Identification of volatile components from oviposition and non-oviposition plants of Gasterophilus pecorum (Diptera: Gasterophilidae)

Oviposition by Gasterophilus pecorum on shoot tips of Stipa caucasica is a key determinant of its severe infection of the reintroduced Przewalski’s horse (Equus przewalskii). Volatiles in shoots of grasses on which Przewalski’s horse feeds, including S. caucasica at preoviposition, oviposition, and postoviposition stages of G. pecorum, S. caucasica, Stipa orientalis, and Ceratoides latens at the oviposition stage, and S. caucasica in various growth periods, were collected by dynamic headspace adsorption and analyzed by automatic thermal desorption gas chromatography-mass spectrometry. Among five volatiles with highest relative contents under three sets of conditions, caprolactam and 3-hexen-1-ol,(Z)- were common to all samples. Caprolactam was highest in C. latens at oviposition stage of G. pecorum and lowest in S. caucasica at postoviposition stage, and that of 3-hexen-1-ol,(Z)- was lowest in C. latens and highest in S. caucasica at its oviposition stage. Particularly, in S. caucasica during the three oviposition phenological stages of G. pecorum, 3-hexen-1-ol,acetate,(Z)-, 2(5H)-furanone,5-ethyl-, and 3-hexen-1-ol,acetate,(E)- were unique, respectively, to the preoviposition, oviposition, and postoviposition stages; in three plant species during the oviposition stage of G. pecorum, 3-hexen-1-ol,acetate,(Z)-, 3-hexenal, and 1-hexanol were unique to S. orientalis, acetic acid, hexanal, and 2(5H)-furanone,5-ethyl- to S. caucasica, and 1,3,6-octatriene,3,7-dimethyl-, cis-3-hexenyl isovalerate, and acetic acid hexyl ester to C. latens; in S. caucasica, 2-undecanone,6,10-dimethyl- was unique to the early growth period, acetic acid and 2(5H)-furanone,5-ethyl- to the flourishing growth period, and 3-hexen-1-ol,acetate,(Z)- and 1,3,6-octatriene,3,7-dimethyl- to the late growth period. Furthermore, substances specific to S. orientalis and C. latens were also present in S. caucasica, except at oviposition stage. Our findings will facilitate studies on G. pecorum’s adaptation to the arid desert steppe and its future control.


Results
Volatile contents of S. caucasica shoots during the stages of oviposition by G. pecorum. Overall, 60 volatile compounds were identified in S. caucasica shoots during the preoviposition (I), oviposition (II), and postoviposition (III) stages of G. pecorum. These comprised 16 aldehydes, 14 ketones, 12 esters, 9 alcohols, 3 alkanes, 3 aromatic hydrocarbons, 1 acid, 1 ether, and 1 other. Among them, 35 volatiles were identified in I-L, 36 in II-L, and 37 in III-L. In addition, 18 volatiles were common to I-L, II-L, and III-L; 5 to I-L and II-L; 5 to II-L and III-L; and 2 to I-L and III-L. Ten volatiles were unique to I-L, 8 to II-L, and 12 to III-L ( Table 1). The main chemical classes of I-L, II-L, and III-L were alcohols, esters, and others; alcohols and others; and alcohols and esters, respectively (Fig. 1).
Fourteen ketones were identified from the three stages of S. caucasica. Among them, two, i.e., 5-hepten-2-one,6-methyl-and acetophenone, were common to all three stages; and three, i.e., 2(3H)-furanone,dihydro-5-methyl-, 2-hexanone,4-methyl-, and benzophenone, were common to two of the three stages. The relative contents of ketones were lower in I-L (0.86%) than in II-L (3.65%) or III-L (2.08%) (P = 0.022 and P > 0.05), with www.nature.com/scientificreports/ no significant difference between II-L and III-L (P > 0.05) (Fig. 1D). The content of 2(5H)-Furanone,5-ethyl-was specific to II-L (2.38%), and the relative contents of the other ketones were < 0.7% ( Table 1). The relative content of caprolactam, the only volatile in the class 'others,' was 30.66% and 22.68% in I-L and II-L, respectively; there was no significant difference between values for I-L and II-L (P > 0.05), and both were higher than those for III-L (12.9%) (P = 0.017 and P > 0.05, respectively) (Fig. 1E). The relative content of acetic acid, the only volatile in the class of acids, was lower in III-L (0.61%) than in II-L (3.36%) or I-L (2.14%) (P = 0.022 and P > 0.05, respectively); there was no significant difference between the latter two (P > 0.05). The relative contents of alkanes, aromatic hydrocarbons, and ethers were less than 0.22% ( Fig. 1G-I). These included three alkanes, one in I-L and two each in II-L and III-L; three aromatic hydrocarbons, one of them specific to each stage; and one ether, which was not found in III-L ( Table 1).
