Additive effects of two quantitative trait loci that confer Rhopalosiphum maidis (corn leaf aphid) resistance in maize inbred line Mo17

Summary At least two genetic loci contribute to the high aphid resistance observed in the seedlings of maize inbred line Mo17. One of these loci increases the biosynthesis of defence-related benzoxazinoids.

Rhopalosiphum maidis (corn leaf aphid) is a pest on several economically important monocot crops, including maize, sorghum, wheat, barley, and Miscanthus sinensis (Carena and Glogoza, 2004;Huggett et al., 1999). In the case of maize, all aboveground plant parts are susceptible to R. maidis. Infestation at the seedling stage slows development, reduces plant height, and decreases yield (Bing and Guthrie, 1991). Other damage occurs through tassel infestations, where the accumulation of sticky honeydew can prevent pollen shed, and yield losses of up to 90% have been reported (Carena and Glogoza, 2004;Foott and Timmins, 1973). Additionally, R. maidis can transmit damaging plant viruses, including Maize dwarf mosaic virus and Barley yellow dwarf virus (Saksena et al., 1964;Thongmeearkom et al., 1976).
Parental lines of the maize nested association mapping (NAM) population (McMullen et al., 2009b;Yu et al., 2008; www.panzea.org) vary greatly in their resistance to R. maidis (Meihls et al., 2013). Genetic mapping of this trait using RILs derived from crosses between B73 and inbred lines that are more aphid-sensitive (CML 52, CML69, CML277, and CML322) identified a natural transposon insertion in Bx10c (GRMZM2G023325). This gene encodes a methyltransferase that converts DIMBOA-Glc to HDMBOA-Glc (Meihls et al., 2013). Aphids produce more progeny on plants with a functional methyltransferase and constitutively high levels of HDMBOA-Glc. Maize lines with a transposon insertion in the methyltransferase gene have higher levels of DIMBOA, which, unlike HDMBOA-Glc, can induce callose accumulation and perhaps other maize defence responses (Ahmad et al., 2011). Thus, even though HDMBOA-Glc is more toxic to aphids in vitro than DIMBOA-Glc, plants with higher levels of DIMBOA-Glc and DIMBOA are more resistant to R. maidis (Meihls et al., 2013).
Whereas the Bx10c gene was identified by studying maize inbred lines that are more aphid-sensitive than the reference line, B73, Mo17 is one of the most aphid-resistant inbred lines among the 26 that were tested (Meihls et al., 2013). As both B73 and Mo17 contain the same inactivating transposon insertion in the Bx10c gene (Meihls et al., 2013), this locus would not affect aphid resistance in the well-characterized inter-mated B73 by Mo17 (IBM) RIL population (Lee et al., 2002). Thus, it was hypothesized that genetic mapping with these RILs would identify novel maize aphid resistance QTL. Data presented here show that at least two major loci have additive effects on aphid resistance in inbred line Mo17 relative to B73.
Aphid growth assays R. maidis was obtained from Stewart Gray (USDA Plant Soil and Nutrition Lab, Ithaca, NY) and the colony was maintained on seedlings of maize inbred line B73. For genetic mapping,142 RILs derived from crosses between B73 and Mo17 (a subset of the maize IBM population; Lee et al., 2002) were used for genetic mapping. Ten R. maidis aphids were confined on two-week-old plants with microperforated polypropylene bags (15 cm×61 cm; PJP Marketplace, http://www.pjpmarketplace.com), and the progeny were counted 7 d later. Experiments with the 142 RILs were conducted in triplicate and the results were averaged for QTL analysis. For experiments to assess aphid growth on B73, Mo17, NILs, and W22 Bx1::Ds segregating mutant stocks, bioassays were conducted in a similar manner.
