The NIa-Pro protein of Turnip mosaic virus improves growth and reproduction of the aphid vector, Myzus persicae (green peach aphid)



Many plant viruses depend on aphids and other phloem-feeding insects for transmission within and among host plants. Thus, viruses may promote their own transmission by manipulating plant physiology to attract aphids and increase aphid reproduction. Consistent with this hypothesis, Myzus persicae (green peach aphids) prefer to settle on Nicotiana benthamiana infected with Turnip mosaic virus (TuMV) and fecundity on virus-infected N. benthamiana and Arabidopsis thaliana (Arabidopsis) is higher than on uninfected controls. TuMV infection suppresses callose deposition, an important plant defense, and increases the amount of free amino acids, the major source of nitrogen for aphids. To investigate the underlying molecular mechanisms of this phenomenon, 10 TuMV genes were over-expressed in plants to determine their effects on aphid reproduction. Production of a single TuMV protein, nuclear inclusion a-protease domain (NIa-Pro), increased M. persicae reproduction on both N. benthamiana and Arabidopsis. Similar to the effects that are observed during TuMV infection, NIa-Pro expression alone increased aphid arrestment, suppressed callose deposition and increased the abundance of free amino acids. Together, these results suggest a function for the TuMV NIa-Pro protein in manipulating the physiology of host plants. By attracting aphid vectors and promoting their reproduction, TuMV may influence plant–aphid interactions to promote its own transmission.


For most plant viruses, transmission by phloem-feeding insects, such as aphids, is an essential component of the infection cycle (Ng and Perry, 2004). Thus, aphid feeding behavior can affect pathogen success, and there is likely to be strong evolutionary pressure for these viruses to manipulate plant–insect interactions to optimize their own transmission (Sisterson, 2008; Ingwell et al., 2012). Broadly, insect-mediated virus transmission can be categorized as persistent, semi-persistent, or non-persistent. Whereas persistent viruses are internalized and pass through the gut and hemolymph to the salivary glands, semi-persistent and non-persistent viruses attach transiently to the foregut or mouthparts, respectively (Gray and Banerjee, 1999; Blanc et al., 2011). These different transmission modes suggest that there should also be diverse strategies that viruses use to manipulate host plants and promote transfer by insects (Mauck et al., 2012).

Evidence is accumulating that virus infections influence plant resistance to insect herbivores in both natural and agricultural systems (Eigenbrode et al., 2002; Belliure et al., 2005; Mauck et al., 2010, 2012). Virus infection can alter nutritional composition, defense responses, and other host plant traits that affect the settling, feeding, and dispersal of insect vectors. In the case of non-persistently transmitted viruses, which are acquired rapidly during initial aphid probing to reach the phloem (Hoh et al., 2010; Martiniere et al., 2013), infected host plants are often less suitable for insect growth than healthy controls (Sisterson, 2008; Bosque-Perez and Eigenbrode, 2011; Mauck et al., 2012). Thus, it has been proposed that non-persistent viruses can optimize their transmission by attracting insect vectors using volatile cues, making the infected plants less suitable for long-term feeding, and thereby promoting dispersal to new hosts (Eigenbrode et al., 2002; Ngumbi et al., 2007; Sisterson, 2008; Werner et al., 2009; Mauck et al., 2010, 2012; Bosque-Perez and Eigenbrode, 2011).

Despite the ecological and agricultural importance of tritrophic interactions between viruses, vectors, and hosts, few studies have addressed the molecular mechanisms through which viruses mediate changes in host plant resistance to insect vectors. In the case of the non-persistently transmitted Cucumber mosaic virus (CMV), the virally encoded 2b silencing suppressor protein disrupts jasmonate-mediated defense responses in Arabidopsis thaliana (Arabidopsis; Lewsey et al., 2010). Myzus persicae (green peach aphid), an important CMV vector, reproduces more rapidly on CMV-infected tobacco (Nicotiana tabacum) than on uninfected controls. In addition, plants infected with a CMV 2b deletion mutant are more aphid resistant, a finding that suggests that this protein is required for enhanced aphid performance (Ziebell et al., 2011). In another three-way interaction, Tomato yellow leaf curl China virus, a whitefly-transmitted virus, suppresses plant defenses using the βC1 protein of the viral beta-satellite DNA, thereby benefiting the insect vector. (Zhang et al., 2012).

Turnip mosaic virus (TuMV, a member of the Potyviridae) is a monopartite positive-strand RNA virus that infects hundreds of dicot plant species (Walsh and Jenner, 2002). TuMV is transmitted by M. persicae and more than 80 other aphid species (Shattuck, 1992), making it one of the most damaging viruses for vegetable crops worldwide (Tomlinson, 1987). Unlike most other non-persistently transmitted viruses, TuMV and Potato virus Y, both members of the Potyviridae, can make infected host plants more suitable for M. persicae reproduction (Hodge and Powell, 2008; Kersch-Becker and Thaler, 2013). In wild populations of Arabidopsis, about 20% of the plants have been found to be infected with TuMV (Pagan et al., 2010), showing that Arabidopsis is a natural host for this virus. M. persicae, one of the most important insect vectors for TuMV, readily feeds on Arabidopsis in greenhouses and in nature (Bush et al., 2006; Harvey et al., 2007).

