Control of Western Corn Rootworm (Diabrotica virgifera virgifera) Reproduction through Plant-Mediated RNA Interference

RNA interference (RNAi) in transgenic maize has recently emerged as an alternative mode of action for western corn rootworm (Diabrotica virgifera virgifera) control which can be combined with protein-based rootworm control options for improved root protection and resistance management. Currently, transgenic RNAi-based control has focused on suppression of genes that when silenced lead to larval mortality. We investigated control of western corn rootworm reproduction through RNAi by targeting two reproductive genes, dvvgr and dvbol, with the goal of reducing insect fecundity as a new tool for pest management. The results demonstrated that exposure of adult beetles, as well as larvae to dvvgr or dvbol dsRNA in artificial diet, caused reduction of fecundity. Furthermore, western corn rootworm beetles that emerged from larval feeding on transgenic maize roots expressing dvbol dsRNA also showed significant fecundity reduction. This is the first report of reduction of insect reproductive fitness through plant-mediated RNAi, demonstrating the feasibility of reproductive RNAi as a management tool for western corn rootworm.

phenotype 20 . The bol gene was first described in the molecular genetic analysis of spermatogenesis mutants generated by a P-transposable element in Drosophila 21 . BOL is an RNA-binding protein with an RNA Recognition Motif (RRM) domain 22 . It shares homology with the DAZ (Deleted in Azoospermia) protein outside the RRM domain, and mutations in this protein cause severe sperm reduction in animals 23 . Based on this previous characterization of vgr and bol as essential reproductive genes in other insect species, we chose to evaluate WCR homologs of these genes, dvvgr and dvbol, as targets for reproductive RNAi in WCR. The impact to fecundity by suppressing gene expression was tested by feeding WCR larvae and adults artificial diet incorporating dsRNA derived from vgr or bol (dsvgr and dsbol). In addition, transgenic maize plants expressing dsvgr and dsbol were generated to determine whether fecundity would be affected by larvae feeding on transgenic roots to demonstrate the potential for WCR reproductive RNAi as a transgenic trait. To the best of our knowledge, this is the first report of plant-mediated reproductive RNAi for insect control, and it represents a new rootworm management approach.

Results
Identification and expression of WCR VgR and BOL. We selected two WCR reproductive genes, vgr and bol, based on the following: [1] homology to reproductive genes listed in FlyBase 24 or to RNAi lines showing sterile phenotype 25 ; [2] previously reported to have reproductive functions; and [3] likely expressed in germ cells. A putative full-length cDNA sequence encoding WCR VgR was identified from the transcriptome assembled from WCR adult females by blastP search and compared to other insect VgRs, including Drosophila (DmVgR) 15,26 . VgR proteins belong to the low-density lipoprotein receptor (LDLR) family, which are membrane-bound proteins 15 . The dvvgr (or vgr) cDNA has an open reading frame of 5313 nucleotides (nt) which encodes a large protein of 1770 amino acids (aa) that is predicted to contain multiple domains typically found in other insect VgRs (Fig. 1a). These domains include a signal peptide, five low-density lipoprotein receptor class A (LDLa) repeats, two epidermal growth factor (EGF)-like repeats, five low-density lipoprotein receptor class B (LDLb) repeats, one EGF, two LDLb repeats, and one EGF region, followed by eight LDLa, two EGF, three low homology LDLb repeats, one EGF-like region and a transmembrane region at the C-terminus. Although the identity between DvVgR and DmVgR is only 28%, the overall protein architectures are very similar (Fig. 1a). Overall identity of DvVgR to other insect VgRs ranges from 47% in the Coleoptera order to 26% in the Lepidoptera order ( Supplementary Fig. 1a).
