Introduction
The Mediterranean fruit fly or medfly, Ceratitis capitata (Wiedemann) (Diptera: Tephritidae), is one of the most devastating fruit pests worldwide. In Spain, current medfly control has been based primarily on aerial and ground applications of organophosphate insecticides, especially malathion, mixed with protein baits, although the use of the naturally derived compound spinosad is being promoted as an alternative (Chueca et al., Reference Chueca, Montón, Ripollés, Castañera, Moltó and Urbaneja2007). Chemical approaches affect both the food quality and the environment and can induce resistance, as detected recently in Mediterranean populations that have become resistant to malathion (Magaña et al., Reference Magaña, Hernández-Crespo, Ortego and Castañera2007, Reference Magaña, Hernández-Crespo, Brun-Barale, Couso-Ferrer, Bride, Castañera, Feyereisen and Ortego2008). Effective control of C. capitata requires an area-wide, multitactic pest management program. In recent years, emphasis has been placed on implementing environmentally friendly methods to control medfly in Spain. To this end, different biological control approaches are currently being implemented. One of them involves identifying and conserving polyphagous ground-dwelling predators of the medfly (Urbaneja et al., Reference Urbaneja, Mari, Tortosa, Navarro, Vanaclocha, Bargues and Castañera2006). With respect to C. capitata, three developmental stages can be found in the soil of citrus orchards, late third-instars larvae, pupae and teneral adults, which remain on the soil until they are able to fly. All three stages are susceptible to being preyed upon by ground-dwelling predators.
Generalist predators play a major role in biological control of agricultural pests (Legaspi et al., Reference Legaspi, Legaspi, Meagher and Ciomperlik1996; Symondson et al., Reference Symondson, Glen, Wiltshire, Langdon and Liddell1996, Reference Symondson, Glen, Ives, Langdon and Wiltshire2002a,Reference Symondson, Sunderland and Greenstoneb; Morris et al., Reference Morris, Campos, Kidd and Symondson1999; Sheppard et al., Reference Sheppard, Henneman, Memmott and Symondson2004; Foltan et al., Reference Foltan, Sheppard, Konvicka and Symondson2005). Many ground-dwelling predators have been recorded in citrus orchards located in Valencia, Spain. Rove beetles (Coleoptera: Staphylinidae) are the most abundant-active group, representing 38.6% of the total number of predators collected, followed by spiders (Arachnida: Araneae) (28.9%), earwigs (Dermaptera) (18.0%) and ground beetles (Coleoptera: Carabidae) (12.7%) (Monzó et al., Reference Monzó, Sabater-Muñoz, Urbaneja and Castañera2007). Spiders are important generalist predators in agricultural habitats, but their effects on regulating pest populations are poorly known (Greenstone, Reference Greenstone1999; Hagen et al., Reference Hagen, Mills, Gordh, McMurtry, Bellows and Fisher1999; Marc et al., Reference Marc, Canard and Ysnel1999; Sunderland, Reference Sunderland1999). In a previous study, we found over 50 spider species present in eastern Spanish citrus orchards (Monzó et al., Reference Monzó, Sabater-Muñoz, Urbaneja and Castañera2007), with the most common being the generalist predator Pardosa cribata Simon (Araneae: Lycosidae), representing about 19.3%, present throughout the year on the ground in citrus orchards. We have found that P. cribata is a highly efficient predator on C. capitata under laboratory conditions (Monzó et al., Reference Monzó, Mollá, Castañera and Urbaneja2009). Nevertheless, whether this predator plays a role in regulating C. capitata under field conditions is as yet unknown.
Different approaches have been developed to deal with the inherent difficulties of detecting arthropod predation by generalist predators. Monoclonal antibodies and isoenzyme-electrophoresis have been used successfully to detect prey-specific protein within predators under field conditions (Hagler et al., Reference Hagler, Naranjo, Bradleydunlop, Enriquez and Henneberry1994, Reference Hagler, Naranjo, Erickson, Machtley and Wright1997; Agustí et al., Reference Agustí, Aramburu and Gabarra1999a; Symondson et al., Reference Symondson, Gasull and Liddell1999; Traugott, Reference Traugott2003). Nonetheless, these approaches show specific disadvantages and limitations (Juen & Traugott, Reference Juen and Traugott2005). Polymerase chain reaction (PCR)-based techniques are increasingly becoming recognized as valuable tools in ecological studies and, recently, have been reviewed for predator/prey identification and detection in the context of their contributions to biological control of arthropods (King et al., Reference King, Read, Traugott and Symondson2008).