Callose staining Ten adult aphids were confined to a clip cage for 72 h on the third leaf of two-week-old maize seedlings. Control plants received cages without aphids. Three days later, leaf material within the aphid cage was excised and used for callose staining as described previously (Luna et al., 2011;Ton et al., 2005). Leaves were de-stained in 98% ethanol for at least 48 h until all tissue was transparent, washed in 0.07 M phosphate buffer (pH 9), incubated (stained) for 2 h in 0.07 M pH 9 phosphate buffer containing 0.01% aniline-blue (Sigma, St. Louis, MO), and stored at 4 °C in 0.07 M pH 9 phosphate buffer until microscopic analysis. Observations were performed with an Leica DM5500 epifluorescence microscope equipped with a ×20 immersion objective, a UV filter (BP 340 to 380 nm, LP 425 nm) a Retiga-2000R colour CCD camera, and Qcapture Pro 6.0 acquisition software (Leica, Wetzlar, Germany). Callose spots were quantified on the adaxial side of the leaf segment contained within the clip cage (~146 mm 2 , with or without aphid feeding) and number of callose spots was calculated per mm 2 of leaf tissue.

Benzoxazinoid assays
Groups of ten aphids were confined to individual leaves of twoweek-old maize plants, as described above for the callose assays. The maize leaf tissue within the cage was harvested, weighed, and extracted in 30% methanol:0.1% formic acid:69.9% deionized water (v:v:v). Samples of the extract were analysed by HPLC-absorbance detection using a C18 reverse-phase Luna column (5 μm pore size, 250 × 4.60 mm; Phenomenex, Torrance, California, USA), a Waters 2695 pump system (Waters, Milford, MA, USA), and a Waters 2996 absorbance detector. The HPLC solvents were A, 0.1% v/v formic acid in deionized water, and B, 0.1% v/v formic acid in methanol, with a flow rate of 1 ml min -1 . The following gradient was used for the analytical analysis: 0-15 min, gradient from 20%-30% B; 15-25 min, gradient to 40% B, 25-30 min, 40% B; and 35-40 min, 20% B. Benzoxazinoid abundance was calculated from a standard curve that was produced with authentic standards that were kindly supplied by Gaetan Glauser (University of Neuchatel, Neuchatel, Switzerland).

Bx1::Ds transposon insertion identification
A Ds transposon insertion (B.W06.0775) in exon 4 of the Bx1 gene (GRMZM2G085381) in the W22 genetic background was identified through the Ds project website (http://acdstagging. org; Vollbrecht et al., 2010). Seed stocks segregating this insertion were planted and primers were designed to identify plants carrying homo-or heterozygous Bx1::Ds alleles, using the B73 RefGen_v2 genome browser at MaizeGDB (www.maizegdb.org) in conjunction with Primer3 software (http://www-genome.wi.mit. edu/genome_software/other/primer3.html). DNA was extracted from seedling leaf tissue using a CTAB (cetyltrimethyl ammonium bromide)-based extraction protocol (adapted from Fulton et al., 1995). A Ds 5'-end primer (GTTCGAAATCGATCGGGATA) was used in combination with a primer designed to the chromosomal sequence flanking the B.W06.0775 Ds insertion (TCTTA ACCTCCTGGATGAGTG). The insertion was confirmed using a 20 μl GoTaq PCR reaction (Promega, Fitchburg, Wisconsin) with 4% dimethylsulfoxide, using the following conditions: initial denaturation 94 °C for 3 min; 35 cycles of 94 °C for 30 s, 57 °C for 30 s, 72 °C for 1 min; and final extension 72 °C for 10 min. Seedlings were genotyped in a second PCR assay to distinguish Ds insertion homozygotes from heterozygotes and wild-types. For this assay, a second Ds flanking primer complementary to chromosomal DNA on the other side of the Ds insertion site (GCCAAGAACAACAACCTGGAGC) was designed using the methods described above. The two Ds-flanking primers were used to amplify the wild-type allele in Ds heterozygotes and wild-type plants. The segregating population was subjected to benzoxazinoid analysis by HPLC, as described above. Seed stocks segregating the unstable B.W06.0775 homozygote (in an Acimmobilized W22 genetic background; Conrad and Brutnell, 2005) are available from the Maize Genetics Cooperation Stock Center (http://maizecoop.cropsci.uiuc.edu/) as AcDs-00565.

