and their Effects on Behavior of Strawberry Blossom Weevil, Anthonomus rubi

The strawberry blossom weevil (SBW), Anthonomus rubi, is a major pest in strawberry fields throughout Europe. Traps baited with aggregation pheromone are used for pest monitoring. However, a more effective lure is needed. For a number of pests, it has been shown that the attractiveness of a pheromone can be enhanced by host plant volatiles. The goal of this study was to explore floral volatile blends of different strawberry species (Fragaria x ananassa and Fragaria vesca) to identify compounds that might be used to improve the attractiveness of existing lures for SBW. Floral emissions of F. x a. varieties Sonata, Beltran, Korona, and of F. vesca, were collected by both solid-phase microextraction (SPME) and dynamic headspace sampling on Tenax. Analysis by gas chromatography/mass spectrometry showed the floral volatiles of F. x ananassa. and F. vesca were dominated by aromatic compounds and terpenoids, with 4-methoxybenzaldehyde (p-anisaldehyde) and α-muurolene the major compounds produced by the two species, respectively. Multi-dimensional scaling analyses separated the blends of the two species and explained differences between F. vesca genotypes and, to some degree, variation between F. x ananassa varieties In two-choice behavioral tests, SBW preferred odors of flowering strawberry plants to those of non-flowering plants, but weevils did not discriminate between odors from F. x ananassa and F. vesca flowering plants. Adding blends of six synthetic flower volatiles to non-flowering plants of both species increased the preference of SBW for these over the plants alone. When added individually to non-flowering plants, none of the components increased the preference of SBW, indicating a synergistic effect. However, SBW responded to 1,4-dimethoxybenzene, a major component of volatiles from F. viridis, previously found to synergize the attractiveness of the SBW aggregation pheromone in field studies.


Introduction
The strawberry blossom weevil (SBW), Anthonomus rubi Herbst, (Coleoptera, Curculionidae) is an oligophagous species that feeds and reproduces on rosaceous plants (Popov 1996). Among its hosts are strawberry, Fragaria spp., raspberry, Rubus idaeus L., blackberry, Rubus spp. and rose, Rosa spp. (Hill 1987;Popov 1996). Early in spring, adults move to strawberry and raspberry from overwintering shelters, both inside the cropping area and from perimeter wild host plants (Alford 1984). There, they feed on foliage, flower buds and open flowers. The female weevil usually oviposits a single egg in an unopened bud (Aasen et al. 2004;Jary 1931;Leska 1965) before severing the bud petiole, either partially or completely, preventing further development of the bud. The oviposition period lasts 1-2 months, during which more than 150 buds can be destroyed per female (Easterbrook et al. 2003). The larva develops and pupates inside the withered bud and emerges in late summer. After emergence, the young adult feeds on foliage for a few weeks before moving to an overwintering site (Hill 1987).
For many insects, pheromones are key cues in mate finding (Yew and Chung 2015), and can play an important role in integrated pest control programs (Suckling et al. 2014). The male-produced aggregation pheromone of SBW was identified as a blend of grandlure I, grandlure II and lavandulol in a 1:4:1 ratio by . This pheromone blend is attractive to both sexes and is used as a lure in commercial monitoring of SBW (Cross et al. 2006a, b). However, recent data suggested that, in order to control SBW populations, more effective lure formulations are needed (Baroffio et al. 2018). For a number of pests, the attractiveness of a pheromone can be enhanced by host plant volatiles (Reddy and Guerrero 2004). This has been demonstrated for other weevils of the genus Anthonomus (Dickens 1989;Muniz-Merino et al. 2014). Furthermore, it has been shown that combining a floral volatile, 1,4-dimethoxybenzene, with the aggregation pheromone of A. rubi, gives improved trap catches (Wibe et al. 2014). This compound is now included in commercially available lures.
