SEARCH

SEARCH BY CITATION

Keywords:

  • ethylene;
  • Macrosiphum euphorbiae (potato aphid);
  • Mi-1-mediated resistance;
  • Never ripe (Nr);
  • plant defense response;
  • Solanum lycopersicum (tomato);
  • virus-induced gene silencing (VIGS)

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  • • 
    Resistance to potato aphid (Macrosiphum euphorbiae) in tomato (Solanum lycopersicum) is conferred by Mi-1. Early during both compatible and incompatible interactions, potato aphid feeding induces the expression of ethylene (ET) biosynthetic genes. Here, we used genetic and pharmacologic approaches to investigate the role of ET signaling in basal defense and Mi-1-mediated resistance to potato aphid in tomato.
  • • 
    The effect of potato aphid infestation on ET biosynthesis in susceptible and resistant plants was assessed. Aphid bioassays were performed using plants impaired in ET biosynthesis or perception using virus-induced gene silencing, the Never ripe (Nr) mutant, and 1-methylcyclopropene (MCP) treatment.
  • • 
    A burst of ET was observed after aphid feeding in both resistant and susceptible plants, correlated with an increase in the expression of ET biosynthetic genes. However, impairing ET signaling or biosynthesis did not compromise Mi-1-mediated resistance but it did decrease susceptibility to potato aphid in a compatible host.
  • • 
    ET may not play a significant role in Mi-1-mediated resistance to potato aphids in tomato but modulates the host basal defense, enhancing its susceptibility to the aphid.

Abbreviations:
ACC

1-aminocyclopropane-1-carboxylic acid

ACO

ACC oxidase

ACS

ACC synthase

ET

ethylene

ETR

ethylene receptor

JA

jasmonic acid

MCP

1-methylcyclopropene

Nr

Never ripe

RT-PCR

reverse transcription–polymerase chain reaction

SA

salicylic acid

TRV

tobacco rattle virus

VIGS

virus-induced gene silencing

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

The potato aphid (Macrosiphum euphorbiae L.) is one of the most damaging pests of cultivated tomato (Solanum lycopersicum) worldwide (McKinley et al., 1992; Walgenbach, 1997). Damage includes stunted growth, shoot die-back, malformation of terminal growth, distortion, chlorosis and necrosis of leaves, resulting in defoliation and yield losses. In addition to direct damage caused by the aphid feeding, potato aphids transmit several viruses to tomato (Ng & Perry, 2004) and the honeydew excreted onto foliage and fruit promotes the development of sooty mold, decreasing photosynthesis and fruit quality. Sources of genetic resistance to potato aphids have been found in wild tomato species and introduced into tomato breeding programs (Kennedy, 2005), but limited information exists about the molecular mechanisms underlying aphid resistance (Kaloshian, 2004; Smith & Boyko, 2007).

Plant responses to aphid feeding can be elicited by two different processes (Kaloshian & Walling, 2005). One involves the race-specific gene-for-gene interaction (Flor, 1971) in which the host resistance (R) protein directly or indirectly recognizes a herbivore-derived effector protein, similar to the recognition of many plant pathogens. This recognition event induces specific plant defenses that prevent pathogen or pest establishment. In the absence of R-mediated defense, plant immunity relies on the perception of physical damage and mechanical stress caused by the stylet during the probing for feeding, and/or the perception of chemical cues that can be present in salivary secretion, delivered into the host during aphid probing and feeding (Mithöfer & Boland, 2008). These processes of recognition induce the production of plant signaling molecules that trigger a general stress response, similar to the plant basal defense to phytopathogens (Kaloshian & Walling, 2005).

The tomato gene Mi-1 is the first cloned insect R gene, and provides resistance to certain biotypes of potato aphid (Rossi et al., 1998). It also confers resistance to two other phloem feeders, the tomato psyllids (Bactericera cockerelli; Casteel et al., 2006) and two biotypes of whitefly (Bemisia tabaci; Nombela et al., 2003), and to three species of root-knot nematodes (Meloidogyne spp.; Roberts & Thomason, 1986). The Mi-1-mediated resistance to potato aphids in tomato is characterized by reduced aphid longevity and fecundity, with insects dying as early as 24 h after exposure to resistant plants (Kaloshian et al., 1997). Analysis of honeydew excretion and electronic monitoring of potato aphid feeding behavior have demonstrated that Mi-1-mediated resistance has strong antixenotic effects, dramatically limiting phloem feeding. Although aphids repeatedly probe and are able to access the phloem, no significant amount of fluid is ingested (Kaloshian et al., 2000). Starvation and desiccation are the likely causes of death, as aphids recover when transferred from resistant to susceptible plants (Kaloshian et al., 2000); however, Mi-1-mediated resistance might also have antibiotic effects on aphids (Hebert et al., 2007).

Significant progress in understanding the defense signaling pathways that operate during plant–aphid interactions has been made in the last few years using comparative transcriptome analyses (Thompson & Goggin, 2006; de Vos et al., 2007; Kusnierczyk et al., 2007; Li et al., 2008). Although particular transcript profiles vary substantially among different plant–aphid species combinations, gene expression profiling upon aphid infestation established a common feature indicating that aphids elicit plant defense networks controlled by hormones such as salicylic acid (SA), jasmonic acid (JA), and ethylene (ET), which leads to increased expression of genes encoding pathogenesis-related proteins commonly associated with attack by pathogens (Glazebrook, 2005). The SA signaling pathway often contributes to R-mediated resistance, while JA and ET play a central role in the wound-signaling pathway and plant defense responses induced by necrotrophic pathogens. Not only have these phytohormones been shown to participate in activating parallel defenses against pathogens and pests, but also many events of cross-talk among the SA, JA and ET signaling pathways determine the optimum defense strategy (Rojo et al., 2003). Concomitantly, pathogens and phloem-feeding insects have evolved malicious strategies to cope with these responses and suppress effective plant defenses by manipulating defense signaling networks (Abramovitch & Martin, 2004; Walling, 2008).