At present, DNA-based techniques are the state-of-the-art for gut content analysis, mainly PCR-based approaches (Agustí et al., Reference Agustí, De Vicente and Gabarra1999b; Zaidi et al., Reference Zaidi, Jaal, Hawkes, Hemingway and Symondson1999; Chen et al., Reference Chen, Giles, Payton and Greenstone2000; Agustí et al., Reference Agustí, Shayler, Harwood, Vaughan, Sunderland and Symondson2003a,Reference Agustí, Unruh and Welterb; Sheppard et al., Reference Sheppard, Henneman, Memmott and Symondson2004, Reference Sheppard, Bell, Sunderland, Fenlon, Skervin and Symondson2005; Harper et al., Reference Harper, King, Dodd, Harwood, Glen, Bruford and Symondson2005). The advantages of this approach are: (i) the techniques to develop molecular probes have become cheaper and more available; (ii) candidate target regions have already been sequenced for a great number of insects, and are increasing; and (iii) once prey specific-primers have been published, they can be used in reproducible protocols worldwide.
Efficient molecular gut content analysis is a powerful tool to establish biological control programs targeting arthropod pests (e.g. Morris et al., Reference Morris, Campos, Kidd and Symondson1999; Fournier et al., Reference Fournier, Hagler, Daane, de León and Groves2008). These prey-detection PCR-based techniques have been applied in laboratory conditions to investigate predator-prey trophic interactions; however, applying them to field conditions requires additional work (Harwood & Obrycki, Reference Harwood and Obrycki2005). It appears that, at least in some cases, spiders play an efficient role in limiting pests in agro-ecosystems (King et al., Reference King, Read, Traugott and Symondson2008).
Here, we report on the development of species-specific PCR primers and protocols, utilizing them to detect C. capitata DNA in the guts of P. cribata and, thereby, developing a rapid and suitable protocol to track medfly predation by P. cribata under field conditions.
Materials and methods
Arthropods
Forty-two arthropod species (phytophagous or predaceous) were chosen to develop C. capitata specific primers (table 1). Live arthropods were collected mainly from fields in the Valencia region (Spain). Bactrocera species, except B. oleae (Gemli), were trapped in sticky traps in Queensland (Australia) by the DPI & F Staff. Anastrepha suspensa (Loew) was collected in guava fruits in Florida (USA). A laboratory strain of C. capitata, IVIA2000 (San Andrés et al., Reference San Andrés, Urbaneja, Sabater-Muñoz and Castañera2007), and the Vienna-8 tsl strain from the mass-rearing facility at Caudete de las Fuentes (Valencia, Spain) were used as reference strains and for field assays.
Arthropod species used to screen Cc-specific primers grouped as predator or prey, including order and family level, locality and country of sampling and assigned code.
* , not at taxonomical level; –, not presented in fig. 1. # Cc for the IVIA2000 strain and Ccv8 for the Vienna-8 tsl strain. VLC, Valencia; CS, Castellón; ALM, Almeria; FL, Florida; QL, Queensland.
Primer design
Medfly species-specific primers
Internal transcribed spacer 1 (ITS1) sequences from Tephritidae were retrieved from the GenBank database [AF307848 (C. capitata), AF189690 (Ceratitis rosa Karsch), AF276515 (Bactrocera cucurbitae Coquillett) and AF276516 (Bactrocera dorsalis (Hendel))] and aligned using GeneDoc software (Nicholas et al., Reference Nicholas, Nicholas and Deerfield1997), applying an MSA algorithm implemented in the software with blossom 62 as a scoring table, a constant cost length of 20, a gap open cost of 8 and gap extension cost of 4. Three primers (CcITS1-2dir, CcITS1-2rev and CcITS737) were designed on C. capitata ITS1 sequence using OLIGO v4 primer analysis software (Rychlik, Reference Rychlik1992) (table 2).
