A Floral-Derived Compound Attractive to the Tephritid Fruit Fly Parasitoid Diachasmimorpha longicaudata (Hymenoptera: Braconidae)
- First Online:
- Cite this article as:
- Rohrig, E., Sivinski, J., Teal, P. et al. J Chem Ecol (2008) 34: 549. doi:10.1007/s10886-008-9438-y
Many adult hymenopteran parasitoids, even host-feeding species, consume the nectar of flowering plants. Previous field studies had identified plants attractive (Lobularia maritima L.) and unattractive (Spermacoce verticillata L) to certain opiine braconids (Hymenoptera). Under laboratory conditions, Diachasmimorpha longicaudata (Ashmead), a parasitoid of tephritid fruit fly larvae and representative opiine, responded in flight tunnels to L. maritima but not to S. verticillata. Volatile chemicals of the two flowers were collected and analyzed by using capillary gas liquid chromatography and mass spectral analysis. Acetophenone was isolated from L. maritima but not from S. verticillata. In flight tunnels, D. longicaudata were exposed to 10 concentrations (doses) of acetophenone. Female parasitoids showed a significant attraction to several acetophenone doses, with concentrations of 25 and 50 ng the most attractive. No odor source, either floral or floral-derived, was attractive to male parasitoids. Reliable trapping systems for parasitoid species, particularly species such as D. longicaudata used for augmentative biological control, would be a valuable monitoring tool. At present, there are few, if any, florally derived synthetic lures for attracting hymenopteran parasitoids.
Understanding the population dynamics of target pests and their natural enemies is critically important in Integrated Pest Management (IPM) (Kogan 1998; Suckling et al. 2002). Attraction and trapping techniques are among the simplest means of determining pest and natural enemy distributions and densities in the field, and this information, combined with known insect biology, can help determine what controls should be applied, when and where to use them, and the status of non-target organisms that might be affected (Prokopy 1985). Among the controls best timed and placed through the use of trapping-demographic data is the inundative release of beneficial insects (Jewitt and Carpenter 2001). It is particularly important to monitor the dispersal and survival of the mass-released predators/parasitoids so that adjustments can be made to the releases, and their efficacies estimated.
At present, most pest-insect attractants are based on sex pheromones, and majority of these are female-derived chemicals that capture males (Foster and Harris 1997). Sex pheromones are also emitted by female, and more rarely male, parasitic Hymenoptera (Sivinski and Petersson 1997; Kainoh 1999), and can also be used as attractants (Rice and Jones 1982; Jewett and Carpenter 2001; Suckling et al. 2002). However, compounds associated with hosts, or environments where hosts might be found, may be more generally attractive to a presumably largely non-virgin and sexually unresponsive female parasitoid population. For example, among tephritid fruit fly parasitoids, Fopius arisanus (Sonan) and Diachasmimorpha longicaudata (Ashmead) females are attracted to fruit volatiles (Messing and Jang 1992; Eben et al. 2000; Altuzar et al. 2004), the latter particularly to decaying fruit and associated fungi (Greany et al. 1977). Yet another tephritid parasitoid Psyttalia fletcheri Silvestri responds to decaying fruits and leaves of pumpkins and cucumbers (Messing et al. 1996). Nishida (1956) earlier found that stem tissues of cucurbits are attractive to P. fletcheri, and Messing et al. (1996) suggest that the attractiveness to leaf and stem tissues may be due in part to “green leaf volatiles”, a suite of common leaf-derived compounds known to be attractive to other braconid species (Whitman and Eller 1990).
Flowers of many species of plants produce odors that are also highly attractive to various natural enemies (Andersen 1987; Maini and Burgio 1999; Ventura et al. 2000; Landolt et al. 2001). Although nectar feeding, and to a lesser extent pollen consumption, is common in the parasitic Hymenoptera (Jervis et al. 1993), there has been little if any use of floral-derived compounds as attractants. However, there may be advantages to the use of such compounds. Like host odors, but unlike male-produced sex pheromones, flower volatiles might be more consistently appealing to the mated females that likely constitute the bulk of many parasitoid populations.
