A sugarcane mosaic virus vector for rapid in planta screening of proteins that inhibit the growth of insect herbivores.

Spodoptera frugiperda (fall armyworm) is a notorious pest that threatens maize production worldwide. Current control measures involve the use of chemical insecticides and transgenic maize expressing Bacillus thuringiensis (Bt) toxins. Although additional transgenes have confirmed insecticidal activity, limited research has been conducted in maize, at least partially due to the technical difficulty of maize transformation. Here, we describe implementation of a sugarcane mosaic virus (SCMV) vector for rapidly testing the efficacy of both endogenous maize genes and heterologous genes from other organisms for the control of S. frugiperda in maize. Four categories of proteins were tested using the SCMV vector: (i) maize defence signalling proteins: peptide elicitors (Pep1 and Pep3) and jasmonate acid conjugating enzymes (JAR1a and JAR1b); (ii) maize defensive proteins: the previously identified ribosome-inactivating protein (RIP2) and maize proteinase inhibitor (MPI), and two proteins with predicted but unconfirmed anti-insect activities, an antimicrobial peptide (AMP) and a lectin (JAC1); (iii) lectins from other plant species: Allium cepa agglutinin (ACA) and Galanthus nivalis agglutinin (GNA); and (iv) scorpion and spider toxins: peptides from Urodacus yaschenkoi (UyCT3 and UyCT5) and Hadronyche versuta (Hvt). In most cases, S. frugiperda larval growth was reduced by transient SCMV-mediated overexpression of genes encoding these proteins. Additionally, experiments with a subset of the SCMV-expressed genes showed effectiveness against two aphid species, Rhopalosiphum maidis (corn leaf aphid) and Myzus persicae (green peach aphid). Together, these results demonstrate that SCMV vectors are a rapid screening method for testing the efficacy and insecticidal activity of candidate genes in maize.

Introduction

Maize (Zea mays) is one of the world’s most important cereal crops, serving not only as a food source for humans and livestock, but also as a raw material for the production of ethanol and other industrial products (Ai and Jane, 2016; Chaudhary et al., 2014). The needs of an ever‐expanding population will lead to increasing demands on maize production in the coming years. Therefore, maintaining adequate maize yields will require reducing not only the cost of agricultural inputs, but also the negative impacts of biotic and abiotic stresses that limit maize productivity.

More than 100 species of insect pests feed on maize plants in agricultural fields (McMullen et al., 2009; Meihls et al., 2012). Among these pests, one of the most damaging is Spodoptera frugiperda (fall armyworm; Figure 1a), a lepidopteran species that is indigenous to the Americas but recently has become invasive in Africa and Asia (Food and Agriculture Organization of the United Nations, 2018; Goergen et al., 2016). By consuming all above‐ground plant parts, S. frugiperda larvae reduce photosynthetic area, cause developmental delays and decrease yield.

Figure 1

Insect species used in this study. (a) Spodoptera frugiperda, fall armyworm, (b) Rhopalosiphum maidis, corn leaf aphid and (c) Myzus persicae, green peach aphid

Currently available S. frugiperda control methods, both application of chemical insecticides (Togola et al., 2018) and transgenic maize producing Bacillus thuringiensis (Bt) toxins (Huang et al., 2014; Tabashnik and Carrière, 2017), are becoming less effective as the insects develop resistance. Therefore, there is a need to screen for additional transgenes that can be used to enhance maize resistance to S. frugiperda feeding. Broadly, such approaches can include upregulation of maize defence signalling, overexpression of individual maize defensive proteins and expression of heterologous insecticidal proteins.

Plant elicitor peptides (Peps) trigger anti‐herbivore defence responses (Huffaker, 2015; Huffaker et al., 2013; Poretsky et al., 2020). In maize, ZmPep1 and ZmPep3 up‐regulate defences, at least in part by induction of the jasmonic acid (JA) signalling pathway (Huffaker et al., 2011, 2013). A key step in the JA pathway is the conjugation of JA with isoleucine by JAR1 (JASMONATE RESISTANT 1) enzymes (Koo and Howe, 2009; Staswick et al., 2002) to form JA‐isoleucine. Expression of JAR1a and JAR1b, two of the five predicted JAR genes in maize (Borrego and Kolomiets, 2016), is highly induced by Spodoptera exigua (beet armyworm) herbivory (Tzin et al., 2017). Thus, these maize genes are good targets for overexpression to enhance resistance against S. frugiperda.

Maize and other plants produce ribosome‐inactivating proteins (RIPs) that block ribosome function by depurinating a specific adenine residue of the large ribosomal RNA (Bass et al., 2004; Zhu et al., 2018). These proteins, which are toxic for a variety of insects, including Lepidoptera (Shahidi‐Noghabi et al., 2009) and Hemiptera (Hamshou et al., 2016), have been used previously in transgenic approaches. For instance, the expression of a maize kernel RIP1 in Nicotiana tabacum (tobacco) increased resistance to Helicoverpa zea (corn earworm) feeding (Dowd et al., 2003). The RIP2 gene is expressed in all maize tissues except the kernels (Bass et al., 2004). RIP2 expression was induced by S. frugiperda herbivory and recombinant RIP2 protein decreased caterpillar growth on artificial diet (Chuang et al., 2014a).