DvBOL was identified from the transcriptome assembled from adult WCR testes by a blastP search using Drosophila DmBOL 22 . The dvbol (or bol) cDNA encodes a 694 aa protein with a conserved RRM domain found in DmBOL (Fig. 1b). While the identity of RRM domain between DvBOL and DmBOL is 72% (over 90 aa), the putative DAZ domain of DvBOL is 32% identical to that of DmBOL (over 31 aa). DvBOL is considerably larger than DmBOL protein (694 aa vs 189 aa), but DvBOL is similar in size to Tribolium castaneum TcBOL, which is another family in Coleoptera (694 aa vs 634 aa). There was no other region outside of the RRM domain that was conserved across the analyzed insect BOL proteins in Genbank, but one segment of amino acid sequence (PPPAPYSPMT) present in DvBOL is conserved in BOL proteins from multiple hymenopterans, including Athalia rosae ArBOL-2 isoform, Atta colombica and Diachasma alloeum (Fig. 1b). Overall, the identity of DvBOL to other insect BOLs ranges from 47% in the order Coleoptera to 9% in the order Lepidoptera ( Supplementary Fig. 1b).
VgR and BOL proteins were extracted from adult WCR reproductive tissues and analyzed by western blot. The VgR and BOL proteins were detected at the predicted MWs of 197.8 kDa (ovary) and 75.9 kDa (testes), respectively ( Fig. 1c and Supplementary Fig. 2). Expression of bol and vgr mRNA was analyzed by real-time quantitative reverse transcription PCR (qRT-PCR) using individual whole insects representing different life stages (Fig. 1d), and by in situ hybridization (ISH) methods on specific life stages as well as dissected reproductive tissues from the WCR adults ( Fig. 2, Supplementary Fig. 3). The mRNA expression of vgr showed clear differences depending on the life stage. For example, vgr mRNA expression was about 10-fold higher in the egg and between 2-4 fold higher in pupae and adults than in larvae. On the other hand, expression of bol mRNA was similar across life stages (within 2 fold) by qRT-PCR and it was highly expressed in testis and moderately expressed in the ovary compared to vgr (Fig. 2).
Fecundity of WCR adults exposed to dsvgr or dsbol. The genes of vgr and bol from WCR were evaluated for the impact of suppression on adult beetles in diet feeding assays. Since vgr is highly expressed in the ovary, it is possible that the age of the adult females may influence the effect of dsRNA treatments. Thus, experiments testing the impact of RNAi of vgr gene on fecundity were performed using young (<5 days old) and old adults (>11 days old) (50 female and male pairs of young adults and 50 mated female old adults) exposed to dsRNA in artificial diet (containing water control, 100 ng µl −1 of dsgfp, or dsvgr) for 24 h. Two parameters for fecundity (egg production and egg hatch rate) were assessed for each treatment. These parameters were used to calculate the net reduction in fecundity (NRF). Exposure of young and old adults to dsvgr led to a significant reduction in egg number, egg hatch rate, and vgr expression compared to water and gfp controls (Fig. 3a-c). The overall NRF following dietary exposure to dsvgr was 50.8 ± 20.3% and 75.9 ± 4.4% for old and young adults, respectively. Under similar conditions, exposure of old adults to 100 ng μl −1 dsRNA bol (dsbol), showed no significant change in egg production and egg hatch rate, despite detectable gene suppression ( Supplementary Fig. 4). Further dose-response analysis was performed by exposing old adults to increasing concentrations of dsvgr (0.01, 0.1, 1, 10, and 75 ng μl −1 ) and monitoring fecundity endpoints. Old adults exposed to dsvgr concentrations ranging from 0.1 to 10 ng μl −1 resulted in 46.5 ± 12.5% to 75.4 ± 11.3% lower fecundity, indicating that lower doses were still effective (Supplementary Table 1).