In the course of a previous field comparison, flowers that were attractive (Lobularia maritima L.) and unattractive (Spermacoce verticillata L.) to an economically significant subfamily of Braconidae, the Opiinae, were preliminarily identified (Rohrig 2006). In the present experiments, the flight tunnel responses of D. longicaudata, a widely used tephritid biological control agent and representative opiine, to the two flowering plants and to a compound uniquely identified from L. maritima volatiles were determined. These plants and the model insect are described below.
Spermacoce verticillata, “shrubby false buttonweed” (Rubiaceae) is native to the West Indies, but can be found in Florida and Texas as well as west Africa, the tropical Americas, and the south Pacific. Habitats consist of open or disturbed sandy zones and pinelands where it grows as a shrub. Spermacoce verticillata is a perennial dicot whose small white flowers form dense clusters at the upper stem nodes. Flowers have a mean corolla depth of 1.5 mm, a width of 1.0 mm, and possess a honey guard at the interior base of the corolla (Sivinski et al. 2006). The nectar is a major food source for the mole cricket ectoparasitoids Larra bicolor F. and Larra analis F. (Hymenoptera: Sphecidae) in Florida (Frank and Parkman 1999).
Lobularia maritima, “alyssum” (Brassicaceae), was introduced from the Mediterranean region and now ranges throughout most of the United States including Hawaii. This plant is a hardy, non-weedy, annual dicot herb that flowers consistently from fall through spring in subtropical areas. Small cruciform-stalked white flowers grow in clusters randomly throughout the plant that attract large numbers of parasitic Hymenoptera (Chaney 1998). On average, corollas are 0.67 mm wide by 1.4 mm deep (Sivinski et al. 2006).
Lobularia maritima nectar increases the longevity of several ichneumonoid parasitoids, both in the laboratory and in the field (Johanowicz and Mitchell 2000; Berndt and Wratten 2005). While L. maritima flowers increase the longevity and realized fecundity of the egg parasitoid Trichogramma carverae Oatman and Pinto (Begum 2004), this is only true of the white flower variety (Begum et al. 2004). This was the variety used in the present experiments.
Due to availability and its wide use as a biological control agent, D. longicaudata was chosen to examine the response of a representative opiine to the floral volatile. While a variety of chalcidoids, diapriids, figitids, and ichneumonoids parasitize Tephritidae (e.g., Sivinski et al. 2000), braconids of the subfamily Opiinae are typically the most numerous and diverse members of the guilds that attack frugivorous species (Purcell 1998). Opiinae are solitary, koinobiont, larval/egg-prepupal endoparasitoids of Cyclorrhapha Diptera (Bess et al. 1961; Lopez et al. 1999; Wharton 1999). Several species are considered important regulators of fruit fly populations (Wharton 1989) and have been introduced, and frequently established, throughout the world (Ovruski et al. 2000).
Diachasmimorpha longicaudata is one of the most widely used of these opiine tephritid biological control agents (Ovruski et al. 2000). Adult females use their relatively long ovipositor to parasitize a number of second- and third-instar fruit fly species’ larvae in a wide variety of host fruits (Wharton 1989; Sivinski et al. 2000). The species was originally discovered in the Indo-Philippine region where it attacked Bactrocera spp. (Wharton and Marsh 1978), and in 1947 was introduced into Hawaii for the control of oriental fruit fly, Bactrocera dorsalis (Hendel) (Clancy and Dressner 1952). In 1972, D. longicaudata was established in Florida to control the Caribbean fruit fly, Anastrepha suspensa (Loew), and it subsequently reduced populations by ~40% (Baranowski et al. 1993).
In addition to wide-spread introductions, D. longicaudata has been mass-reared and inundatively released, either alone or in combination with sterile male flies, for the control of Ceratitis capitata (Wong et al. 1991), and various Anastrepha spp. (Sivinski et al. 1996; Montoya et al. 2000). In Florida, such releases suppressed A. suspensa populations by as much as 96% and have been considered as a means of supporting “fly-free” zones to facilitate citrus exports.