Two additional classes of maize proteins with anti‐herbivore activity are proteinase inhibitors and antimicrobial peptides (Campos et al., 2018; Koiwa et al., 1997). Proteinase inhibitors, which are produced by many plant families, impair the growth and survival of insects by disrupting the function of digestive enzymes. Maize proteinase inhibitor (MPI) expression was induced by both caterpillar herbivory and JA signalling (Cordero et al., 1994; Shivaji et al., 2010; Tamayo et al., 2000). Heterologous expression of MPI in rice increased resistance to Chilo suppressalis (striped stem borer) (Vila et al., 2005). Cyclotides are macrocyclic insecticidal peptides with the length of about 30 amino acids and a conserved cystine knot motif containing three disulphide bonds (Campos et al., 2018; Craik et al., 1999; Weidmann and Craik, 2016). Cyclotide Kalata B1 from Oldenlandia affinis decreased the growth of Helicoverpa armigera (corn earworm) larvae by rupturing epithelial cells in the midgut (Barbeta et al., 2008). Cycloviolacins, cyclotides from Viola odorata, negatively affected the probing and feeding behaviour of Myzus persicae (green peach aphid), suggesting that cycloviolacins limit aphid population growth (Dancewicz et al., 2020). Among predicted antimicrobial peptides in maize, a few belong to the cyclotide family (Mulvenna et al., 2006; Noonan et al., 2017), but their efficacy against insects has not been confirmed.

Plant lectins, carbohydrate‐binding proteins that interact with glycoproteins and glycan structures in insect guts, have antinutritional or insecticidal effects (Macedo et al., 2015). For instance, snowdrop lectin (Galanthus nivalis agglutinin; GNA), onion lectin (Allium cepa agglutinin; ACA) and garlic (Allium sativum) leaf lectin reduce nutrient uptake and growth in wide range of insects (Vandenborre et al., 2011). Expression of a maize lectin gene, Jacalin 1 (JAC1), is induced by JA, an indication that it may provide protection against herbivory (Van Damme et al., 2004). There can be additive or even synergistic effects if lectins are co‐expressed or fused to scorpion or spider venom peptides. For instance, the insecticidal efficacy of GNA was increased by fusions to Hvt (ω‐hexatoxin Hv1a, also called ω‐ACTX Hv1a) from Hadronyche versuta (Blue Mountains funnel web spider) (Fitches et al., 2012), ButaIT (Buthus tamulus insect toxin) from Buthus tamulus, also known as Mesobuthus tamulus (Fitches et al., 2010; Wudayagiri et al., 2001), AaIT (Androctonus australis insect toxin) from Androctonus australis (Liu et al., 2016) and δ‐amaurobitoxin‐PIla from Pireneitega luctuosus (Yang et al., 2014).

Scorpion and spider venoms, which contain numerous insecticidal toxins (King and Hardy, 2013; Ortiz et al., 2015), have been explored as sources of insecticidal peptides. UyCT3 and UyCT5, two antimicrobial peptides that are produced in the venom glands of Urodacus yaschenkoi (inland robust scorpion) decrease the fitness of Acyrthosiphon pisum (pea aphid) and the density of the primary symbiont, Buchnera aphidicola, suggesting that those are promising candidates for the production of insect‐resistant transgenic plants (Luna‐Ramirez et al., 2017). Similarly, Hvt from H. versuta is broadly effective against both lepidopteran and hemipteran pests when expressed in transgenic plants (Ullah et al., 2015; Javaid et al., 2016; Rauf et al., 2019). Given the broad insecticidal activity of these spider and scorpion venom proteins, we hypothesized that they would also be effective in reducing S. frugiperda growth.

Testing the effectiveness of transgenes for controlling pest insects on maize is limited by the high cost of maize transformation and the often greater than one‐year timeline that is required to obtain transgenic maize plants for experiments. Therefore, we are proposing an alternate approach, whereby the efficacy of transgenes that enhance maize pest tolerance is tested by transient expression in maize using a virus vector. Sugarcane mosaic virus (SCMV), a positive‐sense single‐stranded RNA virus, which has been adapted for efficient transgene expression in maize (Mei et al., 2019), is attractive vector for such experiments. Genes of interest are inserted between the SCMV P1 and HC‐Pro cistrons in SCMV‐CS3, a newly created plasmid vector, maize seedlings are infected by particle bombardment or Agrobacterium inoculation, and the pest resistance of the infected plants can be assessed after three weeks. Here, we show that virus‐mediated expression of maize defence‐regulating proteins, maize insecticidal proteins and exogenous toxins (Table 1) can reduce the growth of insect pests. We demonstrate the insect‐controlling properties of not only known proteins but also two maize proteins that were not previously confirmed to have anti‐herbivore properties. Furthermore, we show that expression of transgenes using SCMV also is effective in reducing reproduction by two hemipteran pests, Rhopalosiphum maidis (corn leaf aphid; Figure 1b) and M. persicae (Figure 1c).

Table 1

SCMV‐expressed transgenes used in this study

Source species	Gene name	GenBank ID	MaizeGDB ID	Description
Zea mays	Pep1	XM_008670579	GRMZM2G055447	Peptide elicitor 1
Zea mays	Pep3	XM_008670581	GRMZM2G339117	Peptide elicitor 3
Zea mays	JAR1a	NM_001174342	GRMZM2G091276	Jasmonate‐isoleucine conjugating enzyme
Zea mays	JAR1b	NM_001361064	GRMZM2G162413	Jasmonate‐isoleucine conjugating enzyme
Zea mays	RIP2	NM_001137489	GRMZM2G119705	Ribosome‐inactivating protein 2
Zea mays	JAC1	NM_001148875	GRMZM2G050412	Jacalin 1, maize lectin
Zea mays	AMP	NM_001371023	GRMZM2G032198	Cyclotide antimicrobial peptide
Zea mays	MPI	X78988	GRMZM2G028393	Maize proteinase inhibitor
Allium cepa	ACA	DQ255944		Onion lectin
Galanthus nivalis	GNA	M55556		Snowdrop lectin
Hadronyche versuta	Hvt	AJ938032		Spider venom toxin, ω‐hexatoxin‐Hv1a
Urodacus yaschenkoi	UyCT3	JX274241		Scorpion venom toxin, antimicrobial peptide
Urodacus yaschenkoi	UyCT5	JX274242		Scorpion venom toxin, antimicrobial peptide