Relative contents of volatiles in three plant species during the oviposition stage of G. pecorum. During the oviposition stage of G. pecorum, a total of 60 volatiles were identified in S. orientalis (II-D), S. caucasica (II-L), and C. latens (II-T). These comprised 18 esters, 13 aldehydes, 11 alcohols, 10 ketones, 2 alkanes, 2 aromatic hydrocarbons, 1 acid, 1 alkene, 1 ether, and 1 other. Of these, 35 were identified in II-D, 36 in II-L, and 27 in II-T. In addition, 11 were common to II-D, II-L, and II-T, 14 to II-D and II-L, and 2 to II-L and II-T; 10 were unique to II-D, 9 to II-L, and 14 to II-T ( Table 2). The main chemical classes of II-D and II-L were alcohols and others, and those of II-T were alcohols, esters, and others (Fig. 2).

Relative contents of volatiles from S. caucasica in different growth periods. From S. caucasica
at the early, flourishing, and late growth periods (GP1, GP2, and GP3, respectively), a total of 69 volatile compounds were identified. These comprised 17 ketones, 13 aldehydes, 11 esters, 10 alcohols, 4 alkanes, 4 aromatic hydrocarbons, 2 acids, 2 alkenes, 1 ether, and 5 others. Of these, 35 were found in GP1, 36 in GP2, and 40 in GP3. In addition, 11 were common to all three stages, 10 to both GP2 and GP3, 6 to both GP1 and GP2, and 4 to both GP1 and GP3; 14 were unique to GP1, 9 to GP2, and 15 to GP3 ( Table 3). The main chemical classes of GP1 and GP2 were alcohols and others, and those of GP3 were esters and others (Fig. 3).
Four alkanes were identified, and the relative contents of individual alkanes ranged from 0.06% to 0.89%. The relative contents of all alkanes were higher in GP1 (1.56%) than in GP3 (0.15%) (P = 0.022), with no significant difference between GP2 (0.22%) and GP1 or GP3 (both P > 0.05) (Fig. 3G). Two alkenes were found only in GP3; they had a total relative content of 4.76% (Fig. 3J); one, 1,3,6-octatrine,3,7-dimethyl-, accounted for 4.70% of this total. The relative aromatic hydrocarbon and ether contents were < 0.4% (Fig. 3H and I). Four of the former were detected, each unique to one of the three periods; one of the latter was identified, only in GP2 ( Table 3).

Discussion
G. pecorum is one of the most pathogenic species of Gasterophilus spp., which is distributed in Europe, Africa, and Asia. A large number of G. pecorum larvae attached to a horse can cause inflammation and dysphagia and may result in death due to esophageal contraction 22 (pp. 526-528). A previous study 9 showed that G. pecorum was the only Gasterophilus spp. that oviposited on grass. Liu et al. 8 found that G. pecorum oviposited on tips of S. caucasica shoots in KNR, whereas our team also found that S. orientalis of the same genus was rare in the area, and eggs of G. pecorum were not found on its tip. Selection of an oviposition site by G. pecorum might be related to the behavior of Przewalski's horses as well as the availability of food and water 8 . S. caucasica is the dominant plant species in KNR, and S. caucasica and C. latens are the main food sources for Przewalski's horse 12 . During infection by G. pecorum, the proportion of Przewalski's horse and Mongolian wild ass feeding on S. caucasica is increasing (unpublished data). High densities of equine feces and G. pecorum eggs are found in the vicinity of water sources, making these areas the main transmission sites of Gasterophilus myiasis in the local area 8,23 . The Przewalski's horse drinks water daily, and its activity range is restricted to the vicinity of water sources 24,25 . In contrast, Mongolian wild ass is more drought tolerant and has a wider activity range 26 , which may explain why a greater number of Przewalski's horses than of Mongolian wild asses is infected by Gasterophilus spp. The lifespan of adult G. pecorum is 1-4 days; its longest survival time is shorter than that of other species of the genus. Each G. pecorum lays 1,300-2,425 eggs, more than other species of the genus 9 . Chereshnev 27 (pp. 765-768) reported that in Kazakhstan, G. pecorum lays 10-15 eggs at each site of grass, whereas G. pecorum lays 1-10 eggs at each S. caucasica tip in the wild release area of the Przewalski's horse in Xinjiang, with an average of 4 eggs per tip 21 ; hence, G. pecorum is more likely to contaminate the whole pasture. G. pecorum's oviposition strategies including not chasing the host, ovipositing a large number of eggs, and scattering oviposition are important for its adaptation to the arid desert steppe and facilitate its infection of Przewalski's horse.
