Identification of potential early regulators of aphid resistance in Medicago truncatula via transcription factor expression profiling

Authors


Author for correspondence:
Karam B. Singh
Tel: 61 8 9333 6320
Email: karam.singh@csiro.au

Summary

  • Resistance to aphids has been identified in a number of plant species, yet the molecular mechanisms underlying aphid resistance remain largely unknown.
  • Using high-throughput quantitative real-time PCR technology, the transcription profiles of 752 putative Medicago truncatula transcription factor genes were analysed in a pair of susceptible and resistant closely related lines of M. truncatula following 6 and 12 h of bluegreen aphid (Acyrthosiphon kondoi) infestation.
  • Eighty-two transcription factor genes belonging to 30 transcription factor families were responsive to bluegreen aphid infestation. More transcription factor genes were responsive in the resistant interaction than in the susceptible interaction; of the 36 genes that were induced at 6 and/or 12 h, 32 were induced only in the resistant interaction. Bluegreen aphid-induced expression of a subset of these genes was correlated with the presence of AKR, a single dominant gene conferring resistance to bluegreen aphids. Similar transcription factor expression patterns of this subset were associated with bluegreen aphid resistance in other M. truncatula genetic backgrounds, as well as with resistance to pea aphid (Acyrthosiphon pisum).
  • Our results suggest that these transcription factors are among the early aphid-responsive genes in resistant plants, and may play important roles in resistance to multiple aphid species.

Introduction

In nature, plants face enormous challenges from microbial pathogens and insect pests. Among them, phloem-feeding insects, such as aphids, cause serious damage to agriculture by depleting plant nutrients, modifying plant metabolism and vectoring pathogenic viruses (Blackman & Eastop, 2000). Aphid infestation usually causes limited damage to the plant in comparison with that caused by microbial pathogens and chewing insects. Aphids use their stylets to probe intercellularly through epidermal and mesophyll cell layers until they reach the phloem sieve element, where they establish a sustained interaction on suitable hosts (Tjallingii & Esch, 1993). Aphids probably use their specialized feeding tactics to minimize plant defence responses during the infestation process (reviewed by Thompson & Goggin, 2006; Goggin, 2007; Walling, 2008). Moreover, aphids may be able to manipulate plant physiology and, in particular, defence responses through the secretion of saliva into the phloem sieve elements to maintain their successful feeding on the plant (Will et al., 2007). Plant defence against aphids is complex, including premature leaf chlorosis and cell death (Pegadaraju et al., 2005). Such unique interactions represent an intriguing area in the study of plant interaction with other organisms. Understanding the nature of the plant’s ability to resist aphid feeding is of importance to agriculture and to our understanding of plant defence against biotic stresses.

Studies using the model plant Arabidopsis have contributed greatly to our understanding of resistance (R) gene-mediated plant defence, especially against pathogens (Jones & Dangl, 2006), as well as the basal defence mechanisms against aphid feeding (Pegadaraju et al., 2005, 2007; Couldridge et al., 2007; Kusnierczyk et al., 2008). With the identification and characterization of a large number of R genes and their corresponding pathogen effectors, a picture is emerging about the interactions of the two partners, as well as the downstream signalling cascades leading to full disease resistance (Takken et al., 2006; Schwessinger & Zipfel, 2008). Transcription factors (TFs) have been shown or implicated to be key regulators of the plant’s defence response (Hammond-Kosack & Parker, 2003; Gutterson & Reuber, 2004; McGrath et al., 2005; Eulgem & Somssich, 2007; Libault et al., 2007; Onate-Sanchez et al., 2007). Some members of specific TF families have been shown to be key components in the regulation of R gene-mediated disease resistance. For instance, Arabidopsis AtWRKY70 modulates the cross-talk between signalling pathways activating salicylic acid (SA)-dependent and repressing jasmonic acid (JA)-dependent responses, and is an essential element for both basal and full R gene-mediated resistance against the oomycete Hyaloperonospora parasitica (Li et al., 2004; J. Li et al., 2006; Knoth et al., 2007). TFs have also been implicated recently in the regulation of plant response to herbivory. Loss of function of Nicotiana attenuate WRKY3 or WRKY6 promotes the susceptibility of the plant to the chewing insect, Manduca sexta (Skibbe et al., 2008). In addition, overexpression of a rice TF gene, WRKY89, increases plant resistance to the white-backed planthopper Sogatella furcifera, a sap-sucking insect (Wang et al., 2007). Several cDNA or oligo-microarray experiments have shown that TFs are responsive to aphid feeding (De Vos et al., 2005; Levy et al., 2005; van Oosten, 2007; Kusnierczyk et al., 2008; Li et al., 2008). However, the limited sensitivity of these technologies has been a disadvantage in the accurate measurement of the low transcript level typical of TF genes. High-throughput quantitative reverse transcription-polymerase chain reaction (qRT-PCR) remains a sensitive and efficient method available for the quantification of large numbers of low-abundance TF genes. Such platforms have been established for the model plants Arabidopsis, rice and Medicago truncatula and have shown to be significantly more sensitive than microarrays (Czechowski et al., 2004; Caldana et al., 2007; Kakar et al., 2008).