Results
Aphid nymph production was measured on 142 RILs of the maize IBM mapping population. Analysis of these data showed significant QTL on chromosomes 4 and 6 ( Fig. 1A), accounting for 15% and 27% of the total variance in aphid resistance, respectively. QTL for the survival of adult aphids that were placed on the RILs were found at the identical positions on chromosomes 4 and 6 (data not shown). For both loci, the allele conferring higher aphid resistance came from the Mo17 parent in the cross. The effects of the two QTL are additive and aphid reproduction on RILs that have the Mo17 allele on both chromosomes 4 and 6 is significantly lower than on RILs that have only one or the other Mo17 allele (Fig. 1B).
B73×Mo17 NILs (Eichten et al., 2011) were used to confirm the effects of the aphid resistance QTL on chromosomes 4 and 6. In experiments with NILs that have Mo17 segments of the chromosome 4 QTL region introgressed into B73, lines b037, b040, and b070 exhibited reduced aphid reproduction relative to B73, whereas lines b117 and b152 did not (Fig. 1C). This defined an interval between markers AC213878 and AC204415 (1.35 Mbp on chromosome 4; maize RefGen v2, www.maizegdb.org) as the position of the Mo17 aphid resistance locus. In contrast, aphid reproduction on NILs with introgressions of the B73 allele into the Mo17 genetic background (lines m011, m031, and m072) was similar to reproduction on Mo17 (Fig. 1C), suggesting that there are additional aphid resistance loci acting in the overall Mo17 genetic background.
In the case of the chromosome 6 QTL, only introgressions of the Mo17 allele into B73 were available. These lines showed an intermediate aphid reproduction phenotype (Fig. 1D) and, on average, aphid reproduction on lines with a segment of Mo17 chromosome 6 introgressed into B73 was approximately 0.5 nymphs per week lower of that on inbred line B73 (0.67 ± 0.09 vs. 1.17 ± 0.13 progeny per week; P<0.05, t-test). The effect of this introgression is similar to the magnitude of the chromosome 6 effect on aphid resistance that was observed in the QTL mapping (Fig. 1B).
The chromosome 4 aphid resistance QTL coincided with a region of the maize genome that contains at least eight genes of the benzoxazinoid biosynthesis pathway (Gierl and Frey, 2001). To determine whether benzoxazinoid content can influence R. maidis reproduction, DIMBOA abundance was mapped as a quantitative trait using 128 IBM RILs, a subset of the 142 that were used for aphid resistance mapping. This identified significant QTL on chromosomes 4 and 5, accounting for 11% and 8% of the variance in this trait, respectively ( Fig. 2A). In both cases, the high-DIMBOA allele came from the Mo17 parent in the cross. QTL for DIMBOA-Glc abundance, which is highly correlated with DIMBOA abundance, are located at the same two chromosomal positions (data not shown). Aphid reproduction on the IBM population RILs was negatively correlated with DIMBOA content (r=-0.361, P<0.01, Pearson correlation; Fig. 2B). Previous research has shown that R. maidis feeding has no significant effect on B73 benzoxazinoid content (Meihls et al., 2013). Similarly, in the current study, no significant changes were observed in the abundance of DIMBOA, DIMBOA-Glc, or HDMBOA-Glc in Mo17 response to aphid feeding (P>0.05, t-test).
Both an aphid resistance QTL (Fig. 1A) and a high-DIM-BOA QTL ( Fig. 2A) were mapped to a region of chromosome 4 containing DIMBOA biosynthetic genes. Association mapping with maize inbred lines linked DIMBOA abundance to genetic variation in Bx1, the first gene in the pathway (Butrón et al., 2010), suggesting that Bx1 by itself could affect aphid resistance. To test this hypothesis, a Ds transposon knockout mutation of Bx1 was identified in the W22 maize genetic background. Homozygous Bx1::Ds progeny, confirmed by PCR-based genotyping, had no detectable DIMBOA (Fig. 3A) and aphid reproduction was significantly increased relative to wild-type W22 (Fig. 3B). In contrast, segregating wild-type siblings of the Bx1::Ds plants were not significantly different from wild-type W22 in either their DIMBOA content or their aphid resistance.
