A comparison of molecular screening versus dissection and rearing methods to estimate parasitism of Lygus nymphs by Peristenus wasps: size matters

Abstract

Several broadleaf crops on the Canadian prairies suffer economic damage by Lygus bugs (Hemiptera: Miridae), populations of which are suppressed by parasitoids in the genus Peristenus (Hymenoptera: Braconidae). We used three different methods to assess levels of this parasitism for different instars of Lygus collected in southern Alberta, Canada. Screening Lygus for Peristenus DNA using a molecular polymerase chain reaction (PCR) method identified levels of parasitism ranging from 35 (second and third instars) to 47 (fifth instars) per cent. With dissection, parasitoids were recovered from 13 (second instars) to 44 (fifth instars) per cent of Lygus. For Lygus collected in the field and reared in the lab, Peristenus emerged from about 22 per cent of individuals. Our results show that use of PCR or dissection for fifth-instar larvae provides comparable estimates of parasitism. For earlier instars, PCR identifies levels of parasitism undetected using dissection – that is, 2.7-fold more for second instars in the present study. For the purposes of pest management and conservation biological control, dissection can provide adequate estimates of parasitism to inform a decision to reduce insecticides to protect Peristenus parasitoids.

Introduction

Feeding by Lygus bugs (Hemiptera: Miridae) causes economic damage to several broadleaf crops in North America. Affected crops include seed alfalfa, Medicago sativa (Fabaceae), faba bean, Vicia fava (Fabaceae), and canola, Brassica rapa and Brassica napus (Brassicaceae) (Otani and Cárcamo 2011). Lygus adults overwinter in leaf litter, emerge in spring to feed on the meristematic tissue of noncrop plants, and then move onto crops when the plants become suitable for feeding and oviposition. Lygus populations are bivoltine in southern regions of Canada and univoltine further north. Economically important Lygus species include L. lineolaris (Palisot de Beauvois), L. borealis (Kelton), L. elisus (Van Duzee), and L. keltoni (Schwartz), which vary in relative abundance with region, crop type, and stage (Cárcamo et al. 2002).

Lygus damage varies with crops. In canola, Lygus causes economic losses by reducing seed weight (ca. 200 kg/ha) when numbers reach more than three bugs per sweep at the early pod stage (Cárcamo et al. 2024). In faba beans, Lygus may not impact seed weight (Cárcamo, unpublished data), but even low abundance of one or fewer per sweep (Cárcamo, unpublished data) can downgrade seed quality by causing necrotic spots on seeds (Smith et al. 2019). In seed alfalfa, Lygus bugs are part of the pest complex that includes alfalfa weevil (Hypera postica) and alfalfa plant bug (Adelphocoris lineolatus), and together, they can drastically reduce yield (Soroka and Murrell 1993). Insecticides are often the only option for control, but those products lose efficacy when insecticide resistance develops (Menke 1954; Snodgrass 1996; Snodgrass et al. 2009; Dorman et al. 2020). For this reason, there has been sustained interest in the use of parasitoids as biocontrol agents to target the Lygus pest complex (Craig and Loan 1984; Broadbent et al. 2002, 2013; Seymour et al. 2005; Haye et al. 2006; Day 2007).