Methods and Materials
Diachasmimorpha longicaudata were obtained from a >5-yr-old colony at the USDA-CMAVE and reared as described by Sivinski et al. (1996). Parasitoids were maintained in a climate-controlled room with a temperature range of 21–24°C and a relative humidity of 65–80%.
Preliminary chromatographic examinations of volatile chemicals collected from cut and attached flower were similar. Therefore, due to the handling ease, flowers used for compound identification were cut from the plant stem as close to floral tissue as possible, and the cut end was surrounded with cotton wool soaked in water. Flower heads were placed in a glass volatile collection chamber (Heath and Manukian 1992). This system consists of a chamber (30 cm long and 4 cm outside diameter [OD]) that has a sintered glass frit at the upwind end and a joint outlet with a single-port collector base. Humidified charcoal filtered air was pushed into one end of the chambers and over the flower heads. Air was pulled out the other end via a vacuum system. Air exiting the chamber passed through a volatile collection filter that contained 50 mg of Super-Q® (Refined Technologies, Woodland, TX, USA) to collect and hold any volatiles. Afterwards, collection filters were eluted with three aliquots of 100 µl methylene dichloride to remove volatile compounds.
Extracts of samples were analyzed by capillary gas liquid chromatography (GC) using a Hewlett Packard 5890 (Palo Alto, CA, USA) equipped with a cool on-column injector and flame ionization detector. The column, a 30 m × 0.25 mm (i.d.) SE-30 capillary column (Alltech Assoc., New Gloucester, ME, USA), was attached to a 10-m length of 0.25-mm (id) deactivated fused silica as a retention gap, which was in turn attached to a 10-cm length of 0.5 mm id deactivated fused silica in the injector. Helium (linear flow velocity 18 cm/sec) was used as a carrier gas. The oven temperature and injector temperatures were programmed to go from 60°C (held for 5 min) to 200 °C at a rate of 10°C/min. Mass spectral analysis was accomplished with an HP6890 GC equipped with a DB-5 MS® column (30 m × 0.25 mm ID × 0.25 µm,) linked to an HP 5973 mass spectrometer. Both electron impact (70 eV) and chemical ionization (isobutene reagent gas) spectra were obtained. Helium was the carrier gas, and a splitless injector (injector temperature of 240°C, split valve delay of 0.5 min) was used. The oven temperature was held at 35°C for 1 min, then programmed to increase at a rate of 10°C/min to 230°C, which was held for 10 min. The ion source temperature was 230°C. Tentative identifications were made by comparison of fragmentation patterns with patterns available in the NIST-MS library and libraries developed at USDA-CMAVE (Gainesville, FL, USA). Identifications were confirmed by comparison of chromatographic retention times and mass spectra of natural compounds with those of commercially available standards analyzed on the same instrument.
To determine if male and female D. longicaudata had a sensory response to L. maritima volatiles (see results), extracts were analyzed with a GC interfaced to both flame ionization (FID) and electroantennograph detectors. In this manner, antennal responses were matched with FID signals for compounds eluting from the GC. Volatile extracts were prepared in the manner described above, and 1-µl aliquots were analyzed on a Hewlett-Packard (HP) 5890 Series II gas chromatograph equipped with an HP-5 column (30 m × 0.32 mm ID × 0.25 mm) (Agilent, Palo Alto, CA, USA). The oven temperature was held at 40 °C for 5 min, then programmed to increase to 10°C /min to 220°C and held at this temperature for 5 min. Helium was used as a carrier gas at a flow rate of 2.0 ml/min. A humidified air stream was delivered over the antenna is at 1 ml/min.
The antennae were excised by grasping the scape with jeweler’s forceps (No. 5, Miltex Instrument Company Inc, Switzerland) at the base, and the antennae were removed from the head of either male or female wasps. The extreme distal and proximal ends of the antennae were held between gold electrodes in conductivity gel (Syntech, Hilversum, The Netherlands). The electroantennal detector (EAD) and FID signals were concurrently recorded with a GC-EAD program (Syntech GC-EAD 2000, Hilversum, The Netherlands), which analyzed the amplified signals on a PC.
