Resistance of Wheat Accessions to the English Grain Aphid Sitobion avenae

The English grain aphid, Sitobion avenae, is a major pest species of wheat crops; however, certain varieties may have stronger resistance to infestation than others. Here, we investigated 3 classical resistance mechanisms (antixenosis, antibiosis, and tolerance) by 14 wheat varieties/lines to S. avenae under laboratory and field conditions. Under laboratory conditions, alatae given the choice between 2 wheat varieties, strongly discriminated against certain varieties. Specifically, the ‘Amigo’ variety had the lowest palatability to S. avenae alatae of all varieties. ‘Tm’ (Triticum monococcum), ‘Astron,’ ‘Xanthus,’ ‘Ww2730,’ and ‘Batis’ varieties also had lower palatability than other varieties. Thus, these accessions may use antibiosis as the resistant mechanism. In contrast, under field conditions, there were no significant differences in the number of alatae detected on the 14 wheat varieties. One synthetic line (98-10-30, a cross between of Triticum aestivum (var. Chris) and Triticum turgidum (var. durum) hybridization) had low aphid numbers but high yield loss, indicating that it has high antibiosis, but poor tolerance. In comparison, ‘Amigo,’ ‘Xiaoyan22,’ and some ‘186Tm’ samples had high aphid numbers but low yield loss rates, indicating they have low antibiosis, but good tolerance. Aphid population size and wheat yield loss rates greatly varied in different fields and years for ‘98-10-35,’ ‘Xiaoyan22,’ ‘Tp,’ ‘Tam200,’ ‘PI high,’ and other ‘186Tm’ samples, which were hybrid offspring of T. aestivum and wheat related species. Thus, these germplasm should be considered for use in future studies. Overall, S. avenae is best adapted to ‘Xinong1376,’ because it was the most palatable variety, with the greatest yield loss rates of all 14 wheat varieties. However, individual varieties/lines influenced aphid populations differently in different years. Therefore, we strongly recommend a combination of laboratory and long-term field experiments in targeted planting regions to identify varieties/lines that consistently show high resistance to S. avenae infestation.

Introduction evaluation of wheat tolerance is costly, and time-consuming, with inconsistent result being produced from field experiments [23,31,49,52].
Here, we conduct a comprehensive evaluation on the resistance mechanisms of 14 wheat varieties using the 3 classical resistance mechanisms. This research is expected to provide baseline information about the mechanisms that cause resistance in these varieties, and a theoretical basis on breeding resistant wheat varieties to reduce the impacts of aphids for integrated management of aphid.