Due to its ability to systemically infect two well studied model plants, Arabidopsis and Nicotiana benthamiana (Sanchez et al., 1998; Martín Martín et al., 1999; Bombarely et al., 2012), TuMV has become a model for studying potyvirus–host interactions (Walsh and Jenner, 2002). As with other potyviruses, TuMV codes for 11 proteins, 10 of which are translated as a single polyprotein before cleavage by the P1, HC-Pro, and NIa proteases (Mavankal and Rhoads, 1991; Verchot et al., 1991; Chung et al., 2008). To provide further insight into the molecular biology of plant–virus–vector interactions, we investigated how TuMV infection of Arabidopsis and N. benthamiana affects the reproductive success of its aphid vector, M. persicae, and the role of specific viral proteins in vector–plant ecology. Our results show that TuMV infection suppresses host plant defenses, elevates the free amino acid content of the phloem, promotes aphid settling, and increases aphid reproduction. Furthermore, the likely role of the viral protein NIa-Pro in vector–plant ecology was identified by cloning 10 TuMV genes, expressing each one individually in N. benthamiana, and assessing M. persicae performance and behavior, and physiological changes in the host plants.


In initial experiments, we compared M. persicae fecundity on uninfected and virus-infected N. benthamiana plants. Similar to the findings of previously reported turnip experiments (Hodgson, 1981), infection with both wild type TuMV and an engineered strain that expressed green fluorescent protein (GFP; TuMV–GFP) (Lellis et al., 2002) increased M. persicae fecundity (Figure 1). In contrast, two other viruses, Potato Virus X (PVX), which is not aphid-transmitted, and CMV, which is non-persistently transmitted by M. persicae, did not increase aphid reproduction (Figure 1).

Figure 1.

Aphids reproduce better on TuMV-infected plants.

Number of progeny produced by a single aphid after 9 days on mock-inoculated or virus-infected N. benthamiana. WT = wild type, GFP = green fluorescent protein, TuMV = Turnip mosaic virus, CMV = Cucumber mosaic virus, PVX = Potato virus X. Numbers under the X-axis labels indicate sample sizes. A–DDifferent letters indicate significant differences (P < 0.05) by analysis of variance (anova) and Tukey's HSD post-hoc test.

Association of non-persistent viruses with aphid stylets is transient and virus attachment is thought to be most efficient immediately after the initiation of aphid feeding (Mauck et al., 2012). However, reverse transcription polymerase chain reaction (RT-PCR) to detect TuMV RNA in groups of five aphids (Figure S1a,b) or individual aphids (Figure S1c) from well established colonies on TuMV-infected N. benthamiana showed that they carry the virus with high frequency. Among individual aphids, 100% (10 out of 10) were found to be viruliferous by RT-PCR. Efficient transmission by an established aphid colony also is demonstrated by the fact that 100% of uninfected plants that were placed adjacent to an established M. persicae colony on TuMV-infected N. benthamiana were infected subsequently with TuMV (Figure S2). Given the similar effects of wild type TuMV and TuMV–GFP on M. persicae reproduction (Figure 1), all further experiments to characterize this interaction were conducted with TuMV–GFP. This approach facilitated tracking of the virus infection and made it possible to quantify aphid growth on leaves that were infected, but which did not yet show visible virus infection symptoms.

The number of progeny produced on TuMV-infected Arabidopsis and N. benthamiana was significantly higher than on mock-inoculated plants (Figure 2a,b), and twice as many aphids were in the adult stage on TuMV-infected plants after 6 days (Figure 2c,d). Adult aphids transferred onto N. benthamiana, which is a less suitable host than Arabidopsis, showed a higher survival rate on TuMV-infected plants than on controls (Figure 2e). Among the surviving aphids, weight gain was more rapid on virus-infected plants (Figure 2f). Aphids did not have to complete their entire life cycle on TuMV-infected plants to experience reproductive benefits; adult M. persicae that moved onto infected N. benthamiana produced more progeny over 48 h than those that moved onto mock-infected controls (Figure 2g). Alate aphids, which can move further to new host plants, often are produced in response to crowding (Debarro, 1992). Consistent with a more rapid population growth on TuMV-infected plants, alate aphids were more abundant in colonies on virus-infected N. benthamiana than on control plants (Figure 2h).

Figure 2.

TuMV infection improves aphid performance on Arabidopsis and N. benthamiana.

Effects of TuMV infection on M. persicae biology. Number of progeny produced by a single aphid after 9 days on TuMV-infected or mock-inoculated Arabidopsis (a), and N. benthamiana (b). The percentage of aphids in the adult stage after 6 days of development on TuMV-infected or mock-inoculated Arabidopsis (c), and N. benthamiana (d).

(e) Aphid survival after 1 week on healthy, TuMV-infected, or mock-inoculated N. benthamiana.

(f) Weight of aphids after 4 days on N. benthamiana with and without TuMV infection.