Primer sequences (5'–3') and PCR conditions.
Ta, annealing temperature.
Universal primers
A pair of primers (18S_lo1270 and 18S_up1060) was designed on a conserved region of 18S rDNA alignment from sequences AF096450 (C. capitata), AF012518 (Cicindela sedecimpunctata Klug), Z97594 (Forficula auricularia Linneus.), X89492 (Leptothorax acervorum (Fabricius)), EF012870 (Lasius (Acanthomyops) californicus (Wheeler)), AF487712 (Myzus persicae (Sulzer)) and AF062961 (Tetranychus urticae Koch) obtained as described above.
Amplification conditions
Each primer pair was used in 20 μl volume reactions, containing 300 nM dNTPs (Eppendorf AG, Hamburg, Deutschland), 1×DNA pol buffer (Biotools B&M labs S.A., Madrid, Spain), 3 mM MgCl2 (Biotools), 0.75 u DNA polymerase (Biotools), 10 pmol each primer and 10 ng of total DNA. Amplification profile was one denaturation step at 94°C for 2 min, 40 cycles at 94°C for 30 s, 50°C for 30 s and 72°C for 15 s, followed by a final extension at 72°C for 150 s. Amplification was performed in a Mastercycler® ep gradient-S thermal cycler (Eppendorf). PCR products were run in a 2% agarose D-1 low EEO (Pronadisa, Sumilab S.L., Madrid, Spain) gel in 0.5×TBE buffer, stained with ethidium bromide and visualized under UV light.
DNA extraction
All P. cribata specimens were frozen (−80°C, 20 min) and dissected, then total DNA was extracted from the opisthosoma following the ‘Salting-out’ protocol (Sunnucks & Hales, Reference Sunnucks and Hales1996), adding fresh Proteinase-K at 100 μg ml−1 after tissue homogenization. The prosoma of each spider was retained in order to compare the results with an immunological assay being developed in parallel (Monzó et al., unpublished data). Other specimens (table 1) were subjected either to the protocol of Latorre et al. (Reference Latorre, Moya and Ayala1986) or the ‘salting-out’ as described above for total DNA extraction of complete specimens. Total DNA was finally dissolved in 20–100 μl LTE-R (10 mM Tris-HCl pH 8.0, 0.1 mM EDTA pH 8.0, 6 μg ml−1 RNase A), depending on specimen size. DNA integrity was verified by gel electrophoresis in 1% agarose gel and concentration adjusted to 5–10 ng μl−1 for PCR amplification.
Test for cross-reactivity and sensitivity
Primer specificity was tested by assaying CcITS primer pairs with total DNA from several individuals of the two strains of C. capitata, several closely related species and a wide range of arthropods (prey and predators) potentially present in the same citrus orchards (table 1). Furthermore, we have included other tephritids species, considered exotic in Spain (except B. oleae), but potentially present in other regions where C. capitata coexist with these tephritids.
Sensitivity was determined by assaying both CcITS primer pairs with ten-fold dilution of C. capitata total DNA starting with 10 ng till 1:10−10 dilution.
Detection period
Live adult specimens of P. cribata were collected from citrus orchards close to the Instituto Valenciano de Investigaciones Agrarias (IVIA) in individual 150-ml containers. Spiders were starved (water was supplied daily on soaked cotton) for a seven-day period at 25°C and 16:8 h (L:D) photoperiod. After starvation, one medfly adult from the IVIA2000 colony was offered to each spider, and time was set to zero from the moment of medfly capture. Spiders were allowed to feed on medfly over a three-hour period (ingestion period). Afterwards, any remaining prey were withdrawn, and spiders were frozen immediately (t=0) or maintained in starvation (given only water) for 6, 12, 24, 36, 48, 72 or 96 h at 25°C and 16:8 h L:D until freezing at −80°C for subsequent molecular assay. Twenty replicates per time were conducted (except for 96 h where 19 spiders were available). Additional spiders (n=20) were starved for seven days and frozen for use as negative controls in the PCR.
