Behavioral Responses of Thrips hawaiiensis (Thysanoptera: Thripidae) to Volatile Compounds Identified from Gardenia jasminoides Ellis (Gentianales: Rubiaceae)

Thrips hawaiiensis is a common thrips pest of various plant flowers with host preference. Plant volatiles provide important information for host-searching in insects. We examined the behavioral responses of T. hawaiiensis adults to the floral volatiles of Gardenia jasminoides Ellis, Gerbera jamesonii Bolus, Paeonia lactiflora Pallas, and Rosa chinensis Jacq. in a Y-tube olfactometer. T. hawaiiensis adults showed significantly different preferences to these four-flower plants, with the ranking of G. jasminoides > G. jamesonii > P. lactiflora ≥ R. chinensis. Further, 29 components were identified in the volatile profiles of G. jasminoides, and (Z)-3-hexenyl tiglate (14.38 %), linalool (27.45 %), and (E3,E7)-4,8,12-trimethyltrideca-1,3,7,11-tetraene (24.67 %) were the most abundant. Six-arm olfactometer bioassays showed that T. hawaiiensis had significant positive responses to (Z)-3-hexenyl tiglate, linalool, and (E3,E7)-4,8,12-trimethyltrideca-1,3,7,11-tetraene tested at various concentrations, with the most attractive ones being 10−3 μL/μL, 10−2 μL/μL and 100 μg/μL for each compound, respectively. In pairing of these three compounds at their optimal concentrations, T. hawaiiensis showed the preference ranking of (Z)-3-hexenyl tiglate > linalool > (E3,E7)-4,8,12-trimethyltrideca-1,3,7,11-tetraene. Large numbers of T. hawaiiensis have been observed on G. jasminoides flowers in the field, which might be caused by the high attraction of this pest to G. jasminoides floral volatiles shown in the present study. Our findings shed light on the olfactory cues routing host plant searching behavior in T. hawaiiensis, providing important information on how T. hawaiiensis targets particular host plants. The high attractiveness of the main compounds (e.g., linalool, (E3,E7)-4,8,12-trimethyltrideca-1,3,7,11-tetraene, particular (Z)-3-hexenyl tiglate) identified from volatiles of G. jasminoides flowers may be exploited further to develop novel monitoring and control tools (e.g., lure and kill strategies) against this flower-inhabiting thrips pest.


Introduction
Thrips hawaiiensis Morgan (Thysanoptera: Thripidae) is a common flower-dwelling thrips pest of various horticultural plant species [1,2]. Native to the Oriental and Pacific regions, T. hawaiiensis Insects 2020, 11 is nowadays distributed in Asia, America, Africa, Australia, and Europe due to the expansion of international trade in fresh flowers, fruits, and vegetables [3][4][5][6]. In the field, T. hawaiiensis can attack a large number of plant species such as banana, mango, citrus, apple, tobacco, coffee, tea, horticultural plants and vegetables [6][7][8]. Therefore, it has become an important agricultural pest globally. Chemical insecticides have always been the primary tool for T. hawaiiensis control, especially the high number of specific treatments applied on banana and mango crops [9][10][11]. T. hawaiiensis showed the potential to rapidly develop resistance to insecticides (e.g., spinetoram) under laboratory selection [9,10]. Early detection of thrips is important for growers to decide when best to apply chemical insecticides or when to alter them, which would help to limit the frequency of insecticide applications, delaying the development of insecticide-resistance. Understanding the host-location behavior in thrips would be helpful for the development of monitoring tools for their early detection.
The color, shape and volatiles associated with different plant species were considered to provide vital cues for thrips and help them to search and locate more suitable host plants [12][13][14][15][16]. Many studies have evaluated the olfactory responses of thrips, e.g., the western flower thrips Frankliniella occidentalis Pergande (Thysanoptera: Thripidae), indicating that both floral volatiles and non-floral odors were attractive to this pest [14,[16][17][18]. Further, it was reported that rose volatile compounds allowed the design of new control strategies for western flower thrips [19]. So far little information has been mentioned on the behavioral responses of T. hawaiiensis to plant volatiles.
As a polyphagous flower-inhabiting thrips in China, T. hawaiiensis is always the dominant thrips pest on many banana and mango orchards during their flowering stages [9][10][11]20]. In addition, T. hawaiiensis varies in its population size on different flower host plants and vegetable crops at the flowering stage and particularly showed a preference for Gardenia jasminoides Ellis (Gentianales: Rubiaceae) flowers [21,22]. These results indicated that host plant flowers had a great attractiveness to T. hawaiiensis. To contribute to the knowledge on the role of volatile compounds in the host-searching behavior of T. hawaiiensis, the olfactory responses of adult insects to the odors of different flower plants including G. jasminoides, Gerbera jamesonii Bolus (Campanulales: Asteraceae), Paeonia lactiflora Pallas (Ranales: Ranunculaceae), and Rosa chinensis Jacq. (Rosales: Rosaceae) and to the main components identified from the preferred flower plant (G. jasminoides) were studied. These results may also help in developing new monitoring tools and other control options, which could be implemented in integrated pest management (IPM) strategies against T. hawaiiensis.