Results

SCMV‐GFP does not affect S. frugiperda growth on maize

A previously described SCMV cloning vector (Mei et al., 2019) was modified to produce SCMV‐CS3 (Figure 2a), and the mEGFP coding sequence (Zacharias et al., 2002) was placed in the multiple cloning site to produce SCMV‐GFP. An empty vector control virus (SCMV‐EV) and SCMV‐GFP were used to inoculate one‐week‐old maize seedlings. Both SCMV‐GFP and SCMV‐EV caused mosaic symptoms within two weeks after rub‐inoculation and continued to spread in newly emerging leaves of the infected plants (Figure 2b). Three weeks post‐inoculation, the transcript levels of GFP relative to maize endogenous control genes, Actin and EF1‐α, was quantified using quantitative RT‐PCR. The GFP expression level was significantly higher in plants infected by SCMV‐GFP than SCMV‐EV (Figure 2c). Infected leaves were examined using confocal microscopy, and the green fluorescence signal was only detected in leaves infected with SCMV‐GFP (Figure 2d). In these short‐term experiments, we did not observe a significant reduction in the growth of SCMV‐infected maize plants. However, older maize plants, which had been infected for more than four weeks, showed leaf mosaic symptoms associated with the SCMV infection. Neither S. frugiperda larval survival (Figure 2e) nor larval growth (Figure 2f) differed significantly between plants infected by SCMV‐GFP and SCMV‐EV. Therefore, SCMV‐GFP was used as a transgene‐expressing virus control treatment in subsequent experiments to test the efficacy of insect growth‐inhibiting proteins (Table 1) against S. frugiperda, R. maidis and M. persicae (Figure 1a–c).

Figure 2

Use of a sugarcane mosaic virus (SCMV) to transiently express GFP in maize. (a) Schematic diagram of the SCMV‐CS3 cloning vector. A multiple cloning site (MCS) was inserted between P1 and HC‐Pro genes of SCMV. 35S: cauliflower mosaic virus 35S promoter; T: NOS terminator. (b) Mosaic infection symptoms of SCMV vector in maize three weeks post‐inoculation. (c) Transient overexpression of GFP in plants infected by SCMV‐empty vector (EV) or ‐GFP. Gene expression was determined by qRT‐PCR three weeks post‐inoculation. Means ± SE of N = 5, ***P < 0.001, t‐test. (d) Confocal image of foliar GFP expression. (e,f) S. frugiperda survival and larval growth on plants infected by EV or SCMV‐GFP. Two‐day‐old caterpillars were fed on infected plants for one week and caterpillar mass was determined. Means ± SE of N = 10, EV, empty vector; NS, non‐significant; MCS, multiple cloning site

Expression of endogenous maize genes using SCMV enhances S. frugiperda resistance

We cloned the 69 bp sequences of the maize Pep1 and Pep3 defence elicitors (Huffaker et al., 2011, 2013) into pSCMV‐CS3 and inoculated seedlings of sweet corn variety Golden Bantam. Pep1 and Pep3 expression was confirmed by qRT‐PCR three weeks post‐inoculation (Figure 3a,b). Relative to SCMV‐GFP, infection with SCMV‐Pep1 and SCMV‐Pep3 caused increased transcript accumulation of maize proteinase inhibitor (MPI; Figure 3c), a JA‐induced marker gene with anti‐herbivore activity (Chuang et al., 2014b). Pep1 and Pep3 expression decreased the growth of S. frugiperda larvae on Golden Bantam maize by 25% and 51%, respectively, compared to GFP control plants (Figures 3d and S1), but did not affect survival of the larvae (Table S3). Similar results were obtained for Pep3 using sweet corn (P39) and field corn (B73) inbred lines (Figure 3e,f), showing that the effect on S. frugiperda growth is not specific to Golden Bantam.

Figure 3

Gene expression and Spodoptera frugiperda larval growth on plants expressing maize defence genes. (a,b) Transient overexpression of Pep1 and Pep3 in Golden Bantam plants infected by SCMV‐GFP, ‐Pep1 or ‐Pep3. (c) The expression level of MPI in plants expressing SCMV‐encoded GFP, Pep1 or Pep3. Gene expression was determined by qRT‐PCR three weeks post‐inoculation. (d) Performance of S. frugiperda on Golden Bantam plants expressing GFP, Pep1 or Pep3. (e,f) Performance of S. frugiperda on P39 and B73 plants expressing GFP or Pep3. (g–j,m,n) Transgene expression in Golden Bantam plants infected with SCMV‐GFP, ‐JAR1a, ‐JAR1b, ‐RIP2, ‐JAC1, ‐ AMP or ‐MPI. (k) Expression level of MPI in Golden Bantam plants expressing GFP, JAR1a or JAR1b. (l,o) Performance of S. frugiperda on plants expressing GFP, JAR1a, JAR1b, RIP2, JAC1, AMP or MPI. Five caterpillars were confined on each infected plant three weeks post‐inoculation and average caterpillar weights on each plant were measured one week later. Means ± SE of N = 5 for gene expression, N = 10–12 for insect bioassays, *P < 0.05, **P < 0.01, ***P < 0.001 relative to GFP control, Dunnett’s test