Lygus are parasitised by a complex of species in the genus Peristenus (Hymenoptera: Braconidae). Adult wasps oviposit in early-instar Lygus nymphs in the spring or summer, with the third-instar parasitoid larva emerging from older juvenile hosts to form a cocoon, pupate in the soil, overwinter, and then emerge as an adult (Brindley 1939; Loan 1965). Research on the use of Peristenus species as biological control agents has been hampered by a lack of taxonomic tools and knowledge on species diversity within the genus (Goulet and Mason 2006). Currently, membership in this species complex for the Canadian Prairies is recognised as containing P. dayi Goulet, P. mellipes (Cresson), and P. howardi Shaw (Zhang et al. 2017). In southern Alberta, P. dayi has a spring activity period on an unknown host, but it attacks mainly Adelphocoris lineolatus (Hemiptera: Miridae) and, to a lesser extent, first-generation nymphs of Lygus. Peristenus mellipes has a broader phenology from late spring to mid-summer and occurs mostly on the first Lygus generation (Loan 1965; Goulet and Mason 2006). The third species, P. howardi, emerges later in the season to attack Lygus nymphs of the second generation (Fernández et al. 2018). In the Pacific Northwest of the United States of America, P. howardii is bivoltine, or it is represented by two species (Seymour et al. 2005). In eastern Canada, P. mellipes has distinct temporally separated populations on the first and second generations of Lygus lineolaris (Carignan et al. 2007; Zhang et al. 2017), which were previously recognised as distinct species (P. mellipes and P. pseudopallipes (Loan); Goulet and Mason 2006).

Levels of parasitism by Peristenus can be assessed by dissecting Lygus nymphs to recover immature parasitoids, rearing Lygus for parasitoid emergence, or by using molecular polymerase chain reaction (PCR) methods to detect parasitoid DNA in the host. Dissection is the most convenient method (Day 2007) but is best done using fresh nymphs and is more prone to overlooking parasitoids at earlier stages of development. Host rearing requires holding nymphs for extended periods on suitable plant material until parasitoid emergence, and some nymphal mortality occurs during the process, affecting results. Use of PCR is time-consuming and requires specialised equipment and training; however, nymphs can be frozen or preserved in ethanol for processing, and PCR can identify higher levels of parasitism than either dissection or host rearing does. For Lygus collected in alfalfa fields near Saskatoon, Saskatchewan, Canada, Peristenus parasitism percentage was estimated by each method to be 71 per cent via PCR, 59 per cent via dissection, and 25 per cent via host rearing (Ashfaq et al. 2004). However, molecular screening does not necessarily provide the best estimates of parasitism. Dissection and PCR methods provided similar estimates of parasitism for Lygus in New York state, United States of America (Tilmon et al. 2000) and for mirid bugs in Europe (Gariepy et al. 2008). In their European study, Gariepy et al. (2008) suggested that parasitoids could be detected using PCR within the host at earlier stages of development, but they did not present data to support this assertion. Verifying whether the assertion is correct will aid understanding of parasitism in Western Canadian Lygus populations.

In the present study, we compared estimates of parasitism by Peristenus for different Lygus instars obtained using dissection versus PCR. To our knowledge, this type of comparison has not been done previously. We also assessed parasitism for some of the same collection sites using host rearing.

Materials and methods

Field collections

Samples of Lygus nymphs (mainly L. keltoni, with minor representation by L. elisus and L. borealis) were collected by sweep net from different sites in southern Alberta from 16 June through 26 July 2022 and from 15 June through 17 August 2023. Previous studies by Fernández et al. (2018) identified these periods as the ideal times to sample the primary Peristenus parasitoids occurring on the first and second generations of Lygus bugs when older instars are sufficiently abundant. Host plants included a mix of field and forage crop and weedy species (Table 1), documented during a more extensive ongoing survey, based on dissections, to be reported elsewhere (Pawluk, Ibarra-Galvis, Roy, and Cárcamo, unpublished data). The range of dates and host plants, shown in Table 1, allowed for recovery of different instars and an expected range of low to high levels of parasitism.

Table 1. Cultivated and uncultivated plants where Lygus nymphs were sampled to estimate parasitism by Peristenus wasps using dissection, rearing, and PCR screening in southern Alberta, Canada

Dissection method

Field-collected nymphs were stored in a 0.5-L plastic container with romaine lettuce at 12 °C for 24–48 hours before dissection. Ethanol diluted to 20–50% was used to anesthetise the nymphs in Petri dishes immediately before dissection with insect pins under a stereoscope at 10–20× magnification. The number of nymphs dissected for a sample ranged from 8 to 74. The number of parasitoid larvae were recorded and stored in 95% ethanol.