Aphids
In the wheat field population of Shaanxi Province, China, more than 95% aphids are S. avenae, and less than 5% are bird cherry-oat aphid Rhopalosiphum padi L. and greenbug aphid Schizaphis graminum (Rondani) [3]. Therefore, we used S. avenae aphids in our experiment.
The stock population of S. avenae aphids used in the antixenosis experiment was originally obtained from a single wingless individual that was found on a winter wheat plant in the experimental fields. The stock population was maintained in a separate cage on wheat seedlings (variety 'Xiaoyan22') in a climate chamber. Over the 3-year period, each variety/line had a total of 6 replications in 6 fields/sections. Each replicated block of each variety/line covered a 20-m 2 area. The blocks were 4 m long, and consisted of 20 rows with 25 cm spacing (total 5 m). Before planting, 10 kg/666.7 m 2 of nitrogen (ammonium nitrate) and 25 kg phosphate fertilizer (potassium dihydrogen phosphate) were applied to the experimental plots. An additional top-dressing of 10 kg/666.7 m 2 nitrogen was added at the stem elongation stage (March) of winter wheat. To avoid assessing the mutual effect between different wheat varieties/lines planted in neighboring plots, data were only collected from a 12 m 2 (3 m × 4 m) area in the center of each repeated block. Natural infestation by adult S. avenae alatae was counted 3 times in the 12 m 2 center area of each repeated block, with 1 week separation between the end of March to mid-April, This is because, during this period, alatae migrated from wheat fields in southern China [53,54].
The second antixenosis experiment was performed under laboratory conditions from September 2011 to November 2011. In this experiment, aphid alatae were allowed to choose among the 14 wheat varieties. Five seeds from each variety/line were planted in separate plastic pots (i.e., 1 pot per variety). The 14 pots containing wheat seedlings at the 2-leaf stage were randomly arranged in a circle, and were covered with a 2 m × 2 m × 1.5 m rectangular gauze cage to stop the alatae from escaping. One-hundred adult aphid alatae were deprived of food for 24 h, and were then released into the center of the experimental plants in the cage. The number of adult aphid alatae that established on each variety/line was counted and recorded after 1, 2, 4, 8, and 24 h. This experiment was repeated 3 times. The randomly layout of pots within the three repeats was different.
The third antixenosis experiment was a palatability experiment, which was performed under laboratory conditions from February 2012 to May 2012. In this experiment, aphid alatae were allowed to choose between 2 wheat varieties. The 14 wheat varieties/lines were combined as 91 pairs (14 × 13 Ä 2). The leaves of each paired combination of wheat leaves were cut to a length of 3 cm, and were placed on opposite sides of a 9-cm-diameter Petri dish. The bottom of the Petri dish was covered with filter paper, on which 2 ml distilled water was added to keep the leaves moist. Subsequently, 20 alatae aphids that had been deprived of food for 24 h were released in 2 batches onto the center of the Petri dish. After 1, 2, 4, 8, and 24 h, the number of aphids that had established on the leaves of each cultivar/line was counted and recorded. This experiment was repeated 3 times.

Antibiosis experiments
Antibiosis is the antagonistic association between 2 species, in which at least 1 of the 2 species is adversely affected. It is a heritable quality possessed by a plant that adversely affects the life history or biology of the insect [50,51]. In this study, the antibiosis experiments focused on population dynamics, and were performed between 2003 and 2006 in the same fields/sections that were used for the antixenosis experiments.
Data collection was performed in the 12 m 2 central area of each repeated block. Five 1 m long rows were sampled per block. In this antibiosis experiment, all aphids (including alatae, apterae [wingless] adults, and nymphaea) from 5 points in each block were counted once a week from mid-March to the end of May (a total of about 10 times). Direct counts were used before the end of April, and scale counts, which were the average aphid number of 10 heads multiplied by count wheat head number in 1 m long rows, were used following the end of April. The maximum aphid number (MAN, or aphid peak number) and the aphid cumulative count (ACC) were integrated to elucidate aphid population development. The sum of the MAN and ACC at the 5 points in each repeat block was used as a single sample. The index of aphid susceptibility (IAS) was counted as: IAS = the number of aphids at each point/ the average number of aphids at all points combined [2] IAS 0.3 indicated high resistance, 0.3 < IAS 0.6 indicated intermediate resistance, 0.6 < IAS 0.9 indicated low resistance, 0.9 < IAS 1.2 indicated low susceptibility, 1.2 < IAS 1.5 indicated intermediate susceptibility, and IAS ! 1.5 indicated high susceptibility.

Wheat tolerance experiment in the field
Tolerance is a genetic trait of a plant that enables it to tolerate higher pest populations than a susceptible cultivar before damage occurs [50,51]. At plant maturity, the grain yield, the yield components, and their loss ratios were measured. The yield components included the number of heads in sampled 1 m rows, kernels per wheat head (the average of 30 heads), and the weight of 1,000 kernels (KW). The theoretical yield (TY) was calculated. The actual grain yields (AY) per 1 m row for all 5 sampled points were measured for each block. The TY, AY, and KW loss ratios were calculated using the following formula: loss ratio = (control value -actual value)/ control value [2] The control points were 1 m long rows. These points were located outside the 12-m 2 core area, but within the 20-m 2 area of each block. The control points were sprayed 3-5 times with imidacloprid (2.5% WP 8000X, 20 g a.i./hm 2 ; Shijiazhuang Yaoyuan Pharmaceutical Technology Co., Ltd., Shijiazhuang, Hebei, China) at the start of April and end of May to eliminate all aphid infestation. When spraying, both sides of the control line were protected with cardboard to prevent the pesticide from reaching the other wheat plants [2,3,28,29].