Effect on adult WCR fecundity by larval exposure to dsvgr or dsbol. We also investigated if larval exposure to dsvgr or dsbol would cause a similar effect on fecundity as observed by adult exposure. We exposed 11-day-old 3 rd instar larvae to WCR larval diet mixed with 50 ng µl −1 of dsbol, dsvgr, dsgus (control) or sterile water. Relative mRNA expression of vgr and bol from larvae were measured by qRT-PCR and ISH. Exposed WCR larvae were allowed to complete life-cycle development and emerged adult beetles were collected, sexed and used for subsequent fecundity studies (egg production and egg hatch rate). Oral ingestion of dsbol and dsvgr at the 3 rd instar larval stage resulted in a significant reduction in adult beetle egg production and hatch rate compared to dsgus and water controls (Fig. 4). The estimated NRF relative to water control was 80.9% (±2.7) and 42.3% (±10.2) for dsbol and dsvgr, respectively. Subsequent dose-response experiment performed on 3 rd instar using a range of dsbol concentrations (0.1, 1, 10, and 50 ng µl −1 ) showed that 10 ng µl −1 was a high enough concentration to cause a significant reduction in emerged adult egg production, egg hatchability (Fig. 5a,b) and NRF (87.8 ± 4.8%). The 10 ng µl −1 was also high enough to reduce mRNA expression in both 3 rd instars (Fig. 5c) and pre-oviposition adults (Fig. 5d). In contrast, dsvgr reduced fecundity by only 12.5 ± 14.2% at 10 ng  Table 2). Target-specific knockdown of bol and vgr was confirmed in 3 rd instar larva collected 2 days after 50 ng µl −1 dsRNA exposure ( Supplementary Fig. 5a,b) and this suppression was maintained in pre-oviposition adults for at least 25 days after beetle emergence ( Supplementary Fig. 5c Fig. 9a) could result in a reduction of WCR fecundity, four-week-old maize T1 transgenic seedlings were infested with WCR eggs in the greenhouse. Beetles emerged from the infested transgenic plants were collected and maintained in the laboratory to assess fecundity. Exposure of neonates to transgenic plants expressing dsvgr or dsbol transcripts had no significant effect . Fecundity and mRNA expression of adult Diabrotica virgifera virgifera exposed to dsRNA vgr (dsvgr). For young adults (<5 days old; 50 pairs) and for old adults (>11 days old; 50 mated females) were exposed individually to sterile deionized water (control), 100 ng μl −1 gfp dsRNA (dsgfp), or 100 ng μl −1 vgr dsRNA (dsvgr) in artificial diet for 24 h. Exposed beetles were used to assess fecundity. Least square means pairwise comparison P-values: >0.05 (ns; not significant), <0.05*, <0.01**, <0.0001***. (a) Number of eggs laid per female per day (mean ± SE; n = 13 to 14). Treatment main effect P = 0.002; Age main effect P = 0.0028; Treatment *age interaction P > 0.05. (b) Percent egg hatchability by treatment and age group (mean ± SE; n = 13 −14; 12-226 eggs/day). Treatment main effect P < 0.001; Age main effect P > 0.05; Treatment *age interaction P > 0.05. (c) Gene suppression analysis of WCR old adults 4 days after dsRNA treatment and young adults 8 days after dsRNA treatment (mean ± SE; n = 4). Relative expression of vgr mRNA by qRT-PCR assay is shown for each treatment using rps10 as a reference and normalized to vgr expression in water control. on the adult emergence compared with non-transgenic plants (Supplementary Fig. 10a and b). However, WCR larvae feeding on the three transgenic plant lines expressing the dsbol transcript resulted in a significant reduction in egg production and egg hatch rate of the emerged adult beetles ( Fig. 6a and b), which resulted in an overall NRF ranged from 84.1 ± 5.8 to 95.3 ± 2.2%. This effect was mainly due to a significant reduction in egg hatch rate (80.8 ± 5.6-92.7 ± 2.7%) when compared to non-transgenic control (Fig. 6b). WCR larvae feeding on eight transgenic plant lines expressing the dsvgr transcripts (three different fragments) showed less impact on fecundity as there were only two lines where emerged adults showed a significant reduction in egg production ( Fig. 6c) but none of them influenced egg hatch rate (Fig. 6d). Molecular analyses confirmed that both long dsRNA transcripts and siRNA were expressed in the transgenic lines (Supplementary Fig. 9b and Supplementary Table 3).