Flight Tunnel Bioassays
To determine the response of D. longicaudata to various odor sources, parasitoids were observed in three identical flight tunnels. These flight tunnels were constructed of clear Plexiglas and measured 128 cm long × 32 × 32 cm, and were located inside two climate-controlled greenhouses. Natural sunlight illuminated the chambers. Although the greenhouses were climate controlled, extremes of outside temperatures caused the internal temperature to fluctuate between 21° and 27°C and relative humidity between 60% and 85%. Airflow and air speed were maintained at 0.3 to 0.4 m/s by using a variable speed fan at the downwind end. This was the speed that stimulated the most flight in D. longicaudata in previous investigations (Messing et al. 1997). Odors were released into the anterior end of the tunnel. Two separate chambers that contained either an odor source or a blank-air control were housed outside of the flight tunnel. Chambers were constructed from 114-l all-glass aquaria (31 × 61× 61 cm). Lids were constructed of clear Plexiglas and fit tightly to prevent outside air from entering the chambers. Purified air was passed through the sample chambers at 0.3 m/s regulated with adjustable flow meters (Aalborg Instruments, Monsey, NY, USA) and into the flight tunnel. Each chamber was connected to a trap inside the front end of the tunnel. The traps were located midway between the top and bottom of the tunnel, and were constructed from clear cylindrical plastic vials (9 × 5 cm) placed horizontally. Both cylinders had orange colored snap tight lids and had a 1.3-cm hole in the center to allow air to exit. This provided an opening for the parasitoids to enter in search of the odor source.
Traps, washed initially with a mild detergent, were changed with each new odor source, and the tubing connecting the traps to volatile-source container was washed with detergent after each replication.
In all flight tunnel experiments, 25 presumably mated (sexes held together for 3–7 d) parasitoids were used per day. All parasitoids had access to water, but food was withheld for 24 hr before use in experiments to enhance response to potential food sources. Flight tunnels were checked hourly from 0900 to 1700 hours. A parasitoid inside a trap was recorded as a positive response, and was removed and returned to the tunnel after counting. Each day, the tubes connecting the odor chambers to the traps were switched to prevent positional effects.
Ripe mango fruit, Mangifera indica L., and the flowering plants were tested as odor sources. Mangos are attractive to D. longicaudata in flight tunnel tests (Eben et al. 2000) and were used as a positive control. Flowering plants were in full bloom and contained in 2-l plastic pots. Each of the three odor sources was tested individually against a blank control of clean odorless air. There were 10 replicates for males alone and 10 for females alone, each replicate lasting a day. In addition, there were 10-d-long replicates of 25 females and 25 males together with L. maritima as the odor source. This was done to ensure that any sexual dimorphism in response was not due to undetermined environmental factors present during separate-sex replications.
Acetophenone was loaded on 11-mm sleeve stopper natural red rubber septa by dissolving the synthetic material at various doses in methylene chloride and filling the large well of the septa with 100 µl of solution. Septa were air-dried for 24 hr and stored in a freezer, except during use in the wind tunnels. During flight tunnel tests, septa were placed in glass airtight containers (2 l) that served as the odor chambers.
All volatile loaded septa were run against blank septa loaded with 100 µl of methylene chloride and air-dried as a control. To find the dose that elicited the greatest response, various concentrations were exposed to parasitoids. An initial dose of 1 mg did not produce a response, and increasingly lower doses were tested until there was no significant response. Each dose was initially run for 3 d in sequential decreasing order from highest to lowest. Only females were in the wind tunnel. Acetophenone doses used were as follows: 1 mg, 100 µg, 50 µg, 10 µg, 1 µg, 100 ng, 50 ng, 25 ng, 10 ng, 1 ng. If any response was seen, then that volatile dose was run for 10 d with females and 10 d with males.
