Doryctobracon areolatus (Hymenoptera, Braconidae) a parasitoid of early developmental stages of Anastrepha obliqua (Diptera, Tephritidae)

Natural parasitism of Doryctobracon areolatus (Szépligeti) (Hymenoptera: Braconidae) on various development stages of Anastrepha obliqua (Macquart) (Diptera: Tephritidae) attacking Spondias mombin L. fruits was studied under field conditions. We collected 120 fruits of S. mombin from which we got 495 A. obliqua larvae of different instars. A total of 88% of these larvae were parasitized. Within the parasitized cohort, the first-instar of D. areolatus was frequently detected in all 3 larval stages (L1 = 94.3%, L2 = 98.1%, and L3 = 100%), and the rest (i.e., L1 = 5.7%, L2 = 1.8%) corresponded to the presence of eggs. In fruits with controlled infestation and cage-induced parasitism under field conditions, D. areolatus oviposited in mature eggs and recently hatched larvae of A. obliqua with comparable frequencies. Seven preimaginal stages of D. areolatus were observed during their development, which was completed in 27 days. It is concluded that D. areolatus has the capacity to oviposit in embryo eggs and neonate larvae of A. obliqua and that its first-instar larvae (with three distinct sizes) are capable of synchronizing their development with the development of the host larvae. This finding represents the first report of a native parasitoid attacking eggs or neonate larvae of a tephritid in the Neotropics. The implications of this finding are discussed within the context of the competitive interactions of this species with other parasitoid species under sympatric conditions, as well as the relevance for developing laboratory rearing methods for biological control purposes. JHR 46: 91–105 (2015) doi: 10.3897/JHR.46.5586 http://jhr.pensoft.net Copyright Félix D. Murillo et al. This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. RESEARCH ARTICLE Félix D. Murillo et al. / Journal of Hymenoptera Research 46: 91–105 (2015) 92


Introduction
The native parasitoid guild that attacks fruit flies of the genus Anastrepha Schiner in the Neotropics is mainly composed of a group of solitary koinobiont endoparasitoids (primarily Braconidae and Figitidae) that oviposit in the host larvae and emerge from the pupae.The genus Doryctobracon constitutes 27% of the parasitoid species and shares a closely related evolutionary history with Anastrepha (López et al. 1999, Ovruski et al. 2000).
Another smaller group of Anastrepha parasitoids are pupal idiobionts that attack their hosts when they are in the soil and is represented by five species within the genera Coptera and Trichopria (Diapriidae) and three polyphagous species, Pachycrepoideus vindemiae (Rondani), Spalangia cameroni Perkins, and S. endius Walker (Pteromalidae) (Ovruski et al. 2000).The parasitoid guild of Anastrepha spp. in the Neotropics represents an important source of species with potential to be used in biological control programs against native tephritids (Aluja et al. 2003).However, native Anastrepha parasitoids attacking eggs or early larval developmental stages have not been reported.The only report of a parasitoid attacking Anastrepha eggs in the Americas correspond to Wharton et al. (1981), who found a small level of parasitism of the introduced Fopius arisanus (Sonan), apparently on A. striata (Schiner) eggs.
According to Aluja et al. (2009) species of the genera Doryctobracon, such as D. crawfordi (Viereck) and D. areolatus (Szépligeti), exhibit great potential as biological control agents and should be examined from the point of view of mass rearing projects.D. areolatus has been reported as a solitary koinobiont endoparasitoid that attacks third instar larvae of Anastrepha spp., both in native and exotic commercial fruits, with a wide distribution from Florida to Argentina (Hernández-Ortiz et al. 1994, López et al. 1999, Ovruski et al. 2000, Sivinski, et al. 2000, Aluja et al. 2003, 2009).This species frequently shows field dominance among concurrent parasitoid species attacking Anastrepha spp.(López et al. 1999, Sivinski et al. 1997, 2000;Aluja et al. 2003, Ovruski et al. 2004).The presence of diapause (Aluja et al. 1998, Ovruski et al. 2004) and an extrinsic capacity to find patches with low host density (Sivinski et al. 1998) allows an ample distribution in regions with low plant diversity (Eitam et al. 2004).
In the central region of Veracruz, Mexico, D. areolatus is the most abundant parasitoid species attacking Anastrepha obliqua (Macquart) in Spondias spp.(Cabrera et al. 2006), where it has frequently been observed at the beginning of the fruiting season.This suggests that this species could parasitize earlier developmental stages in relation to sympatric parasitoid species.Normally, parasitoids attacking eggs and first instar larvae of their hosts become more competitive than those attacking later stages (Wang and Messing 2002, Wang et al. 2003, Wang et al. 2008, Argov et al. 2011), showing important potential to be used as biocontrol agents.Therefore, our aims in this study were: 1) to determine if D. areolatus parasitize immature stages of A. obliqua and 2) to characterize its developmental stages during each phase of host development.