Male-specific effect of dsbol exposure shown by reciprocal crossing. The bol gene was discovered to impact male spermatogenesis in Drosophila 21 , as well as other insects 20,27 . To further characterize dvbol, a reciprocal crossing experiment was conducted to assess the effect on fecundity if only one of the adults (either the male or female) in the mating pair was exposed to dsbol. Eleven-day-old 3 rd instar larvae (n = 3780) were exposed to 50 ng µl −1 dsbol or water (control) for one day and allowed to complete their development and to emerge as adults. Four reciprocal crossing combinations were assessed: exposed dsbol males mated with exposed dsbol females (bol ♂ x bol ♀); exposed dsbol males mated with unexposed females (bol ♂ x H 2 O ♀); unexposed male mated with exposed dsbol female (H 2 O ♂ x bol ♀), and unexposed males mated with unexposed females ( The mating of exposed dsbol males with unexposed females (bol ♂ x H 2 O ♀) significantly reduced both egg number and egg hatch (Fig. 7), while the mating of unexposed males to exposed dsbol females (H 2 O ♂ x bol ♀) had no significant effect on fecundity (Fig. 7). The results of the reciprocal crossing experiments and the observed suppression of bol mRNA in testes ( Supplementary Fig. 7a) together support dvbol's male-specific role in WCR reproduction.

Discussion
Protection against rootworm in transgenic plants has been demonstrated using insecticidal RNAi targets 8,10-12 . We postulated that targeting reproductive genes using RNAi (reproductive RNAi) would be another approach using transgenic maize plants for WCR control and management. RNAi of reproductive genes has shown great promise for pest management with mosquitoes (Aedes aegypti) 27 and with oriental fruit fly (Bactrocera dorsalis) 28 . These studies generated sterile males by feeding or injecting dsRNA against essential testis genes into the insect which resulted in reduced fecundity 20,27,28 . We in this study have shown that reproductive RNAi can be an effective approach for controlling WCR reproduction. Reduction of fecundity was observed when RNAi exposure occurred to WCR adult beetles as well as 3 rd instar larvae. More significantly, this effect on fecundity reduction was extended to WCR exposed to transgenic maize plants. We selected bol and vgr genes because they have been extensively characterized in their role in reproduction across multiple insect species. We have further characterized their tissue expression across developmental life stages and observed an expression profile (timing and tissue) consistent with their expected reproductive function. Furthermore, silencing of dvvgr and dvbol in WCR did not cause growth inhibition or mortality (adult beetle emergence) in diet-or transgenic plant-based assays. This allowed us to assess the effect of gene suppression directly on WCR fecundity. The observed fecundity reduction of emerged beetles that had fed on the roots of transgenic plants demonstrates the feasibility of WCR control using reproductive RNAi.