We targeted previously investigated maize insect resistance genes (JAR1a, JAR1b, RIP2 and MPI) and two predicted resistance genes (AMP and JAC1) for overexpression using SCMV. A previously uncharacterized gene (GRMZM2G032198) encoding a maize cyclotide antimicrobial peptides was designated as AMP. Expression of both AMP and JAC1 was induced by S. frugiperda herbivory (Figure S2), suggesting that these genes are involved in maize insect resistance. The expression levels of JAR1a, JAR1b, RIP2, JAC1, AMP and MPI were significantly higher in plants infected by the corresponding SCMV constructs than in SCMV‐GFP control plants (Figure 3g–j,m,n). SCMV‐JAR1a and SCMV‐JAR1b also significantly increased MPI expression (Figure 3k), confirming upregulation of JA‐related defence pathways. More importantly, the growth of S. frugiperda larvae was up to 60% lower on plants expressing JAR1a, JAR1b, RIP2, JAC1, AMP and MPI than on Golden Bantam control plants infected with SCMV‐GFP (Figure 3l,o). Although the larval growth rate was reduced in these experiments, larval survival was not significantly affected (Table S3).

Expression of scorpion insecticidal proteins reduces S. frugiperda growth

To determine whether SCMV can be used to express heterologous insect resistance genes in maize, we cloned UyCT3 and UyCT5, which encode U. yaschenkoi venom toxins, into the SCMV vector. Transgene expression in maize was confirmed by qRT‐PCR (Figure 4a,b). Both SCMV‐UyCT3 and SCMV‐UyCT5 reduced S. frugiperda weight on P39 and B73 plants compared to SCMV‐GFP control plants (Figure 4c,d). Relative to control plants, neither SCMV‐UyCT3 nor SCMV‐UyCT5 significantly reduced the larval survival rate (Table S4).

Figure 4

Gene expression and larval growth on P39 plants expressing heterologous insecticidal genes. (a, b) Transient overexpression of UyCT3 and UyCT5 in P39 plants infected by SCMV‐GFP, ‐UyCT3 or ‐UyCT5. Gene expression was determined by qRT‐PCR three weeks post‐inoculation. (c) Performance of S. frugiperda on P39 plants expressing GFP, UyCT3 or UyCT5. (d) Performance of S. frugiperda on B73 plants expressing GFP or UyCT3. Five caterpillars were confined on infected plants three weeks post‐inoculation. Average caterpillar weights on each plant were measured one week later. Means +/‐ SE of N = 5 for gene expression, N = 7‐14 for insect bioassays, *P < 0.05; **P < 0.01; ***P < 0.001 relative to GFP control, Dunnett’s test

Expression of fusion proteins using SCMV has additive effects on S. frugiperda

Fusion of spider and scorpion neurotoxins with plant lectins can improve their toxicity (Fitches et al., 2012; Liu et al., 2016; Rauf et al., 2019). We investigated this effect in maize using SCMV constructs. Because SCMV vectors produce a polyprotein precursor from which functional proteins are cleaved by the NIa protease (Mei et al., 2019), we generated SCMV constructs with and without an NIa cleavage site between the venom toxin and lectin to determine whether the proteins are more efficacious separately or as fusion proteins. The expression levels of H. versuta toxin (Hvt) and A. cepa agglutinin (ACA) genes were significantly higher in plants infected by each corresponding SCMV constructs than SCMV‐GFP control (Figure 5a,b), and the presence of protease cleavage site between Hvt and ACA did not affect Hvt and ACA transcript accumulation. Expression of single proteins did not significantly decrease S. frugiperda growth relative to the GFP control (Figure 5c). However, the expression of fusion protein by SCMV‐Hvt‐ACA or two individual proteins of Hvt and ACA by SCMV‐Hvt‐cleavage site‐ACA reduced the larval growth by 39% and 46%, relative to the GFP control (Figure 5c), respectively, suggesting an additive effect from the expression of the two genes.

Figure 5

Gene expression and larval growth on plants expressing heterologous insecticidal genes. (a,b) Transient overexpression of Hvt and ACA in plants infected by SCMV‐GFP, ‐Hvt, ‐ACA, ‐Hvt‐ACA or ‐Hvt‐*‐ACA. * represents a protease cleavage site. (c) Performance of Spodoptera frugiperda on plants expressing GFP, Hvt, ACA, Hvt‐ACA or Hvt‐*‐ACA. (d,e) Transient overexpression of GNA and RIP2 in plants infected by SCMV‐GFP, ‐RIP2, ‐GNA or ‐RIP2‐GNA. (f) Performance of S. frugiperda on plants expressing GFP, RIP2, GNA or RIP2‐GNA. (g,h) Transient overexpression of GNA and Hvt in plants infected by SCMV‐GFP, ‐GNA, ‐Hvt or ‐Hvt‐GNA. (i) Performance of S. frugiperda on plants expressing GFP, GNA or Hvt‐GNA. Gene expression was determined by qRT‐PCR three weeks post‐inoculation. Five caterpillars were confined on each infected plant three weeks post‐inoculation, and average caterpillar weights on each plant were measured one week later. Means ± SE of N = 5 for gene expression and N = 10–12 for insect bioassays. Different letters indicate significant differences, P < 0.05, ANOVA followed by Tukey’s HSD test