Roles for SA and JA in tomato defenses against potato aphids have been investigated recently using pharmacologic and forward genetic approaches. Transgenic NahG tomato lines, expressing salicylate hydroxylase that fail to accumulate SA, show increased susceptibility to potato aphids during the compatible (susceptible) interaction and Mi-1-mediated resistance is abolished during the incompatible interaction (Li et al., 2006). Genetic studies in Arabidopsis unraveled pleiotropic changes in defense signaling of NahG plants, such as a decrease in ET production after inoculation with the bacterial pathogen Pseudomonas syringae pv. tomato, that are unlikely to result from the low SA content (Heck et al., 2003). Interestingly, the Mi-1-mediated resistance phenotype in NahG tomato can be rescued by application of an SA analog, indicating that SA is an important component of the Mi-1-mediated resistance (Li et al., 2006).

Wound- and JA-inducible defense marker genes Proteinase inhibitor 1 (Pin1) and Pin2 are induced transiently in both compatible and incompatible tomato–potato aphid interactions, suggesting a role for JA signaling in defense (Martinez de Ilarduya et al., 2003). Alteration of JA perception using the jasmonic acid insensitive 1 (jai1-1) mutation in resistant or susceptible tomato impairs neither Mi-1-mediated resistance nor the basal defense to potato aphids when insects are confined to the plants (Bhattarai et al., 2007b). By contrast, it does affect aphid behavior given a choice of host. JA mutants are more colonized by potato aphids than wild-type plants, indicating that JA might have an indirect role in plant defense by affecting the selection of the host by the aphids.

An early ET burst is frequently observed after plants are attacked by pathogen or insects but the role of ET is largely controversial (van Loon et al., 2006). In tomato, ET has been associated with the induction of host defense responses (O'Donnell et al., 2001; Diaz et al., 2002) as well as with the promotion of disease and pathogen virulence (Lund et al., 1998; Cohn & Martin, 2005; Balaji et al., 2008). Similarly, increases in ET production by aphid feeding have been reported in various plant–aphid interactions and are associated with both susceptibility and resistance (Dillwith et al., 1991; Anderson & Peters, 1994; Miller et al., 1994; Argandoña et al., 2001).

ET production during pathogen infection is largely controlled at the transcriptional level, through regulation of genes encoding the two committed steps of ET synthesis catalyzed by 1-aminocyclopropane-1-carboxylic acid (ACC) synthase (ACS) and ACC oxidase (ACO) (Argueso et al., 2007). Both ACS and ACO are encoded by multigene families, which are differentially regulated by various developmental and environmental factors. Modulation of ET signaling can also occur at the level of perception. In tomato, there are six ET receptors (ETR1–6) and each has a distinct pattern of expression throughout development and in response to external stimuli (Klee & Tieman, 2002). Only ETR3 (also known as nr) and ETR4 are induced by pathogen infection in association with increased ET synthesis (Ciardi et al., 2001). In addition, ETR3 and ETR4 are negative regulators of ET responses and exhibit functional redundancy in tomato (Ciardi et al., 2000; Tieman et al., 2000).

Although a role for ET has been established in diverse tomato–pathogen interactions, the involvement of ET has not been investigated during interaction with phloem-feeding insects. We demonstrate in this study that a burst of ET and an increase in expression of ET biosynthetic genes occur in tomato during both compatible and incompatible interactions with potato aphids. To functionally assess the role of ET in potato aphid defense in tomato we concomitantly used genetic and pharmacologic approaches to impair ET perception in susceptible and Mi-1-resistant plants. In addition, because host responses were correlated to changes in ET production, we targeted genes involved in ET biosynthesis for silencing. We observed that reducing ET sensitivity by attenuating the hormone perception or impairing ET biosynthesis decreased susceptibility during compatible interaction but did not compromise Mi-1-mediated resistance.

Materials and Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Plant material and growth conditions

Tomato (Solanum lycopersicum L.) lines used in this study were: near-isogenic cultivars (cvs) Motelle (Mi-1/Mi-1) and Moneymaker (mi/mi), cv. VFN (Mi-1/Mi-1), the Never ripe (Nr) mutant (mi/mi Nr/Nr) and the wild-type parent cv. Pearson (mi/mi nr/nr). Except for virus-induced gene silencing (VIGS) experiments, plants were grown in a pesticide-free glasshouse inside large plant cages to avoid insect infestations, with temperatures ranging from 22 to 26°C (Bhattarai et al., 2007b). After germination, plants used for VIGS experiments were grown in growth chambers at 19°C until the aphid bioassay (Bhattarai et al., 2007a). All experiments were performed with 7-wk-old plants unless otherwise stated.

Genetic crosses and homozygous (Mi-1 Nr) plant selection

Genetic crosses were performed between cv. VFN (Mi-1/Mi-1 nr/nr) and the Nr mutant (mi/mi Nr/Nr). Homozygous F2 plants were selected for ET insensitivity as described by Lanahan et al. (1994). Selected seedlings were transplanted in organic planting mix (Sun Gro Horticulture, Bellevue, WA, USA), and the presence of the Nr mutation was further confirmed by PCR-based genotyping using allele-specific primers to distinguish between the Nr mutant allele and the wild-type nr allele. A common reverse primer Nr-nr-R (5′-ATGGCATCCACAAAGCACATTC-3′) was used in combination with either the Nr-specific forward primer Nr-Mut-F (5′-TGCTGTAGCCTACTTTTCCATTCT-3′) or the nr-specific forward primer nr-WT-F (5′- TGCTGTAGCCTACTTTTCCATTCC-3′). Hot start PCR was used (94°C for 4 min, 35 cycles (94°C for 30 s, 64.8°C for 30 s, and 72°C for 45 s), and 72°C for 8 min). Homozygous plants for the Nr mutation were then genotyped for the presence of Mi-1 by PCR-RFLP using the REX-1 marker as described previously (Williamson et al., 1994). Bulked seeds from selfed F3 populations, homozygous for Mi-1 and the Nr mutation, were used for further evaluation.

