Introduction

Lygus spp., particularly L. hesperus Knight and L. elisus Van Duzee, (Heteroptera: Miridae) are the most serious pests of alfalfa, Medicago sativa, grown for seed in the Pacific Northwest and California. In addition to direct yield reductions caused by feeding on alfalfa flowers and seeds (Sorenson 1939), insecticides used to manage Lygus spp. can indirectly cause further reductions by negatively impacting the activity of the alfalfa leafcutting bee, Megachile rotundata (F.) (Hymenoptera: Megachilidae), the principal pollinator of alfalfa seed in the Pacific Northwest (Peterson et al. 1992). Moreover, insecticide-resistant Lygus populations have been reported in the Pacific Northwest (Xu and Brindley 1994). Establishment of an effective biological control program for Lygus spp. would benefit alfalfa seed production by reducing direct damage to alfalfa seed, minimizing the disruption of pollinators through reduced insecticide use, and delaying or preventing the development of insecticide resistance.

Native and introduced Peristenus spp. (Hymenoptera: Braconidae: Euphorinae) are known to parasitize Lygus spp. in the northeastern USA (Day 1996) and the Canadian prairies (Braun et al. 2001). Until recently, little information was available concerning parasitism of Lygus spp. in the Pacific Northwest (Mayer et al. 1998). Earlier surveys reported that Lygus spp. collected in Idaho and Utah were parasitized by P. pallipes (Curtis) (Clancy and Pierce 1966; Musebeck et al. 1951). More recently, a braconid wasp was found to be parasitizing a high percentage of L. hesperus nymphs collected in Idaho. Although this parasite is morphologically similar to P. pallipes and P. pseudopallipes (Loan), which are common in the northeastern USA, it was determined to be a new species, P. howardi Shaw, apparently native to the Pacific Northwest (Mayer et al. 1998; Day et al. 1999).

Because P. howardi has been found to parasitize a high percentage of both the first and second generations of L. hesperus in some Idaho and Washington locations, it may be a potentially important biological control agent for Lygus spp. in alfalfa seed and other seed, vegetable, fruit and forage crops in the Pacific Northwest (Mayer et al. 1998). Little is known about the biology, distribution, and extent of P. howardi parasitism of Lygus spp. in alfalfa seed fields and nothing is known about the effects of crop and pest management practices on its biological control potential. Research efforts have been hindered in that the reduced morphology of euphorine parasites renders P. howardi larvae indistinguishable from other parasites of mirids in the genera Peristenus and Leiophron Nees (Day and Saunders 1990). Furthermore, although P. howardi is apparently multi-voltine, only a small percentage of the larvae do not diapause prior to pupation and adult emergence. This diapause results in a 9–10 month delay in the recovery of data related to percentage parasitism and species composition from field research programs. Additionally, mortality of parasites during the rearing process for species identification can be 40% or higher resulting in a significant and unavoidable loss of data important to understanding the within-season impact of P. howardi (Day 1994). A reliable method for detecting Peristenus parasitism of Lygus spp. occurring in the Pacific Northwest that incorporates positive species identification, particularly that for P. howardi, would benefit research aimed at studying the ecology and biology of the Lygus-Peristenus host-parasite interactions and may prove useful for monitoring some of the biological control components of future IPM programs designed to control Lygus damage in alfalfa seed and other Pacific Northwest crops.

Tilmon et al. (2000) developed a two-step molecular method that uses the polymerase chain reaction (PCR) followed by restriction endonuclease digestion of amplimers to detect and identify several Peristenus spp. parasitizing L. lineolaris (Palisot de Beauvois) nymphs. Using a slightly different approach, Erlandson et al. (2003) developed species-specific PCR primers that allowed for the identification of Peristenus and Lygus spp., but found that their procedure was less sensitive in detecting Peristenus DNA than that of Tilmon et al. (2000). Neither of these studies included P. howardi, considered the most important Peristenus parasite of Lygus spp. in the Pacific Northwest (Mayer et al. 1998; Day et al. 1999). Therefore, we decided to modify the methods of Tilmon et al. (2000) to allow for the definitive separation of P. howardi from several other Peristenus spp. based on restriction endonuclease digestion of PCR amplimers. The specific purpose of this modification was to provide same-season identification of P. howardi from parasitized Lygus spp.

