Silencing of a Germin-Like Gene in Nicotiana attenuata Improves Performance of Native Herbivores^,

Abstract

Germins and germin-like proteins (GLPs) are known to function in pathogen resistance, but their involvement in defense against insect herbivores is poorly understood. In the native tobacco Nicotiana attenuata, attack from the specialist herbivore Manduca sexta or elicitation by adding larval oral secretions (OS) to wounds up-regulates transcripts of a GLP. To understand the function of this gene, which occurs as a single copy, we cloned the full-length NaGLP and silenced its expression in N. attenuata by expressing a 250-bp fragment in an antisense orientation with an Agrobacterium-based transformation system and by virus-induced gene silencing (VIGS). Homozygous lines harboring a single insert and VIGS plants had significantly reduced constitutive (measured in roots) and elicited NaGLP transcript levels (in leaves). Silencing NaGLP improved M. sexta larval performance and Tupiocoris notatus preference, two native herbivores of N. attenuata. Silencing NaGLP also attenuated the OS-induced hydrogen peroxide (H₂O₂), diterpene glycosides, and trypsin proteinase inhibitor responses, which may explain the observed susceptibility of antisense or VIGS plants to herbivore attack and increased nicotine contents, but did not influence the OS-elicited jasmonate and salicylate bursts, or the release of the volatile organic compounds (limonene, cis-α-bergamotene, and germacrene-A) that function as an indirect defense. This suggests that NaGLP is involved in H₂O₂ production and might also be related to ethylene production and/or perception, which in turn influences the defense responses of N. attenuata via H₂O₂ and ethylene-signaling pathways.

Germins and germin-like proteins (GLPs) were first identified in a search for germination-specific proteins in wheat (Triticum aestivum; Thompson and Lane, 1980; Grzelczak and Lane, 1984) and constitute a large plant gene family. They occur as glycoproteins frequently retained in the extracellular matrix by ionic bonds. Most are very stable oligomers (Bernier and Berna, 2001; Lane, 2002). They are functionally diverse but structurally related to members of the cupin superfamily, which includes isomerases, cyclases, dioxygenases, sugar- or auxin-binding proteins, and monomeric or dimeric globulin seed storage proteins, such as phaseolin (Dunwell et al., 2000; Druka et al., 2002). Hence, germins and GLPs are known to play a wide variety of roles as enzymes, structural proteins, or receptors (Bernier and Berna, 2001). As enzymes, germins have oxalate oxidase activity (Lane et al., 1993; Lane, 2000) and some GLPs have superoxide dismutase (SOD) activity (Bernier and Berna, 2001), both of which can produce hydrogen peroxide (H₂O₂) in plants and therefore may function in defense responses (Lamb and Dixon, 1997). For example, H₂O₂ may reach levels in plant tissues that are toxic to microbes (Peng and Kúc, 1992) and herbivores (Ramputh et al., 2002). H₂O₂ may contribute to the structural reinforcement of plant cell walls (Bolwell et al., 1995); trigger lipid peroxide, salicylic acid (SA; Leôn et al., 1995), and ethylene (van Breusegem et al., 2001) synthesis; and activate program cell death (van Breusegem et al., 2001). Moreover, H₂O₂ plays a central role in signal transduction cascades that coordinate various defense responses, such as the elicitation of hypersensitive responses and the synthesis of pathogenesis-related proteins, phytoalexins, proteinase inhibitors, and polyphenol oxidases (Greenberg et al., 1994; Hammond-Kosack and Jones, 1996; Orozco-Cárdenas et al., 2001).

Various studies have demonstrated that germins and GLPs modulate a plant's responsiveness to abiotic and biotic stresses (Dunwell et al., 2000; Lane, 2002). For example, germins are highly expressed during germination of wheat and barley (Hordeum vulgare) seeds and after pathogen attack on mature leaves (Zhang et al., 1995; Berna and Bernier, 1999). The wheat germin (OXALATE OXIDASE) enzyme expressed in soybeans (Glycine max; Donaldson et al., 2001) and hybrid poplar (Populus euramericana; Liang et al., 2001) is able to degrade oxalic acid and confers resistance to oxalic acid-generating pathogens. In sunflowers (Helianthus annuus), overexpression of wheat germin influences some defense-related transcripts and increases resistance to pathogens (Hu et al., 2003). Transient overexpression and silencing GLPs in barley and wheat have demonstrated that members of GLP subfamily 4 function in resisting pathogens (Schweizer et al., 1999; Christensen et al., 2004). Hence, while the exact biological functions of germins and GLPs remain unclear, it is becoming increasingly clear that they play a role in pathogen resistance (Lane, 1994, 2002). In contrast, much less is known about their role in herbivore resistance. To date the best evidence comes from the heterologous expression of wheat germin in maize (Zea mays), which enhanced resistance to the European corn borer Ostrinia nubilalis (Ramputh et al., 2002).

Here we examine the influence of GLPs on herbivore-induced direct and indirect defenses in Nicotiana attenuata, a post-fire annual native of the Great Basin desert of California, Nevada, Idaho, and Utah (Goodspeed, 1954; Wells, 1959), whose herbivore-induced responses have been intensively studied with ecological, chemical, and molecular approaches (Baldwin, 2001; Kessler and Baldwin, 2002). N. attenuata recognizes feeding by the larvae of its specialist sphingid herbivore, Manduca sexta; such recognition is illustrated by Manduca-induced patterns of hormone signaling (jasmonic acid [JA], ethylene, SA), secondary metabolite accumulation (responsible for both the plant's direct and indirect defenses), and transcript accumulations (Baldwin, 2001; Baldwin et al., 2002; Heidel and Baldwin, 2004). Given that herbivore-induced defense responses have been well characterized and that M. sexta caterpillar feeding up-regulates a germin-like gene (Hermsmeier et al., 2001), N. attenuata represents an ideal system with which to examine the effects of GLPs on the entire herbivore-induced defense response, including signaling pathways, and on the direct and indirect defenses that these pathways elicit.