The objective of this study was to identify other hostplant compound(s) that might be combined with the SBW aggregation pheromone to improve attraction of SBW. We aimed: (i) to determine whether SBW is able to discriminate between volatiles from flowering versus nonflowering strawberry plants of two species, Fragaria x ananassa and F. vesca; (ii) to determine whether SBW prefers volatiles of flowering F. x ananassa over those of flowering F. vesca; (iii) to identify components of the volatile floral blends of F. x ananassa and F. vesca; and (iv) to determine whether selected components of these floral volatiles elicit behavioral responses by SBW in a laboratory bioassay when tested individually or in blends.

Insects and Plants
Adult SBW were collected from strawberry fields in SE Norway (N59.66, E10.69) in mid-May and transported to the laboratory at the Royal Institute of Technology, Stockholm. Both sexes were kept together on potted plants of non-flowering Fragaria x ananassa Duchesne (Rosales: Rosaceae), variety Sonata, enclosed in 400 cm 3 plastic cups covered with nylon mesh. A moistened piece of cotton at the bottom of the cup served as a water source. The day before experiments, weevils were sexed and kept individually in plastic cups without plants. Sex was determined by the presence of a thorn on each intermediary coxa of male weevils (Innocenzi et al. 2002).

Sampling and Analysis of Floral Volatiles
We used solid-phase microextraction (SPME; Rout et al. 2012) to sample volatiles of three varieties of F. x ananassa β-P h e l l a n d r e n e M T 1 2 0 7 L , R I t r t r 2 ± 2 a 9 ± 4 a § 6 (Z)-β-O c i m e n e M T 1 2 3 9 R C 0 0 0 t r § Methyl benzoate AR 1602 RC 13 ± 2 a 9 ± 5 a 8 ± 1 a 12 ± 6 a § 2 5 3 , 6 , 6 -T r i m e t h y l -2 -n o r p i n a n o n e O M T 1 6 1 8 L , R I 0 0 0 1 1 ± 6 26 Acetophenone AR 1630 RC tr 3 ± 1 a 1 ± 1 a 13 ± 13 a 27 1-Ethenyl-4-methoxy-benzene AR 1661 RC 3 ± 2 b 26 ± 8 a 0 0 28 3-Ethyl-benzaldehyde AR 1690 RC 8 ± 3 b 5 ± 2 b 6 ± 2 b 24 ± 5 a 2 9 G e r m a c r e n e D S T 1 6 9 6 R C 0 0 0 2 ± 1 § and flowers of different genotypes of F. vesca using the same fiber type, headspace volume, and temperature during sampling. An internal standard was used to check for saturation of fibers. This is a sensitive technique and provided quantitative data for statistical comparisons among varieties and species. Even so, comparison of quantities of different compounds in the same sample is not possible without use of labelled standards, because of different affinities of the fiber for compounds and different vapor pressures of compounds (SPME guidelines 2020). Thus, dynamic headspace sampling with an internal standard was also used for more reliable quantification of compounds and release rates. S P M E s a m p l i n g w a s c a r r i e d o u t w i t h polydimethylsiloxane-divinylbenzene-coated fibers (65 μm; Supelco, Sigma-Aldrich group, PA, USA) (Vas and Vekey 2004), desorbed at 250°C for 2 min in a gas chromatograph (GC) injector prior to sampling. A single, ready-to-flower bud from a potted strawberry plant was placed in a glass chamber (30 cm 3 ) through an opening at the bottom, along with a filter paper (1 cm 2 ) treated with 100 ng of pentadecane per 10 μl of cyclohexane as internal standard. The opening was carefully sealed with aluminum foil. After the bud had opened, the SPME fiber was introduced close to the flower through a second opening in the chamber. Sampling of volatiles was carried out from 08.00 until 18.00 h covering the main period of emission (Ceuppens et al. 2015). After collection, the fiber was removed and desorbed directly in the injector of the gas chromatograph/mass spectrometer (GC/MS). Volatiles were collected from individual flowers of F. ananassa variety Sonata (N = 3), F. ananassa variety Beltran (N = 4), F. ananassa variety Corona (N = 4), and F. vesca (N = 9).