Materials and Methods

Specimens of adult L. Hesperus, P. howardi, P. digoneutis Loan, P. conradi Marsh, P. pallipes, and P. pseudopallipes were authoritatively identified by and obtained from W.H. Day (Beneficial Insects Research Laboratory, USDA-ARS, 501 S. Chapel Street, Newark, DE 19713). They were preserved in 95% ethanol and held at −25°C until used for DNA extraction. DNA was extracted from all insect species using the Qiagen DNeasy^® Purification System kit (Qiagen Inc.,  www.qiagen.com) and a modification of the Qiagen protocol for isolation of genomic DNA from insects. A whole, adult insect was placed in an autoclaved 1.5 ml microcentrifuge tube containing 180 µl of buffer ATL (from kit), 20 µl of 20 mg/ml proteinase K (from kit), and 40 µl of 10 mg/ml RNase A (Product No. R6513, Sigma-Aldrich, Inc.,  www.sigmaaldrich.com) and homogenized by hand with disposable microtube pestles. The homogenate was incubated in water bath at 55°C for 4 hours, after which 200 µl of buffer AL (from kit) was added and the tube vortexed. The homogenate was incubated at 70°C for 10 minutes after which 200 µl of 100% ethanol was added and the tube vortexed. The entire mixture was transferred to a DNeasy^® spin column in a 2 ml collection tube and centrifuged at 6000 × g for 1 minute. The spin column was transferred to a new collection tube, 500 µl of buffer AW1 (from kit) added, and again centrifuged at 6000 × g for 1 minute. The spin column was again transferred to a new collection tube, 500 µl of buffer AW2 (from kit) added, and the centrifugation repeated. Finally, the spin column was transferred to an autoclaved 1.5 ml microcentrifuge tube (with lid removed), 50 µl of nuclease-free water added, incubated for 30 minutes at room temperature, and centrifuged at 6000 × g for 1 minute. The eluate containing extracted DNA was stored at 4°C until used as a PCR template. Negative controls were prepared in parallel with all extractions by performing the above procedure without insect material included.

To estimate the DNA yield from the above extraction procedure, we devised a microplate procedure based on quantification of DNA by spot testing using ethidium bromide (Sambrook and Russell 2001). Ten µl of 2 mg/ml ethidium bromide in nuclease-free water were added to the appropriate wells of a conical-bottom, polystyrene ELISA plate followed by addition of 5 µl of extracted DNA. Standards were prepared from E. coli genomic DNA (Product No. D2001, Sigma-Aldrich, Inc.) in nuclease-free water at concentrations of 0, 1.25, 2.5, 5, 10, and 20 µg/ml and included in every assay. The plate was placed on a transilluminator emitting UV light at 312 nm and the wells of interest were photographed using a digital camera (CoolPix^® 990, Nikon, Inc.,  www.nikon-coolpix.com) with the room lights off. The image was transferred to a computer and converted to gray scale using Paint Shop Pro^® version 7.04 (Jasc Software, Inc.,  www.jasc.com). The gray scale image was opened in SigmaScan Pro^® version 5.0 image analysis software (SPSS, Inc.,  www.systat.com) and the density in the center of each well representing samples and standards was measured using the built-in point intensity tools of the software. The densities of the standards were adjusted by subtracting the density for the 0 µg/ml concentration from those for all concentrations and the results were fitted to an exponential function for estimating the DNA concentrations of the extracted samples.