To elucidate the effects of GLPs on herbivore-induced responses in N. attenuata, we down-regulated the expression of the herbivore-induced GLP gene by expressing a 250-bp fragment from the 3′-untranslated region (UTR) of the gene in an antisense (as) orientation with an Agrobacterium-based transformation system (Krügel et al., 2002) and in a Tobacco rattle virus-based virus-induced gene silencing (VIGS) system that had been optimized for N. attenuata (Saedler and Baldwin, 2004). We elicited leaves from three asGLP lines, one wild-type line, and VIGS (pTVGER)/non-VIGS plants (pTV00) by treating wounds with M. sexta oral secretions (OS; these OS elicit all of the measured herbivore-specific responses in N. attenuata; Halitschke et al., 2001, 2003; Roda et al., 2004) and measured the magnitude of the subsequent JA and SA bursts as well as H₂O₂ production and the accumulation of herbivore-induced defense metabolites in N. attenuata, such as nicotine (Winz and Baldwin, 2001), trypsin proteinase inhibitors (TrypPIs; van Dam et al., 2001; Glawe et al., 2003), volatile organic compounds (VOCs; Halitschke et al., 2000; Lou and Baldwin, 2003), caffeoylputrescine, chlorogenic acid, and diterpene glycosides (DTGs; Keinanen et al., 2001; Lou and Baldwin, 2003). The performance of M. sexta larvae (a leaf-tissue feeder) and the preference of a mirid Tupiocoris notatus (a cell-content feeder) on plants that had silenced NaGLP expression were also determined.

RESULTS

Sequence Analysis of a cDNA Encoding NaGLP

A 773-bp PCR product of NaGLP was obtained using the primer pair NaGLPup and NaGLPlow (see details in “Materials and Methods”). The full-length cDNA of NaGLP was subsequently obtained with this PCR product and combined with the previously reported NaGLP fragment (AW191813). The cDNA (AY436749) is 876 bp with a 5′-UTR of 2 bp, an open reading frame of 663 bp, and a 3′-UTR of 211 bp. The amino acid sequence deduced from the open reading frame revealed that NaGLP encodes a protein of 220 amino acids (Fig. 1) with a calculated M_r of 23,290 and a theoretical pI of 8.99 (Editseq). Alignments of the amino acid sequence of N. attenuata with other sequences available in databases revealed similarities of 84% to potato (Solanum tuberosum) GLP (M.A. Campbell, L.M. Hermann, and J. Colvin, unpublished data), 72% to tomato (Lycopersicon esculentum) GLP (K. Sasaki, K. Amako, and M. Takahashi, unpublished data), and 70% to Nectarin Ia (Carter and Thornburg, 2002) and Nectarin Ib (Carter et al., 1999), which have SOD activity (Carter and Thornburg, 2000). A phylogenetic tree deduced from the sequence alignment of 70 GLPs revealed that NaGLP is a new member of GLP subfamily 2 (Carter and Thornburg, 2000).

Figure 1.

Alignment of deduced amino acid sequences of GLPs from N. attenuata (NaGLP, AY436749), S. tuberosum (StGLP, AF067731), L. esculentum (LeGLP, BAA25197), Nectarin Ia (AAK95664), and Nectarin Ib (AAF03355). Missing amino acids are indicated by dashes. The three conserved regions, boxes A, B, and C, are framed. The KGE sequence, highlighted in black, suggests that GLPs participate in the exchange of information between the outside and the inside of cells.

NaGLP possesses the characteristics common to plant germins and GLPs (Lane et al., 1991; Bernier and Berna, 2001): three highly conserved oligopeptides (boxes A, B, and C; Fig. 1), two Cys residues (Cys-35 and Cys-50) known to form disulfide bonds, and three His residues (His-110, His-112, and His-155) involved in binding a metal ion (Fig. 1). Moreover, two potential N-glycosylation sites (Asn-71 and Asn-72; Jaikaran et al., 1990) and a KGE tripeptide, a potential characteristic with which the proteins may be located in the extracellular matrix and may participate in the exchange of information between the outside and the inside of the cells (Bernier and Berna, 2001), were found in NaGLP. Thus, NaGLP may be an apoplastic glycoprotein.

Southern-blot analysis revealed that the NaGLP gene is present in the N. attenuata genome as a single copy (Supplemental Fig. 3).

Methyl JA and H₂O₂ Elicitations Up-Regulate Expression Levels of NaGLP mRNA

As previously described (Hermsmeier et al., 2001), NaGLP mRNA expression is high in roots, low in leaves, and higher in younger leaves than in older leaves (Fig. 2, left inset and A). Exogenous application of methyl JA (MeJA) and H₂O₂ up-regulated expression of NaGLP mRNA, while wounding and treatment with methyl SA (MeSA) and indole-3-acetic acid (IAA) did not (Fig. 2B).

Figure 2.