As the presence of a cleavage site between the two protein components did not increase efficacy (Figure 5c), we tested insecticidal activity of fusion proteins without viral protease cleavage sites in subsequent experiments. SCMV constructs were made with a maize defence gene, RIP2, and a spider insecticidal protein, Hvt, fused to G. nivalis agglutinin (GNA). We confirmed that each gene was expressed in plants infected by the corresponding construct (Figure 5d,e,g,h). Although the expression of GNA alone did not affect S. frugiperda larval growth, infection with SCMV‐RIP2 and SCMV‐RIP2‐GNA reduced S. frugiperda growth by 28% and 27%, respectively (Figure 5f). In the case of Hvt and GNA, neither gene by itself significantly reduced caterpillar growth. However, the SCMV‐Hvt‐GNA fusion construct significantly decreased S. frugiperda larval growth compared to GFP control protein (Figure 5i). Relative to controls, larval survival was not significantly affected by these SCMV constructs (Table S5). Together, these results indicate that fusion proteins combining lectins and maize defence proteins or venom toxins can improve resistance against S. frugiperda.

Expression of maize and scorpion genes enhances resistance to phloem‐feeding herbivores

To investigate whether our SCMV constructs also provide protection against phloem‐feeding insects, we conducted aphid bioassays using P39 plants infected with a subset of the previously tested constructs: SCMV‐GFP, SCMV‐Pep3, SCMV‐RIP2, SCMV‐UyCT3, SCMV‐UyCT5, SCMV‐AMP and SCMV‐JAC1. GFP expression did not affect aphid numbers compared to the empty vector control (Figure 6a,b). By contrast, expression of maize defence proteins and scorpion toxins significantly decreased progeny production by both R. maidis (Figure 6a,c) and M. persicae (Figure 6b,d) compared to SCMV‐GFP control.

Figure 6

Total number of aphids on P39 plants expressing maize defence genes and scorpion toxin genes. (a,c) Rhopalosiphum maidis (b,d) Myzus persicae. Eight R. maidis adults or ten M. persicae adults were confined on plants infected by SCMV‐EV, ‐GFP, ‐Pep3, ‐RIP2, ‐UyCT3, UyCT5, AMP or JAC1 three weeks post‐inoculation. Surviving aphids and their progeny were counted one week later. Means ± SE of N = 8–10, *P < 0.05, **P < 0.01, ***P < 0.001 relative to GFP control, Dunnett’s test

Discussion

Artificial diet assays are commonly employed for testing the oral efficacy of novel insecticidal proteins against insect herbivores (Fitches et al., 2012; Panwar et al., 2018; Yao et al., 2003). However, growth inhibition on artificial diet does not always correlate well with the effects that are observed when the same insecticidal proteins are subsequently expressed in transgenic plants (Khan et al., 2020). Both the context of the surrounding plant tissue and the localization of the insecticidal proteins in the plants could affect their toxicity against insect herbivores. Therefore, rather than pre‐screening insecticidal proteins by cloning in microbial systems, purification and artificial diet assays, we propose that transient expression using a viral vector such as SCMV will be a more effective approach for rapidly testing the in planta efficacy of novel insecticidal proteins.

A possible limitation of the current study is that, although we have confirmed transgene expression, we have not verified the location and abundance of each individual protein in SCMV‐infected maize plants. However, in previous experiments, easily visualized proteins [GFP, GUS (β‐glucuronidase) and NLuc (nano luciferase)] were clearly overproduced in maize after gene expression using SCMV vectors (Beernink et al., 2021; Mei et al., 2019). Production of BAR (bialaphos resistance) using an SCMV vector conferred herbicide resistance to virus‐infected plants (Mei et al., 2019), indicating that there was successful protein production. Similarly, cloning of genes that encode insecticidal or defence‐inducing proteins into SCMV as part of the current study caused reduced S. frugiperda caterpillar growth. This is a strong indication that there is not only gene expression but also protein production from these SCMV constructs, and that SCMV is an effective vector for testing the in planta effectiveness of proteins that control S. frugiperda and other insect herbivores.

Reduced weight gain of S. frugiperda in response to SCMV‐mediated expression of Pep1, Pep3, JAR1a, JAR1b, RIP2, MPI, UyCT3 and UyCT5 as single‐gene constructs is consistent with previous reports of these genes providing protection against insect herbivory. Additionally, AMP and JAC1, two maize genes that are up‐regulated in response to S. frugiperda feeding (Figure S2), reduced caterpillar weight gain when overexpressed in maize. Although transcriptomic studies have identified numerous maize genes that are up‐regulated in response to arthropod feeding (Bui et al., 2018; Guo et al., 2019; Pan et al., 2020; Song et al., 2017; Tzin et al., 2015, 2017; Wang et al., 2017; Yang et al., 2020; Zhang et al., 2016), the majority of these genes have not been investigated for their roles in plant defence against herbivory. This is at least in part due to the time and cost of creating maize lines that have individual genes overexpressed. Transient expression using SCMV, as we have done for AMP and JAC1, will accelerate the process of testing the defencing functions of maize genes that are induced in response to herbivory. However, SCMV infection itself changes the abundance of numerous maize proteins (Chen et al., 2017). Thus, the possible effects of SCMV infection on the function of insect resistance proteins will need to be considered when selecting candidate genes for SCMV‐mediated expression. Moreover, the effects of SCMV‐mediated overexpression always should be evaluated with a proper control treatment such as SCMV‐GFP and/or SCMV‐empty vector.