Data analysis
There were 2 fields in 2004, 3 sections in 2005, and 1 field in 2006. The results for each field or section were assumed to be independent in different years; thus, the 6 fields/sections were treated as a single factor in the field experiments when using ANOVA. The number of adult aphid alatae was very small in early spring; thus, the total number of alatae obtained from each repeated block was combined as a single sample for each wheat variety/line. Then, two-way (14 wheat varieties and 6 fields/sections) ANOVA without duplication was used in the first antixenosis field experiment. The number of aphid alatae was analyzed using a repeatable one-way ANOVA in the second and third antixenosis laboratory experiments. MAN and ACC were transformed using natural logarithm transformation to ensure homogeneity of variance. The loss ratios of TY, AY, and KW were transformed using arcsine transformation to reduce variance. IAS were not transformed. Then, these parameters were analyzed using a two-way (14 wheat varieties and 6 fields/sections) ANOVA with the duplication of random factors model.
All of the parameters were analyzed using ANOVA with General Linear Model (GLM). The means were separated by Tukey's test under α = 0.05 (Data Processing System, DPS, software). All figures were drawn with SigmaPlot 12.1 (Systat Software Inc., Chicago, IL, USA).

Results and Analysis
Antixenosis experiment

Antibiosis field experiment
The abundance patterns of the S. avenae population were similar among all 14 wheat varieties/ lines in all 6 fields/sections over the 3-year study period (Fig 4). Aphids were first detected on wheat seedlings in mid-March. Population numbers accelerated in mid-April. The MAN occurred in the first half of May each year. Population numbers declined in late May, and had disappeared by early June when the wheat was harvested.
In the ANOVA-based random model with 2 factors (14 wheat varieties and 6 fields/sections) variation existed among the 14 wheat varieties/lines with respect to MAN and ACC

