The soybean aphid is one of the most important insect pests of soybean in the North Central United States (Ragsdale et al. 2007). The feeding injury caused by soybean aphids can result in reductions of photosynthetic rates, plant height, the number of pods per plant, seeds per pod, seed coat quality, seed weight, and increased pod abortion (Lin et al. 1993, Macedo et al. 2003, Beckendorf et al. 2008). Yield losses of 40–50% have been reported (Wang et al. 1994, Ragsdale et al. 2007). Indirect injuries from aphid transmitted plant viruses are also associated with soybean aphid feeding (Hill et al. 2001).

In the United States, chemical control with insecticides is the most common management strategy employed to control soybean aphid populations (Myers et al. 2005, Ragsdale et al. 2007, Magalhaes et al. 2008, Chandrasena et al. 2011, Hodgson et al. 2012, Krupke et al. 2017). Over a period of ∼6 years, the pest potential of this species throughout the North Central United States affected conventional management practices, increased scouting activities, increased production costs (Ragsdale et al. 2007), and increased insecticide use (Song and Swinton 2009, Ragsdale et al. 2011).

In recent years, neonicotinoid-treated soybean seeds have been planted over large areas as a prophylactic treatment targeting several early-season soybean pests (Magalhaes et al. 2008, Johnson et al. 2008). In addition, neonicotinoids represent an important active ingredient of several foliar-applied insecticides registered for use on soybean. This widespread and chronic exposure over consecutive generations and all instars of soybean aphid is likely to increase the risk for resistance development (Magalhaes et al. 2008). Although the potential for soybean aphid to cause significant loss is high and chemical control represents the only management option in this region, there is a general lack of information on the evolution of insecticide resistance in this pest (Chandrasena et al. 2011). Insecticide resistance increases costs of pest management, human health, and environmental risks (Pimentel et al. 1992). Conversely, the rational and judicious use of insecticides may mitigate the economic and environmental impacts of chemical control, reduce the frequency of resistance alleles among pest populations, and increase profitability (ffrench-Constant and Roush 1990, Pimentel et al. 1992).

Insecticide resistance in soybean aphids have been reported in China, where resistance was reported at low frequency to the organophosphate insecticide, omethoate (Qian et al. 2012). In another study, laboratory selection with the pyrethroid lambda-cyhalothrin resulted in increased levels of resistance (76.6-fold) (Xi et al. 2015). In addition, the results from the synergism experiments, suggested high risk for cross-resistance to chlorpyriphos, acefate, esfenvarelate, cyfluthrin, cypermethrin, bifenthrin, carbofuran, and methomyl. However, in other aphid species, resistance to neonicotinoids has become common due to a reliance on foliar and seed treatments with these compounds (Bass et al. 2015). Resistance to thiamethoxam has been reported in Myzus persicae (Bass et al. 2011, Bass et al. 2015), Aphis gossypii (Herron and Wilson 2011), and Brevicoryne brassicae (Ahmad et al. 2013). In M. persicae, neonicotinoid resistance has been associated with multiple mechanisms, including the overexpression of a single cytochrome P450-dependent monooxygenase (CYP6CY3), and a conserved mutation in the insect β1-subunit of the nicotinic acetylcholine receptor (nAChR) known as R81T (Puinean et al. 2010, Bass et al. 2011, 2014) which is the target site for neonicotinoid insecticides. The R81T amino acid substitution was also found to be present in the β1-subunit of the nicotinic acetylcholine receptor of an imidacloprid-resistant strain of A. gossypii and in other field populations from Eastern Asia (Koo et al. 2014).

Pests resistant to insecticides of different modes of action, or with several mechanisms of resistance associated with a single mode of action, represents major limitations to chemical control (Jeschke et al. 2010, Fuentes-Contreras et al. 2013, Yu 2014, Bass et al. 2015). For this reason, management strategies based on insecticides require resistance monitoring to ensure efficacy and longevity (Forrester 1990, Siegfried et al. 2007). Monitoring for insecticide resistance is a fundamental component of insecticide resistance management (IRM) programs that provide early detection of resistance (Siegfried et al. 2007, Guedes 2017) and inform proactive pest management strategies that minimize selective pressures and prevent or at least delay field control failures (Mota-Sanchez et al. 2006, da Silva et al. 2012, Caballero et al. 2013, Guedes 2017).