Expression of GLP by qPCR (A–D) and RNA gel-blot (left inset) analysis in three N. attenuata asGLP T2 line (A860-19, A871-16, and A911-10) plants and one wild-type, untransformed line (WT) plant elicited by the following treatments. A, Leaves at nodes 0 undergoing the source-sink transition (L0) and +2 (L + 2), and roots (Roots) of untreated wild-type plants. B, Leaf node +2 of wild-type plants, 24 h after leaves at nodes +2 and +3 were treated with 20 μL of pure lanolin (LC) or 20 μL of lanolin containing 150 μg of MeJA (MJ), MeSA (MS), or IAA; wounded (WD); wounded and treated with 40 μL of 20 mm sodium phosphate buffer (pH 6.5; Buf), 12.5 mm Glc in the buffer (Glc), 25 units mL⁻¹ GOX in the buffer (Gox), or GG, which generates H₂O₂ in situ; or left untreated (C). C, Roots of untreated plants of the three asGLP T2 lines and wild-type plants. D, Leaves growing at node +2 of three N. attenuata asGLP T2 lines and wild-type plants, 4.5 h after leaves at node +2 were wounded four times at 30-min intervals or left untreated (C). Puncture wounds were immediately treated with 6 μL of M. sexta OS (OS). Left inset, The northern-blot results for GLP expressions in leaf node +2 (L + 2) and roots (Roots) of untreated wild-type plants. Right inset, diagram of rosette-stage plant. Asterisks indicate significant differences between members of a pair (each treatment versus its corresponding control; asGLP line versus wild type; P < 0.05, Student's t test).

Silencing NaGLP Gene Decreases the Elicited Increases in H₂O₂

To understand whether the function of NaGLP is related to H₂O₂ production, we measured H₂O₂ concentrations before and 15, 30, and 60 min after OS elicitation in wild-type plants (Fig. 3, inset). From this analysis, we learned that H₂O₂ levels are maximally elicited at 30 min. We measured H₂O₂ levels 30 min after OS elicitation in one N. attenuata wild-type line and three asGLP lines (A860-19, A871-16, and A911-10), all of which showed lower constitutive and/or induced NaGLP mRNA levels in roots and leaves than those in the wild-type line (Fig. 2, C and D). All transformed lines contained a single insertion of the transgene (Supplemental Fig. 3) and were normal diploids as determined by flow cytometry (Bubner et al., 2006). The result showed that M. sexta-OS-treated leaves of asGLP line A911-10, which had a significant herbivore phenotype, produced significantly less H₂O₂ than did similarly treated wild-type leaves (Fig. 3), suggesting the possible involvement of NaGLP in H₂O₂ production. This was further confirmed by H₂O₂ analysis of OS-elicited plants in the VIGS experiment. VIGS effectively silenced NaGLP expression (Fig. 4), which decreased elicited H₂O₂ levels in proportion to the levels of NaGLP expression 30 min after elicitation (R² = 0.4238, df = 8, P < 0.05; Fig. 4, insets A and B).

Figure 3.

Mean (+1 se) H₂O₂ concentrations in node +2 leaves of three N. attenuata asGLP T2 lines (A860-19, A871-16, and A911-10) and one wild-type, untransformed line (WT), 30 min after the leaves were wounded and treated with 20 μL of M. sexta OS, the time of maximum H₂O₂ accumulation after OS-elicitation (inset). Asterisks indicate significant differences between members of a pair (each asGLP line versus wild type, P < 0.05, Student's t test).

Figure 4.

VIGS of NaGLP. Mean (+1 se) expression levels of GLP in node +1 leaves of pTVGER and pTV00 plants, by qPCR analysis, 30 min after node +1 leaves were wounded and treated with 20 μL of M. sexta OS (OS) or left untreated (C). Inserts: A, Mean (+1 se) H₂O₂ concentrations in node +2 of pTVGER and pTV00 plants, 30 min after node +2 leaves were wounded and treated with 20 μL of M. sexta OS. B, Relationship of H₂O₂ concentration to expression level of GLP in pTVGER and pTV00 plants. Asterisks indicate significant differences between members of a pair (pTVGER versus pTV00; P < 0.05, Student's t test) or a significant relationship (D, P < 0.05); letters (a, b, c, and d) indicate significant differences among treatments (P < 0.05, Fisher posthoc test).

Silencing the NaGLP Gene Improves Herbivores' Performance/Preference

M. sexta caterpillars gained more mass on asGLP lines A860-19, A871-16, and A911-10 than on wild-type lines (Fig. 5A). By day 6, the masses of caterpillars fed on A860-19, A871-16, and A911-10 were 132.2%, 126.5%, and 159.1% of those feeding on wild type, respectively, although statistically significant increases in larval mass compared to wild type were found only in those that fed on A911-10 on day 6 (Fig. 5A). A similar result was observed in plants whose expression of NaGLP mRNA was silenced by VIGS (Fig. 7, inset A). By day 10, the masses of caterpillars fed on pTVGER plants were significantly (t = −2.428, df = 49, P = 0.0189) higher than those feeding on pTV00 plants (Fig. 7). However, caterpillars gained less mass on pTVGER plants than on (82.26% of those at 10th d) pTVLOX3 plants in which the expression of NaLOX3, a key gene related to herbivore resistance that encodes the specific lipoxygenase, which supplies hydroperoxide substrates for JA biosynthesis in N. attenuata (Halitschke and Baldwin, 2004), was silenced by VIGS (Fig. 7).

Figure 5.

A, Mean (+1 se) mass of 20 replicate M. sexta larvae fed on individual plants from each of the three N. attenuata asGLP T2 lines (A860-19, A871-16, and A911-10) and one wild-type, untransformed line (WT) plants, 2, 4, and 6 d after the larvae hatched and were placed on plants. B, Percentage of T. notatus adult and nymphs per plant on A860-19 versus wild-type, A871-16 versus wild-type, and A911-10 versus wild-type pairs, 4 d after four-replicate plant pairs were exposed to 80 to 100 insects. Asterisks indicate significant differences between members of a pair (each asGLP line versus wild-type lines, P < 0.05, Student's t test).

Figure 7.