PCR reactions were carried out in a PowerBlock^® I thermocycler (Ericomp, Inc., San Diego, CA) equipped with a heated lid using a protocol similar to that described by Tilmon et al. (2000), but adjusted to increase the volumes of all components included in each PCR reaction mixture. A portion of all DNA extracts was diluted 1:5 in nuclease-free water for use as the template in PCR reactions. The primers C1-J-2252 (Tilmon et al. 2000) and TL2-N-3014 (Simon et al. 1994) were synthesized by Integrated DNA Technologies, Inc. ( www.idtdna.com), diluted to 6.4 µM in nuclease-free water, and stored in 50 µl aliquots at −25°C. Each 50 µl PCR reaction contained 5 µl of each 6.4 µM primer, 10 µl of 5 mM MgCl₂, 25 µl of Promega PCR Master Mix (Promega Corporation,  www.promega.com), and 5 µl of diluted template DNA. Amplification was carried out in 35 cycles of 94°C for 60 seconds, 52°C for 60 seconds, and 72°C for 90 seconds. PCR products were electrophoresed in 10 cm, 1% gels (1:1 agarose: Synergel [FMC Corporation, Rockland, ME]) in 1× TBE buffer at 57 volts for 3 hours using a Sigma Model E0638 horizontal submarine electrophoresis unit (Sigma-Aldrich,  www.sigmaaldrich.com).

PCR products from amplification of P. digoneutis, P. howardi, P. pallipes, and P. pseudopallipes DNA were sequenced in a LI-COR ( www.licor.com) 4000L automated sequencing machine available through the University of Idaho Automated DNA Sequencing Facility. Amplified DNA from P. conradi was sequenced commercially by SeqWright DNA Technology Services, Houston, TX. Sequences were verified in both directions to maximize accuracy. Consensus sequences were derived from the chromatograms using PHRED (Ewing et al. 1998) for base calling, PHRAP (Gordon et al. 1998) for sequence assembly, and CONSED (Gordon et al. 1998) for sequence finishing and the resulting sequences were aligned using ClustalX v. 1.83 (Chenna et al. 2003). A phylogenetic tree based on the Hasegawa-Kishino-Yano substitution model (Hasegawa et al. 1985) was generated using PAUP 4.0b10 (Swofford 1998) with the inclusion of the sequence from Apis mellifera L. (Crozier and Crozier 1993) as a reference and the GenBank sequence for L. lineolaris (Palisot) (AF189240) as an outgroup.

Restriction sites were mapped in these five sequences to identify a restriction endonuclease that would produce species-specific restriction fragment-length polymorphisms (RFLPs) in the respective PCR products. We chose SfcI (restriction site: 5'…C^∇TRYAG…3') because the predicted fragment lengths provided the clearest separation of P. howardi from the other Peristenus species. PCR products obtained from amplification of all five Peristenus species were digested with SfcI following the protocol supplied with the enzyme (New England Biolabs, Inc.,  www.neb.com). All 20 µl digestion reactions were carried out in 0.6 ml PCR tubes containing 9 µl of PCR product, 2 µl of NE buffer 4 (50 mM K-acetate, 20 mM Tris-acetate, 10 mM Mg-acetate, 1 mM dithiothreitol, pH 7.9), 8 µl of 0.25 mg/ml BSA, and 1 µl of SfcI. Reactions were incubated for 1 hour at 25°C and the digestion products electrophoresed in 3% Low Range Ultra Agarose (Bio-Rad Laboratories,  www.bio-rad.com) gels in 1× TBE buffer at 57 volts for 4 hours.

Results

The Qiagen DNeasy Purification System proved to be very efficient in extracting DNA suitable for PCR amplification from adult Peristenus spp. Microplate quantification revealed that the extraction procedure produced 65–190 ng DNA in the 50 µl of final eluate. We subsequently found that this yield range remained consistent when extracting DNA from Peristenus larvae dissected from parasitized Lygus spp (data not shown). Consequently, the PCR amplifications in these experiments were seeded with approximately 1–4 ng of Peristenus template DNA.

The approximately 760 base pair PCR products from P. conradi, P. digoneutis, and P. pallipes were identical to those produced by Tilmon et al. (2000) using our DNA extracts with the same primers and amplification protocol (Figure 1). Similar PCR products were obtained using template DNA from P. howardi and P. pseudopallipes, indicating that the same region of the cytochrome oxidase I gene in these species was amplified. No PCR products were detected for L. hesperus or the negative extraction control. Consensus sequences of our amplimers for all five Peristenus spp. are deposited in GenBank under accession numbers AY626370, AY626371, AY626372, AY626373, and AY626374. ClustalX alignment of our sequences for P. conradi, P. digoneutis, and P. pallipes with the approximately 820 base pair sequences published by Tilmon et al. (2000) (AF189243, AF189241, AF189242) resulted in 99.5, 98.3, and 98.3% homology, respectively.