The detection of changes in the insecticide susceptibility levels of pest populations commonly relies on the use of validated and sensitive in vivo bioassays (Siegfried et al. 2007, Magalhaes et al. 2008, Snodgrass et al. 2008, Galdino et al. 2011, Gore et al. 2013). Insecticide bioassays can guide the selection of the most cost-effective, efficient, and selective products and commonly serve as a basis for classification and labeling of insecticides (ffrench-Constant and Roush 1990). Although in vitro assays can be an important and complementary tool to classic in vivo bioassays (Siegfried and Ono 1993a,b; Moores et al. 1996; Foster et al. 2000; Criniti et al. 2008), they are not a replacement as they are normally specific to only one resistance mechanism (Nauen et al. 2015).

In vivo bioassays are commonly used to calculate the lethal concentration required to kill 50% of the insects tested (LC₅₀) but may vary for a variety of reasons including bioassay method (Robertson et al. 2007), arithmetical method (Saganuwan 2011), rearing conditions (Beranek 1974, Cueto et al. 2006), insect taxa (Stark and Sherman 1989), physiological development (Qu et al. 2015), prestress exposure (Maltby 1999), and symbiont composition (Ghanim and Kontsedalov 2009). In aphids, intraclonal genetic variation (Shufran et al. 2003) and epigenetics (Field and Blackman 2003) are also known to be associated with variation in responses to insecticides.

Bioassay methods should be tailored to aspects of a specific insect/insecticide system (Galdino et al. 2011), and it is essential to understand the inherent characteristics of the system (Gerami et al. 2013). Variation in bioassay response is always present and can even be observed between response curves obtained from the same sample, between samples of the same population and from samples of different populations (Robertson et al. 1995, Schaub et al. 2002). More importantly, significant differences between two populations with non-overlapping CL, do not always indicate resistance (Robertson et al. 2007). For this reason, estimating the 95% limits of intrapopulation variation is considered crucial for monitoring studies (Siegfried et al. 2007) and has been shown to guide the differentiation in response to toxicants between insect pest populations (Robertson et al. 1995).

The use of bioassays to monitor soybean aphid susceptibility to thiamethoxam is necessary to support the development and validation of a standard methodology for resistance monitoring in the North Central United States The present study was performed between 2012 and 2015 and documents the lethal and sublethal effects of systemic and contact exposure, as well as detection of potential shifts in response of soybean aphid populations to thiamethoxam.

Materials and Methods

Rearing Conditions

In total, 26 soybean aphid populations were obtained from collaborators between 2012 and 2014 as part of a North Central Soybean Research Program project (Table 1, Fig. 1). Infested leaves were gathered from soybean fields at different locations in South Dakota, North Dakota, Minnesota, Wisconsin, Iowa, Indiana, Michigan, Ohio, and Nebraska. In addition, two laboratory populations were obtained from USDA-APHIS National Biological Control Laboratory in Niles, MI, and from USDA-ARS laboratory in Brookings, SD. One potentially susceptible soybean aphid strain, originally collected in 2005, was maintained under pesticide-free laboratory conditions at Bayer Crop Science (Morrisville, NC). A subset from this strain, named laboratory reference colony (lab) was started in 2013 at the University of Nebraska and used as a reference strain in this study.

Prior to the establishment of soybean aphid colonies, the samples were inspected for parasitoid presence and quarantined until no parasitized aphids were found. An average of 200 aphids were used to establish each soybean aphid colony after quarantine. In addition, a sample of 100 aphids from each field population collected in 2012 were combined into a single colony, named Field_Pooled1 and in 2014 a set of 200 aphids were used to establish Field_Pooled2 population (Table 1). In total, 28 soybean aphid colonies were used as a source for all bioassays performed from 2012 to 2015. Soybean aphid populations collected in 2012 were bioassayed from 2012 to 2013, and those collected in 2014 were bioassayed from 2014 to 2015. The colonies were maintained on pesticide-free (cv KS4202) V2 soybean (Fehr et al. 1971) at 24 ± 3°C, 70 ± 5% relative humidity, and photoperiod of 16:8 (L:D) h inside individual plant growth chambers (model I-35VLX Percival Scientific Inc., Boone, IA) and transferred to non-infested seedlings every week.

Table 1.

Source description of A. glycines populations, year of collection, collection sites, and origin

Fig. 1.

Location of sites in North Central United States, where clones of soybean aphid monitored for thiamethoxam susceptibility were collected. For collection information, refer toTable 1.