Host plant volatiles may induce gravid insects to land 28 . The species and quantities of volatiles released by a host plant are influenced by a variety of factors, such as the species, tissue, and organ, physiological state, phenological state, circadian rhythm, and environment 17,[29][30][31][32][33] . In this study, 35 volatiles were identified from GP1, I-L, and II-D, 36 from II-L/GP2, 37 from III-L, 40 from GP3, and 27 from II-T. Insects can detect or identify only a small portion of the volatiles released by their host plants. For example, Bruce and Pickett 20 reported that in a complex food-source plant, volatile mixtures typically include 3-10 candidate key compounds for host recognition.
The relative contents of alcohols were significantly higher in II-L than in I-L or III-L, but significantly lower in II-T than in II-L or II-D, and in GP3 than in GP1 or GP2. Furthermore, 3-hexen-1-ol,(Z)-was one of two volatiles common to all samples, and its relative content was highest in II-L and lowest in II-T. Additionally, 3-hexen-1-ol,(Z)-is strongly attractive for adult Agrilus planipennis Fairmaire, 1888 (Coleoptera: Buprestidae) and has been used in trapping pests and monitoring population dynamics in forests 34,35 . Cui et al. 36 reported that the relative content of 3-hexen-1-ol,(Z)-was high in leaves of Malus halliana Koehne, 1890 and Malus domestica Borkh., 1803, the host plants for adult Agrilus mali Matsumura, 1924 (Coleoptera: Buprestidae). Also, the combination of 3-hexen-1-ol,(Z)-with egg-yellow and light-green sticky trap plates reportedly enhances attraction of A. mali 37 . Finally, 3-hexen-1-ol,(Z)-is also a plant lure used by Batocera horsfieldi Hope, 1839 (Coleoptera: Cerambycidae) and Anoplophora chinensis Forster, 1771 (Coleoptera: Cerambycidae) 38 .
The relative contents of esters were significantly lower in II-L than in I-L or III-L, and in II-L than in II-D or II-T, but they were significantly higher in GP3 than in GP1 or GP2. Among them, 3-hexen-1-ol,acetate,(Z)-, which was unique to I-L, II-D, and GP3, was one of the top five. The attractant effect of purple prism traps with 3-hexen-1-ol,(Z)-for adult A. planipennis was greater than those with unbaited controls, but addition of 3-hexen-1-ol,acetate,(Z)-did not yield a synergistic effect 35 . However, 3-hexen-1-ol,(Z)-and 3-hexen-1-ol,acetate,(Z)-are synthesized by fatty acid derivatization 39 , and 3-hexen-1-ol,(Z)-can be esterified by an alcohol acyltransferase to 3-hexen-1-ol,acetate,(Z)- [40][41][42] . In this study, the low relative content of esters in II-L was caused by the absence of 3-hexen-1-ol,acetate,(Z)-, possibly as a result of inhibition of alcohol acyltransferase activity.
Caprolactam The relative contents of aldehydes were similar in all samples. Among them, 3-hexenal, one of the top five aldehydes, was detected in all samples except II-L and II-T. Hexanal, which was detected only in II-L, III-L, and GP1, was also one of the top five, and this compound elicits a strong electroantennogram response by both male and female adult Protaetia brevitarsis Lewis, 1879 (Coleoptera: Scarabaeidae) 47 . Hexanal is a strong attractant for Sphaerophoria menthastri Linnaeus, 1758 (Diptera: Syrphidae), Plutella xylostella Linnaeus, 1758 (Lepidoptera: Plutellidae), and Chrysopa septempunctata Wesmael, 1841 (Neuroptera: Chrysopidae) 48,49 , and its content is positively correlated with the number of Locustoidea insects 50 .