Statistical analyses were conducted by using SAS programming (SAS Institute Inc. 2001). Polynomial regression analysis (PROC GLM) was used to examine flight responses over time and the interactions of “time of day ×volatile source” on response. Analysis of variance (PROC ANOVA) followed by means separation through the Waller test was employed to compare the mean responses to various doses. Note that the protocol for all wind tunnel experiments included the return of captured females to the population inside the tunnel. While this was a simple means of controlling for insect density and environmental history, the summed captures could include multiple responses by a particular insect. This led us, when comparing a particular volatile to its control, to employ the most conservative method available, the nonparametric “paired sign test” (Zar 1974) where a consistently greater response over 10 replicates to a particular source yields a P value of 0.01. Mean and variance data should be interpreted in terms of “responses” and not “insects responding”.
Flight Tunnel Bioassays
Males showed no significant response to any odor source. In several instances, they were caught in isolation traps, but in general, they exhibited occasional walking and flying behaviors at the rear of the flight tunnel away from the odor source.
When both males and females were simultaneously exposed in the wind tunnel to L. maritima flower volatiles, no males entered the isolation chambers, but female response was similar to the single-sex experiment with L. maritima (mean of 0.9 [SE 0.07] responses/hourly observation period; paired sign test, greater response to volatile source equal to 10 of 10; P = 0.01).
These experiments provide additional evidence that not all flowers are equally attractive to any one particular species of parasitoid (Jervis et al. 1993). In the field, L. maritima caught more Braconidae, specifically unidentified Opiinae, than did S. verticillata (Rohrig 2006). This difference was substantiated in flight tunnel tests that found a significant flight response by the opiine D. longicaudata to L. maritima but not to S. verticillata. In addition to this corroborative evidence in support of the generality that there is variance in the diversity of parasitoids attracted to flowers, there were several novel/unusual findings. These included: (1) the suggestion that there may be broad phylogenetic preferences for certain flowers, i.e., the common response of at least two allopatric opiines to the pair of flower species (Rohrig 2006); (2) the identification of a compound from a flower volatile that is attractive to a hymenopteran parasitoid; and (3) the sexual dimorphic response to L. maritima, with only females attracted to either the complete flower volatiles or the isolated compound unique to L. maritima.
Direct comparisons of parasitoid olfactory response to different live flowers can lead to ambiguous results. Although the sizes of the plants used and the overall area of flowers present were similar (Rohrig 2006), the amounts of volatiles released may have been different. Even within a species, odor concentration can vary among flower parts and change diurnally as well as seasonally (Bergstrom et al. 1995). The identification of the attractive compounds present in floral odors can help overcome this obstacle. Once isolated, compounds can be presented in fixed amounts, thereby reducing variation in the odors encountered by the test insects. Thus, the identification of acetophenone as a compound both unique to L. maritima and attractive to D. longicaudata supports the difference in attractiveness observed between the two whole plants. The isolation of acetophenone also allowed estimation of optimal release rates.
Diachasmimorpha longicaudata had a sexually dimorphic response to all volatiles that elicited flight; male parasitoids were not attracted to any. This was true when the sexes were tested both separately and together, so that unnoticed environmental cues present during single-sex experiments cannot be responsible for the sexual difference. Males did exhibit oriented flight and entered the isolation traps in response to virgin females + fruit + honey solutions in experiments conducted in the same wind tunnels (C. S., unpublished data). Thus, the lack of response in the present experiments was due to the volatiles used and not because of male inability to perform in the flight tunnel environment.
Messing and Jang (1992) had previously found that female D. longicaudata responded to various host fruit stimuli to a greater extent than did males. Similar sexually dimorphic responses to plant materials are not restricted to hymenopteran parasitoids. For instance, females of the tachinid Eucelatoria bryani Sabrosky are attracted to many more plant volatiles, including those of flowers, than are males (Martin et al. 1990).