DNA from each spider was tested in triplicate by PCR (Agustí et al., Reference Agustí, Shayler, Harwood, Vaughan, Sunderland and Symondson2003a). Each sample scored as zero was tested with the 18S primer pair to assess whether PCR failure was due to a lack of C. capitata DNA.
For each pair of primers, data of positive detections were subjected to Probit analysis using Proc Probit in PC SAS version 9.1.3. Chi-square (χ2) tests were performed to determine the fitting of data to the Probit model. Detectability half-lives (post-ingestion time during which 50% of positives were still detectable) were obtained (Chen et al., Reference Chen, Giles, Payton and Greenstone2000). In addition, a χ2 test of equality was performed to assess whether there were statistical differences between the two primer-pair data sets, by testing whether the slopes and intercepts of each data fitting were the same.
Field assay
The field test was conducted in a 1 ha clementine citrus orchard located in Bétera (UTM X722106 Y4388610; Z30 m altitude) surrounded by other orchards, with spontaneous natural cover crop and drip-irrigation system. High populations of P. cribata had previously been documented in this plot (Monzó et al., Reference Monzó, Sabater-Muñoz, Urbaneja and Castañera2007). About 40,000 Vienna-8 tsl sterile male pupae were deployed in the orchard to simulate a high natural medfly infestation (R. Argilés & I. Pla, personal communication). Twenty adult P. cribata were randomly captured by hand in the central area of the plot at 24, 48, 72, 96 h and seven days after the release of medfly pupae. All collected individuals were immediately taken to the laboratory and frozen for later DNA extraction. DNA from each spider was tested by PCR using the CcITS737 primer pair in triplicate and scored, as previously mentioned, depending on the number of positives. The percentage of samples with a positive PCR reaction was obtained for the entire assay and for every capturing day. To obtain data about the daily emergence rate (FTD) of Vienna-8 medfly adults, four modified McPhail traps were placed in the orchard, baited with trimedlure® and a tablet of the insecticide dichlorvos® (Biagro, S.L., Valencia, Spain).
Results
Primer design and cross-reactivity
The CcITS1-2 primer pair (CcITS1-2dir and CcITS1-2rev) was first designed on ITS-1 region alignment. CcITS1-2rev was designed on a conserved region of 5.8S rDNA close to ITS-1 end. This primer pair amplified a 330-bp fragment from C. capitata DNA, whereas no such product was detected in other tephritids or other arthropods tested (fig. 1a). However, this primer pair produced a smear background in all samples, except in negative controls, suggesting that the conserved CcITS1-2rev primer also binds to P. cribata DNA and to other non-target arthropods. A third primer, CcITS737, was then designed on a C. capitata specific region, which in combination with CcITS1-2dir primer (CcITS737 primer pair) gives a 130-bp fragment only detectable in C. capitata DNA (fig. 1b). The universal primer pair Univ18SrDNA was used to demonstrate C. capitata specificity of ITS-1 primers (fig. 1c), as well as to discriminate between unsuccessful CcITS amplification (absence of target-DNA) and a lack of DNA in the PCR reaction (target and non-target DNA).
Ceratitis capitata species-specific amplification pattern of ITS primer sets with 18SrDNA control on several arthropods. (a) Specific amplification of ITS1_2 primer pair. (b) Specific amplification of ITS737 primer pair. (c) Control amplification with Univ18SrDNA universal primer pair. M, 100 pb ladder; c-, PCR negative control. Samples 1 to 21 correspond to species coded in the same order in table 1 (Pc, Fru, Mba, Fau, Emo, Cca, Psr, Cmo, Tur, Mpe, Ago, Asp, Asu, Bol, Btr, Bne, Bmu, Bfr, Cro, Cc and Ccv8).
Both primer pairs were tested for sensitivity, detecting as little as 0.1 pg of C. capitata DNA (fig. 2).
Sensitivity of CcITS primer sets on C. capitata DNA serial dilutions. M, molecular marker 100 pb ladder; c-, PCR negative control.
Detection period
The proportion of positive detection data fitted the assumptions of the Probit model for both pairs of primers (χ2=0.2055, df=3, P=0.977 for CcITS1-2 and χ2=0.1519, df=3, P=0.985 for CcITS737) (fig. 3). Detectability half-lives were of 78.25 h and 78.08 h for CcITS1-2 and CcITS737, respectively. No statistical differences were obtained between both half-lives; indeed, when testing the equality for both sets of data, statistical differences were found for neither (χ2=1.9407, df=1, P=0.1636).