Species-specific 51 and ratio-specific 52 odor recognition are the means by which phytophagous insects recognize host plants. The selection ratio of male G. pecorum to S. caucasica in different growth periods was basically similar, but its female tended to select S. caucasica at early and flourishing periods and the ratio to select the latter is the highest among the three growth periods. The ratio of females preference for S. caucasica over C. latens was higher, whereas males' selection of the two plant species was similar (unpublished data). Further studies are needed to clarify whether the preference of female G. pecorum for GP1 and GP2 is a result of specific www.nature.com/scientificreports/ components or ratio-specific components. Our findings will enable G. pecorum electroantennogram response studies and the development of G. pecorum attractants, which could reduce the incidence of severe infection by G. pecorum of Przewalski's horse. Moreover, such attractants could enable wild release, rather than the current artificial semi-wild release, of Przewalski's horse. Experimental design. Due to special life cycle and biological characteristics of G. pecorum, it is difficult to define the oviposition period: the mature larvae of G. pecorum are discharged from the body with equine feces and then pupate into the ground and become adults. Adults of the fly need to mate and lay eggs immediately, because their life span is very short, only 1 to 4 days. When larvae of the fly are collected in the field, the relatively concentrated time when larvae are discharged from the body is judged according to the number of larvae, which is the peak of larvae discharge, and external environment temperature during the period is recorded. In this work, based on the mature larvae status of G. pecorum collected in the field in 2019, including the peak of larvae discharge, the local ambient temperature, the prediction formula for the pupa development history of the fly 55  Volatile collection. For each duplicate of a sample, 100.00 g of fresh plant shoots were weighed and placed in an oven bag (Reynolds, Richmond, VA) to collect volatiles. The plant materials were dried at 65℃ for 48 h to determine their dry weight after the collection of volatiles was completed.

Materials and methods
Before volatile collection, the air in the oven bag removed out using the QC-1S air sampler (Beijing Municipal Institute of Labour Protection, Beijing, China). The oven bag was refilled with air filtered through the activated carbon of a drying tower, which was connected through polytetrafluoroethylene tubes to an air sampler and an activated absorption pipe (CAMSCO, Houston, TX) filled with Tenax TA (60/80 mesh; Alltech, Deerfield, IL) to form a closed system. The air was recycled to collect volatiles for 5 h at a flow rate of 1 L/min. After one sampling, the absorption pipe was sealed and kept at − 20℃ until gas chromatography-mass spectrometry (GC-MS) analysis.
Automatic thermal desorption gas chromatography-mass spectrometry analysis. Volatiles collected in the absorption pipes were enriched using a Turbo Matrix 650 Automatic Thermal Desorber (Perki-nElmer, Waltham, MA) with a two-stage heating program. The carrier was high-purity helium. The volatiles in the adsorption pipes were desorbed at 260℃ for 10 min, and then reabsorbed in the cold trap (− 30℃), which was heated to 300℃ at a rate of 40℃/s for 5 min, and finally moved into the GC through a capillary transfer line (250℃). The conditions of the Clarus 600 Gas Chromatograph (PerkinElmer) were as follows: DB-5MS UI chromatographic column (30 m • 0.25 mm • 0.25 μm; Agilent Technologies, Santa Clara, CA); initial temperature of 40℃ for 2 min, increased to 180℃ at 6℃/min, followed by an increase to 270℃ at 15℃/min for 3 min. The conditions of the Clarus 600 T Mass Spectrometer (PerkinElmer) were as follows: electron ionization at 70 eV; mass scan range of 30 to 500 m/z; and interface and ion source temperatures of 250℃ and 230℃, respectively.
The volatiles were analyzed by TurboMass 5.4.2 GC-MS software (PerkinElmer, Shelton, CT). The volatiles were identified by matching their retention times, characteristic ions, and mass spectra with the NIST 08 library (National Institute of Standards and Technology, Gaithersburg, MD). The relative contents of individual volatiles were calculated by the area normalization method. Data analysis. Data analysis was performed using SigmaPlot Version 12.5 and SPSS Version 22.0 software.
The normality of the distribution and the homogeneity of the data were examined by the Shapiro-Wilk test and Levene's test, respectively. One-way analysis of variance (ANOVA) or the t-test was used for quantitative data. Nonparametric testing was used for data that did not meet the requirements for normality and homogeneity after transformation. All statistical tests were performed at a 5% significance level, and data are expressed as mean ± standard error (SE).