(g) Number of progeny produced by a single aphid after 2 days on TuMV-infected or mock-inoculated N. benthamiana.

(h) Number of alates produced on leaves of mock-inoculated or TuMV-infected N. benthamiana after 3 weeks. *P < 0.05, Student's two-tailed t-test. A,BDifferent letters on bars indicate significant differences, P < 0.05, by analysis of variance (anova) and Tukey's HSD post-hoc test. Numbers in bars indicate sample sizes; mean ± standard error (SE) for all data.

In choice tests, both apterous and alate M. persicae settled on TuMV-infected N. benthamiana in preference to on uninfected controls (Figure 3). A similar preference for TuMV-infected plants in complete darkness shows that visual cues were not involved in host plant choice by apterous aphids (Figure 3). Alate aphids, however, did not differentiate between TuMV-infected and uninfected plants that were kept in complete darkness (Figure 3), suggesting either less efficient orientation in the dark or greater use of visual cues in host plant choice.

Figure 3.

Aphids prefer TuMV-infected plants.

M. persicae arrestment on TuMV-infected N. benthamiana growing under (a) continuous light, or (b) continuous dark conditions. Groups of 10 aphids were allowed to disperse onto pairs of a mock-inoculated and TuMV-infected plants, and were then counted after 24 h (mean ± standard error (SE) of = 24, *P < 0.05, chi-squared test).

To determine whether there is a general suppression of insect defenses in TuMV-infected plants, e.g. by inhibition of jasmonate-mediated defenses, we conducted experiments with two chewing herbivores, the crucifer-feeding specialist Pieris rapae (white cabbage butterfly) and the generalist Trichoplusia ni (cabbage looper). P. rapae weight gain on TuMV-inoculated plants was significantly reduced compared with mock-inoculated controls (Figure 4), this result suggested changes in tissue consumption due to either defense induction or altered phagostimulant content. T. ni growth, on the other hand, was not affected significantly (Figure 4).

Figure 4.

Lepidopteran growth is not improved by TuMV.

Weight of P. rapae and T. ni caterpillars feeding on leaves of mock-inoculated or TuMV-infected plants (mean ± standard error (SE), numbers in bars indicate sample sizes, *P < 0.05, Student's two-tailed t-test).

Because TuMV infection enhances aphid fecundity and development, we investigated whether one or more individual TuMV proteins can elicit this response through transient expression in N. benthamiana (Figure 5a; confirmation of gene expression by RT-PCR is shown in Figure S3). The fecundity of M. persicae was unaffected on plants that transiently expressed six of the 10 tested TuMV genes, was decreased significantly by the expression of HC-Pro, 6K1, and VPg, and was increased significantly by NIa-Pro expression (Figure 5a). Both transient NIa-Pro expression in N. benthamiana (Figure 5b) and expression in stable transgenic Arabidopsis (Figure 5c) increased M. persicae fecundity in a manner similar to that in actual TuMV infection. The settling preference of apterous aphids on virus-infected plants (Figure 3) also was recapitulated on plants that expressed NIa-Pro after 1 and 24 h (Figure 5d). In these experiments, expression of NIa-Pro mRNA in transgenic plants was comparable with levels of viral RNA in TuMV-infected plants (Figure S4). However, this transcript abundance may not reflect NIa-Pro protein abundance, which would be influenced by the different translation modes of the two expression systems and was not quantified in this experiment. To confirm that actual accumulation of NIa-Pro protein can influence aphid reproduction, the experiment was repeated with a FLAG-tagged version of the protein, which also increases aphid fecundity (Figure S5).

Figure 5.

Effects of single TuMV protein expression on M. persicae fecundity.

(a) Relative change in M. persicae fecundity on Agrobacterium-infiltrated N. benthamiana plants that transiently express 10 TuMV proteins from the cauliflower mosaic virus 35S promoter compared with empty vector (EV) control plants (mean ± standard error (SE), P < 0.05, Dunnett's test). Buffer = control plants infiltrated with the MgCl2 infiltration buffer and no Agrobacterium.

(b) Number of progeny produced by aphids on N. benthamiana expressing NIa-Pro, the EV control, TuMV-infected plants and mock-inoculated controls (mean ± SE, A–Cletters represent significant differences of analysis of variance (anova) and Tukey's HSD post-hoc test).

(c) Number of progeny produced by aphids on Arabidopsis expressing NIa-Pro, the EV control, TuMV-infected plants and mock-inoculated controls (mean ± SE, A–Cletters represent significant differences by anova and LSD post-hoc tests).

(d) Aphid settling preference after 1 and 24 h for leaves of N. benthamiana expressing NIa-Pro or the EV control (mean ± SE, *P < 0.05, chi-squared test) under dark conditions. Numbers in bars represent sample sizes.