Study area
The study was conducted in the coastal region of central Veracruz, which is characterized by high densities of A. obliqua hosts, such as mango (Mangifera indica L.), native Spondias species and guavas (Psidium guajava L.).Fruit samples were collected from trees located in backyard orchards and marginal zones, which provide resources for the presence of flies and parasitoids all year round.This zone is located between 19°00' and 18°55' North latitudes and 96°10' and 96°13' West longitudes, with a mean altitude of 18.5 m.a.s.l.The climate is semi-humid, with a mean annual rainfall of 1,358 mm and a very marked rainy season from June to September.The highest mean monthly temperature (29.1 °C) occurs in the month of June and the lowest mean monthly temperature (21.4 °C) occurs in the month of January (SMN 2010).

Determination of the natural parasitism of A. obliqua larvae by D. areolatus under field conditions
From September to October 2013, hog plum (Spondias mombin L.) fruits were collected from four sites, three in the locality of "El Copital" and one in "El Mangal", municipality of Medellín de Bravo, from four to five trees per site.Fruits were collected directly from the trees (36 fruits, 30%) and from the ground surrounding the trees (84 fruits, 70%).Each sample consisted of 10 fruits per site.Samples were in three sampling dates separated by seven days to cover the fruiting season of Spondias spp.A total of 120 fruits were dissected.
Anastrepha obliqua larvae were extracted from each of the fruits the same day they were collected.Larval instars were categorized based on the width of the cephalic capsule and the body length (mean ± SE) (Carroll and Wharton 1989).Larvae were dissected immediately after collection, and the frequencies of the immature stages of D. areolatus, or any other parasitoid species, were recorded following descriptions by Aluja et al. (2013) and Murillo et al. unpublished data.
Photographs were captured with a Motic Plus 2.0® camera connected to a Carl Zeiss Smz -168® stereomicroscope.The D. areolatus immatures inside the A. obliqua larvae were measured using Motic Imagen Plus 2.0 ® software.The percentage of parasitized larvae was calculated, and frequencies of immature D. areolatus stages per larval instar of A. obliqua were determined.

Induction of D. areolatus parasitism on A. obliqua eggs and recently hatched larvae
Wild A. obliqua flies were collected as larvae from infested S. mombin fruits in the field.Upon completion of their development, the larvae were placed in containers with sterile sand for pupation.They were maintained under these conditions until adult emergence.Adults were maintained with water and food (sugar plus hydrolyzed yeast in a 3:1 ratio) until they were sexually mature.
Hog plum (S. mombin) fruits were previously protected from natural infestation by bagging clusters of young fruits using 30 × 20 cm brown paper bags.A total of 30 bags (≈10 fruits/bag) were used to protect ≈300 fruits.The fruits were subsequently collected, taking their maturity into account to allow for experimental infestation.
Infestation on the previously protected fruits was induced by exposing the fruits to A. obliqua gravid females (8-10 days old) in Plexiglass cages (20×20×20 cm) placed on a table in the field.Two clusters with five to eight fruits were placed in each cage together with 10 female flies and remained in the field at a mean temperature of 28.2 °C (range: 23.2-36.1)and a mean RH of 81.6% (range: 55.1-95.3).Flies were maintained for six hours in each cage and dead flies were replaced.
Anastrepha obliqua eggs were exposed to the parasitoid in the same type of cages 24, 48, and 72 hours after fly oviposition in the fruit (≈egg age), in order to cover the different egg stages before larval hatching.Twenty 7-day-old D. areolatus females were placed in each cage for three hours, time enough to locate and oviposit in the exposed eggs.Immediately after exposure, 35 A. obliqua eggs and 15 newly hatched larvae were extracted from the fruits.The eggs were characterized as either yolk-egg or embryo-egg (after Chapman 2013, pp: 358-407).All of the A. obliqua individuals in the egg and larval stages were dissected to characterize and record the immature stages of D. areolatus.