PCR method

In preliminary work, we compared the methods described by Tilmon et al. (2000) and Erlandson et al. (2003) to detect Peristenus developing in Lygus hosts. Both methods use primer pairs specific to Peristenus DNA. However, the Tilmon method uses a primer pair that amplifies a region of the mitochondrial gene cytochrome c oxidase 1 (CO1), whereas the Erlandson method uses a primer pair that amplifies the internal transcribed spacer region between the 5.8S and 18S nuclear rRNA genes (ITS2). When testing the two methods using the same sample template DNA, we found the Tilmon method to be more sensitive for detecting Peristenus DNA, an observation that Erlandson et al. (2003) also made. For this reason, we used the Tilmon method in the present study.

Field-collected nymphs were preserved in 95% EtOH shortly after they were collected. Before DNA extraction, nymphs were surface-sterilised by three washes with distilled water. The DNA was then extracted from individual nymphs using DNeasyVR Blood and Tissue kits (Qiagen Group, Hilden, Germany) as per manufacturer protocols. Amplification of Peristenus DNA was performed using the Peristenus Ia primer in conjunction with the primer, TL2-N-3014 (Fig. 1) and methods as described by Tilmon et al. (2000).

Figure 1. Polymerase chain reaction results for 20 samples of second-instar Lygus screened for the presence of Peristenus parasitoids using the primers Peristenus Ia and TL2-N-3014. When present, Peristenus DNA amplifies to form a band of 820 bp. Samples 1–4 and 8 (very faint) produced bands of the correct size and were considered positive for Peristenus (red arrows). All other samples were considered negative. No sample was run on the last lane (to the right of the negative control).

Host rearing method

Field-collected nymphs of mixed instars (second to fifth) were reared in 0.75-L (≤ 50 nymphs) or 5-L (> 50 nymphs) containers equipped with meshed screens at the bottom to contain the larger Lygus juveniles and allow Peristenus larvae to fall onto a container with fine, moist vermiculite in which they formed cocoons (Whistlecraft et al. 2010). Nymphs were fed romaine lettuce twice weekly until they died or reached the adult stage. Rearing took place at room temperature (22 °C), under a photoperiod of 16:8 hours light:darkness and relative humidity of approximately 40%. Parasitoid larvae were allowed to mature in the rearing containers for the remainder of the summer. In September and October, the vermiculite in each Petri dish was searched, and the number of cocoons were counted using a stereoscope at 10× magnification.

Statistical analysis

For comparison between dissection and PCR methods, we paired samples of Lygus for instar, host plant, and date and location of collection. The mean ± number of nymphs per sample was as follows: dissection – 12.6 ± 1.0 (range: 8–74); PCR – 9.9 ± 0.05 (range: 7–10). The number of parasitised individuals per sample was converted into percentage parasitism. We then tested for an effect of method on percentage parasitism using Wilcoxon signed-rank tests (critical P-value = 0.05). We repeated this analysis for second, third, fourth, and fifth instars and for all samples combined.

For comparison between dissection and host-rearing methods, we used a different set of Lygus samples. These samples were paired for instar (several instars/sample), host plant, and date and location of collection. The mean ± number of nymphs per sample was as follows: dissection – 27.4 ± 2.0 (range: 10–50); rearing – 47.7 ± 10.3 (range: 10–162). The number of parasitised individuals per sample was converted into percentage parasitism. We then tested for an effect of method on percentage parasitism using Wilcoxon signed-rank tests (critical P-value = 0.05).

With each Wilcoxon signed-rank test, we also performed a Pearson correlation to assess how well estimates of percentage obtained by using dissection predicted results obtained by using either PCR or host rearing.

All analyses were performed using Systat, version 13 (Systat Software, Inc, San Jose, California, United States of America; https://systatsoftware.com/).