Dynamic headspace sampling (Millar and Haynes 1998) was carried out on individual flowers of F. ananassa variety Sonata (N = 3) and F. vesca (N = 3). A single, ready-to-flower bud was placed in the type of glass jar used for SPME sampling. Charcoal-purified and humidified air (50 ml.min −1 ) was supplied by a diaphragm vacuum pump (NMP 830 KNDCB; KNF Neuberger Inc., Freiburg, Germany), and pulled through a glass collection tube containing Tenax TA adsorbent (50 mg; 60/80 mesh; Sigma-Aldrich AB, Sweden) by a second pump. Volatiles were collected from 08.00-18.00 h as above. After sampling, traps were extracted with 250 μl of redistilled diethyl ether, and 20 ng of pentadecane in cyclohexane was added as internal standard. Samples were concentrated under a gentle flow of nitrogen and analyzed on the same day as collected.
After sampling, flowers were detached from the peduncle, placed in a glass beaker and kept in a thermostat at 60°C for 72 h for determination of dry weight. For both sampling  techniques, volatiles were also collected from empty glass chambers. Samples were analyzed using a Varian 3400 GC coupled to a Finnigan SSQ 7000 MS (Thermo-Fisher Scientific, USA). A DB-Wax fused silica capillary column (30 m length, 0.25 mm id, 0.25 μm film thickness; Supelco-Sigma-Aldrich group, USA) was used. The column oven was programmed from 40°C for 3 min, then at 4°C.min −1 to 200°C, then at 10°C.min −1 to 230°C, and held for 9 min. The split/splitless injector temperature was 225°C and the splitless period lasted for 60 s. Helium was used as carrier gas with an inlet pressure of 70 kPa. The transfer line temperature was 235°C. Electron ionization mass spectra were determined at 70 eV with an ion source at 150°C. Chromatographic profiles of volatiles were compared and compounds in large amounts, relative to those in blank samples, identified by comparison of mass spectral data and retention indices with synthetic standards (Table 2), along with data from NIST version 2.0 mass spectral search program (National Institute of Standards and Technology, USA). Relative amounts of compounds were determined as areas under chromatographic peaks. Absolute amounts of compounds trapped by aeration were quantified by applying standard calibration curves derived from pentadecane at 0.5 ng, 1 ng, 10 ng and 50 ng.

Behavioral Tests
Behavioral responses of SBW to natural and synthetic odors were tested in two-choice olfactometers. The test area was illuminated with a quartz metal halide lamp ( H P I -T P l u s 4 0 0 W ; P h i l i p s , A m s t e r d a m , t h e Netherlands) placed 180 cm above the olfactometers. The olfactometers comprised three layers of acrylic plastic (each layer 0.5 cm thick) sandwiched together with an arena cut out in the middle layer and consisting of a central zone (2 × 2.5 cm) with two tapered arm zones (4 cm length from the air inlet to the central zone and 0.4 cm to 2.5 cm width at the inlet and the central zone, respectively) (Hambäck et al. 2003) (Fig. 5a). One vacuum diaphragm pump delivered stimuli in purified and humidified air to the arms of the olfactometer and another pump was connected via Teflon tubes to the top so as to withdraw air at ca. 3 ml.sec −1 . Four olfactometer trials could be run simultaneously, testing an odorant of the same type. Before each trial, a weevil was allowed to acclimatize inside the olfactometer for 3 min. During delivery of the stimulus, the weevil's position in the arena was noted at 30 s intervals for 15 min, giving a total of 30 recordings for each individual weevil. The responsiveness of a SBW was assessed by calculating the percentage of records in each arm. Each weevil was only used once in an experiment, and weevils that were inactive in the central zone for more than 5 min during trials were excluded from the analysis. Between trials, the olfactometers were washed with water and a mild detergent.