Construct generation for virus-induced gene silencing

We used tobacco rattle virus (TRV)-based VIGS to repress candidate genes. The TRV-VIGS constructs used to silence the tomato ETR and ACS genes were developed from expressed sequence tag (EST) clones (Table 1) using pTRV2 and the Gateway cloning system described by Liu et al. (2002a). The tomato ESTs were amplified by PCR using the Gateway-compatible primers TomEST-attB1 (5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTCCCCCGGGCTGCAGGAATTC-3′) and TomEST-attB2 (5′-GGGGACCACTTTGTACAAGAAAGCTGGGTGGTACCGGGCCCCCCCTCGAG-3′), which anneal to the vector containing ESTs. The resulting PCR products were recombined into the donor vector pDONR207 (Invitrogen, Carlsbad, CA, USA), and subsequently into the pTRV2 Gateway-compatible vector pYL279. The identity of all clones in pTRV2 was confirmed by sequencing and the TRV-VIGS clones were transformed into Agrobacterium tumefaciens strain GV3101.

Table 1.  Tomato (Solanum lycopersicum) expressed sequence tags (ESTs) and related constructs used for virus-induced gene silencing (VIGS)
ESTaConstruct nameGenes targeted by VIGSNucleotide identitycEffective silencing
  • a

    EST identification numbers are from the SOL Genomics Network (http://www.sgn.cornell.edu/).

  • b

    ACS4 and ACS5 could not be amplified from leaf tissue.

  • c

    Length of the longest continuous stretch, where the sequences are identical between the gene and the EST.

  • ACC, 1-aminocyclopropane-1-carboxylic acid; ACS, ACC synthase; ETR, ethylene receptor; TRV, tobacco rattle virus.

SGN-E333189TRV-ETRETR4725Yes
ETR6 25No
SGN-E258763TRV-ACSIACS1A 23No
ACS1B 29Yes
ACS2608Yes
ACS4b 31
ACS6 21Yes
SGN-E365788TRV-ACSIIACS1A 69Yes
ACS1B 70Yes
ACS2 28Yes
ACS4b 22
ACS5b 24
ACS6 38Yes

Agrobacterium tumefaciens-mediated virus infection

Cultures of A. tumefaciens strain GV3101 containing each of the constructs, empty vector pTRV2, or pTRV1 (Liu et al., 2002b) were grown as previously described (Li et al., 2006) using 50 µg ml−1 kanamycin and 25 µg m l−1 rifampicin. Bacteria were resuspended in infiltration buffer at an optical density (OD600) of 1. Cells were incubated at room temperature for 3 h before use. An equal volume of pTRV1 A. tumefaciens culture was mixed with pTRV2 culture before infiltration. The abaxial side of leaflets of 2–3-wk-old seedlings was infiltrated with the A. tumefaciens cells (agroinfiltration) using a 1-ml needleless syringe.

1-Methylcyclopropene (MCP) and ethylene treatments

SmartFresh (0.14% 1-methylcyclopropene (MCP)) was obtained from AgroFresh Inc. (Philadelphia, PA, USA). Plants were enclosed in an airtight container and MCP gas was produced directly in the container by injecting water onto the powder through a septum using a syringe. Tomato plants were treated for 24 h with MCP released to a final concentration of c. 0.1 µl l−1. For ET treatment, tomato plants were placed in the airtight container and exposed to ET gas (California Tool & Welding Supply Company, Riverside, CA, USA) for 18 h. The gas was injected into the container through a septum with a syringe, to a final concentration of 1 or 10 µl l−1. Potassium hydroxide was included in the container to prevent carbon dioxide accumulation during both MCP and ET treatments (de Wild et al., 2003). Untreated control plants were held in air. Treated plants were aerated for 2 h before use in aphid bioassays. The airtightness of the container was monitored by measuring the concentration of n-butene gas using high-performance liquid chromatography (HPLC) 24 h after injection. A 17% leakage in 24 h was detected. To compensate for this loss, 20% more MCP and ET was used in subsequent experiments.

Measurement of ethylene production

To measure ET production, single tomato leaflets were placed in glass Mason jars with the cut stem inserted in 1× Murashige & Skoog (Invitrogen/GibcoBRL, Carlsbad, CA, USA) supplemented with 30 g l−1 sucrose and 1% agarose. The headspace in the jar was 200 ml. Two hours later, leaflets were infested with c. 300 apterous adults and nymphs of potato aphids. Jars were covered with a double layer of cheesecloth and placed in the glasshouse inside a large plant cage. Infested and noninfested jars were sealed with a perforated lid containing a rubber syringe cap to allow collection of generated ET (Supporting Information Fig. S1). After a 12-h incubation period, 0.9 ml of headspace was sampled from each jar and the ET content was measured using a 6850 series gas chromatography system (Hewlett-Packard, Palo Alto, CA, USA) equipped with an HP Plot alumina-based capillary column (Agilent Technologies, Palo Alto, CA, USA). Tissue fresh weight was measured for each sample.

RNA blot analyses

Total RNA was isolated from frozen tissues using hot phenol, and subjected to RNA gel blot analyses as described previously (Bhattarai et al., 2007b). The tomato EST clone cTOA29O3 was used to probe for E4, and a PCR product amplified from genomic DNA with gene-specific primers was used to probe for the ET synthetic gene ACO3 (Table S1). An 18S ribosomal DNA (18S rDNA) clone was used as control to ensure equal loading and transfer. Probes were labeled with 32P-α-dCTP, using the Prime-A-Gene labeling kit (Promega, Madison, WI, USA). Hybridization was carried out overnight at 42°C in 50% (v/v) formamide, and final washes were performed at 65°C in 0.5 × saline sodium citrate (SSC) and 0.1% sodium dodecyl sulfate (SDS) (w/v).

Semiquantitative reverse transcription–polymerase chain reaction (RT-PCR)

Total RNA was isolated from frozen tissues using either hot phenol (Verwoerd et al., 1989) or TRIzol® Reagent (Invitrogen, Carlsbad, CA, USA). Twenty micrograms of total RNA was treated with the RQ1 RNase-free DNase (Promega, Madison, WI, USA), and the DNase was subsequently removed by phenol/chloroform extraction. First-strand cDNAs were synthesized from 2.5 or 5 µg of DNase-treated RNA using Super-Script II reverse transcriptase (Invitrogen). For PCR, the different transcripts were amplified (94°C for 3 min, cycled (94°C for 45 s, annealing (Table S1) for 30 s, and 72°C for 1 min), and 72°C for 5 min) from 1 µl of cDNA in a 25-µl reaction using gene-specific primers (Table S1; this study; Kevany et al., 2007; Bhattarai et al., 2008). The tomato ubiquitin Ubi3 gene was used as a control. To check for the absence of genomic DNA contamination, 200 ng of DNase-treated RNA was used as template.