Two likely virus-induced changes in the host plants that could promote aphid growth are improved nutritional content and reduced defenses. As aphids primarily take up nitrogen in the form of free amino acids, we examined the free amino acid content of mock-inoculated and TuMV-infected leaves of Arabidopsis and N. benthamiana. Free amino acid content was significantly higher in virus-infected leaves than in mock-inoculated controls (Figure 6a). Although most amino acids increased in virus-infected leaves, changes in specific amino acids differed quantitatively between host plants (Figure S5), and may reflect underlying differences in the phloem transport of amino acids in these two species. Free amino acid content also was increased in NIa-Pro-expressing leaves (Figures 6b and S6), suggesting that such changes in the NIa-Pro-expressing plants could improve aphid performance.

Figure 6.

TuMV and NIa-Pro effect on amino acid content.

(a) Free amino acid changes in leaves of mock-inoculated or TuMV-infected plants (mean ± standard error (SE) of = 6).

(b) Free amino acid changes in leaves of plants that expressed NIa-Pro or the empty vector (EV), (mean ± SE of = 6) *P < 0.05, Student's two-tailed t-test.

To test the alternate hypothesis, that TuMV infection compromises plant defenses, we measured two well characterized plant responses to aphid feeding: (1) expression of CYP81F2, which encodes a cytochrome P450 that contributes to 4-methoxyindol-3-yl-methylglucosinolate synthesis and aphid defense in Arabidopsis and other crucifers (Kim and Jander, 2007; Clay et al., 2009; De Vos and Jander, 2009; Pfalz et al., 2009); and (2) callose accumulation, which is observed as a response to aphid feeding in many plant species (Dreyer and Campbell, 1987; Walling, 2000; Botha and Matsiliza, 2004; Meihls et al., 2013). Consistent with a suppression of aphid-specific plant defense responses by TuMV, aphid-induced CYP81F2 expression was reduced significantly by TuMV infection (Figure 7a). We also measured callose accumulation in TuMV-infected and mock-inoculated plants, with and without M. persicae infestation. Aphid feeding increased callose deposition in both Arabidopsis (Figure 8a,b) and N. benthamiana (Figure 8c). However, this induction was reduced significantly in TuMV-infected plants (Figure 8a–c). CYP81F2 expression and callose deposition also was measured in NIa-Pro-expressing Arabidopsis leaves. NIa-Pro expression reduced aphid-induced callose deposition by more the 50% compared with control Arabidopsis (Figure 8a) and was consistent with what was observed in experiments with the whole virus. However, aphid-induced expression of CYP81F2 was not inhibited in plants that expressed NIa-Pro constitutively in Arabidopsis (Figure 7b).

Figure 7.

CYP81F2 expression is involved in virus enhanced aphid fecundity. Relative CYP81F2 expression was measured by quantitative RT-PCR in (a) mock-inoculated plants and TuMV-inoculated, with or without aphid feeding for 48 h and (b) in leaves of plants expressing NIa-Pro or the empty vector with and without M. persicae feeding for 24 h. Mean ± standard error (SE) of = 4. A–CDifferent letters indicate significant differences by analysis of variance (anova) and Tukey's HSD post-hoc test, arbitrary units, with expression in mock-inoculated plants set to 1.

Figure 8.

TuMV and NIa-Pro reduce callose accumulation.

(a) Callose deposition in Arabidopsis plants that were mock-inoculated, TuMV-infected, expressing NIa-Pro, or transformed with the empty vector (EV), with and without M. persicae infestation (mean ± standard error (SE) of = 4–6, A–Ddifferent letters indicate significant differences by analysis of variance (anova) and Tukey HSD post-hoc test).

(b) Images of Arabidopsis callose deposition, detected by aniline blue staining.

(c) Callose deposition in mock-inoculated and TuMV-infected N. benthamiana, with and without M. persicae infestation (mean ± SE of = 4–5, A,Bdifferent letters indicate significant differences by anova and Tukey's HSD post-hoc test).


Our results suggest that the NIa-Pro protein contributes to the improved M. persicae growth and the increased settling that is observed on TuMV-infected plants (Figure 5). In addition to being the main protease that cleaves the TuMV polyprotein, NIa-Pro probably has other functions, including non-specific DNase activity (Anindya and Savithri, 2004; Rajamäki and Valkonen, 2009). During potyvirus infection of plant cells, NIa-Pro can be transported to the nucleus as a fusion with VPg, which is adjacent to NIa-Pro in the viral polyprotein and has a nuclear localization signal (Schaad et al., 1996). Thus it has been hypothesized that, once inside the nucleus, NIa-Pro and/or VPg interferes with host defenses (Riechmann et al., 1992; Anindya and Savithri, 2004; Beauchemin et al., 2007; Rajamäki and Valkonen, 2009). However, as we observed an increase in aphid performance on plants that expressed NIa-Pro alone (Figure 5), it is likely that transport to the nucleus is not required for suppression of plant defenses against aphids.