Although expression of Hvt, ACA and GNA increased resistance in other plant‐insect interaction studies (Ullah et al., 2015; Liu et al., 2016; Vandenborre et al., 2011), expression of these single‐gene SCMV constructs in maize did not significantly reduce S. frugiperda weight gain. Nevertheless, expression of gene fusions, Hvt‐ACA and Hvt‐GNA, reduced S. frugiperda weight gain, indicating that there are additive effects of the spider venom and the lectin (Figure 5c,i). This is consistent with other studies showing that the toxicity of spider and scorpion toxins was improved by combining with a lectin, which may facilitate the transfer of the venom proteins across the gut lumen (Fitches et al., 2002; Javaid et al., 2016; Nakasu et al., 2014; Rauf et al., 2019). In addition to these synergistic effects on larval growth, it is likely that the stacking of multiple toxic proteins with different modes of action in one viral construct will delay the development of resistance in insects (Head et al., 2017; Ni et al., 2017). In contrast to Hvt‐lectin fusions, we did not observe an additive effect when RIP2 was linked to GNA (Figure 5f). This difference may be attributed to the differing origins of the toxin proteins expressed in the SCMV constructs. Whereas spider venoms like Hvt are injected directly into the haemolymph, the maize RIP2 protein would be consumed orally by lepidopteran larvae. As an endogenous maize insecticidal protein, RIP2 may bind to an as yet unknown receptor and thereby enter the midgut cells and/or the insect haemolymph.

Our observation of increased defence gene expression (Figure 3c) and reduced S. frugiperda weight gain (Figure 3d‐f) on plants infected with SCMV‐Pep1 and SCMV‐Pep3 is consistent with experiments showing that pre‐treatment of maize plants with Pep1 and Pep3 increases JA levels, defence gene expression and defensive metabolites, leading to reduced growth of S. exigua larvae (Huffaker et al., 2011, 2013). Under normal circumstances, peptide signalling is initiated when plasma membranes are disrupted and elicitor peptides are released from the cytoplasm into the apoplastic space (Bartels and Boller, 2015). Receptors on neighbouring intact cells recognize these peptides and elicit downstream defence pathways (Lori et al., 2015). However, in our experiments, the expression of MPI was induced in plants infected by SCMV‐Pep1 and SCMV‐Pep3, even before initiation of caterpillar feeding (Figure 3c). Pathogen attack can cause plant proteins without secretory signals to be released into the apoplast (Agrawal et al., 2010) and is possible that SCMV infection initiated Pep1 and Pep 3 signalling in this manner.

Induced defences, in particular those regulated by the jasmonic acid pathway, typically are turned on after perception of insect herbivory by maize and other plants (Erb and Reymond, 2019; Howe and Jander, 2008). However, initiation of jasmonate‐regulated defences takes time and some lepidopteran herbivores may have the ability to suppress jasmonate signalling (Chen et al., 2019a). Therefore, targeted initiation of maize defence responses by expression of regulatory proteins in SCMV may be an approach for increasing pest resistance. Such virus‐mediated induction could be deployed in maize fields when there is the specific threat of insect pests such as S. frugiperda. However, the negative growth effects that may be associated with longer‐term SCMV infection of mature maize plants have not been fully investigated. Future research could be focused on engineering SCMV to be less pathogenic or identifying other maize viruses that lead to asymptomatic infections in mature plants.

Currently available insect‐resistant transgenic maize varieties, in particular those expressing Cry, Cyt or Vip genes from B. thuringiensis (Bravo et al., 2011; Chakroun et al., 2016), are not effective against phloem‐feeding insects such as R. maidis (corn leaf aphid; Figure 1b) or M. persicae (green peach aphid; Figure 1c). Although the toxicity of Bt toxins can be increased by incorporating peptide sequences that bind to aphid guts (Chougule et al., 2013), this has not yet resulted in a commercially viable Bt toxin directed at hemipteran herbivores. Thus, there is an interest in identifying additional insecticidal proteins that can be expressed in the plant phloem to enhance aphid resistance. Our demonstration that transient expression of both endogenous maize proteins and insecticidal proteins from other species can reduce reproduction of R. maidis and M. persicae (Figure 6) suggests that SCMV‐mediated overexpression can be used to screen proteins for their effectiveness against hemipteran pests of maize.

Whereas overexpression of endogenous maize proteins is likely to be considered low risk in a food crop, other constructs that we have tested would have difficulty gaining regulatory approval for commercialization. Some of the tested spider venom proteins have insect‐specific toxicity (Fletcher et al., 1997) and have been approved by United States Environmental Protection Agency (Bilgo et al., 2017). However, further studies are required to investigate whether they would have negative effects on vertebrates when incorporated into transgenic food crops. Additional consumer‐related concerns include dietary restrictions in some segments of the population. For instance, maize expressing an arthropod protein might not be considered suitable for a vegetarian diet.

Our results demonstrate the utility of SCMV‐mediated overexpression for screening the efficacy of proteins that reduce insect growth on maize plants. Virus‐mediated transient expression assays included genes encoding maize regulatory proteins, endogenous maize defensive proteins and non‐maize insecticidal proteins. The main advantage of the SCMV overexpression system is a timeline that makes it possible to test the effectiveness of single‐ and multi‐gene constructs in actual maize plants in less than two months. Although the main focus of our efforts was a lepidopteran herbivore, S. frugiperda, we also showed efficacy of SCMV constructs against two aphid species, R. maidis and M. persicae, suggesting that the SCMV‐mediated transient expression approach will be broadly useful for experiments with both chewing and piercing/sucking herbivores of maize. Proteins that inhibit insect growth after transient gene expression using SCMV infection of maize are good candidates for further experiments to determine their effectiveness in stable transgenic maize plants.