Characterization of D. areolatus development and morphological changes
Development of D. areolatus eggs and larvae was individually photographed and measured.To follow the development of D. areolatus in A. obliqua pupae, mature A. obliqua larvae were obtained from presumably infested fruit collected in the field.These larvae were placed in 100-ml plastic containers with sterile sand as a substrate to facilitate pupation.Three cohorts of 50 A. obliqua pupae were examined, and 3 to 5 pupae per day were dissected from 0 to 12 days of growth.D. areolatus individuals and their developmental stages were recorded for each A. obliqua pupa.
The frequencies of the different immature developmental stages and the characteristics of D. areolatus were calculated for each immature stage of A. obliqua.All of the observations of the organisms were conducted using the above-mentioned microscope.

Data analysis
Chi-square test was used to compare the number of D. areolatus individuals observed at each developmental stage with the expected number, using SPSS Statistic 17.0.(SPSS Inc., 2008).Measurements of the cephalic capsules of A. obliqua larvae are given as the mean ± (SE).The proportions of the immature stages of D. areolatus for each A. obliqua egg and larval stages are presented as observed numbers, and D. areolatus development is presented as numbers of individuals and percentages.
D. areolatus was the dominant parasitoid species (93.1%), and only Utetes anastrephae (Viereck) (5.4%) was found as the second most dominant parasitoid (Figure 1a).The remaining 1.5% (third-instar larvae) was parasitized by both species and no apparent advantage was observed for either species, except for the occasional larger size of D. areolatus larvae (Figure 1b).D. areolatus and U. anastrephae larvae found together were first-instar larvae, which were easily distinguishable from each other, primarily because of the larger sizes of the cephalic capsule and the jaws of U. anastrephae (see Fig. 1).
The mean ± (SE) of the widths of the cephalic capsules and the body lengths, respectively, of A. obliqua larval instars were 0.09 ± 0.001 mm and 0.90 ± 0.05 mm for the L 1 , 0.37 ± 0.03 mm and 4.67 ± 0.3 mm for the L 2, and 0.63 ± 0.004 mm and 9.16 ± 0.3 mm for the L 3 .The numbers of the developmental stages of D. areolatus recorded in the various stages of naturally parasitized A. obliqua are given in Table 1.Embryo-eggs and a high frequency of early first-instar larvae were detected in A. obliqua first-instar larvae.In second-instar larvae, the presence of D. areolatus eggs was minimal, with a higher frequency of intermediate first-instar larvae.In the third-instar larvae, nearly all of the D. areolatus were late first-instar larvae.The relationship between the immature stages of D. areolatus and the larval states of A. obliqua was significant (χ 2 4 = 800.9,P < 0.0001).

Presence of D. areolatus eggs in recently hatched A. obliqua larvae
In the controlled infestation experiment, A. obliqua yolk-eggs were not parasitized, which indicates that D. areolatus did not parasitize eggs without a formed embryo (Table 2).However, parasitism was detected in A. obliqua embryo-eggs.A recently laid D. areolatus egg (with yolk in its interior) on an A. obliqua embryo can be observed in Figure 2 and is folded in the embryo's interior, given that both structures are of a similar length.Seven recently hatched A. obliqua larvae that were parasitized by D. areolatus eggs were dissected; five were still in the yolk stage and two in the embryo stage (Figure 3a).Superparasitism by D. areolatus was recorded in two out of seven recently hatched A. obliqua larvae, and six parasitoid eggs were recorded from one larva (Figure 3b).
Fifteen mature first-instar A. obliqua larvae were dissected.Of these, 13 were parasitized with early first-instar larvae and 2 with embryo-eggs of D. areolatus.Yolk-eggs of D. areolatus were not found (Table 2).The relationship between the immature stages of D. areolatus and the immature stages of A. obliqua was significant (χ 2 4 = 22.4,P < 0.0001).