Aphids and other phloem-feeding insects obtain nitrogen primarily in the form of free amino acids, which constitute only about 2% of the plant nitrogen content (Lam et al., 2003; Lemaitre et al., 2008). As phloem free amino acid content closely tracks cytosolic free amino acid content (Riens et al., 1991; Winter et al., 1992; Lohaus et al., 1994), the measurement of amino acids in whole leaves is a good proxy for the amino acids that the aphids are able to acquire from the phloem. In response to both TuMV infection and NIa-Pro expression, there were significant increases in foliar free amino acid content (Figure 6). Although specific changes in individual amino acids differed among host plants and between TuMV and NIa-Pro expression (Figure S6), the cumulative effect of increased amino acids may be more important to aphids. Buchnera aphidicola, obligate bacterial endosymbionts of aphids, can synthesize essential amino acids that are deficient in the aphid diet (Hansen and Moran, 2011; Macdonald et al., 2012). Although increased phloem amino acid content can improve the nutrient content of the plant, there must be additional factors that limit aphid growth. For instance, increased uptake of amino acids by M. persicae on the Arabidopsis ant1 amino acid transporter mutant did not increase aphid reproduction and suggests that available nitrogen did not limit aphid growth in this experiment (Hunt et al., 2006).

Virus-mediated changes in plant defense against aphids are also important for the ecology of this interaction. Callose deposition in response to aphid feeding was reduced by both TuMV infection and transgenic NIa-Pro expression (Figure 8). However, although aphid induction of CYP81F2 mRNA was reduced in TuMV-infected plants compared with wild type, this effect was not observed in Arabidopsis plants that stably expressed NIa-Pro (Figure 7). Thus, although we cannot rule out limitations of the experimental design, it is possible that NIa-Pro contributes to some, but not all, of the plant defense suppression that is mediated by TuMV.

Despite a decrease in plant defenses and an increase in free amino acid content, P. rapae grows less well on TuMV-infected Arabidopsis. However, P. rapae is a crucifer-feeding specialist that has co-opted glucosinolates and perhaps other plant defenses as attractive signals, and caterpillars grow less well on Arabidopsis plants in which glucosinolate-mediated defenses are inhibited (Barth and Jander, 2006). Thus, if crucifer-specific defenses are reduced in TuMV-infected plants, P. rapae might grow less well. However, it should be noted that CYP81F2, which catalyzes the conversion of indole-3-yl-methylglucosinolate to 4-hydroxy-indole-3-yl-methylglucosinolate to provide defense against M. persicae (Kim and Jander, 2007; Pfalz et al., 2011), does not affect resistance against lepidopteran herbivores (Pfalz et al., 2009). Free amino acids, which constitute only a small portion of the total nitrogen in plant leaves (Lam et al., 2003; Lemaitre et al., 2008), are likely to be less important for caterpillars that consume the entire leaf than for aphids that feed specifically from the phloem.

Most viruses express proteins that inhibit one or more components of RNA silencing, a common anti-viral defense mechanisms in plants and other eukaryotes (Csorba et al., 2009; Fraile and García-Arenal, 2010). Previous studies with CMV, a non-persistently transmitted plant virus, showed that infection increased M. persicae survival on tobacco (Lewsey et al., 2010). Whereas the CMV 2b silencing suppressor protein was implicated in enhancing aphid performance (Ziebell et al., 2011), TuMV HC-Pro reduced aphid reproduction (Figure 5). These differing effects could be the result of functional differences between the two proteins. Whereas CMV 2b has a dual mode of silencing inhibition, sequestering siRNAs and interacting with AGO1 to prevent RISC (RNA-induced silencing complex) assembly, HC-Pro only prevents RISC assembly (Roth et al., 2004; Csorba et al., 2009; Burgyan and Havelda, 2011). Our observation of increased aphid resistance in plants that express TuMV HC-Pro (Figure 5) is consistent with a previous report showing that plants that express Tobacco mosaic virus HC-Pro are significantly more resistant to multiple unrelated pathogens (Pruss et al., 2004).

TuMV is somewhat unusual among non-persistently transmitted viruses, which generally make plants less suitable as feeding sites for their aphid vectors (reviewed by Mauck et al., 2012). Aphids that land on a plant rapidly acquire non-persistently transmitted viruses during their initial probing to reach the phloem (Hoh et al., 2010; Martiniere et al., 2013) and, if the feeding site is not suitable, move elsewhere and transmit the viruses to a new host. Our observation that M. persicae from well established colonies on TuMV-infected plants are viruliferous (Figure S1) and transmit viruses to neighboring plants (Figure S2) suggests that TuMV, unlike many other non-persistently transmitted viruses (Sisterson, 2008; Mauck et al., 2012), is acquired efficiently not only during initial aphid probing to reach the phloem but also during stable feeding. This difference in virus uptake may make it advantageous for TuMV, to promote long-term feeding on host plants and thereby a more rapid increase in the aphid population. Future research will determine whether other potyviruses, e.g. Potato virus Y (Kersch-Becker and Thaler, 2013), follow a similar strategy in promoting aphid growth by means of the NIa-Pro protein.