In the first experiment, weevil preference to the olfactometer arms without olfactory cues was tested to confirm lack of bias in the olfactometer. In the second experiment, weevil responses to odors from flowering versus non-flowering strawberry plants were evaluated. For the non-flowering plants, buds or flowers were removed 2 d prior to the experiment. A single potted flowering and non-flowering F. x ananassa plant were placed separately in a polyester cooking bag (25 × 40 cm; Toppits, Minden, Germany), while 4-5 potted flowering and non-flowering F. vesca plants (collected at N59 o 21.412, E18 o 17.118) of total leaf area approximately equal (estimated by visual evaluation) to that of the F. x ananassa plant were placed together in another polyester cooking bag. One F. x a. var. Sonata flower corresponded to four F. vesca flowers, based on dry weight. The pots were covered with aluminum foil in order to minimize soil volatiles in the headspace. Purified and humidified air was delivered at a ca. 12 ml.sec −1 from the bottom of the bag, and volatiles were collected at the top of the bag.
In the third experiment we tested preference for flowering F. vesca versus flowering F. x a. var. Sonata plants, using five wild strawberry plants with eleven flowers and one garden strawberry plant with three flowers, respectively.
In the fourth experiment, the responses of SBW to synthetic chemicals added to non-flowering plants were tested against plants alone. Compounds were dissolved in cyclohexane and 10 μl of solution applied to 2 cm 2 of filter paper in a glass vial (10 cm 3 ) positioned close to the plant. Single compounds were tested at a dose of 100 ng. Two six-component blends of synthetic compounds were made to represent floral bouquets of F. x ananassa var. Sonata (FAS) and F. vesca (FV), r e s p e c t i v e l y . T h e F A S b l e n d c o m p r i s e d 4methoxybenzaldehyde, benzaldehyde, benzyl alcohol, methyl salicylate, limonene and decanal at loadings of 200, 15, 51, 20, 9 and 40 ng, respectively. The FV blend comprised αmuurolene, benzaldehyde, benzyl alcohol, methyl salicylate, limonene and decanal at loadings of 30, 25, 45, 60, 70 and 80 ng, respectively. The loading ratios and amounts of synthetic odorants were selected to emit profiles similar to those of strawberry flowers. Five components in these two synthetic blends were common, consistent with the observation that SBW showed no preference for flowering plants of either species. All five compounds were previously shown to elicit electrophysiological responses from the antennae of SBW by Bichão et al. (2005a). Furthermore, these compounds represent different classes of floral volatiles and would be economically feasible for practical control use. In addition, the FAS blend included 4-methoxybenzaldehyde, the major component in floral volatiles from cultivated strawberry, while the In the fifth experiment, the responses of SBW to 4methoxybenzaldehyde added to non-flowering F x ananassa var. Sonata plants at doses of 10 ng, 100 ng and 1000 ng were investigated.

Data Analysis
To monitor saturation of SPME fibers, the amounts of pentadecane adsorbed were compared by Mann Whitney U test, using Statistica software version 6.0.
To compare amounts of floral volatiles between F. x ananassa varieties and F. vesca, data were analyzed by Kruskal-Wallis a non-parametric test followed by a Conover-Iman test using R (version 4.0.2) and Rstudio (version 1.3.959).
To assess and visualize associations among odor blends of strawberry flowers sampled by SPME, a multidimensional scaling (MDS) analysis with a Bray-Curtis index was performed on absolute amounts, expressed as areas under chromatographic peaks, using R (version 4.0.2) and Rstudio (version 1.3.959), with the metaMDS function in the vegan package (version 2.5-6). The results were visualized using ggplot2 (version 3.3.2). Prior to analysis, the data were square root transformed. Amounts of volatiles were also used to show degree of similarity of odor bouquets between F. x ananassa varieties and F. vesca by cluster analysis, based on Euclidian distance using Statistica software version 6.0.
The behavioral responses of SBW in two-choice olfactometer tests were analyzed by nonparametric Wilcoxon matchedpairs signed-ranks test using the Statistica software version 6.0.