Materials and methods

Plants and insects

Maize (Zea mays) plants, sweet corn variety Golden Bantam (West Coast Seeds, British Columbia, Canada) and inbred lines P39 and B73, were grown in a maize mix [0.16 m3 Metro‐Mix 360 (Scotts, Marysville, OH, USA), 0.45 kg finely ground lime, 0.45 kg Peters Unimix (Griffin Greenhouse Supplies, Auburn, NY, USA), 68 kg Turface MVP (Banfield‐Baker Corp., Horseheads, NY, USA), 23 kg coarse quartz sand and 0.018 m3 pasteurized field soil]. All plants, including those used for SCMV propagation and insect bioassays, were maintained in a growth chamber at 23 °C with a 16:8 h light:dark cycle. Unless specified otherwise, Golden Bantam maize was used for the described experiments.

Eggs of S. frugiperda were purchased from Benzon Research (Carlisle, PA, USA) and maintained on an artificial diet (Fall Armyworm Diet, Southland Products Inc, Lake Village, AR, USA) in an incubator at 28 °C. A colony of a genome‐sequenced R. maidis lineage (Chen et al., 2019b) was maintained on maize (Golden Bantam or P39), and a colony of a previously described tobacco‐adapted strain of M. persicae (Ramsey et al., 2007) was maintained on Nicotiana tabacum (tobacco) at 23 °C under 16:8 h light:dark cycle. Both R. maidis and M. persicae were originally collected in USDA‐ARS greenhouses by Stewart Gray (Robert W. Holley Center for Agriculture & Health, Ithaca, NY, USA).

Cloning of candidate genes into Sugarcane mosaic virus for protein expression

The pSCMV‐CS3 expression vector used in this work was derived from pSCMV‐CS2 (Mei et al., 2019), which was modified to contain the CS3 restriction sites in the multiple cloning site between the P1 and HC‐Pro cistrons (Figure 2a). The modified pSCMV‐CS2 genome plus the flanking 35S promoter and NOS terminator was amplified with SuperFi polymerase (Thermo Fisher Scientific, Waltham, MA) using primers DCPacI 1380 F and DCPacI 1380 R (Table S1). The pCAMBIA1380 backbone (www.cambia.org) was amplified with SuperFi polymerase using primers 1380F and 1380R (Table S1). The two PCR fragments were subsequently assembled into pCAMBIA1380‐SCMV‐CS3, hereafter referred to as pSCMV‐CS3, by Gibson Assembly (New England Biolabs, Ipswich, MA). pSCMV‐GFP was created by amplifying the mEGFP coding sequence (Zacharias et al., 2002) using SuperFi polymerase with the GFP‐PspOMI and GFP‐SbfI primers (Table S1). The resulting amplicon was digested with PspOMI and SbfI, gel purified and ligated into similarly digested pSCMV‐CS3.

Genes encoding maize defence regulators and protein toxins were amplified from the B73 cDNA template with gene‐specific primers containing restriction sites at the 5’ end for cloning into the SCMV vector. PCR products were gel purified, digested with the corresponding restriction enzymes and cloned into the SCMV vector (Figure 2a). The UyCT3 and UyCT5 genes (codon optimized for maize), as well as the maize Pep1 and Pep3 genes, were synthesized by Genscript Biotech (Piscataway, NJ, USA) and cloned into the SCMV vector. To generate SCMV with Hvt, ACA, GNA, RIP2, Hvt‐ACA, Hvt*‐ACA, RIP2‐GNA and Hvt‐GNA, primers listed in Table S1 and DNA synthesized by Genewiz (South Plainfield, NJ, USA) were used as template and the resulting PCR products were gel purified and cloned into PspOMI‐SbfI cut SCMV by following the NEBuilder HiFi DNA Assembly method (NEB, Ipswich, MA, USA). Primers, restriction sites and cloning methods used in this study are listed in Table S1.

Inoculation of maize with SCMV constructs

SCMV constructs were delivered into maize plants by particle bombardment using a Biolistic PDS‐1000/He system (Bio‐Rad, Hercules, CA, USA) as described previously (Mei and Whitham, 2018). Five µg of the plasmid DNA was coated on 3 mg 1.0 μm diameter gold particles, and the coated gold particles were distributed onto five microcarriers and allowed to air dry. Plants were placed in the dark 12 h before the particle bombardment. Two leaves of one‐week‐old plants, randomly oriented so that the adaxial or abaxial surface was targeted, were bombarded with using 1100 psi rupture discs at a distance of 6 cm (between stopping screen and leaves; Figure S3). Successfully infected plants were obtained with a frequency of 5% to 40%, in particle bombardment experiments.

Agroinoculation was used to initiate maize infections with the SCMV constructs containing Hvt, ACA and UyCT3, as described previously (Beernink et al., 2021). The constructs were transformed into the Agrobacterium tumefaciens strain GV3101, and an A. tumefaciens suspension with optical density at 600 nm (OD600) = 1.0 in infiltration buffer (200 μm acetosyringone, 10 mm MES, pH 5.6 and 10 mm MgCl2) was injected above the coleoptile node of one‐week‐old plants. Successfully infected plants were obtained with a frequency of 1% to 5%, in agroinoculation experiments.