Results and discussion

Of the three methods examined, PCR detected overall higher levels of parasitism than did dissection, which detected higher levels of parasitism than did host rearing. When PCR was compared with dissection, the difference was significant (P ≤ 0.028) for second, third, and fourth instars and for all instars combined, but no significant differences were found for fifth instars (Fig. 2; Table 2). This pattern is explained easily: parasitoids in early-instar nymphs are more often present as eggs or first instars and are more likely to be overlooked when hosts are dissected. When dissection was compared with host rearing, mean (± standard error) percentage parasitism was 31.1 (± 4.6) versus 19.8 (± 4.7), respectively (Wilcoxon signed-rank test: Z = –1.834, P = 0.067). This same order of sensitivity was reported for a survey of Lygus parasitism in Saskatchewan that did not assess parasitism of different host instars (Ashfaq et al. 2004).

Figure 2. Parasitism by Peristenus for different instars of Lygus. Each data point is based on a sample of Lygus for which a subset of individuals (typically 10) was assessed for parasitism either by host dissection or using polymerase chain reaction.

Table 2. Percentage parasitism (mean ± standard error) by Peristenus wasps for different instars of Lygus bugs. Comparisons are made between paired samples for which the host was either dissected or screened for Peristenus DNA using a polymerase chain reaction (PCR) method (Wilcoxon signed-rank test)

* Number of pairwise comparisons for samples typically containing 10 individuals (dissection: range of 8–74; PCR: range of 7–10).

Pearson’s correlation values provide further insight into the values of the three methods used in the present study. For estimates of parasitism obtained using dissection versus PCR of samples within instars and combined across instars, values ranged from 0.633 to 0.804 (P ≤ 0.002; Fig. 2). Thus, even though dissection underestimated parasitism, the two methods similarly ranked samples by percentage parasitism. In contrast, the value for estimates of parasitism obtained using dissection versus rearing was 0.316 (P = 0.116): percentage parasitism estimated from reared hosts did not reflect parasitism in paired samples of dissected hosts, and rearing therefore should be used only when adult Peristenus specimens are required.

Use of PCR versus dissection also clarifies when the host is parasitised (Table 1). With PCR, estimates of parasitism were unaffected by instar stage (Kruskal–Wallis test statistic = 1.838, df = 3, P = 0.607). This finding is consistent with reports that parasitism of hosts normally occurs in second and third instars (Loan 1965). With dissection, estimates of parasitism increased with instar (Kruskal–Wallis test statistic = 9.353, df = 3, P = 0.025), which could incorrectly be interpreted as evidence that cohorts of Lygus become more parasitised with maturity. In the present study, we collected at multiple time periods for the first and second generations of Lygus bugs and pooled all the results shown in Table 2. Because of this, we expected no significant differences among instars, as confirmed by the statistical test of samples from the PCR screening. With limited dates of sampling for a given generation of Lygus nymphs, one would expect major differences of parasitism among instars depending on when the parasitoid adults were active and the age distribution of nymphs, as shown in Goulet and Mason (2006). In our case, both methods suggested that we sampled at the early spectrum of the oviposition period because the highest incidence of parasitism occurred in the older fourth- and fifth-instar nymphs. A peak of parasitism in the younger instars would indicate sampling of nymphs too late in the season for a given site. This type of data is further complicated by the tendency of Lygus to be attracted to plants during their reproductive stage when flowering is more advanced, so that temperature alone may not accurately predict their phenology (Cárcamo, unpublished data). Detailed phenological studies among adjacent crops planted at different dates would elucidate these interactions.