Expression of dvbol and dvvgr showed different patterns in larvae and suppression of these genes had differential effects on fecundity between larval and adult stages of WCR. The fecundity reduction of dsvgr was high with exposure to adult females but low with 3 rd instar larvae. In contrast, dsbol's impact on fecundity was only observed by larval exposure (artificial diet and transgenic root). It has been reported that rootworm RNAi is more effective at larval stage than adult stage 29 . There are several possible explanations for the differences in fecundity that were measured from silencing of the two targets. One explanation may be differences of the sex-specific and age-dependent functions of the genes. WCR vgr is a female-specific gene that is highly expressed in ovaries which are fully developed in adult female beetles (post-emergence from pupae). In this study, vgr mRNA expression was found to be low at multiple larval stages but high in eggs and adults as measured by both qRT-PCR and ISH. As a result, dsvgr exposure to adult females would coincide with VgR's function in transport of nutrients (including lipids and carbohydrates) to growing embryos within oocytes 15 , disruption of which would have a significant negative impact on fecundity. No effect on fecundity due to dsvgr larval exposure would be expected during developmental stages when vgr expression is low and VgR is not contributing to egg development. These results also suggest that dsvgr either does not persist from exposure at 3 rd instar larvae to adults or the dose accumulated during this time is not sufficient to significantly suppress the high levels of vgr expression in adults. In comparison, bol expression was observed to be relatively constant throughout all WCR life stages and was also detected in germline cells 20,30 (Supplementary Fig. 3c). In the planarian Schmidtea mediterranean, two bol paralogs have been identified; Smed-bol1 is required for meiotic progression, and Smed-bol2 is required for the maintenance of the earliest male germ cells 31 . In higher animals (invertebrates and primates), ancestral bol was duplicated during evolution and 2-3 BOL family members (BOL, DAZ, and DAZL) have been shown to have more specified roles during germ cell development 23,32 . However, only one bol gene has been identified in the genome of WCR and Tribolium. It is possible that different transcript isoforms of a single invertebrate bol may have different functions throughout male germ cell development. Suppression of bol in WCR larvae is likely to interfere with one or more germline cell functions, including maintenance, renewing and pre-meiotic division 21 , leading to fecundity reduction in WCR adults. The reciprocal cross experiment of bol dsRNA confirms that WCR bol has a male-specific effect (Fig. 7), even though bol mRNA is expressed in the ovary (Fig. 2).
A second explanation for the observed differences in WCR response may be the difference of dsvgr and dsbol accumulation in transgenic maize plants. An impact on adult fecundity was observed because of exposure of 3 rd instar larvae to bol and vgr dsRNA using artificial diet bioassays. The dose administered in the diet bioassays was significantly higher than the expression levels found in transgenic plants, in part due to endogenous plant RNAi dicer activities 33 . The dsRNA accumulation in transgenic roots was in the range of 0.6 to 11 pg mg −1 fresh weight (Supplementary Table 3). This level was near the dsRNA LC 50 threshold needed to elicit RNAi responses in Coleoptera 10 . However, it is unclear but conceivable that continuous exposure to a low dose of dsRNA over the full larval development period, as it is the case for larvae feeding on transgenic roots, can compensate for a single high dose exposure at 3 rd instar as in the diet assay. In the case of dvbol, a single diet exposure to 3 rd instar larvae of 10 to 50 ng µl −1 dsbol produced comparable effects to WCR larvae exposed to roots of transgenic plants expressing dvbol dsRNA, from egg hatch to pupation. In contrast, continuous larval exposure to dsvgr from transgenic roots was either not sufficient to maintain suppression of dvvgr into adults or because dvvgr is highly expressed in ovaries, the amount of dsvgr expressed in roots was inadequate to reduce dvvgr expression in adults below the threshold needed to exert a significant effect on fecundity. The absence of an RNA-dependent RNA polymerase (RdRp) in insects 34 implies that a RdRp-dependent amplification mechanism to spread silencing is lacking in insects. Consequently, cumulative dose and timing of exposure relative to target gene expression in the host organism is an important factor for successful suppression of reproductive genes both in an artificial diet system and through plant-mediated RNAi. These results suggest that although both dvvgr and dvbol are good targets for reproductive RNAi, dvvgr is more effective at the adult stage and dvbol is more effective during the larval stage. In a transgenic plant approach, this would suggest that dsvgr would be most effective if expressed in aerial tissues (pollen, silk, leaf) where adult WCR preferentially feed, while dsbol would be most effective when expressed in root tissue for larval exposure. Targeting suppression of both dvvgr and dvbol may be an interesting approach to maximize fecundity reduction since it would impact both males and females.