Discussion
Our study demonstrates major differences in the crop yield loss rates and IAS of the 14 wheat varieties/lines among repeats, fields, and years. Previous studies also detected great disparity among areas and years for crop yield loss rates and IAS of the same germplasm [2,32,43,[55][56][57]. These differences arise because S. avenae is not evenly distributed within fields, but they form aggregations [58][59][60][61], with generally weak spatial associations and some edge effects [62]. Consequently, wheat varieties that were identified as being resistant may not necessarily be so; rather, these varieties may have been outside the core aphid aggregation area within each field. Thus, the screening and identification of resistant wheat varieties should reflect these aspects in the field. However, several experimental repeats conducted on a long period of time are required to confirm our observations. Alternatively, artificially infesting fields might prevent aggregations forming, resulting in a more evenly distributed spread of aphids within fields.
In the UK, France, Denmark, Romania, and Chile [63][64][65][66][67][68], there is low genetic differentiation among S. avenae populations across different geographical regions. In contrast, S. avenae exhibits high genetic differentiation across different geographical regions in China, and may be divided into different biotypes [69][70][71][72]. For instance, most S. avenae that are found in the wheat fields of northern China originate from winged offspring that have migrated north from wheat fields in southern China [53,54]. Some offspring also originate from eggs that overwinter locally or from adults in non-extreme cold years [73][74][75]. Consequently, aphids originate from multiple geographical populations in spring, resulting in high genetic differentiation. This issue may also cause the highly different resistance levels detected in crops across fields and years.
Our laboratory results show that there was a lack of consistency in aphid alatae attraction when given the choice between 14 wheat varieties versus 2 wheat varieties. Visual and olfactory discrimination is the first step used by insects to locate host plants. The grey glume, green head, and green stem of wheat plants have been reported to be significantly positively correlated with plant resistance to S. avenae [34,76]. After the alatae land on the host leaf, gustatory discrimination is the next step, whereby individuals select the host plant based on taste. Our results showed that the alatae completed this process within 8 h. Differences in the level of selection between the 2 experiments may be because vision, olfaction and taste are used together when choosing among the 14 wheat varieties in field and a 2 m × 2 m × 1.5 m rectangular gauze cage, whereas taste is preferentially used when choosing between 2 wheat varieties in a 9-cm-diameter Petri dish.
Our MAN and ACC field results showed that 'Amigo' was the highest susceptible variety to S. avenae in 2004, but that it was the resistant in 2005 and 2006. Only 13.5% of alatae on average chose to feed on 'Amigo' when given the choice of 2 varieties. 'Amigo' contains a resistant gene to S. graminum biotypes B and C [18,19]. 'Amigo' tends to be resistant to most aphid populations in different regions of China, and across different years [37,70,77]. However, this variety is susceptible in some geographic regions in some years in China [35,70] and Germany [29,77]. These results support the high genetic differentiation of S. avenae in China [69][70][71][72], and suggest Amigo has different resistance level to different biotypes or geographical population of S. avenae. Therefore, 'Amigo' could be used to distinguish the biotypes or geographical population of S. avenae in China [70].
We found that the average IAS of S. avenae on all wheat varieties/lines ranged from 0.4 to 0.8. Of note, fewer alatae chose the 4 German wheat varieties ('Ww2730', 'Astron', 'Xanthus', and 'Batis'). Liu et al. reported that the IAS of S. avenae is about 1.0 on 'Batis' and 'Ww2730' at the adult plant stage in other fields in Yangling, Shaanxi Province [34]. In Germany, 'Astron,' 'Xanthus,' and 'Ww2730' were resistant to S. avenae populations, whereas 'Batis' was susceptible [28,78]. 'Astron,' 'Batis,' and 'Xanthus have a thick wax powder layer on the ear and leaf, while 'Ww2730' has thick tomentulose on the ear and leaf. In China, these 4 German varieties mature about 3 weeks later than Chinese wheat varieties. The electrical penetration graph (EPG) technique shows that S. avenae salivated more in the sieve elements of 'Batis' than 'Ww2730,' and spent more time passive sucking in the sieve elements of 'Batis' than 'Ww2730,' and 'Ww2730' also has other factors in the epidermis, mesophyll, and phloem that restrict feeding by aphids [79]. Thus, the resistant mechanisms of these 4 German varieties include antibiosis and non-preference to S. avenae.