Mean (+1 se) mass of 30 replicate M. sexta larvae fed on pTVGER and pTV00 plants, 4, 7, and 10 d after the larvae hatched and were placed on nodes +1 and +2 leaves of 15 plants each with two larvae. Inserts: A, Mean (+1 se) expression levels of GLP in node +1 leaves, by qPCR analysis, 1 d after leaves from nodes +1 and +2 were attacked by one freshly hatched caterpillar. B and C, Mean (+1 se) DTG (B) and TrypPI (C) concentrations in node +2 leaves of pTVGER and pTV00 plants, 4 d after leaves from nodes +1 and +2 leaves were attacked by one freshly hatched caterpillar. Asterisks indicate significant differences between members of a pair (pTVGER versus pTV00; P < 0.05, Student's t test).

When pairs of N. attenuata genotypes differing in NaGLP levels (A860-19 versus wild type; A871-16 versus wild type; A911-10 versus wild type) were exposed to a T. notatus colony for 4 d, a clear preference for the A911-10 line over the wild-type line was observed, although there was no preference between A860-19 and wild type, and between A871-16 and wild type (Fig. 5B).

Silencing the NaGLP Gene Decreases DTGs and TrypPI and Increases Nicotine, But Does Not Influence the JA and SA Bursts or the Release of VOCs

To analyze the effects of NaGLP on defense responses, we measured the JA and SA bursts, TrypPI, nicotine, DTGs, and VOCs in all wild-type and asGLP lines after plants were elicited with M. sexta OS or MeJA. No significant differences were found between wild-type and asGLP lines in JA and SA levels or in the VOCs released, while significant differences in DTG, TrypPI, and nicotine concentrations were observed (Fig. 6; Supplemental Fig. 4). DTG levels were reduced in all asGLP lines, especially in line A911-10 where significantly lower DTG levels compared to the wild-type line were found in treatments involving caterpillar feeding, lanolin, MeJA, and OS (Fig. 6). Antisense suppression of NaGLP also significantly decreased TrypPI levels in water, lanolin, and MeJA treatments (Supplemental Fig. 4D). Unlike DTGs and TrypPI, nicotine concentrations in asGLP lines increased in treatments involving water and M. sexta OS (Supplemental Fig. 4C). The result, antisense suppression of NaGLP decreased DTG and TrypPI levels, was confirmed in the VIGS experiment. Significantly lower concentrations of DTGs (Fig. 7, inset B) and TrypPI (Fig. 7, inset C) were observed in pTVGER plants compared to control (pTV00) plants.

Figure 6.

Mean (+1 se) DTG concentrations in leaves at nodes 0 and +2 of three N. attenuata asGLP T2 lines (A860-19, A871-16, and A911-10) and one wild-type, untransformed line (WT) plants, 4 d after leaves at nodes +2 and +3 were treated with 20 μL of lanolin containing 150 μg of MeJA (MJ) or 20 μL of pure lanolin (LC), wounded and treated with 40 μL of M. sexta OS or 40 μL of deionized water (W), attacked by first-instar M. sexta caterpillars (Cat; two larvae per leaf, only leaf node +2 was harvested), or left untreated (C). Asterisks indicate significant differences between members of a pair (each asGLP line versus wild type, P < 0.05, Student's t test).

DISCUSSION

Here we provide a whole-plant analysis of the role of an endogenously expressed GLP in a plant-herbivore interaction. Plant-herbivore interactions are played out across a range of spatial scales, from the cellular to the whole-plant level, and involve the elicitation of both direct and indirect defenses. Our analysis of the N. attenuata GLP spans this range of spatial scales and includes analysis both of herbivore performance and preference, the signals (SA, JA, and H₂O₂) known to mediate transcriptional responses and downstream changes in secondary metabolites that, in turn, mediate herbivore resistance in wild-type plants and of three independent lines of antisense transformed plants and VIGS plants. This holistic analysis of a GLP's role in plant-herbivore interactions suggested that NaGLP is involved in H₂O₂ production and can influence the defense responses of N. attenuata to two of its adapted insect herbivores via the H₂O₂-signaling pathway. Each of these tentative conclusions deserves further discussion in the context of the growing literature on GLP function.

A 3′-UTR fragment of NaGLP was first discovered using reverse transcription-PCR differential display and found to be up-regulated in N. attenuata plants that were attacked by its adapted herbivore, M. sexta (Hermsmeier et al., 2001). We express this fragment in an antisense orientation to down-regulate expression of the endogenous gene. We report the full-length sequence of NaGLP and find it occurs as a single copy in the N. attenuata genome (Supplemental Fig. 3) that is expressed in particular tissues and regulated by stress factors, as is characteristic of GLPs from other species (Bernier and Berna, 2001; Lane, 2002). NaGLP levels are expressed strongly in roots and weakly in leaves; this expression pattern is similar to that of barley germin but different from that of wheat germin (Grzelczak et al., 1985; Hurkman and Tanaka, 1996), and up-regulated by M. sexta-OS (Figs. 2D and 4), MeJA, and H₂O₂ treatments but not by wounding, IAA, or MeSA (Fig. 2B). That wounding did not elicit the expression of NaGLP might be due to a rapidly waning response, which is sustained (Fig. 2D) or transient (Fig. 4) when elicitors from M. sexta OS are added to the wounds. Transcripts were measured 24 h after elicitation and the wound-elicited H₂O₂ appears to wane by 3 h (data not shown). These expression patterns differ slightly from those of other plants' germins and GLPs. For example, wheat germin gf-2.8 is up-regulated by auxin, fungal infection, and wounding but not by MeJA, SA, or H₂O₂ (Lane et al., 1991; Berna and Bernier, 1999). Barley germin is developmentally regulated by salt stress and by treatments with SA, MeSA, MeJA, abscisic acid, and IAA (Hurkman and Tanaka, 1996). MeJA and wounding elicited increases in Atriplex lentiformis GLP (AlGLP) mRNA, but IAA, SA, and H₂O₂ treatment did not (Tabuchi et al., 2003). It has recently been shown that H₂O₂ induced expression of germin-like genes in Arabidopsis (Arabidopsis thaliana; Desikan et al., 2001) and in N. attenuata (Halitschke et al., 2003). These results are consistent with the hypothesis that each GLP gene is expressed in particular tissues, perhaps only in a small subset of cells (Bernier and Berna, 2001), and that wounding, jasmonates, and H₂O₂ are consistent elicitors of increased expression.