Previous research demonstrated that DIMBOA is required for callose induction in maize (Ahmad et al., 2011) and that R. maidis resistance is correlated with both DIMBOA content and callose formation (Meihls et al., 2013).To determine whether this trait influences aphid resistance in the IBM population, callose production was measured in the ten most aphid-resistant and the ten most aphid-sensitive RILs that were identified in the analysis shown in Fig. 1A. On average, the aphid-resistant RILs had three times as many callose spots in response to aphid feeding as the aphid-sensitive RILs (8.1 ± 1.2 vs. 2.6 ± 0.5 spots per mm 2 , P<0.01, t-test). Callose content in response to aphid feeding was positively correlated with DIMBOA content (r=0.695, P<0.01, Pearson correlation; Fig. 4B), consistent with prior reports that DIMBOA is required for aphid-induced callose formation (Ahmad et al., 2011;Meihls et al., 2013). Aphid reproduction was negatively correlated with callose accumulation (r=-0.649, P<0.01, Pearson correlation; Fig. 4C), an indication that elevated callose accumulation is part of a maize defence response against aphids.
Aphid-induced callose formation was measured in B73, Mo17, and selected NILs to confirm that aphid resistance associated with the chromosome 4 QTL (Fig. 1A) is associated with callose formation. Consistent with the lower aphid progeny production on Mo17 (Fig. 1), there was elevated callose production in this line relative to B73 (P<0.05, t-test; Fig. 4C). NIL b040, which has an introgression of the Mo17 allele of the chromosome 4 QTL into B73, has an intermediate level of callose production.
On average, NILs m011, m031, and m072, which have introgressions of the B73 QTL on chromosome 4 into M017, exhibit significantly lower levels of aphid-induced callose production than Mo17 (343 ± 49 vs. 1532 ± 292 spots per mm 2 , P<0.05, t-test; Fig. 4C). Benzoxazinoid assays were conducted to determine whether this reduced callose production is associated with lower DIMBOA accumulation owing to the introgression of the B73 Bx chromosomal region into Mo17 in these NILs. Two of the NILs (m031 and m072) had a DIMBOA content that was significantly lower than that of Mo17 and similar to that of B73. Despite the low DIMBOA content and callose formation, R. maidis reproduce quite poorly on m011, m031 and m072 (Fig. 1C). This observation is consistent with the hypothesis that higher aphid resistance in these three NILs relative to B73 is independent of DIMBOA-induced callose production associated with the chromosome 4 QTL (which is derived from B73 in the NILs), and is instead controlled by other loci in the overall Mo17 background of the NILs.

Discussion
Results presented here, together with previous findings (Meihls et al., 2013), show that both synthesis and catabolism of DIMBOA-glucoside contribute to aphid resistance in maize (Fig. 5). DIMBOA content of maize seedlings, which is highly correlated with the abundance of the precursor DIMBOA-glucoside, is an essential contributing factor for callose formation and R. maidis resistance (Ahmad et al., 2011;Meihls et al., 2013). Maize lines with a functional Bx10c gene (e.g. CML52, CML69, CML277, and CML322) have constitutive DIMBOA-Glc catabolism to HDMBOA-Glc and are aphid-sensitive. Both B73 and Mo17 contain a Doppia-like transposon insertion that inactivates Bx10c and thereby confers higher aphid resistance. In the case of Mo17, elevated biosynthesis further increases DIMBOA abundance and thereby causes this line to be even more aphid-resistant than B73 (Fig. 5). Fine-scale mapping of a DIMBOA QTL in the maize IBM population to a specific cis-regulatory region more than 100 kbp upstream the Bx1 gene, together with the longer persistence of Bx1 gene expression in developing Mo17 seedlings (Zheng et al., unpublished results), suggests this as a likely cause of elevated DIMBOA content, callose accumulation, and aphid resistance in Mo17 relative to B73.