Aphid colony and bioassays

Rearing of Mi-1-avirulent potato aphid (Macrosiphum euphorbiae Thomas) maintained as a clonal colony, production of age-synchronized apterous aphid colonies and the different types of aphid bioassays performed in the present work have been described previously (Li et al., 2006; Bhattarai et al., 2007b) and are briefly outlined below. All experiments were performed at least twice. For time-course transcript analyses, c. 50 apterous adults and nymphs of potato aphid were caged onto individual leaflets of tomato. Three cages per plant and two plants were used for each time-point/genotype combination. Infected leaflets collected at indicated time-points were cleared of aphids, pooled, frozen in liquid nitrogen, and stored at −80°C.

In the choice assay, six plants per genotype were arranged in a large cage in a glasshouse. Aphid source plants were randomly distributed among the experimental plants and removed 2 d before evaluation. Plants were evaluated 12 d after exposure to potato aphids by counting the number of insects on the two most infested leaflets per plant.

Two distinct no-choice assays were performed. In the no-choice assay conducted for VIGS with resistant plants, c. 50 apterous adults and nymphs of potato aphids were caged onto individual leaflets (4–5 wk after agroinfiltration). At least four cages per plant, and five to eight plants per TRV-VIGS construct were used. Plants were evaluated by counting the number of aphids on each caged leaflet when aphids were dead on the resistant control plants.

In the second no-choice assay, age-synchronized aphids were used. One to three 1-d-old apterous adult aphids were caged onto a single leaflet. Aphid survival and fecundity were monitored on a daily basis. To evaluate aphid fecundity, first-instar progeny were counted and removed daily. The number of plants per treatment/genotype combination and number of cages per plant are described in the figure legends. Individual leaflets from the VIGS plants were collected at the end of the experiments for gene expression analysis.

Statistical analyses

Discrete values were transformed before statistical analysis. Analysis of variance (ANOVA) was performed using the statistical package sas (v9.1 for Windows; SAS Institute Inc., Cary, NC, USA). Significant differences between group means were assessed using the Tukey HSD test. In time-course experiments, data were analyzed independently day per day. Analysis of ET production data was performed using a two-sample t-test for pairwise comparisons among treatments irrespective of genotypes.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Aphid feeding triggers ET biosynthesis in tomato

The Mi-1-mediated resistance to potato aphid is developmentally regulated, with plants becoming resistant at c. 5 wk of age (Kaloshian et al., 1995). To assess ET production in response to aphid feeding, we developed an explant culture system using individual leaflets of 7-wk-old plants (Fig. S1). Initially, we confirmed the Mi-1 resistance phenotype within this system. At 12 h after infestation, aphids colonized both cvs Moneymaker (mi/mi) and Motelle (Mi-1/Mi-1), which are susceptible and resistant, respectively (data not shown), while a clear difference in aphid colonization was observed between susceptible and resistant genotypes at 24 h after infestation (Fig. S1), indicating that the Mi-1-mediated resistance was functional in the explants.

Using the explant culture system, ET production was assessed for both genotypes in noninfested leaflets or in leaflets challenged with potato aphids. The gas was allowed to accumulate for a 12-h period from 0 to 12 h post infestation (hpi) or from 12 to 24 h post infestation. There was no significant difference in ET production in noninfested leaflets of cvs Moneymaker and Motelle (Fig. 1a). During potato aphid infestation, both susceptible and resistant tomato cultivars showed increased ET production (Fig. 1a). However, the ET burst was observed earlier in susceptible leaflets (12 hpi) compared with the resistant leaflets (24 hpi), suggesting differential control of ET biosynthesis in these two genotypes in response to aphid feeding.

image

Figure 1. Effect of potato aphid (Macrosiphum euphorbiae) infestation on ethylene (ET) biosynthesis in the near-isogenic tomato (Solanum lycopersicum) cultivars (cvs) Motelle (Mi-1/Mi-1) and Moneymaker (mi/mi). (a) ET production was measured in individual leaflets of 7-wk-old plants (open bars, cv. Moneymaker; closed bars, cv. Motelle). ET was allowed to accumulate for a period of 12 h, from 0 to 12 h post infestation (hpi) or from 12 to 24 hpi, and measured using gas chromatography. A different set of samples was used for each time-point. ET production was calculated based on tissue fresh weight (FW). The experiment was performed twice with similar results. Data representing means with SE from a single experiment are presented (n = 5). Bars with different letters denote a significant difference at P < 0.05. (b) Expression of ET biosynthetic genes in leaflets of 7-wk-old plants, caged with c. 50 apterous adults and nymphs. Three cages per plant and two plants were used for each time-point/genotype combination. Leaflets were collected at 0, 6, 24 and 48 h after aphid infestation and tissues from equivalent samples were pooled before RNA extraction. Expression of 1-aminocyclopropane-1-carboxylic acid (ACC) oxidase (ACO) genes and ACC synthase (ACS) genes was determined by semiquantitative reverse transcription–polymerase chain reaction (RT-PCR) using gene-specific primers (Supporting Information Table S1). The PCR cycles were 29, 35 and 28 for ACO1, ACO3 and ubiquitin (Ubi3), respectively, in the first panel, and 40, 35, 28 and 25 for ACS1A, ACS2, ACS6 and Ubi3, respectively, in the second panel.