Sequence analysis confirmed that NaGLP belongs to the cupin superfamily of proteins and is a new member of GLP subfamily 2 (Carter and Thornburg, 2000). Several GLPs, for example, Nectarin I (Carter and Thornburg, 2000) and moss Barbula unguiculata GLP (Yamahara et al., 1999), are known to function as SODs, which catalyze the dismutation of superoxide anion to yield H₂O₂. Our measures of endogenous H₂O₂ levels revealed that silencing NaGLP by antisense (in one line, A911-10) or many individually silenced plants by VIGS reduced H₂O₂ levels, and a positive relationship between the transcript level of NaGLP and H₂O₂ concentration was found (Figs. 3 and 4, insets A and B), suggesting that NaGLP is involved in the production of H₂O₂. However, the biochemical role of NaGLP as a SOD needs to be confirmed by future heterologous expression studies.

Compared to control plants, plants with silenced NaGLP had lower TrypPI levels but higher nicotine concentrations (Figs. 6 and 7; Supplemental Fig. 4). It has been reported that M. sexta-OS treatment did not elicit higher nicotine levels in wild-type plants compared to water treatment of puncture wounds because of an ethylene burst (Kahl et al., 2000). This burst inhibits the expression of putrescine N-methyltransferase genes that catalyze the N-methylation of putrescine in the first committed, and likely regulatory, step of nicotine biosynthesis (Winz and Baldwin, 2001). This lack of increased nicotine production occurs despite the fact that OS treatment elicits a dramatic JA burst (Lou and Baldwin, 2003). In contrast, ethylene is known to synergize the JA-elicited increase in TrypPIs (for review, see Koiwa et al., 1997). Hence, silencing NaGLP might decrease ethylene production and/or perception in plants in response to OS elicitation. Clearly additional work to fully characterize ethylene production and perception in NaGLP-silenced lines will be required before this putative cross-talk can be placed on firm experimental footing. Such an effort would likely identify other elicited secondary metabolites that are also influenced by such signaling cross-talk. For example, the similarity in responses in DTGs (Fig. 7), which are elicited by MeJA (Lou and Baldwin, 2003), suggests that ethylene signaling might also synergize DTG elicitation in wild-type plants.

Silencing NaGLP clearly increased the performance of M. sexta larvae (Figs. 5A and 7), a leaf chewer, and the preference of T. notatus (Fig. 5B), a cell-content feeder. The lower H₂O₂ (Fig. 3), DTG (Figs. 6 and 7, inset B), and TrypPI levels (Fig. 7C; Supplemental Fig. 4D) could account for these differences. Importantly, the differences among the antisense lines in their relative expression of elicited NaGLP transcripts correlated with the differences in herbivore performance and relevant secondary metabolite levels (i.e. the changes of these parameters in A860-19 and A911-10 were much stronger than those in A871-16). Moreover, the concentrations of DTGs and TrypPI in pTVGER plants were significantly lower than those in pTV00 plants, which is consistent with the mass gain of larvae feeding on these two groups of plants (Fig. 7). Antisense expression and the VIGS experiments established that the TrypPIs of N. attenuata are powerful herbivore-resistance traits (Glawe et al., 2003; Zavala et al., 2004), and DTGs are a group of secondary metabolites in N. attenuata strongly correlated with M. sexta larval performance (Lou and Baldwin, 2003). Since H₂O₂ is directly toxic to microbes (Peng and Kúc, 1992) and herbivores (Ramputh et al., 2002), the improved performance/preference for herbivores in the antisense lines may also be due to decreases in H₂O₂ pools.

In summary, we suggest that the role of NaGLP in the herbivore-induced defense response might be as follows: Attack by herbivores, for instance, M. sexta, results in a JA burst in N. attenuata plants. This burst subsequently up-regulates the expression levels of NaGLP and thus enhances H₂O₂ levels, and might influence ethylene production and/or perception. Through H₂O₂- and/or ethylene-signaling pathways, secondary metabolites, such as DTGs and TrypPI, accumulate and finally defend against the herbivore attack. Given that in N. attenuata the signaling pathways of JA, SA, H₂O₂, and ethylene are all activated by M. sexta attack, the discovery of lines with silenced NaGLP is exciting. These lines provide a means of endogenously adjusting H₂O₂ signaling during herbivore attack, which represents a promising avenue for future research into the role of cross-talk among herbivore-induced signaling pathways in shaping plants' defense response.

MATERIALS AND METHODS

General Plant Growth

An inbred genotype of Nicotiana attenuata Torr. Ex Wats. (synonymous with Nicotiana torreyana Nelson and Macbr.; Solanaceae), originally collected from southwestern Utah in 1988, and its transformed lines were used for all experiments. Seeds were sterilized and germinated on agar after soaking with a 1:50 (w/v) dilution of liquid smoke (House of Herbs). Ten-day-old seedlings were planted into soil in Teku pots (Waalwijk) and, once established, transferred to 1-L pots in soil and grown in the glasshouse at 26°C to 28°C, under 16 h light supplemented by Philips Sun-T Agro 400 or 600 W Na lights.