Chromosome 5 also contains a QTL with a significant effect on DIMBOA abundance ( Fig. 2A). However, unlike the chromosome 4 DIMBOA QTL, this was not associated with a QTL affecting aphid resistance (Fig. 1A). As the chromosome 5 QTL does not include any known benzoxazinoid biosynthesis genes, it may be a regulatory locus affecting the benzoxazinoid pathway and perhaps other maize defence responses. Thus, the absence of a significant aphid resistance QTL in this region could be explained by opposing effects of as yet unknown maize defence responses that are regulated by either by the chromosome 5 benzoxazinoid QTL or perhaps by nearby loci on chromosome 5 that were not resolved using the current genetic mapping approach. Whereas introgression of the Mo17 chromosome 4 QTL into B73 decreased aphid reproduction (NILs b037, b040, and b070), the converse introgression (NILs m011, m031, and m072) did not have the opposite effect (Fig. 1C). Aphid resistance associated with the Mo17 genome seems to override any benefits that the B73 chromosome 4 QTL allele provides for aphid reproduction. Given the relatively low aphid-induced callose formation (Fig. 4C) and low DIMBOA content (Fig. 4D) in NILs m031, and m072, the observed aphid resistance that is provided by the Mo17 contribution to these NILs is likely to be independent of benzoxazinoid biosynthesis.
Consistent with the hypothesis of DIMBOA-independent aphid resistance mechanisms in Mo17, the genetic mapping interval of the chromosome 6 aphid resistance QTL (Fig. 1A) contains neither known benzoxazinoid biosynthesis genes, nor genetic evidence for a DIMBOA-related QTL ( Fig. 2A).
Thus, it is likely that the observed aphid resistance associated with Mo17 chromosome 6 involves something other than DIMBOA accumulation. The significant, but relatively weak   5. Model of aphid defence regulation by benzoxazinoid metabolism in maize seedlings. Both synthesis and catabolism regulate the abundance of DIMBOA-glucoside. Genetic variation in the biosynthetic pathway affects DIMBOA-glucoside abundance (Butrón et al., 2010), and both Mo17 and B73 have a transposon insertion in Bx10c, which encodes a constitutively expressed DIMBOA-Glc catabolic enzyme (Meihls et al., 2013). DIMBOA-Glc is a precursor for DIMBOA, which is required to trigger callose formation and perhaps other aphid defence responses in maize. correlation between DIMBOA concentration and aphid reproduction observed in experiments with B73×Mo17 RILs (Fig. 2B) also is consistent with the hypothesis of additional resistance mechanisms acting in Mo17. At this point, however, we cannot rule out the possibility that as yet unknown benzoxazinoid modifications that affect aphid resistance are influenced by the chromosome 6 QTL.
As benzoxazinoids are deleterious to a large variety of insect herbivores (Niemeyer, 2009), natural variation in Bx1 and/or other benzoxazinoid biosynthesis genes in maize bin 4.01 probably plays a key role in maize defence, not only against R. maidis but also other insect herbivores. Similar to R. maidis, Rhopalosiphum padi (birdcherry-oat aphid) produced more progeny on a bx1 mutant than on near-isogenic wild-type Bx1 maize (Ahmad et al., 2011). A QTL for maize DIMBOA accumulation was previously mapped to bin 4.01 using several RIL sets (Butrón et al., 2010), and association mapping indicated that Bx1 polymorphisms were the most likely cause of the phenotypic variation. QTL for the amount of damage caused by two lepidopteran herbivores, O. nubilalis and O. furnacalis (Cardinal et al., 2006;Jampatong et al., 2002;Krakowsky et al., 2004;Xia et al., 2010), also are located in bin 4.01 and may be related to natural variation in benzoxazinoid content. Specialist herbivores, such as Diabrotica virgifera and Spodoptera frugiperda, can be insensitive to variation in benzoxazinoid content  or have specific DIMBOA detoxification enzymes (Glauser et al., 2011). Thus, these insects may be insensitive to benzoxazinoid abundance or might even be attracted to plants with higher benzoxazinoid content.
The chromosomal locations of the two R. maidis resistance QTL described here (Fig. 1A) are different from those that have been mapped in previous studies (Meihls et al., 2013;So et al., 2010). The chromosome 6 QTL is particularly interesting, as further genetic mapping may lead to the discovery of novel aphid resistance mechanisms in maize. By identifying the underlying genetic basis of this R. maidis resistance QTL and others that may be found in future genetic mapping studies, it will be possible to develop new tools for improving the insect resistance of commercial maize lines through breeding or transgenic approaches.