M. persicae behavior on TuMV-infected plants may influence the dynamics of virus transmission in the field. It was recently demonstrated that altered volatile profiles can attract aphids to CMV-infected plants (Mauck et al., 2010). Similarly, our results show that M. persicae prefer to settle on both TuMV-infected and NIa-Pro-producing plants (Figures 3 and 5d). The absence of host plant preference by alate aphids in the dark (Figure 3) may indicate greater use of visual cues by alate than by apterous M. persicae. Because TuMV-infected plants are more suitable for aphid reproduction (Figure 1), more aphids will be available for virus transmission, including larger numbers of alates (Figure 2h) that facilitate more widespread dispersal (Debarro, 1992). Thus, by promoting aphid settling and reproduction, TuMV probably also benefits itself by making more aphids available for subsequent virus transmission to new host plants.

Together, our results provide insight into the molecular mechanisms that underlie an important ecological interaction. By demonstrating increased M. persicae growth on TuMV-infected plants and showing that this phenotype can be recapitulated by expression of a single virus protein, NIa-Pro, we provide evidence that TuMV infection alters plant defense against aphids. As potyviruses are among the most widely distributed plant viruses, and plants in natural settings are frequently infected, virus-mediated suppression of plant defenses may play an important role in plant–aphid interactions.

Experimental Procedures

Plants and growth conditions

N. benthamiana seeds were obtained from Peter Moffett (Université de Sherbrooke, Quebec, Canada). Arabidopsis seeds were obtained from the Arabidopsis Biological Resource Center ( All plants were grown in Cornell mix [by weight 56% peat moss, 35% vermiculite, 4% lime, 4% Osmocoat slow-release fertilizer (Scotts, Marysville, OH, USA,, and 1% Unimix (Scotts)] in 20 × 40-cm nursery flats in Conviron growth chambers. The light intensity in the growth chambers was 200 mmol m−2 sec−1 at 23°C with 50% relative humidity and a 16 h:8 h day:night cycle. The same growth conditions were used in all subsequent experiments. Plants used for all experiments were 1 month old, unless otherwise noted. All experiments were conducted at least two times, with varying numbers of biological replicates per treatment per experiment.

TuMV infection

TuMV–GFP was propagated from infectious clone p35TuMVGFP (Lellis et al., 2002). To prepare inoculum, TuMV–GFP-infected N. benthamiana leaves were weighed, ground in four volumes of 20 mm sodium phosphate buffer (pH 7.2), filtered through Organza Mesh cloth, and frozen in aliquots at −80°C. For inoculations, one leaf from each plant was dusted with carborundum (Sigma, St. Louis, MO, USA, and rub-inoculated with TuMV–GFP leaf extract using a cotton-stick applicator. A corresponding set of control plants was dusted with carborundum and mock-inoculated with a cotton-stick applicator that was soaked in uninfected N. benthamiana leaf extract, prepared in the same manner as the virus-infected extract. One week after inoculation, an ultraviolet (UV) light lamp (UV Products, Upland, CA, Blak Ray model B 100AP) was used to identify fully infected leaves that contained the TuMV–GFP virus. All plants were 2–3 weeks old at the time of infection. Inoculations with other viruses (wild type TuMV, CMV, and PVX) were performed with the same methods.


Aphid experiments were conducted with a tobacco-adapted red strain of M. persicae that was obtained from Stewart Gray (USDA Biological Integrated Pest Management Research Laboratory, Ithaca, NY, USA, Aphids were reared on tobacco (Nicotiana tabacum) with a 16 h/8 h photoperiod at 24°C /19°C (150 mmol m−2 sec−1; 50% relative humidity). P. rapae were from a colony maintained by the Jander laboratory, which is descended from 20 insects collected in the wild on the Cornell University campus in July, 2008. Trichoplusia ni eggs were purchased from Benzon Research (Carlisle, PA, USA,

Virus transmission studies

Two groups of five M. persicae individuals (apterous and alates, separately) were fed on N. benthamiana N. benthamiana infected with wild type TuMV for 1 week. RNA was extracted from pooled aphid tissue and cDNA was synthesized. The presence of TuMV was verified by RT-PCR with primers for the NIa-Pro gene. In a separate experiment, wild-type TuMV-infected N. benthamiana plants were each infested with 10 apterous aphids. After 48 h, five aphids per plant were collected and the viruliferous status was verified as above. To assess infection frequency in individual aphids, aphids were fed on N. benthamiana infected with wild-type TuMV and uninfected control plants. After 48 h, 10 single aphids were collected from infected plants and the viruliferous status was verified by RT-PCR as above. Control aphids were collected from N. benthamiana plants that were not infected with TuMV.

To assess transmission to neighboring plants, aphid colonies were established on N. benthamiana that had been infected with wild type TuMV for 2 weeks. Three days later, when the aphids had become established, two plants that were infected with TuMV and infested with aphids, were placed in an arena surrounded by eight healthy plants. Aphids from the viruliferous colonies were allowed to move freely among plants for 2 weeks. At the end of the experiment the apical leaves of the healthy plants were collected for infection verification. RNA was extracted from plant tissue and cDNA synthesized as below. TuMV infection was verified by RT-PCR with primers for the NIa-Pro gene (Table S1).