There are several possible explanations of why female, and not male D. longicaudata, are attracted to acetophenone. For instance, females may have unique nutritional needs that flower nectar provides. Acetophenone possess a flowery smell and is used in the perfume industry in fragrances such as vanilla, honeysuckle, and jasmine (USEPA 1987). It is emitted as a volatile by several flowers including Centaurea scabiosa L., a species frequently visited by butterflies (Andersson et al. 2002), Elsholtzia argyi Dong (Peng and Yang 2005), and Calanthe sieboldi Decne (Awano et al. 1997). In the laboratory, D. longicaudata feeds avidly on juices of fruits that would be commonly encountered while searching fallen fruit for hosts (Sivinski et al. 2006). Ground-foraging female parasitoids, however, would also be in the general vicinity of flowers and could visit them to obtain additional other nutrients (Koptur 2005). The relatively ambiguous laboratory evidence that flowers are an important adult food sources for D. longicaudata does not offer strong support for this conclusion (Sivinski et al. 2006). Maximum longevities were significantly extended in the presence of flowers (including the presently examined species), but there was no difference between the mean longevities of parasitoids provided with flowers and those given only water.
Females might also mistake acetophenone for a host-fruit volatile or a male-produced pheromone. It has been isolated from Rambutan fruit (Nephelium lappaceum L.; Ong et al. 1998) and guava pulp (Psidium guajava L.) (Idstein and Scherier 1985), and is also produced by a number of diverse insects as a semiochemical (Kohnle et al. 1987; Schulz et al. 1993; Aldrich et al. 1995; Birkett et al. 2004). The latter hypothesis likely can be discarded since neither acetophenone nor closely related compounds are present in the male-produced sex pheromone (Nancy Epsky, unpublished data).
However, female D. longicaudata might mistake the flower volatile for a fruit cue. Tephritids use odor to locate fruit, and the same ability would be useful to fruit fly parasitoids (Eben et al. 2000). Males of both fruit flies and their natural enemies may not have been subject to the same selection pressures as females to develop sensitivity to fruit odor stimuli. For example, when 4-ethyl-acetophenone in the peel of navel oranges (Citrus sinensis Osbeck) was used as an odor source in electroantennogram studies of Ceratitis capitata (Weid.), the Mediterranean fruit fly, it elicited a significant voltage spike in females, but not in males (Hernandez et al. 1996). Levinson et al. (2003) found that female C. capitata sensilla were significantly more responsive to orange odor than male sensilla. Electroantennagrams showed that male D. longicaudata sensed acetophenone, but it did not elicit a behavioral response.
Regardless of the significance of the acetophenone response, it may lend itself to improved traps and ultimately better control procedures. There are no efficient synthetic lures and traps available to monitor opiine parasitoid populations in the field (Messing 1992). At present, most monitoring and delineation of fruit fly parasitoids is accomplished through laborious and time-consuming fruit sampling (Montoya et al. 2000; Sivinski et al. 2000). An improved trapping system would prove valuable in several ways. It could be used to determine the presence of parasitoids in a given area as well as help understand their population dynamics. Traps could provide important wasp dispersal and patch retention information that might improve the application of parasitoids for inundative biological control. Knowing when and where to release parasitoids, as determined by previous trapping in various locations, could be crucial to their application in new sites, whether the goal is permanent establishment or mass releases.
Field testing of acetophenone is needed to examine its usefulness. Although only female wasps were attracted, females are the agents of mortality, and their distribution in time and space is generally of primary interest. It may be that acetophenone by itself will not be sufficiently attractive, but it could prove to have a role in a more efficacious mixture of compounds that could include other floral volatiles, fruit odors (Eben et al. 2000), fungal/ bacterial odors (Greany et al. 1977), and pheromones.
We thank Dr. Ru Nguyen, Dr. Nancy Epsky, and Dr. Rob Meagher, for advice and review. We also thank the late Barbara Dueben for help on the chemical and bioassay aspects of this work. Tim Holler and Amy Moses provided a constant supply of parasitoids. We also thank Gina Posey for help with computer issues and the preparation of the graphs.