Results

Detached-Leaf Bioassay

There was a significant interaction between year of monitoring and the mean LC₅₀ of thiamethoxam generated from all tested populations (F_{1, 24} = 10.55; P = 0.0034). The mean LC₅₀ increased from 2012 to 2014 (t = -7.98; d.f. = 24; P = 0.0034) (Fig. 3A). The lethal effects of thiamethoxam measured after 7 d of systemic exposure are shown in Tables 4 and 5 for 2012 and 2014, respectively. In 2012, the field strains SD4, WI1, WI2, and Field_Pooled1 differed significantly from the LC₅₀ of the lab colony with RRs of 5.7, 3.6, 6.3, and 3.7-fold, respectively. The high chi-square calculated for SD3, WI3, and NE populations indicate a significant departure (α = 95%) from the probit model. In 2014, the LC₅₀ of all tested populations were significantly higher when compared with the LC₅₀ of the lab colony. The mortality data of MI2, MI3, IA4, and ND4 did not fit the probit model as indicated by high chi-square values (Table 5). In 2012 the field strains of SD4, WI1, WI2, ND2, and Field_Pooled1 and the laboratory strain of SD1 were significantly higher than the LC₉₀ of the lab colony (Table 4). In 2014, all populations differed from the LC₉₀ of the lab colony with RRs ranging from 3.0- to 9.2-fold (Table 5).

Fig. 2.

Mean LC₅₀ of A. glycines populations exposed to thiamethoxam in contact vial bioassay. Mortality was measured after 24 h of exposure. Error bars represent the standard error of the mean. Means with different letters between columns indicate significant differences (P < 0.05) between field and laboratory populations across and within years of monitoring. Field populations (2012) n = 3,960, Field populations (2014) n = 5,400, Susceptible laboratory colony (2012) n = 720, Susceptible laboratory colony (2012) n = 4,320.

Table 2.

Thiamethoxam contact vial bioassays performed on all group ages of A. glycines collected from soybean fields, North Central United States in 2012

Table 3.

Thiamethoxam contact vial bioassays performed on all group ages of A. glycines collected from soybean fields, North Central United States in 2014

Fig. 3.

Mean LC₅₀, and EC₅₀ of field A. glycines populations exposed to thiamethoxam in systemic detached-leaf bioassays. Mortality (A) and percent population growth inhibition relative to controls (B) were recorded after 7 d. Error bars represent the SEM. LC₅₀ and EC₅₀ means with different letters between columns indicate significant differences among treatments (P < 0.05). Field populations (2012) n = 12,856, Field populations (2014) n = 22,557.

Table 4.

Thiamethoxam detached-leaf systemic bioassays performed on all group ages of A. glycines collected from soybean fields, North Central United States in 2012

The sublethal effects of thiamethoxam measured as the concentration to cause 50% inhibition of population growth (EC₅₀) generated from all tested populations indicated a significant interaction between the year of monitoring and mean percentage population growth inhibition (F_{1, 24} = 16.78; P = 0.0004). An increase of EC₅₀ values were observed from 2012 to 2014 (6.03 ± 1.47%) (t = –4.10; d.f. = 24; P = 0.0004) (Fig. 3B). The EC₅₀s of all soybean aphid populations to thiamethoxam for 2012 and 2014 are shown in Tables 4 and 5, respectively. In 2012, the field strains from SD4, WI1, WI2, WI3, and Field_Pooled1 were significantly higher than the EC₅₀ of the lab colony. In 2014, the EC₅₀ of all tested populations differed significantly from the EC₅₀ of the lab colony with RRs ranging from 2.5 to 15.8.

Discussion

Both bioassay techniques employed in this monitoring study were shown to be sensitive in detecting shifts in insecticide susceptibility among soybean aphid populations. The glass-vial bioassay has been used to monitor and detect resistance to insecticides in a multitude of insect-insecticide systems, including sap-sucking pests, such as Bemisia tabaci (Prabhaker et al. 1996), Bemisia argentifolii (Sivasupramaniam et al. 1997), stink bugs complex (Willrich et al. 2003, Snodgrass et al. 2005, Nielsen et al. 2008) and other aphid species, such as Myzus persicae (Shean and Ranshaw 1991), Aphis craccivora, Rhopalosiphum maidis (Tang et al. 2013), and Diuraphis noxia (Bayoun et al. 1995). To a lesser extent than the vial assay, toxicological studies were performed using detached-leaf bioassays with Aphis glycines (Magalhaes et al. 2008) and Cerotoma trifurcata (Tietjen et al. 2017). Nonetheless, other similar systemic uptake bioassays have been extensively used in other pest programs (Cahill et al. 1996, Prabhaker et al. 1996, Prabhaker et al. 2005, Castle et al. 2014, Matsuura and Nakamura 2014). In general, the glass-vial bioassay was shown to be reliable, more rapid, and requires less insecticide when compared with the detached-leaf method. Another advantage of the glass-vial bioassay is that it allows resistance monitoring of populations collected directly from the field (McCutchen et al. 1989) because the vials can be pretreated and no insect rearing is necessary prior to experiments (Sivasupramaniam et al. 1997). However, the systemic detached-leaf bioassay is more representative of the route of exposure encountered under field conditions. This assay also allows the measurement of chronic effects of sublethal concentrations at a population level (Magalhaes et al. 2008). The disadvantages of the systemic bioassay include increased labor and dependence on plant material, and it is much longer in duration. Although these attributes of systemic bioassays limit utility for standardized resistance monitoring, the assay still is able to detect shifts in susceptibility levels, such as performed by Magalhaes et al. (2008) and Tietjen et al. (2017).