Insects and Plants
Mixed populations of T. hawaiiensis collected from various host plant species in the Nanming District, Guiyang area (26 • 34 N, 106 • 42 E) of Guizhou Province, China were used to establish a laboratory colony [21]. The independent colony was continuously reared for more than three generations on bean pods, Phaseolus vulgaris L. (Fabales: Leguminosae) in plastic containers [21,23]. The containers were kept in a climate-controlled room at 26 ± 1 • C, 65 ± 5% RH and a 14:10 h light:dark photoperiod.
G. jasminoides, G. jamesonii, P. lactiflora, and R. chinensis were grown in greenhouses in the nursery of Guiyang University, Guizhou Province, China. The greenhouses were maintained pest free by covering vent openings with insect-proof nettings. No pesticides were used in the whole plant growing season. Flowers at anthesis with intact petals were collected for olfactory tests and analysis of volatiles.

Behavioural Responses of T. hawaiiensis to Flower Volatiles
The olfactory responses of T. hawaiiensis were tested in a Y-tube olfactometer using the method described in Cao et al. [15,23]. Two types of two-way comparisons were made: (1) each plant flower versus clean air (CA); and (2) all possible flowers pairing. The flow rate was 300 mL/min. All bioassays were conducted between 8:00 a.m. and 6:00 p.m. in a room under 26 ± 1 • C, 65 ± 5% RH, and 1000 lux illumination conditions. For each comparison, 50 females that were 2-3 days old were tested. Thrips were starved for 4 h before the bioassay and the flower material (15.0 g) was replaced after every 10 tested individuals.

Collection and Analysis of Volatile Organic Compounds (VOCs)
Flower volatiles were collected and analyzed as described in Cao et al. [16]. Flower material (0.3 g) excised from a given host plant was kept in a glass bottle (200 mL) for 2 h prior to capturing the volatiles emitted using a solid-phase microextraction fiber (a~50/30 µm DVB/CAR/PDMS StableFlex fiber). Volatiles were extracted for 40 min at 80 • C then the fiber head was quickly removed. The collected volatiles were analyzed by gas chromatography mass spectrometry (GC-MS) (HP6890/5975C, Agilent Technologies, CA, USA). The chromatographic column was a ZB-5MSI 5% phenyl-95% dimethylpolysiloxane elastic quartz capillary vessel column (30 m × 0.25 mm × 0.25 µm). The gas chromatograph was operated at an initial temperature of 40 • C for 2 min then increased at 5 • C/min to 255 • C, which was maintained for 2 min. To identify compounds, we compared the mass spectra of compounds with those in databases (Nist 2005 and Wiley 275) and their constituents were confirmed through coinjections with authentic standards.