Download figure to PowerPoint

The pathway of ET biosynthesis is controlled by the modulation of both ACS and ACO activities and transcriptional regulation of ACS and ACO gene family members is central for increased ET production in response to abiotic stress and pathogen attack in tomato (Cohn & Martin, 2005; Castagna et al., 2007; Balaji et al., 2008). The transcript abundance of two family members of ACO and three family members of ACS was monitored using semiquantitative RT-PCR in leaves of tomato cvs Moneymaker and Motelle, following potato aphid infestation in time-course experiments (Table S1). Transcripts of ACO1 and ACO3 were weakly detectable in noninfested leaflets of both tomato cultivars (Fig. 1b). Appreciable levels of ACO1 transcripts started to accumulate in infested leaflets within 6 h after aphid infestation in both susceptible and resistant plants, and transcript abundance remained high at 24 and 48 h. Transcripts of ACO3 also accumulated in response to aphid feeding within the first 6 h after infestation in both cvs Moneymaker and Motelle (Fig. 1b). However, ACO3 expression peaked at 6 h after aphid infestation in susceptible cv. Moneymaker plants at levels not seen in resistant cv. Motelle plants until 48 h after aphid infestation. The same pattern of gene induction was consistently observed in an independent experiment (data not shown). Transcripts of ACS1A and ACS2 were detected either weakly or not at all in noninfested leaflets of both tomato cultivars, while ACS6 was constitutively expressed (Fig. 1b). Aphid feeding induced gene expression for all three genes in both susceptible and resistant plants, although the trends of induction were different among genes. In both tomato cultivars, ACS2 transcripts started to accumulate in infested leaflets within 6 h after aphid infestation, and transcript abundance remained high at 24 and 48 h. By contrast, expression of ACS1A and ACS6 was transiently induced in both susceptible and resistant plants. ACS6 transcript abundance was increased at 6 h after infestation, remained elevated at 24 h, and then declined to near control levels at 48 h. ACS1A expression was dramatically increased at 6 h after aphid infestation, and declined markedly at 24 h before reaching basal levels of expression at 48 h. Although ACS1A, ACS2 and ACS6 were all induced by the aphid, ACS1A induction was stronger in susceptible cv. Moneymaker compared with the resistant cv. Motelle, while the regulation of ACS2 and ACS6 was not appreciably influenced by the plant genotype.

ETR3-mediated ET signaling is not required for Mi-1 resistance but contributes to susceptibility in a compatible host

To assess the role of ET in aphid defense in tomato, we used the only identified ET receptor mutant, Never ripe (Nr), which is insensitive to ET (Rick & Butler, 1956; Lanahan et al., 1994). Nr is a co-dominant mutation arising from a single base substitution in the tomato ETR3 gene (Wilkinson et al., 1995). The wild-type parent of Nr, cv. Pearson, does not contain the Mi-1 gene. In order to investigate the role of ETR3 in Mi-1-mediated resistance to potato aphids, the Nr mutation was introduced into the resistant cv. VFN (Mi-1/Mi-1 nr/nr) by genetic crosses. The presence of Nr in the crosses was first assessed based on insensitivity of the seedlings to ET (Fig. S2a). ET sensitivity in the wild-type cv. VFN seedlings exposed to ACC was characterized by short, thick hypocotyls with exaggerated hooks and reduced root elongation. The growth-inhibiting effect of ACC was reduced in Nr and VFN × Nr seedlings. Moreover, VFN × Nr seedlings heterozygous for the Nr mutation displayed an attenuated response compared with the homozygous Nr parent, consistent with the co-dominance of Nr-associated ET insensitivity. To confirm the selection of Nr, we developed an allele-specific PCR-based genotyping (Fig. S2b). Seedlings were further evaluated for the presence of the Mi-1 gene, using the linked marker REX-1 (Williamson et al., 1994; Fig. S2b). Finally, to determine whether a decrease in ET susceptibility in the homozygous VFN × Nr seedlings was accompanied by a decrease in ET perception in mature plants, the expression of the ET-regulated genes E4 (Lincoln & Fischer, 1988) and ACO3 (Diaz et al., 2002) was examined by RNA blot analysis in leaves of 4-wk-old VFN × Nr (Mi-1/Mi-1 Nr/Nr) and parental types held in air or treated with ET. Treatment with an ET perception irreversible inhibitor, MCP (Sisler, 2006), before the ET treatment was also included as a control for the lack of ET-dependent gene induction. The endogenous expression levels of E4 in controls of all four tomato lines were very low and transcripts of ACO3 were not detectable (Fig. 2). The transcript abundance of both genes increased in all ET-treated samples, regardless of genotype, but inductions in plants homozygous for Nr were strongly attenuated compared with the wild-type parents. By contrast, MCP pretreatment totally abolished ET-induced gene expression. As observed in previous studies, we confirmed that Nr is not entirely insensitive to ET, but rather is severely impaired in ET perception (Aloni et al., 1998; Clark et al., 1999). Similarly, the VFN × Nr (Mi-1/Mi-1 Nr/Nr) plants were also strongly impaired in ET perception (Fig. 2).

image

Figure 2. Transcript accumulation of ethylene (ET)-inducible genes 1-aminocyclopropane-1-carboxylic acid (ACC) oxidase 3 (ACO3) and E4 in leaflets of tomato (Solanum lycopersicum) cultivars (cvs) Pearson (mi/mi nr/nr), Never ripe (Nr) (mi/mi Nr/Nr), VFN (Mi-1/Mi-1 nr/nr), and VFN × Never ripe (Mi-1/Mi-1 Nr/Nr) in response to exogenous application of ET. Two plants per genotype of 4-wk-old tomato were treated for 18 h with 1 µl l−1 ET (1) or 10 µl l−1 ET (10), or pretreated with 1-methylcyclopropene (MCP) before being induced with 10 µl l−1 ET (M); C, untreated controls. Leaf tissues collected from the same plant genotype/treatment combination were pooled and frozen. For each sample, 20 µg of total RNA was used.