Our results identified 'Tm' (Triticum monococcum) as a resistant germplasm in all repeats over the 3 years. The resistance mechanism of this variety is antibiosis and non-preference. Seventy percent of S. avenae alatae preferentially fed on 'Xinong1376' after 24 h when given the choice between 2 wheat varieties. 'Xinong1376' contained large numbers of aphids, and had high yield loss rates in most repeats. This result was similar to that obtained by Li et al. and Liu et al. [35,37]. 'Xinong1376' is an early maturing wheat cultivar that is used as a susceptible control variety to S. avenae in China [25,26,37,77,80,81]. The development time of S. avenae on'Xinong1376' was shorter compared to other varieties, with greater fecundity and intrinsic rate of increase than other varieties [77].
We found that 67.8% of S. avenae alatae preferentially fed on 'Xiyan22' after 24 h when given the choice between 2 wheat varieties. However, in the field the average number of aphids found on this variety greatly varied across the 3 years. Since the end of the 20 th century, 'Xiyan22' has been a leading wheat cultivar in central Shaanxi Province. S. avenae must pierce more cell layers in the mesophyll of 'Xiaoyan22' compared to 'Ww2730' and 'Batis' [79]. Consequently, the aphid must salivate more before it can passively ingest the sieve elements of 'Xiaoyan22' compared to 'Xinong 979' and 'Aikang58' [82]. This difference leads to S. avenae that develop on 'Xiaoyan22' having a longer development time [77,79], lighter body weight, and less offspring than those that develop on 'Xinong 979' and 'Aikang58' [82]. However, after 9 days feeding by S. avenae, 'Xinong979' and 'Aikang58' had lower fresh and dry matter loss rates than 'Xiaoyan22' [82]. This result indicates that 'Xiaoyan22' has excellent tolerance to S. avenae.
We detected high variation in the IAS and yield loss rate variations of '186Tm,' 'Tp,' 'Tam200,' and 'PI high' in different fields and years, similar to that detected for '90-10-35.' The resistance of some wheat varieties to S. avenae may be controlled by incomplete dominant genes [24] or dominant major genes [36,80,[83][84][85]. However, for most intermediate or low resistance wheat varieties, resistance to S. avenae is regulated by a quantitative trait locus (QTL), which, in turn, is controlled the by micro effect of multiple genes, with resistance levels being influenced by environmental factors and aphid adaptability.
Our results showed that the yield loss rate of some wheat varieties/lines broadly varied across years. However, such variation was less for 'Amigo,' 'Xiaoyan22,' and some sample points of '186Tm' compared to other varieties subject to similar or higher aphid numbers ( Fig  6). In general, wheat yield loss is not noticeable at 4 aphids/tiller, but rapidly increases from 4 to 10 aphids/tiller in flowering and grain-filling period. However, in our study, 15 to 40 aphids/ tiller caused less than 10% damage when compared to controls (without aphids) in some year. Thus, yield loss rate is relative to aphid number, damage duration time, and damage period [86]. Our results show that the extent of damage varies across years and with respect to the inherited characteristics of different wheat varieties/lines.
Screening for resistant germplasm in the laboratory may produce different results to field screening. Field screening is used to determine the resistance level of wheat germplasm to aphids based on natural aphid numbers that occur on wheat varieties [2]. This technique closely reflects natural conditions, and is suitable for the preliminary screening of large quantities of wheat germplasm materials. Resistance index score units are 0.25 in some references [36,76] and 0.3 in others [25,34,35,39]. Some studies have investigated the IAS in relation to the flowering and filling stage [30,39,42,87], while others have investigated population dynamics throughout the entire wheat growth period [3,38,88,89]. The MAN and ACC of aphid populations may differ due to different climatic conditions in different years, coupled with changes in the aggregation distribution of aphids in fields. Consequently, to guarantee that the sample error is not too large, the sample size should be large enough for the wheat field being tested [58,59]. For instance, in our study, the repeat block area used for planting each wheat variety was 20 m 2 , with the sample area for the 5-point sampling being 0.25 m 2 , in which 80-130 wheat plants were planted.
In conclusion, our study demonstrates that the detected resistance of wheat varieties to S. avenae aphids may differ between laboratory and field experiments, due to the different settings. In addition, under field conditions, the resistance of different varieties differs across fields and years due to the way in which aphids aggregate, as well as differences in the genetic mixing of aphid stocks across years and climatic conditions. Therefore, we strongly recommend a combination of laboratory and long-term field experiments in the target planting regions to identify varieties/lines that consistently show high resistance to S. avenae infestation. Our study suggests that 'Amigo', 'Astron,' 'Xanthus,' and 'Batis' use antibiosis as the resistant mechanism. '98-10-30,' 'Astron,' 'Xanthus,' 'Ww2730,' and 'Tm' also use antibiosis. 'Amigo,' 'Xiaoyan22,' and some '186Tm' samples had good tolerance. Aphid population size and wheat yield loss rates greatly varied in different fields and years for '98-10-35,' 'Tp,' 'Tam200,' 'PI high,' and other '186Tm' samples, all these germplasm should be considered for use in future studies.
Supporting Information S1