The potential of using RNAi-based transgenic plants to suppress WCR reproduction provides a useful alternative approach for rootworm control that is complementary to current insect control strategies. Reproductive interference has been exploited previously as a tool for insect control 35 including tactics for manipulating pheromones, irradiation, pathogens, symbionts and genetic techniques 36 . The production of male sterile insects through sterile insect technique (SIT) by RNAi has also been proposed as a method of control 36 particularly through the silencing of testis-expressed genes. WCR reproductive RNAi mediated through the transgenic delivery of dsRNA is not a stand-alone option for rootworm control, given that root protection is not expected from RNAi targeting these two reproductive genes. However, when combined with other pest management tools (and insecticidal traits) and given sufficient time, it may generate a significant benefit by suppressing pest populations, increasing the effectiveness of insect resistance management and improving root protection. For example, WCR reproductive RNAi pyramided with a WCR insecticidal active would lower the reproductive fitness of any adults that survive from larvae feeding on roots making it less likely to contribute to the development of resistance against the insecticidal active expressed in the transgenic plant. The release of sterile pink bollworm (Pectinophora gossypiella) in combination with Bt cotton significantly reduced the population of pink bollworm to the extent that insecticide sprays were no longer required 37 . Although this is an extreme example, it does show the effectiveness of population suppression through reproductive control as a component for integrated rootworm management. In conclusion, dvvgr and dvbol are suitable gene targets for WCR reproductive RNAi. Double-stranded RNA targeting bol expressed in transgenic maize plants can down-regulate gene expression in WCR and effectively reduce the fecundity of WCR adult beetles. Transgenic plant mediated-reproductive RNAi offers a new tool for pest management which can be used with insecticidal traits to potentially enhance trait durability and efficacy.

Methods
Transcriptome assembly and protein analyses. The D. virgifera virgifera transcriptome was assembled as previously described 38 . Briefly, cDNA prepared from eggs, neonates, midguts of third instar larvae, testes from male and ovaries from female adults were sequenced by Illumina paired-end and 454 Titanium sequencing technologies. De novo transcriptome assemblies were performed using the Trinity method and the pooled assembly resulted in 81,277 contigs. The transcripts of dvvgr and dvbol were identified from transcriptome data set described above by blastP search using insect VgRs and BOL 15,22,26 . WCR VgR and BOL protein domains were predicted by InterProScan 39 . VgR and BOL proteins were extracted from dissected female and male corn rootworm reproductive tissues (Supplemental Methods A), and western blot analysis was used to detect the VgR and BOL protein in extracted WCR female and male reproductive tissues, respectively ( Supplementary Fig. 2).  Table 4) were amplified from gBlock using Taq DNA polymerase (Integrated DNA Technologies, Inc. Coralville, Iowa) with a pair of gene specific primers or produced by overlapping extension by PCR using four complementary DNA oligodeoxyribonucleotide (oligo) primers 40 . The gene-specific primers also contained promoter sites for T7 RNA polymerase at the 5′ end of each primer or at external oligos for overlapping extension (Supplementary Table 5). The PCR product served as the template for dsRNA synthesis by in vitro transcription (IVT) using an MEGAscript kit (Life Technologies, Carlsbad, CA).

Plant expression vectors and transformation.
To demonstrate rootworm efficacy in planta, dvvgr and dvbol gene fragments were assembled into a suppression cassette designed to express dsRNA targeting the vgr or bol gene. The silencing cassette consisted of two 155 base pair stretches of bol (or vgr) and an intervening truncated maize ADH intron1 designed to support assembly into a dsRNA ( Supplementary Fig. 9a). The constructs of bol and vgr were transformed via Agrobacterium tumefaciens into a commercial maize elite-inbred line, PHR03 41 . T0 maize transformants were transferred to soil and backcrossed with a PHR03 inbred line to generate T1 progeny.