Download figure to PowerPoint

To investigate whether the ET signaling pathway plays a role in tomato defense against aphids, the interactions of potato aphid with the Nr mutant, the VFN × Nr tomato line and the wild-type parents tomato cvs Pearson and VFN were compared in two distinct bioassays. Initially, tomato plants were screened in choice assays in which aphids could choose among plants and which therefore allowed the determination of plant-specific aphid feeding preferences. Twelve days after infestation, aphid abundance was significantly greater in susceptible cv. Pearson and Nr plants compared with Mi-1-resistant cv. VFN and VFN × Nr plants (Figs 3a, S3). No significant difference was observed in the number of aphids on Mi-1-resistant plants irrespective of the presence of the Nr mutation. By contrast, the number of aphids on Nr plants was significantly lower than on the wild-type parent cv. Pearson. In no-choice assays, when 1-d-old adult aphids were used in leaf cages, from day 2 onward, significant differences in the per cent daily aphid survival were observed in Mi-1-resistant cv. VFN and VFN × Nr plants compared with the susceptible genotypes Nr and Pearson (Fig. 3b). Aphid viability was not affected by the presence of the Nr mutation in susceptible or resistant plants. The average number of progeny per aphid per day was assessed over a 3-d period, when c. 50% of the insects had died on resistant plants (Fig. 3b), and was found to be significantly lower on resistant plants compared with susceptible plants lacking Mi-1 (Fig. 3c). No significant difference in fecundity attributable to the presence of the Nr mutation was observed in either susceptible and resistant genotypes.

image

Figure 3. Potato aphid (Macrosiphum euphorbiae) bioassays of tomato (Solanum lycopersicum) cultivars (cvs) Pearson (mi/mi nr/nr), Never ripe (Nr) (mi/mi Nr/Nr), VFN (Mi-1/Mi-1 nr/nr), and VFN ×Never ripe (Mi-1/Mi-1 Nr/Nr). Seven-week-old plants were used in aphid choice (a) and no-choice (b, c) assays. Experiments were performed twice with similar results. Data representing means with SE from a single experiment are presented. Bars with different letters denote a significant difference at P < 0.05. (a) Six plants per genotype were used in the aphid choice assay. Twelve days after exposure to insects, the number of aphids on the two most infested leaflets per plant was counted and averaged. Data represent means for n = 6. (b, c) Twelve plants per genotype were used in the no-choice assay. A single leaflet per plant was infested with three 1-d-old apterous adult aphids. Aphid survival (b) and fecundity (c) were monitored daily. Data represent means for n = 12. The average daily aphid reproduction rate presented (c) is calculated for the first 3 d.

Download figure to PowerPoint

Mi-1-mediated resistance to aphids is not compromised in tomato attenuated in ET perception

To overcome potential functional redundancy among tomato ET receptors mediating ET signaling in response to aphids, we used MCP to impair ET perception and TRV-VIGS to target ETR4 for silencing. The chemical MCP functions as a competitive inhibitor of ET and its attachment to the receptors is essentially irreversible (Sisler, 2006). The duration of MCP effects on ET perception in tomato plants has not been critically studied. Thus, the efficiency of the MCP treatment in blocking ET receptors over time was assessed in tomato. Expression of the ET-inducible gene E4 was examined in leaves of tomato cv. Moneymaker treated with MCP and subsequently induced with ET 1–5 d later. Pretreatment of tomato with MCP lowered basal expression of E4 and prevented ET-induced E4 transcript accumulation for 1 d (Fig. 4a), indicating that ET perception in tomato leaves was successfully blocked. However, 2 d after MCP treatment, c. 20–25% of the E4 induction was recovered and this continued to increase over the rest of the 5-d period analyzed.

image

Figure 4. Effect of 1-methylcyclopropene (MCP) treatment of tomato (Solanum lycopersicum) cultivars (cvs) Motelle (Mi-1/Mi-1) and Moneymaker (mi/mi) on E4 expression induced by ethylene (ET) in leaves, and on potato aphid (Macrosiphum euphorbiae) demography. (a) Efficiency of the MCP blocking of ET perception was assessed by monitoring the expression of E4 after induction by ET. Seven-week-old cv. Moneymaker plants (+MCP/+ET) were pretreated with MCP, and two plants were treated daily for 18 h with 10 µl l−1 ET before harvest. Leaf tissues were pooled and frozen. Tissues from untreated plants (−MCP/−ET) or plants only induced by ET (−MCP/+ET) were used as controls. Total RNA (25 µg) for each sample was used for RNA blot analysis. The blot was hybridized sequentially with E4 and an 18S rDNA probe used to normalize expression. (b, c) Seven-week-old plants of cvs Motelle and Moneymaker treated with MCP (+MCP) or untreated (−MCP) were used in the no-choice aphid assay. The MCP-treated plants were aerated for 2 h before caging. Six plants per genotype/treatment combination and two leaflets per plant were infested with a single 1-d-old apterous adult aphid. Aphid survival (b) and fecundity (c) were monitored daily. The experiment was repeated a few days apart. Data representing combined data sets are presented (n = 12) and error bars are SE. The average daily aphid reproduction rate presented (c) is calculated for the first 3 d. Bars with different letters denote a significant difference at P < 0.05.

Download figure to PowerPoint

An aphid no-choice assay was conducted to determine whether initial inhibition of ET perception using MCP could affect aphid survival and fecundity on susceptible and resistant tomato cvs Moneymaker and Motelle, respectively. One-day-old adult aphids were caged onto individual leaflets of MCP-treated and untreated control plants, and aphid survival and fecundity were monitored on a daily basis for 14 d, until all aphids died. The number of progeny per aphid per day was assessed over a 3-d period, when c. 50% of the insects had died on the resistant plants (Fig. 4b,c). Both viability and fecundity were significantly reduced on resistant cv. Motelle plants compared with susceptible cv. Moneymaker, irrespective of the MCP treatment. Within both cultivars, viability and fecundity of aphids were not significantly affected by the MCP treatment in the first 5 d or later (Fig. 4b,c). Although MCP treatment reduced ET perception in plants it did not compromise Mi-1-mediated resistance to aphids or affect the susceptibility of the compatible host. The critical period for aphid survival on resistant plants is between the second and third days after aphid infestation. As partial ET perception is recovered as early as 2 d after exposure to MCP, we do not expect to observe a significant effect of the MCP treatment on aphid survival if a complete blocking of ET perception is required to affect the plant defense to aphids.

As MCP treatment completely blocked ET perception only briefly in tomato with no effect on aphid survival, we investigated an alternate way to eliminate functional redundancy among ET receptors involved in plant defense. We used TRV-VIGS to silence ETR4 in VFN × Nr plants (Table 1). VFN × Nr plants along with parent cv. VFN and Nr plants were agroinfiltrated with TRV-ETR or TRV empty vector and then used in no-choice aphid bioassays. About 50 apterous adults and nymphs of potato aphids were caged onto individual leaflets. Two weeks after infestation, the aphid population grew on susceptible Nr plants while aphids died on cv. VFN and VFN × Nr with and without silencing of ETR4, indicating that silencing this ET receptor in the Nr background did not attenuate Mi-1-mediated resistance to potato aphids (data not shown).