Behavioural Responses of T. hawaiiensis to the Main G. jasminoides Volatile Organic Compounds (VOCs)
The VOC mixture from G. jasminoides was the most attractive to T. hawaiiensis as assessed by the Y-tube olfactometer bioassays. Since linalool, (Z)-3-hexenyl tiglate and (E3,E7)-4,8,12-trimethyltrideca-1,3,7,11tetraene were the most abundant compounds identified in the VOC profile of G. jasminoides, the behavioral responses of T. hawaiiensis to these compounds were tested further in six-arm and Y-tube olfactometer bioassays.

Six-arm Olfactometer Bioassays
The behavioral responses of adult T. hawaiiensis to different doses of linalool, (Z)-3-hexenyl tiglate, and (E3,E7)-4,8,12-trimethyltrideca-1,3,7,11-tetraene) were evaluated in a six-arm olfactometer with the method described by Liu et al. [24]. and Cao et al. [25]. Briefly, the six-arm olfactometer consisted of a central chamber with six arms, each connected to a glass tube that projected outwards at an equidistance, with equal angles (60 • ) between pairs of tubes. Each arm was connected through Teflon tubing to a glass vessel containing a test or control stimulus. For each experiment, equal volumes (25 µL) of each of the five solutions of one compound and mineral oil (used as the control), absorbed onto a filter paper disk (1.0-cm diameter), were used as test and control stimuli, respectively. The airflow was set at 200 mL/min to drive the odor source to thrips. Thrips hawaiiensis (2-3 days old females) were starved for 4 h and introduced in groups (200 individuals per group) into the central chamber with a fine camel hair brush. Within 20 min, insects that entered one arm of the olfactometer were counted as having made a choice for a particular odor, while thrips that did not enter any arm were considered non-responders. After each test, the olfactometer was cleaned, dried and the arms were rotated (60 • ). Each bioassay was replicated six times between 9:00 am and 6:00 pm. In order to eliminate any light bias, a 25-W light was placed in the center 60 cm above the chamber.

Statistical Analyses
All statistical analyses were performed using SPSS 18.0 for Windows (SPSS Inc., Chicago, IL, USA) [26]. The null hypothesis that T. hawaiiensis adults showed no preference for either Y-tube arm (a response equal to 50:50) was analyzed using a chi-square goodness-of-fit test [27]. The number of thrips found in the different arms of the six-arm olfactometer were subjected to Friedman two-way ANOVA by ranks and in the case of significance (p < 0.05) the Wilcoxon signed ranks test was used for separation of means [28].

Statistical Analyses
All statistical analyses were performed using SPSS 18.0 for Windows (SPSS Inc., Chicago, IL, USA) [26]. The null hypothesis that T. hawaiiensis adults showed no preference for either Y-tube arm (a response equal to 50:50) was analyzed using a chi-square goodness-of-fit test [27]. The number of thrips found in the different arms of the six-arm olfactometer were subjected to Friedman two-way ANOVA by ranks and in the case of significance (p < 0.05) the Wilcoxon signed ranks test was used for separation of means [28].

Figure 1.
Behavioral responses of T. hawaiiensis to the volatiles from different flowers. Asterisks indicate highly significant (** p < 0.01) and significant (* p < 0.05) differences in the selectivity of T. hawaiiensis between two odors by 2 test. NS indicates no significant differences (p > 0.05) in the selectivity of T. hawaiiensis between two odors.
Thrips hawaiiensis also showed significant preferences among the flowers when presented with choices. Gardenia jasminoides was more attractive to T. hawaiiensis than G. jamesonii ( 2 = 4.79, df = 1, p Figure 1. Behavioral responses of T. hawaiiensis to the volatiles from different flowers. Asterisks indicate highly significant (** p < 0.01) and significant (* p < 0.05) differences in the selectivity of T. hawaiiensis between two odors by χ 2 test. NS indicates no significant differences (p > 0.05) in the selectivity of T. hawaiiensis between two odors.