Differences in parasitism levels among Lygus instars can be further confounded by host–parasitoid interactions. In Day’s (2007) study with L. lineolaris and P. digoneutis in New Jersey and with L. hesperus and P. howardi in Idaho, United States of America, parasitism was significantly higher in fourth instars than in fifth instars and also in fourth instars than in earlier first and second instars. The low levels of parasitism in the youngest instars was attributed to limited exposure time relative to older instars and was similar to the pattern that we observed in southern Alberta for P. howardi and our Lygus species complex dominated by L. keltoni. According to Day (2007), possible encapsulation and killing of Peristenus larvae by fifth instars of the two Lygus species may explain the pattern of lower parasitism in fifth instars relative to fourth instars. In the present study, PCR screening and dissections estimated higher parasitism in fifth instars than in younger nymphal stages. It therefore appears that encapsulation of our local P. mellipes and P. howardi does not occur to a major extent by the local assemblage of Lygus pests in southern Alberta.

Even with its greater ability to detect parasitoids, the PCR method used in the present study may still underestimate parasitism in Lygus nymphs, depending on the geographic region and the age of the parasitoid egg. The primers used were designed to detect Peristenus DNA and were validated during primer development using P. rubricollis (Thomson) (as P. conradi Marsh), P. pallipes (as either P. dayi, P. mellipes, or P. howardi), and P. digoneutis Loan (Tilmon et al. 2000). In eastern North America, several species of Leiophron (Hymenoptera: Braconidae) also parasitise Lygus nymphs (Goulet and Mason 2006), and studies conducted in that region using the primers we used in the present study would not detect those species. In Alberta, parasitoids of Lygus may include, albeit at low levels, tachinid flies (Clancy and Pierce 1966) that also would go undetected. Furthermore, parasitoid eggs less than three days old may not be detected consistently by PCR methods, as was shown in Gariepy et al.’s (2008) study with European Peristenus. We note that none of the three methods examined in the present study would have detected parasitism of Lygus eggs (Cárcamo et al. 2008) or adults (Day 1995). Eggs may be parasitised by Anaphes, Telenomus, and Polynema wasps at levels below 5% (Cárcamo et al. 2008), and adults can be parasitised, although very rarely, by other wasps and tachinid flies (Clancy and Pierce 1966, Day 1995; Broadbent et al. 2002).

Dissection and rearing, despite their lower accuracy, may still be required in addition to PCR screening to gain a more complete ecological understanding of the trophic interactions of parasitoids and their prey. For taxonomic studies of species that lack prior genomic inventories and primers, rearing parasitoids to the adult stage is the only option to identify them using morphological keys (Goulet and Mason 2006). For species that suffer high mortality during the rearing of the hosts, such as Lygus nymphs, and during the overwintering stage of the parasitoid (Gariepy et al. 2005), the best option is to rear large numbers of individuals under conditions that avoid crowding and to ensure proper moisture during and after overwintering. For native Peristenus of southern Alberta, we observed around 50% mortality during the cocoon stage, even before overwintering (Daniels, Cárcamo, and Ibarra-Galvis, unpublished data). Dissection is also the only method currently available to document cases of multiple parasitism of a Lygus nymph, which we have infrequently observed during our research (Pawluk, Ibarra-Galvis, Roy, and Cárcamo, unpublished data). This type of information, along with the relative size of the parasitoid, which is important for detailed study of phenology, would be missed if we relied solely on the use of PCR screening.

Conclusion

Previous studies by Ashfaq et al. (2004) and Gariepy et al. (2005, 2008) established the superior accuracy of PCR screening over dissection or rearing of Lygus nymphs to estimate Peristenus parasitism. In the present study, we demonstrate that the advantage of the PCR method is specific to the estimation of parasitism in younger nymphs – in particular, second instars. For the largest fifth instars and even for fourth instars, dissection can provide similar estimates at a fraction of the cost and with fewer human resources. For detailed phenological studies requiring assessment of parasitism by all instars of the host, including very small ones, PCR or other molecular screening methods are clearly ideal. If the goal is to assess only baseline levels of parasitism for integrated pest management or before or after release of an exotic biocontrol agent, then repeated time sampling and dissection of the older fourth and fifth instars from all Lygus generations should suffice. Barcoding tools can be further combined to determine the species composition of dissected samples (kept in 95% ethanol or frozen) for fauna with accurate genomic inventories.