Aphid fecundity, development and alate bioassays

To assess the effect of TuMV infection on aphid fecundity, one apterous adult aphid was placed in a 1.5 cm diameter clip cage on a fully infected or mock-inoculated N. benthamiana or Arabidopsis leaf. Infection is not uniform in the host plant, thus only fully infected leaves were used in subsequent experiments. After 24 h, all aphids except one nymph were removed. The single nymph was allowed to develop and progeny were counted after 9–12 days (depending on the experiment) to determine fecundity. To measure aphid development rate, cages with aphids were set up as in the fecundity experiments, but five apterous adults were added to each cage. After 24 h, all insects, except 10 nymphs, were removed from each cage. After 6 days, the number of aphids in the adult stage was recorded in each cage. For measuring alate production, a single nymph was allowed to develop and reproduce on mock- and TuMV-inoculated leaves. After 3 weeks, the number of alates in the colony was counted. Each complete experiment was repeated 2–3 times with at least 10 repetitions (cages) per treatment.

Aphid and caterpillar weight gain bioassays

Groups of 10 1-day-old adult aphids were confined on the leaves of 3-week-old TuMV- or mock-inoculated plants in 1.5 cm diameter clip cages. After 4 days, five adults from each plant were collected. For experiments with P. rapae and T. ni, neonate caterpillars were confined on the leaves of 3-week-old TuMV- or mock-inoculated plants using 100 ml volume mesh-covered cups. Caterpillars were allowed to feed on plants for 7 days before harvesting. Aphids and caterpillars were lyophilized and dry weight was determined using a precision balance.

Aphid preference bioassays

For aphid preference experiments, two N. benthamiana plants (a TuMV-infected and a mock-inoculated plant) were placed in a 30 cm diameter experimental arena, with their positions in the arena randomly assigned. In one experiment, a microcentrifuge tube containing 10 apterous aphids was placed at the height of the base of the plants (10 cm) equidistant from the two test plants. After 1 h or 24 h, the number of aphids on each plant was counted. This study was conducted on two consecutive days with 12 replications per day. Another experiment was performed as above, but with 10 alate aphids in each microcentrifuge tube. This complete set of experiments was performed under constant light in one set of experiments and under constant dark in another set of experiments.

Amino acid assays

For analysis of leaf free amino acids, ~100 mg of plant tissue was collected, weighed, placed in 2 ml microcentrifuge tubes with two 3-mm steel beads, and frozen in liquid nitrogen. Tissue was ground to fine powder using a Harbil model 5G-HD paint shaker (Fluid Management, Wheeling, IL, USA, Ground tissue was extracted with 20 mm HCl (10 μl mg−1 of tissue), the extracts were centrifuged (3800 g for 20 min at RT), and the supernatant was saved for analysis.

Amino acids were derivatized using an AccQ-Fluor reagent kit (Waters, Milford, MA, USA, For derivatization, 2.5 μl extracts were mixed with 17.5 μl borate buffer, and the reaction was initiated by the addition of 5 μl 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate reagent, followed by immediate mixing and incubation for 10 min at 55°C. l-Norleucine was used as an internal standard. From each sample, 10 μl were injected onto a Nova-Pak C18 column using a Waters 2695 pump system, and the data were recorded using Waters Empower Software. Amino acid derivatives were detected using a Waters model 2475 fluorescence detector with an excitation wavelength of 250 nm and an emission wavelength of 395 nm. Solvent A, AccQ-Tag Eluent A, was premixed from Waters; Solvent B was acetonitrile:water (60:40). The gradient used was 0–0.01 min, 100% A; 0.01–0.5 min, linear gradient to 3% B; 0.5–12 min, linear gradient to 5% B; 12–15 min, linear gradient to 8% B; 15–45 min, 35% B; 45–49 min, linear gradient to 35% B; 50–60 min, 100% B. The flow rate was 1.0 ml min−1.

Callose staining

Leaves were collected from host plants 24 h after infestation with 25 M. persicae adults. For virus experiments, aphids were added to 4-week-old plants, 7 days after virus inoculation. For transient expression aphids were added 2 days after infiltration. For visualization of callose, leaves were cleared with 95% ethanol overnight and stained with 150 mm K2P04 (pH 9.5), 0.01% aniline blue for 2 h (Koch and Slusarenko, 1990). The leaves were examined for UV fluorescence using a Leica fluorescence stereoscope (365 nm excitation, 396 nm chromatic beam splitter, 420 nm barrier filter) and the number of callose spots was quantified manually.

RNA isolation

Arabidopsis and N. benthamiana leaves infected with TuMV, were harvested, frozen in liquid nitrogen, and ground to a fine powder as above. Total RNA was extracted from frozen tissue samples using the SV Total RNA Isolation system with on-column DNase treatment (Promega, Madison, WI, USA, RNA integrity was verified using a 1.2% formaldehyde agarose gel (Sambrook et al., 1989). After RNA extraction and DNase treatment, 1 μg of total RNA was reverse transcribed with SMART MMLV reverse transcriptase (Clontech, Mountain View, CA, USA, using oligo-dT12–18 as a primer.