Quantitative Reverse Transcription PCR (qRT-PCR) and
In situ hybridization (ISH). The expression of vgr and bol gene was quantified from WCR eggs, neonates, 1 st , 2 nd , 3 rd instar, pupae, and adults after feeding on diet incorporated with 50 ng μl −1 or different doses of vgr and bol fragment dsRNA. The designs of primers and probe regions are listed in Supplementary Table 6. Gene expression was analyzed using one-step real-time qRT-PCR. The assay was run, with 3 replicates per sample, using a single-plex set up with Bioline Sensifast Probe Lo Rox kit (Taunton, MA) and analyzed using the 2 −ΔΔCt method based on the relative expression of the target gene and reference gene dvrps10. For in situ hybridization (ISH) analyses, target probes were designed by Advanced Cell Diagnostics (Hayward, CA) (listed in Supplementary Table 6). Insect samples were fixed in 10% neutral buffered formalin (4% formaldehyde) for 48 to 72 h and processed as previously reported 11 . Slide images were acquired using a Leica Aperio ® AT2 digital scanner and captured at 40x magnification with a resolution of 0.25 µm pixel −1 .
Fecundity and mRNA expression of adult WCR exposed to dsRNA vgr (dsvgr) or bol (dsbol). The detailed insect rearing methods are described in Supplementary Methods B. Detailed description of fecundity assessment and estimation of net reduction in fecundity (NRF) are described in the Supplementary Methods C and D. Adult WCR insect bioassays were carried out using a diet incorporation methodology by incorporating dsRNA into previously described artificial diet 42,43 . To create the adult artificial diet pellets used in the bioassays, 25 µl of a solubilized solution of dsRNA (dsvgr, dsbol, dsgfp, or dsgus) was mixed with 75 µl of WCR artificial diet and placed in the well of a 96-well microtiter plate, for a final concentration of 100 ng µl −1 of dsRNA. For control diet, 25 µl of sterile deionized water was incorporated into 75 µl of WCR artificial diet per well. Adult beetles from the same batch of eggs were separated into two groups: young adults (<5 days old) and old adults (>11 days old). The following three treatments were compared 1) sterile deionized water (control); 2) gfp dsRNA (dsgfp), and 3) vgr dsRNA (dsvgr). For exposure, individual WCR adult beetles were confined for 24 h in individual wells of 32 cell tray (C-D International, Pitman, NJ) supplemented with a single artificial diet pellet, containing the desired treatment as described above. After 24 h, treated adults were transferred to their respective cages (30 × 30 × 30 cm popup cages with vinyl window; Raising Butterflies LLC, Salt Lake City, UT) to assess fecundity (Supplementary Methods E). Exposure of old adults to 100 ng μl −1 dsRNA bol (dsbol) was described in Supplementary Method F. Fecundity of WCR exposed to dsvgr and dsbol at 3 rd instar stage. 3 rd instar larvae were acclimatized on WCR larval diet 44,45 for 24 h and exposed to a diet containing sterile deionized water (control) or 50 ng μl −1 of vgr dsRNA (dsvgr), bol dsRNA (dsbol) or gus dsRNA (dsgus) (Supplementary Methods G). A total of 18 diet-acclimatized 3 rd instar larvae were added to each well of a 6-well costar plate (Corning Incorporated, Corning, NY) (n = 108 larvae per plate). Four replicate plates were prepared per treatment, for a total of 432 3 rd instar larvae per treatment. After 24 h exposure to dsRNA, larvae were transferred to a pupation dish (clear plastic container 18.7 cm diameter by 7.6 cm height, (Pioneer Plastics, Dixon, KY) filled with Miracle-Gro Garden soil (Scotts Company, Marysville, OH)) and incubated until adult emergence. Prior to the onset of adult emergence, each pupation dish was placed into a separate cage and both food and water source were added. Emerged beetles were counted, sorted by sex and treatment at the end of the 10-day pre-oviposition period. Beetles were randomly picked from male or female cages of the respective treatment to create three replicate cages for the fecundity study (10 to 22 pairs per cage). Each cage received a new oviposition dish every 5 days for 15 days.