The Ceratitis capitata DNA detection probability curves in Pardosa cribata samples after feeding. Lines are fitted Probit model with 95% fiducial limits. Black upright triangles and solid lines, CcITS1-2 primer pair; White upside-down triangles and dotted lines for CcITS737 primer pair.
Field assay
A total of 100 P. cribata individuals were analyzed in the field experiment. Five percent of the spiders tested positive for the entire assay. When analyzing the data per sampling day (fig. 4), this rate rose to 15% at 72 h after artificial infestation. The number of medflies captured in the plot gradually increased from 24 h until 96 h after infestation, when the maximum was registered (84.2±18.3 FTD). From here on, the number of medfly captures decreased and leveled off at approximately 25 FTD.
Percentage of spiders that tested positive with CcITS737 pair of primers and mean (±SE) number of flies captured per trap and day (FTD) in the field test (
, % detections; –◆–, FTD).
Discussion
PCR can be used to detect prey within the gut contents of predators, thereby enabling specific trophic interactions to be studied among soil-dwelling arthropods, which cannot be examined by other approaches (reviewed in Symondson, Reference Symondson2002; Juen & Traugott, Reference Juen and Traugott2005; King et al., Reference King, Read, Traugott and Symondson2008). Accordingly, by using PCR-based prey detection analysis, we were able to track C. capitata predation by the wolf spider, P. cribata, in citrus orchards, where a wide range of alternative preys were also present, such as aphids, spider mites or collembolan.
Specific primers for C. capitata were designed in the ITS-1 rDNA region, due to their presence in multiple copies in each cell, enabling successful detection in predator guts as reported in other arthropod species (Agustí et al., Reference Agustí, De Vicente and Gabarra1999b; Zaidi et al., Reference Zaidi, Jaal, Hawkes, Hemingway and Symondson1999; Hoogendoorn & Heimpel, Reference Hoogendoorn and Heimpel2001) and by the sequence availability in databases. Although this region showed a high level of sequence diversity among tephritids, even at the species level, the C. capitata species-specific primers worked successfully with two non-related strains, the laboratory strain (IVIA 2000) and the sterile strain (Vienna-8), rendering them universal for C. capitata. The specificity of this primer pair for the tephritid C. capitata will allow its use in other regions where it coexists with other tephritid species and on which search of putative predators are planned.
The calculated DNA detectability half-live for both primer pairs is roughly in the range of those reported (96 h to 105.4 h) for other spider species (Harwood et al., Reference Harwood, Phillips, Sunderland and Symondson2001; Ma et al., Reference Ma, Li, Keller, Schmidt and Feng2005). As reported elsewhere, shorter amplicons are more robust than longer amplicons and can be detected for a longer period. In this study, we were able to detect up to 37% of positives 96 h after prey ingestion with the ITS737 primer pair. Certain traits involving both the molecular markers and the predator-prey interactions may help to explain the long detection times obtained. Our amplified PCR products were short enough for long-term detections, 120 bp and 330 bp for each pair of primers, respectively. The shorter the amplified sequences, the longer the detection time (Agustí et al., Reference Agustí, De Vicente and Gabarra1999b, Reference Agustí, De Vicente and Gabarra2000, Reference Agustí, Unruh and Welter2003b; Zaidi et al., Reference Zaidi, Jaal, Hawkes, Hemingway and Symondson1999; Hoogendoorn & Heimpel, Reference Hoogendoorn and Heimpel2001; de León et al., Reference de León, Fournier, Hagler and Daane2006). However, in this assay, the detection time corresponding to the primer pair amplifying the shortest sequence was nearly the same as the primer pair amplifying the longest sequence. As mentioned before, the primers were designed over a multi-copy DNA region, which also contributes to increasing detection times. There seems to be a correlation between sensitivity and detection efficiency, whereby marker sensitivity increases the detection of the prey in the predator gut contents. Thus, de Leon et al. (Reference de León, Fournier, Hagler and Daane2006) reported a sensitivity of 6.0 pg for the