We grew plants in hydroponic culture to analyze GLP mRNA expression in leaves and roots. Seedlings established in Teku pots were transferred to 28-L communal hydroponic boxes with a nutrient solution consisting of 0.292 g/L Peter's Hydrosol (W.R. Grace) and 0.193 g/L Ca(NO₃)₂. After an adaptation period of 5 d, seedlings were transferred to individual 1-L hydroponic chambers containing a no-nitrogen hydroponic solution (Baldwin et al., 1994) with 2 mL of a 1 m KNO3. Plants in the same rosette stage were used in all experiments.

Agrobacterium-Based Transformation and Its Vector Construction

A fragment of the N. attenuata gene for GLP (NaGLP; AW191813; Hermsmeier et al., 2001) was PCR amplified. After digestion with XhoI and BstEII, the 250-bp resulting fragment (see the sequence in Supplemental Fig. 2) was cloned in pRESC20 (see details for the vector construct in Supplemental Fig. 1; also see Zavala et al., 2004), yielding the transformation vector pRESC2GER (10.0 kb) that contained the hygromycin resistance gene hph as a selectable marker. On the T-DNA of the resulting binary plant transformation vector, the NaGLP gene fragment was present in an antisense orientation downstream from 35S promoter and upstream of terminators both from Cauliflower mosaic virus, thus enabling the transcription of NaGLP antisense RNA. This vector was used for the transformation of N. attenuata via an Agrobacterium-mediated transformation procedure (Krügel et al., 2002). To determine the segregation ratios, T1 seeds from transformed plants were sterilized and germinated on hygromycin-containing media. Positive (wild-type seeds on media without antibiotics) and negative (wild-type seeds on media with antibiotics) controls were included. Seedlings that exhibited vigorous growth within 8 to 10 d were tentatively considered to be transformed, and planted in soil and grown in a glasshouse. Nine lines that showed 3:1 segregation ratios in the T1 were selected; these were bred to obtain homozygous lines and screened for GLP mRNA levels by reverse transcription-quantitative PCR (qPCR) analysis (see below). Homozygous lines of T2 seedlings showed 100% resistance to antibiotic media and three homozygous T2 lines, A860-19, A871-16, and A911-10, were chosen for further characterization. Lines A860-19 and A911-10 both showed significantly lower constitutive and OS-inducible GLP mRNA levels than did wild-type plants. Line A871-16 had lower constitutive levels of NaGLP expression but showed a similar level as wild type after OS elicitation (see Fig. 2, C and D).

Plant Treatments

MeJA, MeSA, and IAA Treatments

Plants were treated with 75 μg of MeJA, MeSA, or IAA in 10 μL of lanolin paste per leaf applied to two leaves at nodes +2 and +3 (see Fig. 2, inset, for depiction of leaf nodes). Controls (lanolin) were similarly treated with 20 μL of pure lanolin.

Manduca sexta-OS Treatments

One (at node +2) or two leaves (at nodes +2 and +3) per plant (see details in related experiments) were damaged by rolling a fabric pattern wheel over the leaf surface to create six rows of standardized puncture wounds on each leaf, and 20 μL of OS (diluted 1:5 [v/v] with deionized water) from fourth- to fifth-instar larvae were added to the puncture wounds on each leaf. Controls (water) were wounded and each leaf treated with 20 μL of deionized water. Different volumes of OS were used in other experiments (see below), but previous work (Roda et al., 2004) has found that this variation did not significantly influence the elicited responses.

Mechanical Wounding Treatments

For mechanical damage treatment, two leaves (at nodes +2 and +3) per plant were damaged by rolling a fabric pattern wheel over the leaf surface to create six rows of standardized mechanical wounds for each leaf. Controls (controls) were nonmanipulated plants.

H₂O₂ Treatments

To generate H₂O₂ in situ, Glc oxidase (GOX) and Glc solution were injected into nonwounded leaves (Orozco-Cárdenas et al., 2001). Glc (25 mm) and GOX (from Aspergillus niger; 50 units mL⁻¹) in sodium phosphate buffer (20 mm, pH 6.5) were mixed and then immediately introduced into the leaves at nodes +2 and +3 by pressing a 1-mL syringe onto the leaf surface and twice injecting 200 μL for each leaf. Leaves of control plants at nodes +2 and +3 each received 2× 200 μL injections of sodium phosphate buffer, Glc (12.5 mm), or GOX (25 units mL⁻¹).

Total RNA Extraction and Isolation of the cDNA of NaGLP

Pooled leaf samples were ground under liquid nitrogen and total RNA was extracted with TRI Reagent (Sigma) according to the manufacturer's instructions. Total RNA from MeJA-elicited leaf samples was reverse transcribed into cDNA using SuperScript II RNaseH− Reverse Transcriptase (Invitrogen) according to the manufacturer's instructions. The primer pair NaGLPup (5′-CCATGTCTGCCTTTGGAA-3′) and NaGLPlow (5′-CTCCTTCCTATTTGAACCC-3′) was designed for PCR based on part of the sequence of NaGLP (AW191813) and the sequence of Solanum tuberosum GLP (AF067731), both of which are highly homologous (Hermsmeier et al., 2001). The reaction mixture contained 2 μL of total cDNAs as template, 200 μm of each dNTP, 0.2 μm of each primer, 1 μL of Titanum Taq DNA Polymerase (CLONTECH) with buffer and magnesium supplied by the manufacturer in a total volume of 50 μL. The reactions were carried out in an Eppendorf Mastercycler gradient programmed as follows: 94°C for 2 min, 35 cycles of 94°C for 30 s, 55°C for 30 s, 72°C for 45 s, and a final step at 72°C for 5 min. The 773-bp product obtained was cloned in a pGEM-T Easy vector (Promega) and sequenced. The full-length cDNA of NaGLP was obtained from this PCR product combined with the previously reported part sequence of NaGLP (AW191813; Hermsmeier et al., 2001).