Due to the relatively low infection rate with both particle bombardment and agroinoculation, SCMV constructs in maize were further propagated and used to infect experimental plants by rub‐inoculation. Leaf sap of SCMV‐infected plants was prepared by grinding 0.5 g leaf tissue in 5 mL of 50 mM pH 7.0 potassium phosphate buffer. One‐week‐old maize plants were dusted with 600‐mesh carborundum and mechanically inoculated by rubbing leaf sap from virus‐infected maize plants on two leaves. Successfully infected plants were obtained with a frequency of 50% to 80%, in rub‐inoculation experiments.

Confocal microscopy

Three weeks post‐inoculation, leaf samples were collected from the seventh or eighth leaves of maize plants infected with SCMV‐GFP. The samples were observed at an excitation of 488 nm. The emitted fluorescence signal was monitored from 505 to 545 nm using a SP5 Leica Confocal Microscope in the Plant Cell Imaging Center of Boyce Thompson Institute. A scan of fluorescence across a range of wavelengths (lambda scan) was used to confirm that the observed signal was derived from GFP rather than endogenous maize fluorescence.

Insect bioassays

To determine the effect of defensive proteins on the growth of S. frugiperda, SCMV‐infected plants, three weeks post‐inoculation, were used for caterpillar bioassays. Five two‐day‐old caterpillars were placed on each plant and enclosed using perforated plastic bags (13 × 61 cm, https://www.clearbags.com). The fresh mass of surviving caterpillars was measured one week later. The average mass of caterpillars from each plant was used as a biological replicate when conducting comparisons between maize plants infected with different SCMV constructs. For aphid bioassays, 8 ten‐day‐old apterous adult R. maidis or 10 ten‐day‐old apterous adult M. persicae were placed on each virus‐infected plant 15–18 days after SCMV inoculation and enclosed using perforated plastic bags. The total numbers of aphids on each plant were counted one week later.

RNA extraction, cDNA synthesis, RT‐PCR and quantitative real‐time PCR (qRT‐PCR)

Three weeks post‐inoculation and prior to the insect bioassays, leaf tissue was collected from the seventh or eighth leaves of infected plants, flash‐frozen in liquid nitrogen and stored at −80 °C. RNA was extracted using TRIzol Reagent (Invitrogen, Carlsbad, CA, USA) and treated with RQ1 RNase‐free DNase (Promega, Madison, WI, USA). One microgram of RNA was used to synthesize first‐strand cDNA using High‐Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA) with random primers. To verify the expression of insect resistance genes cloned into the SCMV vector, qRT‐PCR was conducted with gene‐specific primers (Table S1). For the inserts of less than 75 bp (Pep1, Pep3, UyCT3 and UyCT5), one primer was designed to bind to the insert and another primer was designed to bind to the region flanking the cloning site, with the amplification products ranging in size from 100 to 150 bp. The reactions consisted of 5.0 μL of the PowerUp SYBR Green PCR master mix (Applied Biosystems), 0.6 μL primer mix (300 nm for the final concentration of each primer) and 2 μL of cDNA (1:10 dilution with nuclease‐free H2O) in 10 μL total volume. Template‐free reactions were included as negative controls. The PCR amplification was performed on QuantStudio 6 Flex Real‐Time PCR Systems (Applied Biosystems, Foster City, CA, USA) with the following conditions: 2 min at 50 °C, 2 min at 95 °C, 40 cycles of 95 °C for 15 s and 60 °C for 1 min. Primer specificity was confirmed by melting curve analysis. Mean cycle threshold values of duplicates of each sample were normalized using two reference genes, Actin and EF1‐α. Relative gene expression values were calculated using 2‐ΔΔCt method (Livak and Schmittgen, 2001).

Statistical analysis

All statistical analyses were conducted using R (R Core Team, 2017). Data for gene expression and larval mass of S. frugiperda larvae, and aphid fecundity were compared using analysis of variance (ANOVA) followed by Tukey’s test, Dunnett’s tests relative to the GFP control, or t‐tests. Gene expression data were log2 transformed before the statistical analysis to meet the assumptions of ANOVA but untransformed data are presented in the figures. Survival of S. frugiperda larvae was analysed using non‐parametric Kruskal–Wallis tests. Raw data underlying the bar graphs are presented in Tables S2‐S7.

Conflict of interest

The authors declare that they have no conflict of interest.

Author Contributions

SHC, GJ, SPDK and SAW: designed experiments. SHC, RRL, BC and UN: cloned genes and constructed virus vectors. SHC and MB: conducted insect bioassays. SHC and GJ: analysed data. SHC, MB and GJ: involved in manuscript writing. SAW, SPDK and GJ: obtained funding and other essential resources. All authors approved the final manuscript.

Figure S1 Representative pictures of Spodoptera frugiperdacaterpillars raised on P39 maize plants infected with SCMV.

Figure S2 Gene expression of (a) AMPand (b) JAC1in P39 maize plants damaged by Spodoptera frugiperdacaterpillars.

Figure S2 One‐week‐old P39 maize plants were placed in the vacuum chamber of a Biolistic PDS‐1000/He delivery system for inoculation with SCMV constructs.

Click here for additional data file.

Table S1 List of primers and restriction sites.

Table S2 Raw data for the graphs in Figure 2.

Table S3 Raw data for the graphs in Figure 3.

Table S4 Raw data for the graphs in Figure 4.

Table S5 Raw data for the graphs in Figure 5.

Table S6 Raw data for the graphs in Figure 6.

Table S6 Raw data for the graphs in Figure S3.

Click here for additional data file.