mitochondrial cytochrome oxidase subunit I (COI) and 25.0 pg for the COII, and detection efficiency was higher for COI regardless of fragment size. When testing the sensitivity of the markers described here, we were able to detect positives even at 0.1 pg. Due to the starvation adaptations, spiders' gut systems are prepared to store ingested food for long-term periods (Harwood et al., Reference Harwood, Phillips, Sunderland and Symondson2001). This feature also means DNA of the prey can be detected for longer (Greenstone et al., Reference Greenstone, Rowley, Weber, Payton and Hawthorne2007). The quantity of food ingested could also be a factor influencing detection times. Hagler et al. (Reference Hagler, Naranjo, Erickson, Machtley and Wright1997) found a positive effect of meal size on prey detection when carrying out monoclonal-antibody assays on the gut content of three species of insect predators (Geocoris punctipes (Say) (Hemiptera: Lygaeidae), Orius insidiosus (Say) (Hemiptera: Anthocoridae) and Hippodamia convergens Guerin-Maneville) (Coleoptera: Coccinellidae) by indirect ELISA, to detect different ingested quantities of eggs of the pink ballworm Pectinophora gossypiella (Saunders) (Lepidoptera: Gelechiidae). Pardosa cribata is similar in size to its Ceratitis capitata prey; therefore, one medfly provides it with a large amount of food. This means it is stored in the spider's gut system for longer, thereby making DNA detection easier. By contrast, Zaidi et al. (Reference Zaidi, Jaal, Hawkes, Hemingway and Symondson1999) did not find any correlation between quantity of food ingested by the predator and detectability. Our study has also shown that, beyond fragment length, primer design is crucial for amplification success. On average, higher and clearer amplifications were obtained with the shorter primer pair CcITS737 in contrast to Juen & Traugott (Reference Juen and Traugott2005), who found that their shorter primer pair gave the lower amplification bands.
In the field assay, up to 15% of spiders tested positive 72 h after medfly pupae deployment in a citrus orchard. Most medfly captures were recorded between 48 h and 96 h after pupae release, suggesting that most of the adults emerged in the citrus orchard during this period. The maximum rate of positive detections was just 24 h before the maximum FTD recorded, suggesting that P. cribata preyed upon C. capitata more actively when more adults were emerging from the pupae deployed. This is consistent with the predatory behavior established under laboratory conditions (Monzó et al., Reference Monzó, Mollá, Castañera and Urbaneja2009), who showed that P. cribata was able to prey on teneral adults and larval stages of C. capitata, but not on pupae; and predation rate was significantly higher on adults than on larvae.
We would like to highlight that the rates of predation obtained in this work indicate that P. cribata is able to play an important role in multi-tactic strategies, currently required to control C. capitata. The next step to developing such a strategy should study the conservation of P. cribata populations in the citrus ecosystem, such as cover-crop management, thereby improving its potential as a biological control agent.
This is the first study to demonstrate the use of DNA markers to track medfly predation by the wolf spider, P. cribata, in citrus agro-ecosystems, providing a new tool to estimate the role of this spider in biological-control conservation strategies.
Acknowledgements
This work was funded by FEOGA Cooperación, the Conselleria d'Agricultura, Pesca i Alimentació de la Generalitat Valenciana and INIA (RTA03-103-C6-01). We are thankful to N. Agustí (IRTA, SP) for assistance at the beginning of this work and to R. Argilés and I. Pla, (TRAGSA, SP) for providing the Vienna-8 pupae. We are also grateful to A. Melic (Araneae, Sociedad Entomológica Aragonesa, Zaragoza SP) for taxonomical advice, E. Carbonell and J. Pérez-Panadés for statistical analysis, and to A. Azanza, H. Montón, P. Vanaclocha, Ó. Mollá (IVIA, SP) and M. Llavador for technical assistance. We also thank DPI & F Staff for Bactrocera specimens' collection from Queensland (Australia) and P.A. Stansly (University of Florida, USA) for A. suspensa specimens' collection. C.M. was recipient of a PhD grant from the CSIC. B.S.M. is supported by an European Social Fund co-financial contract.