Table 5.

Thiamethoxam detached-leaf systemic bioassays performed on all group ages of A. glycines collected from soybean fields, North Central United States in 2014

Our data suggest that the magnitude of the RRs of all tested populations were similar in range for both assays when significant differences were observed between soybean aphid field populations and the lab colony (Tables 2–5). For example, in both bioassays and years of monitoring and at both lethal concentration levels (LC₅₀ and LC₉₀), the great majority of RRs were categorized as very low (RR = 2 - 10). However, in 2012, the LC₉₀ RRs of SD4 and WI2 populations achieved moderate levels (RR = 21–50).

Regardless of their origin (laboratory, treated, or non-treated field populations) or year of evaluation, soybean aphid populations were highly susceptible to contact and systemic exposure to thiamethoxam. Our results demonstrate significant effects of this insecticide against soybean aphids at lethal concentrations in both bioassays, and at sublethal concentrations in the detached-leaf bioassay. Subtle changes associated with sublethal exposure to insecticides may affect development, reproduction, longevity, feeding behavior, and dispersal (Guedes et al. 2016). Importantly, these subtle changes may also possess stimulatory effects on pest populations. For example, soybean aphids exposed to sublethal concentrations of imidacloprid showed no negative effects as well as higher reproduction and population growth rates when compared with untreated controls (Qu et al. 2015). This stimulatory effect or response to sublethal concentrations of insecticides is defined as hormesis (Wang et al. 2016), and the identification of such changes may help guide pest management decisions and explain insecticide-induced pest outbreaks (Guedes et al. 2016). In our investigation, no indication of hormesis was detected in soybean aphid populations exposed to sublethal concentrations of thiamethoxam, and nymph production was always higher under control treatments (data not shown).

Although the EC₅₀s RRs were all categorized as very low (RR = 2–10), there was a substantial increase in nymph production among soybean aphid populations between the 2 years of monitoring. However, no significant differences were observed in the EC₅₀s calculated for the lab colony during the same period (Tables 4 and 5), suggesting a possible decrease in susceptibility of field populations. This increase in the EC₅₀s could also be associated with other factors, such as the weight of insects (Robertson et al. 2007), seasonal environmental conditions related to the growing season sampling date, host plant quality, and rearing conditions (Godfrey and Fuson 2001). In addition, when aphid survivorship and population numbers at untreated detached-leaves were compared between the lab and field populations across years of monitoring, no significant differences were recorded. However, the number of living aphids of field populations was higher in 2014 (Fig. 4) when compared with 2012. Those results indicate that there is no difference in fitness between the lab and field aphids; therefore, the EC₅₀s calculated in 2014 represent a significant decrease in susceptibility when compared with 2012 estimates. In general, for both bioassay methods, the differences observed in susceptibility to thiamethoxam between 2012 and 2014 revealed increased but small levels of resistance from 2012 to 2014.

The concentration–response curves used to compare susceptibility of different populations are generally considered insensitive for detecting resistance at low frequencies (ffrench-Constant and Roush 1990). Consequently, using a single diagnostic insecticide concentration has shown to be more efficient and practical in detecting resistance (or small shifts in tolerance) at low frequencies (Marçon et al. 2000). However, it is critical that baseline susceptibility data are generated in order to support the calculation of such concentrations. It should be noted that diagnostic bioassays based on a single lethal concentration or lethal dose provide limited information on population level responses (Stark and Banks 2003) as the slope of the response curve cannot be estimated. Therefore, the use of techniques based on bio-demographic parameters that are informative of an insect's population life span, prereproductive period, and fertility rates should be incorporated in toxicological studies (Walthall and Stark 1997). As a result, the detection of the total effects of insecticides at lethal and sublethal levels can be investigated (Kammenga and Riksen 1996, Santos et al. 2016).