Six-Arm Olfactometer Bioassays
In these experiments, the number of insects that entered the control arm connected to the vessel with mineral oil were significantly lower than those of insects found in the arms with the different doses of linalool (Friedman test:

Discussion
Volatile compounds emitted from plants are important cues in the host selection process of phytophagous insects. In a natural plant community, different plant species emit different qualitative and quantitative blends of VOCs, which guide insects to discriminate and locate their host plants [29,30]. For flower thrips pests, a great number of studies have documented that F. occidentalis exhibited significantly positive responses to the volatiles of host plants [13][14][15]17,18], and displayed a significant preference for specific host olfactory cues [16,23]. Therefore, based on the bioactive compounds identified from the host plant volatiles, lures were developed and applied for the monitoring and control of F. occidentalis [19,[31][32][33][34]. In the present study, T. hawaiiensis were significantly attracted to the volatiles from G. jasminoides, G. jamesonii, P. lactiflora, and R. chinensis, and showed olfactory preferences with G. jasminoides > G. jamesonii > P. lactiflora ≥ R. chinensis. This is consistent with our previous field observations during which large numbers of T. hawaiiensis were found in G. jasminoides flowers [21]. Furthermore, the nutritional composition of G. jasminoides flowers could adequately satisfy the nutritional needs of T. hawaiiensis which could have a faster population development when fed on these flowers [35]. Behavioral responses of T. hawaiiensis to the main Ga. jasminoides volatiles. Asterisks indicate highly significant (** p < 0.01) and significant (* p < 0.05) differences in the selectivity of T. hawaiiensis between two odors by χ 2 test.

Discussion
Volatile compounds emitted from plants are important cues in the host selection process of phytophagous insects. In a natural plant community, different plant species emit different qualitative and quantitative blends of VOCs, which guide insects to discriminate and locate their host plants [29,30]. For flower thrips pests, a great number of studies have documented that F. occidentalis exhibited significantly positive responses to the volatiles of host plants [13][14][15]17,18], and displayed a significant preference for specific host olfactory cues [16,23]. Therefore, based on the bioactive compounds identified from the host plant volatiles, lures were developed and applied for the monitoring and control of F. occidentalis [19,[31][32][33][34]. In the present study, T. hawaiiensis were significantly attracted to the volatiles from G. jasminoides, G. jamesonii, P. lactiflora, and R. chinensis, and showed olfactory preferences with G. jasminoides > G. jamesonii > P. lactiflora ≥ R. chinensis. This is consistent with our previous field observations during which large numbers of T. hawaiiensis were found in G. jasminoides flowers [21]. Furthermore, the nutritional composition of G. jasminoides flowers could adequately satisfy the nutritional needs of T. hawaiiensis which could have a faster population development when fed on these flowers [35].
Pollens were considered to be an important factor involved in the preference of thrips for host plant flowers [40][41][42], as F. occidentalis was attracted to the compound of (S)-verbenone identified from the volatiles of pine pollen [18,33]. However, pollens of G. jasminoides flowers were not involved in T. hawaiiensis' olfactory responses in this study. Besides, the host preference of different thrips species were related to the color, shape, nutritional conditions or other physicochemical characteristics of host plants [13,15,17,[43][44][45]. Thus, more related physicochemical characteristics that may influence the behavioral responses of T. hawaiiensis should be studied to comprehensively understand the mechanism of host selection among different flower plants.

Conclusions
Females of T. hawaiiensis exhibited a higher olfactory preference for the VOCs of G. jasminoides flowers over a range of four flower plants. Moreover, insects were significantly attracted by different concentrations of the three main components of G. jasminoides flower VOCs with (Z)-3-hexenyl tiglate being the most attractive. Overall, this study clearly demonstrated that plant volatiles are involved in host-plant selection by T. hawaiiensis even if, to comprehensively understand this mechanism, possible interactions among chemical cues and other physicochemical characteristics remain to be investigated. The kairomonal activity of (Z)-3-hexenyl tiglate, linalool, and (E3,E7)-4,8,12-trimethyltrideca-1,3,7,11-tetraene to T. hawaiiensis females found in this study provides a basis for further electrophysiological, behavioral, and field-trapping experiments to develop semiochemically-based monitoring tools and direct control options for this pest.