Characterization of D. areolatus development
The D. areolatus egg measures ≈ 1 mm long and has an elongated shape and a whitish color with a dark yolk.After 24 hours, the embryo is formed with a claviform appearance and measures ≈ 0.5 mm in length (Table 3 and Figure 4).
The embryo of D. areolatus becomes an early first-instar larva within 24 to 36 hours of oviposition and measures ≈ 0.8 mm in length.After three to four days, an early firstinstar larva grows into an intermediate first-instar larva, with a length of ≈ 1.4 mm.After another three to four days, the larva grows into a late first-instar larva measuring ≈ 1.7 mm in length.This larva almost immediately changes to a second-instar once the host pupa has formed, increasing in size and changing its shape (Table 3 and Figure 4).In recently formed A. obliqua pupae, a higher frequency of late first-instar larvae of D. areolatus in the process of transformation to the second instar stage were observed.In 1-day old pupae, the D. areolatus larva had changed completely to the second instar stage, measuring ≈ 3.5 mm long, losing the cephalic capsule and occupying more than a third of the host pupa.In 3-to-4-day old pupae, the larva develops into the third instar stage, measuring ≈ 6.0 mm long, changing body shape, and occupying all of the host pupa.In 6 day-old A. obliqua pupae, D. areolatus pre-pupae that exhibit eye development have formed.In 8-day old host pupae, the parasitoid pupae are already observed with well-defined structures and genitalia.In 12-day old and older A. obliqua pupae, the parasitoids are found as their complete adult structure (Table 3 and Figure 4).

Discussion
The presence of eggs and larvae of D. areolatus in the interior of eggs and recently hatched A. obliqua larvae represent a novel finding within the native parasitoid guild that attack fruit flies in the Neotropics, because there have been no previous reports of any native parasitoid covering this ecological niche (López et al. 1999, Ovruski et al. 2000).
Among the particular observations regarding this finding under forced conditions, it was notable that D. areolatus oviposit inside embryo-eggs of A. obliqua, depositing a flexible egg that can fold inside the interior of the host embryo, and that the first-instar larvae present a prolonged development with three distinct sizes that synchronize with the development of the host larva and pupa.The low number of embryo-eggs found with egg parasitoids could be explained by the short developmental time of eggs (less than 24 hours).
In Mexico, D. areolatus has been reported to be closely associated with A. obliqua in fruit hosts of the genus Spondias (López et al. 1999, Ovruski et al. 2000, Sivinski et al. 2000), which most likely is favored by the presence of semiochemicals that could allow it to attack early stages of A. obliqua (eggs and recently hatched larvae).This scenario seems similar to that of F. arisanus, which detects marking pheromones and kairomones that emanate from the eggs of its host or from the interaction of the fruit and the host egg (Rousse et al. 2005, 2007, Pérez et al. 2013).
The finding of the parasitism of D. areolatus on eggs and recently hatched larvae of A. obliqua sheds light on two relevant aspects of its role as a natural enemy and biological control agent of fruit flies: 1) its competition and coexistence with other opiine parasitoids, highlighting the exotic species Diachasmimorpha longicaudata (Ashmead) and the native species U. anastrephae (García-Medel et al. 2007, Sivinski et al. 1998, Paranhos et al. 2013, Aluja et al. 2013), and 2) the promising development of its mass rearing under laboratory conditions (Eitam et al. 2003, Aluja et al. 2009).
It has been argued that D. areolatus is an inferior competitor compared to D. longicaudata (Sivinski et al. 1998, Eitam et al. 2004) and to U. anastrephae (Aluja et al. 2013).It has also been suggested that D. longicaudata, given its larger ovipositor, could cause a local extinction by displacing D. areolatus when deprived of free space left by its competitors, as it has been suggested to explain the reduction of the dispersion range in Florida, USA, in the presence of D. longicaudata (Sivinski et al. 1998).For U. anastrephae, it has been suggested that the historic sympatry of D. areolatus and U. anastrephae depends on the ability of D. areolatus to avoid competition with the intrinsically superior competitor by exploiting hosts in larger fruits that are out of the reach of the smaller ovipositor of U. anastrephae (Sivinski et al. 1997, Aluja et al. 2013).However, our new findings suggest that both hypotheses can be reformulated in relation to the biology and oviposition behavior of D. areolatus.
An early action of D. areolatus against immature A. obliqua could represent an ecological advantage that prevents its displacement or local extinction by other competitors such as U. anastrephae and D. longicaudata, since these latter species invariably will attack mature larval stages that could already be parasitized by D. areolatus.According to Wang et al. (2003Wang et al. ( , 2008)), in F. arisanus this earlier attack increases the probability to suppress the invasive larva through starvation or suffocation mechanisms.Field observations seem to support this assertion.Even though D. longicaudata has become established in numerous sites in Mexico where it has been released, its presence in Spondias species is inferior to that of D. areolatus (López et al.1999, Sivinski et al. 2000, Montoya et al. unpublished data).Our data showed that this species also competes successfully against U. anastrephae by parasitizing eggs and recently hatched larvae, which enables D. areolatus to become the dominant parasitoid species in these hosts.
Laboratory studies have reported that D. areolatus is an inferior competitor relative to D. longicaudata and U. anastrephae (Paranhos et al. 2013, Aluja et al. 2013) because larvae of D. longicaudata and U. anastrephae kill larvae of D. areolatus during competition through sequential exposures.However, these studies were conducted using mature host larvae (3rd instar), which, according to our results, presents a disadvantage to D. areolatus.
One difficulty in rearing D. areolatus has apparently been the oviposition stimuli in oviposition units (artificial devices with third-instar host larvae mixed with food) (Eitam et al. 2003), which has not been a problem in the case of D. longicaudata because this species detects its host by larval vibrations when feeding or moving (Lawrence 1981).However, in light of our new findings, new perspectives on laboratory rearing of this species are realized.The use of late-stage eggs or neonate larvae, similar to Fopius arisanus rearing (Harris et al. 1991, Zenil et al. 2004, Rousse et al. 2005, Montoya et al. 2009), should be tested.
Unsuccessful attempts have been made to rear D. areolatus using fruits with third instar host larvae (Eitam et al. 2003).This could be because D. areolatus females need different stimuli, such as chemical signals emitted by host (eggs or young larvae) interacting with fruit volatiles, as has been demonstrated for F. arisanus (Rousse et al. 2007, Pérez et al. 2013).
The preimaginal development of D. areolatus in A. obliqua required approximately 27 days, and 7 preimaginal stages were classified: egg, three larval stages, prepupa, pupa, and pharate adult.The morphological observations of the preimaginal stages of D. areolatus are in agreement with what has been reported for F. arisanus and D. longicaudata (Rousse et al. 2005, Carabajal-Paladino et al. 2010).However, the development of D. areolatus is more akin to F. arisanus because both exhibit a facultative synchronization between their larval development and that of its host, which is reflected in the long period of time for the first-instar larva, which finally changes when its host reaches the prepupal stage (Rousse et al. 2005).This type of development allows the second or third instars of F. arisanus (and probably D. areolatus) to occupy most of the available space inside the pupa, which facilitate the elimination of competing larvae (Wang et al. 2003(Wang et al. , 2008)).
Our study shows that D. areolatus can parasitize A. obliqua eggs and recently hatched larvae, giving an advantage over other parasitoids that attack the later-stage larvae.This finding represents a novel report regarding the oviposition behavior of this species, suggesting that it may occupy an ecological niche that was previously thought empty in the Americas.These findings also open new perspectives for the biological control of fruit flies.If mass rearing methods are developed, this will allow release of the most dominant fruit fly parasitoid species in the Neotropics.