Genetic analysis of soybean aphid population structure in North America has revealed low allelic polymorphism and moderate genotypic diversity (Michel et al. 2009). The narrow differences in the RRs obtained in the present study may reflect this lack of genetic diversity among soybean aphid populations as variability in susceptibility during both years was minimal (Tables 2–5). Furthermore, Michel et al. (2009) reported that the genetic variation in soybean aphid populations is strongly associated with time of collection, but not with geography. As a result, collecting soybean aphids for insecticide monitoring bioassays early in the season may improve the likelihood of detecting small changes in susceptibility that are linked with genotypic variability and minimize effects of clone selection in bioassay responses (Roush and McKenzie 1987, Robertson et al. 2007). For example, Godfrey and Fuson (2001) reported higher tolerance of A. gossypii to bifenthrin, chlorpyrifos, and triazamate when aphids were exposed to early-season conditions in comparison to aphids reared under the late-season conditions. In our study, all field aphid samples were collected between mid-June and early-September (mid-to late-season); however, as described by Wenger and Michel (2013), widespread gene flow during sexual reproduction associated with high interpopulation migration between soybean aphids may reduce the chances for significant differentiation between populations separated by geographic locations.

Fig. 4.

Mean number of living A. glycines exposed to thiamethoxam in systemic detached-leaf bioassays recorded after 7 d of exposure to untreated detached-leafs. Error bars represent the SEM. Different letters between columns indicate significant differences (P < 0.05) between field and laboratory populations across and within years of monitoring.

Because geography does not fully explain variation in response among soybean aphid populations, we combined subsamples of aphids from all field populations into two distinct pooled colonies in an attempt to increase their genetic diversity. The comparison of the LC₅₀s obtained for multiple bioassays of the Field_Pooled populations from 2012 to 2014 represents 1.10- and 3.65-fold increase in resistance for the systemic detached-leaf and contact vial bioassays, respectively. In part, the differences in response observed in our study may be simply associated with natural variation among tested populations, such as observed by Gerami et al. (2013) with A. gossypii in Iran. In detached-leaf bioassays, both the LC₅₀s and EC₅₀s were in the range of concentrations reported in another study with soybean aphids and thiamethoxam (Magalhaes et al. 2008). Herron and Wilson (2011) documented a reduction in susceptibility of A. gossypii to thiamethoxam using treated cotton leaf discs on field collections associated with control failures. Initially, a discriminatory concentration of 0.02 g/liter (20,000 ng/ml) of thiamethoxam were applied for each aphid cohort (25 aphids with 3 replications each), and subsequently, the populations with higher survivorship (%) were submitted to concentration–response bioassays. The authors observed an increase of 13.2-fold in resistance (ranging from 8.8 to 22) in a 2-year period (2007–2009). We observed slightly lower variation in susceptibilty among the tested field populations (RRs ranged from 2.98 to 10.15 in the vial bioassay, and from 1.16 to 6.94 in the detached-leaf bioassays from 2012 to 2014). In our investigation, no previous selection was made to discriminate individuals by percent survivorship such as described in the study by Herron and Wilson (2011). However, it is important to consider that although none of our calculated LC₅₀ RRs were categorized as moderate (RR = 21–50), high (RR = 51–100), or very high (RR = >100), it seems plausible that the high use of this mode of action in the North Central US soybean systems may be associated with the low but apparent increase in resistance observed in our investigation from 2012 to 2014.

Thiamethoxam has been highly effective against a broad-spectrum of crop pests, including sap-sucking arthropods such as aphids. For this reason, it provides a valuable tool in integrated pest management (IPM) and insecticide resistance management (IRM) programs, especially in cases where resistance to other modes of action has adversely affected pest management practices (Nauen and Denholm 2005). To date, no field control failures or resistance to neonicotinoid compounds has been reported in soybean aphids from the North Central United States. However, decreasing susceptibility of thiamethoxam in acute-contact bioassays and decreased inhibition of nymph production after chronic-systemic exposure advise for the importance of reducing the selection pressure imposed by neonicotinoid seed-treated soybean.

The concentration–response data generated in this study will be useful for monitoring susceptibility of soybean aphid field populations to thiamethoxam in future IRM programs. The implementation of such programs may increase the longevity and efficacy of neonicotinoids against soybean aphids and promote more sustainable pest management approaches in North Central US soybean systems.

Our results will also serve as a foundation to develop a standardized technique for monitoring the susceptibility of soybean aphids to neonicotinoid insecticides in North America. Further research is being conducted to elucidate the risks of short-term adaptations, and resistance evolution on soybean aphid field populations, and to identify the presence of a potential fitness-differential after multigeneration of systemic selection with thiamethoxam.