For further assessment of 3 rd instar response to different concentrations of dsbol or dsvgr, a dose-response assay was conducted with four different concentrations (0.1, 1, 10, and 50 ng µl −1 of dsbol or dsvgr in diet). Diet acclimatized 3 rd instar larvae (n = 432) were exposed for 24 h to target doses of dsbol or dsvgr, and were incubated as described above to complete development. Emerged beetles were handled as described above, and at the end of a 10 d pre-oviposition period, four replicate cages containing 10 to 19 pairs were arranged for the fecundity study. Egg collection was performed every 5 days for 15 and 25 days for the dsvgr and dsbol assays, respectively, and egg hatch was assessed over a period of 8 days.
Sex-specific effects of dsbol exposure at 3 rd instar stage via reciprocal crossing. Diet acclimatized 3 rd instar larvae were exposed for 24 h to an artificial diet containing 50 ng µl −1 of dsbol or water control. A total of 3,780 3 rd instar larvae were exposed for each treatment. After 24 h, treated larvae were transferred to individual pupation cup (37 ml capacity translucent plastic cup with clear lids; Dart container corporation, Michigan, USA) filled with moist soil and were incubated to complete development. Emerged beetles were sorted, and placed in cages by sex and treatment. For the fecundity study, four reciprocal crossing combinations (bol ♂ x bol ♀; bol ♂ x H 2 O ♀; H2O ♂ x bol ♀; H 2 O ♂ x H 2 O ♀) were arranged in six replicate cages (n = 18 to 24 pairs per cage). The experiment was conducted for 25 days, and egg production and egg hatch rate was assessed following a similar procedure described above.

Fecundity of WCR feeding on plants expressing vgr and bol dsRNA transcripts. Eight T1
transgenic maize lines expressing the vgr dsRNA transcript (three fragments), and three T1 lines expressing the bol dsRNA transcript were produced and characterized, as described in the Supplemental Methods H and Supplementary Table 3. For each transgenic and non-transgenic (NTG) control line, three T1 plants were transplanted into a plastic pot (for a total of 7 to 11 pots per maize line), and pots were maintained in the greenhouse (27 °C, photoperiod of 15: 9 (L: D) h. Each pot was used as a replicate for exposure. At the V2 growth stage, each pot was infested with 200 WCR eggs. After 30 days, plants/pots were monitored daily for beetle emergence. Adult beetles were collected following previously described methods 46 and were brought to the laboratory every 2 to 3 days, for two to three weeks. Beetles were counted, sexed and kept in cages by treatment. At the end of the pre-oviposition period, females were recounted and randomly picked to create replicate cages (n = 3; 8 to 16 pairs per cage). Experimental cages were maintained as described above for 15 and 25 days, for vgr and bol transgenic plants, respectively. For the vgr treatment, eggs were collected daily or at an interval of 2-4 days for total observation of (n = 9). For the bol treatment, egg collection was performed every 5 days. Data Analysis. Egg numbers and percent egg hatch data was transformed using log10 or arcsine square-root, respectively to satisfy normality and homogeneous variance assumptions. Statistical analyses were performed using PROC GLM or PROC MIXED model. When a significant difference was detected, pairwise comparison of means was performed following the least square mean (LSMEANS) procedure in SAS Enterprise Guide v6.1 (SAS Institute, 2013).
The daily egg numbers and hatch data was considered to be independent and a one-way or two-way ANOVA was performed using PROC GLM. When treatments were replicated and data was collected for multiple time point, the data were analyzed using PROC MIXED procedure. The cage to cage variability was removed from the overall error, using the cage as a random effect. The treatment effect at each time point was estimated and tested at an alpha level of 5%. Insect and plant expression data were subjected to one-way-analysis of variance using JMP (v12. SAS Institute Inc, Cary, NC) followed by Dunnett's post-test. For all analysis, the results were considered statistically significant if the P-value was <0.05. Data availability. The RNAi active target sequences have been deposited in the GenBank of National Center for Biotechnology Information under the accession numbers KY373243 and KY373244.