RNA Gel-Blot Analysis

RNA samples (20 μg) were size fractionated by 1.2% (w/v) agarose formaldehyde gel electrophoresis and capillary blotted onto a nylon membrane (GeneScreenPlus; NEN-DuPont) as described in the manufacturer's instructions. Ethidium bromide staining of the gel prior to blotting revealed rRNA bands, which served as the loading control. After blotting and UV cross-linking, ³²P-labeled probes specific for NaGLP were used for detection. The probe for NaGLP was obtained by PCR of a fragment of NaGLP cDNA, the same sequence as the transgenic insertion (see Supplemental Fig. 2), with random primers following the protocol of rediprime II kit (Amersham Pharmacia Biotech). Hybridization conditions were as follows: Amersham hybridization buffer (Rapid-hyb buffer, Amersham Biosciences); prehybridization at 65°C for 1 h; hybridization at 65°C for 2 h; and washing with 2× SSC and 0.1% SDS (65°C) for two times each with 10 min, followed by 0.1 SSC and 0.1% SDS (65°C) for two times each with 10 min.

Isolation and Blotting of Genomic DNA

Plant genomic DNA was prepared from leaves of N. attenuata using cetyltrimethylammonium bromide (Reichhardt and Rogers, 1994) and was quantified spectrophotometrically at A_{260 nm}. Twenty micrograms of DNA were digested with EcoRI, EcoRV, DraI, or HindIII; electrophoresed on 1% agarose gel (in 1× Tris-acetate EDTA buffer); blotted on a nylon membrane (GeneScreenPlus; NEN-DuPont) with a high-salt buffer (Brown, 1995); and hybridized with a radiolabeled probe specific for GLP (with the same sequence as was used in the RNA gel blot) generated by PCR using the same method as in RNA gel-blot analysis. Hybridization conditions were the same as in the RNA gel-blot analysis.

qPCR

The cDNA template used per well was reverse transcribed from 15 ng total RNA; each sample was replicated three times. The primers and probe used for NaGLP mRNA detection by qPCR are shown in Supplemental Figure 2. The assay using a double dye-labeled probe was performed on an ABI PRISM 7700 sequence detection system (qPCR Core Kit, ; Eurogentec) with 18S RNA for normalization (TaqMan Ribosomal RNA Control Reagents, Applied Biosystems), and the manufacturer's instructions with the following cycler conditions were used: 10 min 95°C; 40 cycles of 30 s 95°C, 30 s 60°C. The relative expression of the target genes was determined using standard curves (Applied Biosystems, 1997).

NaGLP mRNA Expression Analysis

Northern blot and qPCR (Taqman) were used to detect NaGLP mRNA expression in leaves and roots of nonmanipulated hydroponically grown wild-type plants. As previously reported by Hermsmeier et al. (2001), both methods revealed that the NaGLP mRNA expression was high in roots and low in leaves (Fig. 2, left inset and A). Since northern-blot analysis was not sufficiently sensitive to detect changes in GLP mRNA expression in leaves, we used qPCR to analyze the NaGLP mRNA expressions in the remaining experiments.

To analyze the effects of different stresses on expression levels of NaGLP mRNA, wild-type plants were randomly assigned to 10 treatments, each with three replicates: lanolin, MeJA, MeSA, IAA, sodium phosphate buffer, GOX in buffer (Gox), Glc (Glc), Glc plus GOX (GG), wounding, and no manipulation (C). Leaves at node +2 were harvested (roots for GG treatment were also harvested) 24 h after the leaves at nodes +2 and +3 were treated. Total RNA was extracted for all samples, after which NaGLP mRNA expression levels were detected by qPCR. The relative expression of NaGLP mRNA was compared with that in roots of GG-treated plants.

To examine the efficiency of antisense suppression, NaGLP mRNA expression levels in roots of the three asGLP line (A860-19, A871-16, and A911-10) plants and one wild-type line plant (untreated) were measured by qPCR using the level in roots of wild-type plants as a standard. Leaves at node +2 of these lines' plants were harvested, 6 h after the leaves were wounded four times at 30-min intervals, each time with two rows of wounds, and then either treated with 6 μL of M. sexta OS or left untreated. The NaGLP mRNA expression levels in these leaves were also detected by qPCR.

Herbivory Experiment

Freshly hatched M. sexta L. (Lepidoptera: Sphingidae) larvae (eggs from North Carolina State University Insectary, Raleigh, NC) were placed individually on the node +2 leaf of each plant of the three asGLP lines and one wild-type line. Twenty replicate plants from each line (the three asGLP lines and one wild-type line) were used. Larval mass was measured (to 0.1 mg) 2, 4, and 6 d after the start of the experiment.

To determine the colonization preference of Tupiocoris notatus (Hemiptera: Miridae), two potted plants (asGLP line plant and the other, a wild-type plant) were placed in a plastic box into which 80 to 100 mirids were introduced. We monitored the accumulation of the insects on the plants 4 d after exposure to the insect colony. The experiment was replicated four times for each line.

Quantification of H₂O₂

To identify the optimum sampling time for comparison of H₂O₂ concentrations among lines, we first measured the dynamics of H₂O₂ concentrations in wild-type plants after OS elicitation. Leaves at node +2 from wild-type plants were wounded and treated with 20 μL of M. sexta OS, and harvested 0, 15, 30, and 60 min after the treatment. Since H₂O₂ levels are maximally elicited at 30 min (Fig. 3, inset), we compared H₂O₂ levels in the three asGLP lines and one wild-type line at 30 min after the treatment using the same method. Each treatment was replicated three times. Samples were homogenized in liquid nitrogen with a pellet disrupter. The homogenized samples (0.2–0.3 g each) were individually completely mixed with 1 mL of deionized water and the supernatants collected by microcentrifugation of the extract. H₂O₂ concentration was determined using an Amplex Red Hydrogen Peroxide/Peroxidase Assay kit.