Potato aphids require a functional ET biosynthetic pathway for complete virulence

As attenuating ET perception did not compromise aphid resistance in tomato, we tested whether impairing ET production by targeting the biosynthetic genes using TRV-VIGS would affect plant defenses to aphids. The ACS enzyme catalyzes the first committed step and in most cases the rate-limiting step in ET biosynthesis (Yang & Hoffman, 1984). Two constructs were designed, TRV-ACSI and TRV-ACSII, that would enable silencing of the maximum number of ACS genes when combined (Table 1). These two constructs were agroinfiltrated alone or combined into cv. Motelle and Moneymaker plants and distinct no-choice aphid bioassays were conducted for each genotype.

Resistant cv. Motelle plants were infested similarly to the previous TRV-VIGS experiment. Two weeks after infestation, all insects were dead on TRV control plants as well as on plants in which the ACS genes were silenced by either of the constructs alone or by the two constructs in combination (data not shown). Susceptible cv. Moneymaker plants were used in no-choice assays where 1-d-old adult aphids were caged onto individual leaflets of ACS-silenced and TRV control plants. Aphid survival and fecundity were monitored on a daily basis. Aphids died more rapidly on plants silenced with the individual TRV-ACSI and TRV-ACSII constructs and combined constructs than on control plants (Fig. 5a). Although the effect of silencing on aphid survival seems to be more profound with the combined infiltration of TRV-ACSI and TRV-ACSII compared with infiltration of the single constructs, no significant differences were observed among these treatments. No effect on aphid fecundity was observed among the silenced plants compared with controls (data not shown).

image

Figure 5. Aphid (Macrosiphum euphorbiae) survival and transcript levels of four 1-aminocyclopropane-1-carboxylic acid (ACC) synthase (ACS) genes in tomato (Solanum lycopersicum) cv. Moneymaker (mi/mi) leaflets silenced with different tobacco rattle virus (TRV)-ACS constructs. (a) Plants were agroinfiltrated using TRV-ACSI and TRV-ACSII constructs individually or simultaneously infiltrated (TRV-ACSI+II). The TRV control refers to the pTRV2 empty vector. Three weeks after agroinfiltration, four leaflets per plant were caged with a single 1-d-old adult apterous aphid. Five plants per construct were used in this no-choice assay. Aphid survival was monitored daily. The experiment was performed twice with similar results. Data representing means from a single experiment are presented (n = 5). Points of significance between silenced and control plants are denoted by an asterisk (P < 0.05). (b) Transcript accumulation of expressed ACS genes targeted by virus-induced gene silencing (VIGS) in controls and plants co-infiltrated with TRV-ACSI and TRV-ACSII constructs was evaluated by reverse transcription–polymerase chain reaction (RT-PCR) using gene-specific primers (Supporting Information Table S1). PCR cycles are indicated at the top of each panel.

Download figure to PowerPoint

To evaluate the efficiency of silencing in plants infiltrated with TRV-ACSI and TRV-ACSII and co-infiltrated with both constructs, expression of ACS genes was analyzed by semiquantitative RT-PCR using gene-specific primers (Table S1). TRV-ACSI and TRV-ACSII efficiently silenced the predicted target genes (Table 1, Fig. S4): TRV-ACSI silenced ACS1B, ACS2 and ACS6, and TRV-ACSII silenced ACS1A, ACS1B, ACS2 and ACS6. The co-infiltration of both TRV-ACSI and TRV-ACSII constructs resulted in more efficient silencing compared with the infiltration of the constructs individually (Figs 5b, S4). ACS4 and ACS5 were also targeted by one or both TRV-ACS constructs; however, transcripts of both genes could not be amplified from leaf tissue irrespective of silencing and aphid infestation, despite product amplification using tomato DNA as template (Table 1, data not shown).

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Analysis of ET mutants in several plant species has demonstrated a role for ET in both compatible and incompatible interactions, yet the role of ET in pathogenesis varies greatly among pathogens and pests (van Loon et al., 2006). Increased production of ET by aphid feeding has been reported in various plant–aphid interactions and is associated with susceptibility in alfalfa (Medicago sativa) to spotted alfalfa aphid (Therioaphis maculata; Dillwith et al., 1991), in barley (Hordeum vulgare) to Russian wheat aphid (Diuraphis noxia; Miller et al., 1994), and in wheat (Triticum aestivum) to greenbug (Schizaphis graminum; Anderson & Peters, 1994). Conversely, ET signaling is involved in defense responses in barley to greenbug and oat aphid (Rhopalosiphum padi; Argandoña et al., 2001). Furthermore, variation in plant defense by the same host to different species of aphids has been observed in Arabidopsis, where ET is not involved in resistance to green peach aphid (Myzus persicae) and contributes to susceptibility to cabbage aphid (Brevicoryne brassicae; Mewis et al., 2005). Collectively these studies indicate that no consensual role for ET can be established in plant–aphid interactions.

This study investigated the involvement of ET in tomato defense to potato aphids. By comparing the behavior of the aphids on Mi-1 (cv. VFN) and ET-insensitive Mi-1 Nr plants (VFN × Nr tomato line), we concluded that ETR3-mediated ET signaling is not required for Mi-1-mediated resistance to aphids in tomato. Our results also suggest that ET perception in general might not play a significant role in Mi-1-mediated defense. Similar results were obtained using the Nr mutation in 35S::Pto tomato plants, indicating that ET is dispensable for Pto-mediated resistance to P. syringae pv. tomato (Li et al., 2002). In both cases, a role for the ET-signaling pathway in R-mediated resistance cannot be completely ruled out as the Nr mutation is leaky and a low level of ET perception might be sufficient to transduce the signal. We used MCP to overcome the leakiness of Nr or potential functional redundancy among ET receptors. MCP treatment also did not impair Mi-1-mediated resistance to the aphid. However, the MCP treatment completely blocked ET perception only briefly in tomato, with plants recovering partial ET sensitivity in less than 2 d. It was unexpected to find that ET sensitivity recovered so rapidly in leaf tissue, because MCP binds permanently to the receptor (Sisler, 2006). Therefore, our results suggest that the turn-over of the ET receptors in membranes of tomato leaves is fast. Partial restoration of ET signaling may be masking the effect of MCP treatment during the incompatible interaction. Alternatively, ET signaling is not involved in the Mi-1 incompatible interaction with potato aphids.