Figure 1 .
Figure 1.Parasitoids found in naturally parasitized A. obliqua larvae.a First instar larva of U. anastrephae found in a third instar larva of A. obliqua and b Larvae of U. anastrephae and D. areolatus found together in a third instar larva of A. obliqua.Scale bars = 1 mm.

Figure 2 .
Figure 2. D. areolatus parasitizing an A. obliqua embryo.a A. obliqua egg embryo b A. obliqua embryonic egg removed c D. areolatus egg extracted from A. obliqua embryo, and d D. areolatus egg.Scale bars = 1 mm.

Figure 3 .
Figure 3. Parasitized A. obliqua newly hatched larvae a with a D. areolatus egg inside and b with six eggs of D. areolatus inside.Scale bars = 1 mm.

Figure 4 .
Figure 4. Development of immature stages of D. areolatus.a egg yolk b egg embryo c early first instar larva d intermediate first instar larva e late first instar larva f second instar larva g third instar larva h jaw of third instar larva i prepupa j male pupa, and k female pupa with her ovipositor.Scale bars = 1 mm.

Table 2 .
Numbers of D. areolatus developmental stages found in A. obliqua eggs and first instar larvae when S. mombin fruits were infested in a controlled manner.

Table 1 .
Numbers of observed individual stages of development of D. areolatus recorded in the different larval stages of A. obliqua extracted from field-collected hog plums (S. mombin).