Plant-mediated RNAi silences midgut-expressed genes in congeneric lepidopteran insects in nature

Abstract

Background: Plant-mediated RNAi (PMRi) silencing of insect genes has enormous potential for crop protection, but whether it works robustly under field conditions, particularly with lepidopteran pests, remains controversial. Wild tobacco Nicotiana attenuata and cultivated tobacco (N. tabacum) (Solanaceae) is attacked by two closely related specialist herbivores Manduca sexta and M. quinquemaculata (Lepidoptera, Sphingidae). When M. sexta larvae attack transgenic N. attenuata plants expressing double-stranded RNA(dsRNA) targeting M. sexta's midgut-expressed genes, the nicotine-ingestion induced cytochrome P450 monooxygenase (invert repeat (ir)CYP6B46-plants) and the lyciumoside-IV-ingestion induced β-glucosidase1 (irBG1-plants), these larval genes which are important for the larvae's response to ingested host toxins, are strongly silenced.

Results: Here we show that the PMRi procedure also silences the homologous genes in native M. quinquemaculata larvae feeding on irCYP6B46 and irBG1-transgenic N. attenuata plants in nature. The PMRi lines shared 98 and 96% sequence similarity with M. quinquemaculata homologous coding sequences, and CYP6B46 and BG1 transcripts were reduced by ca. 90 and 80%, without reducing the transcripts of the larvae's most similar, potential off-target genes.

Conclusions: We conclude that the PMRi procedure can robustly and specifically silence genes in native congeneric insects that share sufficient sequence similarity and with the careful selection of targets, might protect crops from attack by congeneric-groups of insect pests.

Keywords: CYP6B46; Manduca quinquemaculata; Plant-mediated RNAi; RNA interference; Transgenic tobacco plants; β-glucosidase.

Fig. 1

Experimental system used to evaluate if the PMRi lines generated from Manduca sexta genes can silence M. quinquemaculata homologous genes. a Fifth-instar M. sexta and its closely related species M. quinquemaculata larvae on their native host plant, the wild tobacco Nicotiana attenuata in Great Basin Desert of southwestern Utah. b Schematic overview of plant-mediated RNAi: the binary vector constructed to independently express ~300 bp dsRNA of M. sexta’s target genes such as cytochrome P450 monooxygenase (CYP6B46) and β-glucosidase1 (BG1) in N. attenuata. The trophic transfer of these dsRNA from plant to M. sexta larvae silences their respective target gene expression. PMRi lines generated with M. sexta dsRNA were used to test if the trophic transfer can also silence homologous gene expression in M. quinquemaculata larvae in nature

Fig. 2

Alignments of M. sexta cDNA regions selected to generate PMRi lines with M. quinquemaculata homologs. a M. sexta CYP6B46 fragment cloned to generate inverted-repeat (ir) CYP6B46 showed 98.1% sequence similarity with M. quinquemaculata CYP6B46 and (b) M. sexta BG1 fragment cloned to generate irBG1 showed 96% similarity with M. quinquemaculata BG1. Identical homologous regions of >21 nt were identified in the alignment of M. sexta CYP6B46 and BG1 cDNA regions selected to generate PMRi lines with that of M. quinquemaculata CYP6B46 and BG1, respectively

Fig. 3

Silencing of midgut-expressed CYP6B46 and BG1 genes in wild M. quinquemaculata larvae feeding on WT, PMRi, and nicotine and DTG-depleted plants. a Wild Manduca spp. eggs oviposited on Datura wrightii provided the source of M. quinquemaculata larvae which can be distinguished from M. sexta larvae when the larvae reach the third- instar. b PMRi N. attenuata lines and N. attenuata plants transformed by RNAi to silence: nicotine biosynthesis, by expressing an inverted repeat (ir) construct of the host plant’s putrescine N-methyl transferase (irPMT) and 17-hydroxygeranyllinalool diterpene glycoside (HGL-DTGs) biosynthesis, by expressing an ir construct of geranylgeranyl pyrophosphate synthase (irGGPPS); planted in a field plot in Great Basin Desert of southwestern Utah. c M. quinquemaculata CYP6B46 transcripts (relative to ubiquitin) in various tissues of fourth-instar larvae fed on EV, irCYP6B46 and irPMT plants (midgut: F _2,15 = 7.219 P ≤ 0.006; hindgut: F _2,15 = 6.651 P ≤ 0.008; Malpighian tubules: F _2,15 = 10.604 P ≤ 0.001; n = 6 in all bars). Note that feeding on nicotine-containing WT plants strongly induces the midgut expression of CYP6B46 transcripts and that feeding on the PMRi plants which contain WT levels of nicotine deplete CYP6B46 transcript abundance to levels found in larvae feeding on nicotine-depleted irPMT plants. d M. quinquemaculata BG1 transcripts (relative to ubiquitin) in various tissues of fourth-instar larvae feeding on EV, irBG1 and irGGPPS plants (midgut: F _2,14 = 9.458 P ≤ 0.002; n = 6 in EV, GGPPS and 5 in BG1 group). Note that feeding on HGL-DTG-containing WT plants strongly induces the midgut expression of BG1 transcripts and that feeding on the PMRi plants which contain WT levels of HGL-DTGs deplete BG1 transcript abundance to levels found in larvae feeding on HGL-DTG-depleted irGGPPS plants. Asterisks indicate significant differences between means (± SE) in comparison to EV, determined by one-way ANOVA and Fisher LSD post hoc, which was conducted separately for each tissue

Fig. 4

Silencing of MqCYP6B46 and MqBG1does not silence the most closely related, likely non-target genes. a Alignment of MqCYP6B46 partial coding sequence with MqCYP6B45 showed 83% sequence similarity and one identical homologous region >21 nt was identified. b Alignment of MqBG1 coding sequence with MqBG2 showed 71% sequence similarity and no identical homologous regions >21 nt were identified. Transcripts (relative to ubiquitin) of (c) MqCYP6B45 and (d) MqBG2 in the midguts of fourth-instar EV-, irCYP6B46 and EV-, irBG1-feeding larvae did not show off-target silencing of CYP6B45 and BG2, respectively, (n = 6) Significant differences between means (± SE) in comparison to EV was determined by one-way ANOVA and Fisher LSD post hoc test