A differential role for ET in susceptible tomato hosts was observed depending on the aphid assay used. In a no-choice assay, no difference in aphid survival or fecundity on Nr plants compared with the wild-type parent was observed, indicating that ETR3-mediated ET signaling is not required for basal defense. In the choice assay, however, we observed a greater colonization of the wild-type cv. Pearson plants compared with the ET-insensitive Nr mutant. In this case, the reduced ET perception might affect plant-specific aphid feeding preferences. Aphids are known to employ plant-derived signals, or semiochemicals, in host selection and acceptance after landing and during initial feeding behavior (Walling, 2008). We previously reported that during compatible interaction JA-mediated signaling was not required for basal defense; however, the JA-insensitive jai1-1 mutant plants were more colonized by potato aphids in a choice assay than the wild-type parent, suggesting that changes in the oxylipin signature enhances aphid attraction to the jai1-1 plants (Bhattarai et al., 2007b). Similarly, the emission of plant volatile components or other chemical cues may be modified in the Nr mutant.

In our study, early during both compatible and incompatible interactions a burst of ET was detected after potato aphid infestation, similar to other plant–herbivore and plant–pathogen interactions (Adie et al., 2007; von Dahl & Baldwin, 2007). Because ET regulates the expression of numerous defense-related genes (Adie et al., 2007), increases in ET synthesis could lead to activation of defense responses. Indeed, there is faster and greater accumulation of transcripts of ET biosynthetic genes and higher levels of ET are produced in incompatible interactions of tomato with Xanthomonas campestris (Ciardi et al., 2000), and rice (Oryza sativa) with blast fungus Magnaporthe grisea (Iwai et al., 2006). In this study, however, ET did not appear to play a significant role in the Mi-1-mediated resistance to potato aphids. In addition, ET was produced in both resistant and susceptible plants in response to aphid feeding, suggesting that ET may be involved in basal defense to aphids. The delay observed for the ET burst in resistant cv. Motelle compared with susceptible cv. Moneymaker indicates differential control of ET biosynthesis in these genotypes. Aphid feeding triggers the ET biosynthetic pathway in tomato but, as ACS2 is not induced by ET (Diaz et al., 2002), this response is at least partially independent of ET.

Plants impaired in ET biosynthesis were not compromised in Mi-1-mediated resistance, indicating that ET may be dispensable for Mi-1-mediated resistance. Alternatively, a greatly diminished capacity for ET synthesis may be sufficient to trigger full resistance. By contrast, aphid viability was negatively affected in ACS-silenced susceptible plants. Our experiments indicate that potato aphid feeding on tomato stimulates ET biosynthesis in the host, and ACS-VIGS data in the compatible host provide evidence that this activity is required for their virulence. However, susceptibility to potato aphids was not enhanced in the ET-overproducing tomato mutant Epinastic (Epi; Fujino et al., 1988; S. Mantelin & I. Kaloshian, unpublished data), indicating that, although ET contributes to aphid susceptibility, further increases of ET above wild-type levels do not affect aphid susceptibility. The ET levels produced in wild-type plants appear to be sufficient for full susceptibility. Although the Epi mutation was reported in the Mi-1-containing tomato cv. VFN8, marker analysis of the Epi mutant obtained from both the original source of Epi (Fujino et al., 1988) and from the Tomato Genetics Resource Center (http://tgrc.ucdavis.edu) indicated that the Epi mutant did not contain the Mi-1 gene (S. Mantelin & I. Kaloshian, unpublished data). Therefore, the role of Epi in the Mi-1-mediated resistance was not studied.

As lowering the capacity of the plant to produce ET increased its resistance to potato aphids in the compatible interaction, we demonstrated that ET modulates the basal defense to aphids in tomato. Our work suggests that aphids manipulate the plant to produce more ET. The delay in ET burst in response to aphid feeding in resistant plants suggests that part of the defense response in Mi-1-mediated resistance is to control ET synthesis by regulating expression of ET biosynthetic genes, and thus strengthening the plant basal defense.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

We thank Usha Bishnoi and Kevin Izquierdo for technical assistance, and QiGuang Xie for initial work with Never ripe. We also thank Sarah Aschmann (Air Pollution Research Center, UC Riverside) for the HPLC measurements, Paul Larsen for help with ET measurements (Biochemistry Department, UC Riverside), Harry Klee (University of Florida) for the Never ripe seeds, and Olivia Desmond for comments on the manuscript. The authors are grateful to AgroFresh Inc. for kindly providing MCP. This work was supported in part by grants from the National Science Foundation (grant no. IOB-0543937) and the University of California Agricultural Experiment Station.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Fig. S1 Explant culture system for ethylene (ET) measurement.

Fig. S2 Phenotype and genotype of F1 and parents of the cross VFN × Never ripe.

Fig. S3 Phenotype of tomato leaflets infested with potato aphids in the choice assay (Fig. 3a) carried out with cv. Pearson (mi/mi nr/nr), Never ripe (mi/mi Nr/Nr), VFN (Mi-1/Mi-1 nr/nr), and VFN × Never ripe (Mi-1/Mi-1 Nr/Nr) plants.

Fig. S4 Effect of tobacco rattle virus (TRV) virus-induced gene silencing (VIGS) on transcript levels of 1-aminocyclopropane-1-carboxylic acid synthase 1A (ACS1A), ACS1B, ACS2 and ACS6 in control and silenced cv. Moneymaker (mi/mi) plants.

Table S1 Primers used for gene expression analyses

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office.

FilenameFormatSizeDescription
NPH_2870_sm_FigS1.tif13651KSupporting info item
NPH_2870_sm_FigS2.tif3688KSupporting info item
NPH_2870_sm_FigS3.tif5564KSupporting info item
NPH_2870_sm_FigS4.tif4040KSupporting info item
NPH_2870_sm_TableS1.tif2439KSupporting info item