Quantitative real-time PCR

Arabidopsis plants were inoculated with TuMV, and 14 adult aphids were added to mock-inoculated and TuMV-infected plants. Insects were removed and tissue was collected into liquid nitrogen. Plant RNA was isolated using an SV Total RNA Isolation kit (Promega). Four-week-old Arabidopsis stably expressing the empty vector pMDC32 or NIa-Pro in pMDC32 were used for transgenic plant experiments. Fourteen adult aphids were added as above and tissue was collected after 24 h after the initiation of aphid feeding.

Transcript abundance of CYP81F2 was analyzed with quantitative real-time RT-PCR (qRT-PCR), with eEF1-α (elongation factor 1-alpha, At5 g60390) as an internal standard. eEF1-α was identified from publicly available microarray data as constitutively expressed after herbivory and stable expression was verified across samples using qRT-PCR (Primers; Table S1). Gene-specific primers used for qRT-PCR were designed using Primer-Blast ( with the following criteria: melting temperature of 60°C, PCR amplicon lengths of 90–150 bp yielding primer sequences with lengths of 18–24 nucleotides with an optimum at 21 nucleotides, and guanine–cytosine contents of 40–60% (Table S1). Reactions were carried out using 5 μl of the SYBR Green PCR master mix (Applied Biosystems, Grand Island, NY, USA,, with 800 nm of primer, in the ABI 7900HT instrument (Applied Biosystems). The PCR was initiated by incubation at 95°C for 10 min to activate the enzyme. Then the following cycle was repeated 40 times: 95°C for 15 sec, 60°C for 15 sec, and 72°C for 15 sec. The CT values were quantified and analyzed according to the standard curve method.

DNA manipulations for genetic constructs

PCR products were amplified by gene-specific primers for each TuMV gene (Table S1) flanked by the attB1 and attB2 universal primers and individually cloned using the Gateway cloning kit (XXXclonF and XXXclonR; Table S1; Invitrogen, Carlsbad, CA, USA,, following the instructions of the manufacturer. First, the PCR products were cloned into pDONR207 vector using BP clonase and then they were re-cloned into pMDC32 destination vector with 35S promoter (Curtis and Grossniklaus, 2003), which harbors the double 35S promoter, using LR clonase (Invitrogen). For westerns, NIa-Pro was re-cloned into pGWB511 destination vector with 35S promoter and a FLAG tag (Nakagawa et al., 2007).

Transient expression of TuMV proteins

Recombinant Agrobacterium tumefaciens strain GV3101 cultures that contained each of the 10 TuMV gene constructs individually (primers, Table S1), were grown overnight at 28°C in Luria-Bertani (LB) medium that contained the appropriate antibiotics (pMDC32, kanamycin; pDONR207, gentamicin). Cells were pelleted at 2800 g, resuspended in infiltration medium (10 mm MgCl2), and incubated for 2 h at room temperature. Resuspended cells were infiltrated into the underside of 4-week-old N. benthamiana leaves at an OD600 = 0.2 with a 1-ml needleless syringe. As a control, the p35S: EV (empty vector) construct was also infiltrated into a separate set of plants for each experiment. The plants were incubated in the growth chambers used above for 24 h, after which cages were added to the underside of infiltrated leaves and fecundity or preference experiments were performed as described above. Because not all experiments could be performed at the same time, fecundity is expressed relative to the EV control that was included in each experiment. At the end of the experiments, leaf tissue was collected from three leaves for each construct and the EV. RNA was isolated and cDNA synthesized as described previously. Using the gene-specific primers from cloning (Table S1), expression of each TuMV gene was verified by PCR, and no-reverse transcriptase controls were used to verify that the PCR template was derived from RNA (Figure S3).

Arabidopsis transformation and aphid bioassays with transgenic plants

The transformation vectors harboring p35S:NIa-Pro or the p35S: EV constructs were introduced into Agrobacterium and transferred into wild-type Arabidopsis plants by floral dip transformation (Clough and Bent, 1998). Positive transgenic lines were screened on kanamycin-containing Murashige and Skoog (MS) agar plates (Murashige and Skoog, 1962) and then confirmed by reverse-transcription PCR. Single leaves of 4-week-old Arabidopsis transformed with p35S: EV or p35S:NIa-Pro were used in fecundity, amino acid and callose experiments as described above.

Statistical analysis

For fecundity, development, alate production, and free amino acid experiments t-tests were used to determine significant differences from controls. Analysis of variance (anova) was used to determine if there were comparisons among multiple samples. Post-hoc multiple comparison analyses were conducted using Least Significant Difference (LSD) or Tukey's post-hoc tests. For preference chi-squared tests were performed. All statistical analyses were conducted using JMP analysis software, version 9 (JMP 9.0; SAS, Cary, NC, USA,


This research was supported by US National Science Foundation award IOS-1121788 to GJ and SAW, United States Department of Agriculture award 2010-65105-20558 to GJ, United States Department of Agriculture award 2013-2013-03265 to CLC, an American Society for Plant Biologists Summer Undergraduate Research Fellowship to HND, Binational Agricultural Research and Development Agency award US-4165-08C to SAW, and the Iowa State University Plant Sciences Institute Virus–Insect Interactions group. We thank Neha Pandya for assistance with aphid bioassays.