Nucleotide sequencing of the PCR products for P. conradi, P. digoneutis, P. howardi, P. pallipes, and P. pseudopallipes revealed fragment sizes of 766, 757, 758, 762, and 759 base pairs, respectively (Figure 2)Figure 2.. For these sequences, we found the phylogenetic relationships presented in Figure 3. The same tree topology was obtained by substituting the sequences published by Tilmon et al. (2000) for P. conradi, P. digoneutis, and P. pallipes.

Based on the SfcI recognition sequence, the predicted restriction-fragment lengths are: 96, 165, and 505 for P. conradi, 39, 60, 102, 117, 181, and 219 for P. digoneutis, 102, 155, 165, and 336 for P. howardi; 165, 237, and 360 for P. pallipes; 163, 237, and 359 for P. pseudopallipes. Electrophoretic banding patterns of SfcI restriction fragments produced by digestion of the PCR products were consistent with these predicted fragment lengths (Figure 4), although fragments less than 150 base pairs were very diffuse and difficult to resolve. The choice of SfcI for digestion resulted in the inability to separate P. pallipes from P. pseudopallipes, but P. howardi was easily distinguishable from all other Peristenus spp. tested.

Discussion

We have successfully extended the PCR protocol of Tilmon et al. (2000) to the amplification of DNA from P. howardi and P. pseudopallipes. Adoption and modification of a commercially available DNA extraction procedure (Qiagen DNeasy Purification System) did not impact the PCR results, but significantly improved our ability to more easily process large numbers of field-collected samples. Moreover, replacement of AluI with SfcI permitted the precise identification of P. howardi, the predominant Peristenus species parasitizing Lygus spp. in the Pacific Northwest. The predicted AluI restriction fragments indicated that P. howardi (76, 108, 574) could not be reliably separated from P. pseudopallipes (42, 66, 79, 572) and P. pallipes (42, 66, 80, 574) with agarose electrophoresis. In addition, we found that AluI did not completely digest P. howardi PCR products, further complicating the definitive identification of this species through RFLP analysis with AluI. This supports the speculation of Tilmon et al. (2000) that restriction site variation might result in the inability to identify new parasitoid species within the genus Peristenus. The restriction fragments produced by digestion with either AluI or SfcI did not allow for the separation of P. pallipes and P. pseudopallipes, but SfcI digestion accurately separated P. howardi from all other Peristenus species tested, fulfilling the primary objective of this research.

Our PCR protocol for Peristenus DNA amplification calls for relatively large volumes of all reagents in the reaction mixture. This was an intentional goal to reduce potential pipetting errors when the procedure is applied to routine analysis of large numbers of field-collected samples. This was accomplished by preparing five-fold dilutions of the template DNA and diluting the primers to 6.4 µM. Using 5 µl of undiluted template DNA proved to be excessive as the PCR reaction with this amount of template often produced little or no amplimer. Additional MgCl₂ was necessary as the concentration present in the Promega PCR Master resulted in 1.5 mM MgCl₂ in the final reaction mixture. We found that MgCl₂ concentrations less than 2.5 mM produced inconsistent amplification results.

Peristenus howardi was first detected and described as a new species in 1997 from parasitized L. hesperus collected in Idaho and parasitism rates can reach 80–100% in untreated alfalfa seed fields (Day et al. 1999; JDB, personal observation). However, the host and geographic ranges of P. howardi are largely unknown, as is the potential for use of this native species for biological control of Lygus spp. in seed, fruit, vegetable, and forage crops in the Pacific Northwest. Prior to this report, identification of P. howardi through molecular techniques had not been accomplished. This extension of a species-specific PCR protocol coupled to SfcI endonuclease digestion provides a useful tool for studying Peristenus incidence and distribution within Lygus spp. populations. In addition, it will greatly reduce the time required for sample identification and the loss of data resulting from mortality incurred during the rearing of larval parasites.