Chemical Composition and Variation of Strawberry Floral Odor Blends
Using SPME, 46, 39 and 49 compounds were detected in floral volatiles from F. x ananassa varieties Sonata, Beltran, and Korona, respectively. In the flower headspace of F. vesca, 41 compounds were detected (Table 1). Monitoring saturation of SPME fibers by adding 100 ng of pentadecane as internal standard to the samples revealed no differences in amount of pentadecane trapped on the fibers in blank samples (empty glass jars; median TIC count 17,934,247) versus F. vesca flowers (median TIC count 15,258,935) and versus flowers of F. x ananassa variety Sonata (median TIC count 11,565,433) (Mann Whitney U test, N = 5, P = 0.222 and N = 3, P = 0.071, respectively).
Using the dynamic headspace sampling technique, 49 compounds were detected in floral volatiles collected from F. x ananassa variety Sonata and 18 compounds from F. vesca (Table 2). For both techniques, all samples were from single flowers. GC/MS analyses revealed that floral volatile blends of both species were dominated by aromatic compounds and terpenoids. From the flowers of F. x ananassa, 4methoxybenzaldehyde (p-anisaldehyde) was collected in the largest quantity, and from F. vesca, α-muurolene (Fig. 1,  Tables 1 and 2).
Intensity (arbitrary units. Number has to be multiplied by 10 )  SBW showed no preference for either olfactometer arm when no olfactory cue was present, showing no inherent bias in the bioassay (Fig. 5b). When given the choice between odors of flowering F. x ananassa var. Sonata plants and odors from non-flowering plants, both male and female SBW preferred those from the flowering plant (N = 16, Z = 2.811, P = 0.005 and N = 14, Z = 1.977, P = 0.048, respectively) (Fig. 5c). In further experiments, SBW were not separated by sex. When testing SBW preference to odors of flowering versus non-flowering F. vesca, more weevils chose odors of flowering plants (N = 14, Z = 2.068, P = 0.041). No preference was observed between odors of flowering F. x a. var. Sonata and F. vesca plants (N = 11, Z = 0.044, P = 0.965) (Fig. 5c).
In the final bioassay experiment, SBW did not discriminate between odors released from non-flowering F. x a. var. Sonata strawberry plants and those from non-flowering plants of the same variety with added 4-methoxybenzaldehyde at 10 ng or 100 ng doses (N = 16, Z = 1.172, P = 0.241 and N = 12, Z = 1.579, P = 0.114, respectively). When the dose was increased to 1000 ng, weevils preferred the side with the non-flowering plant alone (N = 10, Z = 2.803, P = 0.005) (Fig. 5e).

Chemical Composition and Variation of Strawberry Floral Odor Blends
The composition of floral volatile emissions from several species of strawberry have been reported previously, including Fragaria x ananassa (Bichão et al. 2005a;Ceuppens et al. 2015;Hamilton-Kemp et al. 1990, 1993Klatt et al. 2013), F. virginiana Duchesne (Ashman et al. 2005, F. vesca (Blažytė-Čereškienė et al. 2017;Wibe et al. 2014) andF. viridis Duchesne (Blažytė-Čereškienė et al. 2017). In those studies, static and dynamic headspace collections as well as hydro-distillation techniques were used to sample floral compounds from cut and intact flowers, making it difficult to compare data. The amounts of volatiles produced by individual strawberry flowers are very small; hence, we used SPME to sample volatile profiles under standardized conditions. This provided a highly sensitive technique able to detect more compounds than in the above studies, and also provided quantitative data for comparisons of varieties and species. Dynamic headspace sampling with an internal standard was also used for more reliable quantification of compounds and release rates. We found that 4-methoxybenzaldehyde (p-anisaldehyde) was the major constituent of the floral volatile blends released by all three F. x ananassa varieties. This contrasts with previous reports in which benzaldehyde was present in the largest amount in volatile emissions of F. x ananassa varieties Darselect, Honeoye (Klatt et al. 2013) and Korona (Bichão et al. 2005a). (E,E)-α-Farnesene and limonene were reported as the major component of floral emissions of the variety Sonata by Klatt et al. (2013) and Ceuppens et al. (2015), respectively. In our study, these compounds were present at lower amounts in Sonata flowers. The reason for the quantitative differences in these studies is unknown.