Secondary Metabolite Analysis

JA and SA Bursts

For JA analysis, OS (20 μL of OS) or water (20 μL of deionized water) were added to the lamina of the leaf at node +2 immediately after six rows of puncture wounds had been created with a fabric pattern wheel. The treated leaves were harvested 0.5 h after treatment (the time of maximum JA accumulation after a single elicitation). For SA analysis, three first-instar M. sexta larvae were placed on the leaf at node +2 and allowed to feed freely. Attacked leaves were harvested at 0, 12, and 24 h after placement of larvae. Three replicate plants for each treatment for each line (three asGLP lines and one wild-type line) were harvested. JA and SA were extracted for analysis by gas chromatography-mass spectrometry with labeled internal standards ([1,2-¹³C]JA and D₄-SA) as described by Heidel and Baldwin (2004).

TrypPI Analysis

Plants from each line (three asGLP lines and one wild-type line) were randomly assigned to five treatments (five plants per treatment): MeJA, lanolin, OS, water, and control. The leaves at nodes +2 and +3 were treated. Two leaves, systemic (leaf node 0, a leaf younger than the first fully expanded leaf) and local (leaf at node +2), were harvested (at 1 pm) 4 d after treatment. TrypPI concentration was measured by radial diffusion assay as described by van Dam et al. (2001) and expressed as nmol per mg of leaf protein.

Nicotine and DTGs

Plants of each line (three asGLP lines and one wild-type line) were randomly assigned to MeJA, lanolin, OS, water, caterpillar damage, and control treatments, each with five replicates. All the treatments except caterpillar damage were treated and harvested as described for TrypPI measures. For caterpillar damage, two first-instar M. sexta were placed on the leaves at nodes +2 and +3 (one larva per leaf) and the leaves at node +2 were harvested 4 d later. Leaf extracts were prepared for analysis of nicotine and DTGs by HPLC as described by Keinänen et al. (2001).

VOCs

Plants of each line (three asGLP lines and one wild-type line) were randomly assigned to four treatment groups (five replicates each): MeJA, lanolin, OS, and water. Plants were treated as described for TrypPI measures and individually placed in 50-L glass chambers; the VOCs released were trapped on super Q (Alltech Associates) traps for 7 h, 24 h after elicitation (the time of maximum release after a single elicitation), and measured by gas chromatography-mass spectrometry (Lou and Baldwin, 2003). VOCs were expressed as percentages of peak areas relative to the internal standard, tetralin, per 7 h of trapping per plant.

VIGS Experiment

Twenty-day-old N. attenuata seedlings of an inbred wild-type line were planted into 1-L pots and placed in a York Chamber under a 16/8 h light/dark, 26°C, 65% relative humidity regime until they were 26 d old. The growth conditions at the start of the experiments were 24 h dark, 20°C, 65% relative humidity for 3 d after inoculation, after which the light was returned to 16/8 h light/dark.

The plants were inoculated with Agrobacterium cells transformed with pBINTRA and one of the lox3, ger, or pds-containing or empty constructs in a 1:1 ratio, following the procedures described by Saedler and Baldwin (2004). Three young fully expanded leaves per plant were inoculated. For the experiment, a set of 80 plants was used: 20 plants were inoculated with empty virus vector construct (pTV00), and 60 plants were separately inoculated with lox3 (pTVLOX3), ger (pTVGER), or a pds (pTVPDS)-containing construct. Ten days after inoculation, when the pTVPDS plants showed strong bleaching (each gene with 20 plants), the experiments were conducted.

Five plants of each pTVGER and pTV00 group were randomly assigned to H₂O₂ measurement. Leaves at node +2 of each plant were wounded, treated with 20 μL of OS, and harvested 30 min after the treatment. To determine expression levels of NaGLP mRNA in these plants, leaves at node +1 of these plants were half wounded (three rows of standardized puncture wounds) and treated with 10 μL of OS, and halves of leaves at node +1 were harvested before and 30 min after the treatment, respectively.

Fifteen plants of each pTVLOX3, pTVGER, and pTV00 group were assigned to herbivore performance and secondary metabolite (DTGs and TrypPI) measurement. Newly hatched larvae were placed on leaves at nodes +1 and +2 of each plant (two larvae per plant) of the three groups, respectively. Leaves at nodes +1 and +2 of each plant were harvested 1 and 4 d after the placement of the caterpillar. Leaves at node +1 were used for NaGLP mRNA expression analysis, and leaves at node +2 were used for DTG and TrypPI measurement (half a leaf for DTG analysis, half for TrypPI). Larval mass was measured (to 0.1 mg) on days 4, 7, and 10 after the start of the experiment. NaGLP mRNA expression levels were detected by qPCR. H₂O₂, DTGs, and TrypPI were analyzed by the methods described above.

Statistical Analysis

Differences in caterpillar performance and caterpillar-elicited transcript levels of NaGLP, H₂O₂, JA, SA, TrypPI, nicotine, DTGs, and VOCs were determined by Student's t tests. Differences in constitutive and M. sexta OS-elicited transcript levels of NaGLP were analyzed by ANOVA. If the ANOVA analysis was significant (P = 0.05), Fisher lsd posthoc tests to detect significant differences between groups were conducted. Data were analyzed with STATVIEW (SAS Institute).

Sequence data from this article can be found in the GenBank/EMBL data libraries under accession number AY436749.