In our study, α-muurolene was the major component in volatiles from flowers of F. vesca. 1,4-Dimethoxybenzene was reported as the major constituent of F. The SPME data were used to carry out MDS analyses, which separated volatile blends released from F. x ananassa and F. vesca flowers. The analyses also explained differences between F. vesca genotypes, and partly explained the variation among the volatile blends from the three varieties of F. x ananassa varieties. Flowers of variety Sonata had the most distinct odor blend. Single cell recordings had previously revealed 58 identified and a few unidentified compounds that elicited responses of SBW olfactory receptors (Bichão et al. 2005a, b). We detected 16 of these compounds in strawberry floral volatile emissions.
Despite the differences in composition of blends of volatiles from the flowers of different species and varieties of strawberry, there were at least 13 common compounds. Five of the most abundant were benzaldehyde, benzyl alcohol, methyl salicylate, limonene and decanal. These were combined with 4-methoxybenzaldehyde and α-muurolene to give blends representative of F. x ananassa (FAS) and F. vesca (FV), respectively, for testing in bioassays. All these compounds elicited electrophysiological responses from antennae of SBW (Bichão et al. 2005a).

Olfactory Preferences of Anthonomus rubi
Our data showed that SBW preferred odors of flowering strawberry over those of non-flowering plants. This was somewhat surprising, as SBW females oviposit in flower buds prior to opening. Possibly, floral volatiles are detected from the bud before opening, or weevils are attracted to the area by neighboring flowers which have already opened. A similar preference for odors released from flowering over nonflowering hosts was also reported for cranberry weevils, Anthonomus musculus Say (Szendrei et al. 2009), which have oviposition and feeding strategies similar to SBW. However, only female A. musculus showed preference to odor blends released by blueberry flowers over those of flower buds (Szendrei et al. 2009). In our study, we did not detect any sex differences in preference.
Addition of a six-component blends of chemicals to nonflowering plants of both cultivated F. x ananassa var. Sonata and wild species F. vesca, increased attractiveness to SBW, relative to non-flowering plants alone. The FAS blend mimicking the blend from cultivated strawberry included 4methoxybenzaldehyde as the major component, while the major component in the FV blend mimicking wild strawberry was α-muurolene. The other five components in these two synthetic blends were the same, consistent with the observation that SBW shows no preference for flowering plants of either species.
However, none of the compounds in the two sixcomponent blends increased the preference of SBW when tested as a single compound added to non-flowering plants. This indicates a synergistic action of the volatiles, a common phenomenon in insect behavioral responses to host plant volatiles (Bruce and Pickett 2011;Richards et al. 2016;Sarkar et al. 2017). Olfactory synergism between floral volatiles is less frequently reported compared to odors of vegetative plant parts, possibly due to the activity of individual floral components rarely having been examined (Metcalf et al. 1995;Richards et al. 2016). 4-Methoxybenzaldehyde, the major component of floral volatiles from F. x ananassa, actually reduced the attractiveness of non-flowering plants when added at the highest dose of 1000 ng. A similar phenomenon was reported by Webster et al. (2010), showing that individual components of attractive blends of host-plant volatiles can have repellent activity when presented at higher than natural doses and outside the context of the natural host blend.
1,4-Dimethoxybenzene increased the attractiveness to SBW of non-flowering F. vesca plants when added as a single compound. This compound was previously identified as a major component of floral volatiles of F. viridis (Blažytė-Čereškienė et al. 2017;Wibe et al. 2014) and increased the attractiveness of SBW aggregation pheromone in field trials (Baroffio et al. 2018;Wibe et al. 2014). This provides encouragement that our bioassay results are relevant to the field. We plan to test the active blends from the bioassays for attractiveness to SBW and/or synergism of the aggregation pheromone in field trials.