The legume M. truncatula has emerged in recent years as an effective model for the study of R gene-mediated aphid resistance (Edwards & Singh, 2006). For example, a single M. truncatula line, Jester, has been shown to contain three independent R genes that confer strong resistance to three different aphid species (Klingler et al., 2005; Edwards & Singh, 2006; Gao et al., 2007a,b, 2008; Klingler et al., 2007; Guo et al., 2009; L. G. Kamphius, S. Guo & K. B. Singh, unpublished). Moreover, another M. truncatula line, A17, has been shown to contain additional R genes to bluegreen aphid (BGA; Acyrthosiphon kondoi) and pea aphid (PA; Acyrthosiphon pisum) (Klingler et al., 2009; Stewart et al., 2009). The aim of the present study was to provide an overview of the TF genes expressed during susceptible and resistant aphid interactions in M. truncatula, with a focus on the identification of TFs specifically involved in the early stages of aphid resistance. The strategy employed involved large-scale profiling of TF gene expression in closely related resistant and susceptible M. truncatula lines after BGA infestation, using qRT-PCR. The expression profiles of 752 putative TF transcripts were analysed at two early time points of the plant–aphid interaction. Comparison of the TF profiles between susceptible and resistant aphid interactions allowed us to identify large changes in TF gene expression associated with resistance following aphid infestation. Members of AP2/EREBP (pathogenesis-related transcriptional factor and ethylene response factor), bHLH (basic helix–loop–helix dimerization), C2H2 (Zn) (Zn-finger, C2H2 type) and WRKY (DNA-binding WRKY) gene families were specifically induced in the resistant plants after BGA infestation. Two genes from the bHLH and WRKY gene families were further analysed and shown to be associated with BGA resistance in other M. truncatula genetic backgrounds and with resistance to PA, suggesting that these TFs may play important roles in aphid resistance.

Materials and Methods

Plant growth

Two Medicago truncatula Gaertn (barrel medic) lines were the primary focus of this study. A17 is relatively susceptible to both BGA (A. kondoi) and PA (A. pisum) compared with its closely related line Jester, which has 89% of its genome derived from A17 and is resistant against both aphid species (Klingler et al., 2005; Gao et al., 2007a, 2008). Resistance to BGA and PA in Jester is conditioned by two independent genes, named AKR and APR (Klingler et al., 2005; Guo et al., 2009). In addition to A17 and Jester, two genotypes specifically developed for the molecular genetic analysis of BGA resistance were tested. A17AKR+ (BGA resistant) is approx. 99% identical to susceptible line A17 and JesterAKR− (BGA susceptible) is approx. 99% identical to Jester. Also included were two other pairs of closely related lines of M. truncatula (cv Cyprus-Caliph and cv Borung-Mogul). The genetic background and origins of all of these lines have been described previously (Klingler et al., 2005; Gao et al., 2007a,b; Guo et al., 2009). The details of seed treatment and plant growth conditions were the same as described by Klingler et al. (2005).

Aphid infestation

Two aphid species were used in this study: BGA (A. kondoi Shinji) and PA (A. pisum Harris). The aphids were sourced and maintained as described previously (Klingler et al., 2005, 2007; Gao et al., 2007a,b, 2008; Guo et al., 2009). Aphids were transferred to experimental plants with a fine paintbrush. Single clones were used for each aphid species to minimize within-treatment variability. Aphid infestation assays were based on Gao et al. (2007a, 2008). Two fully expanded trifoliate leaves from the primary stem of individual plants (age c. 3–4 wk) were each infested with 20 late instars/adults for BGA or 15 late instars/adults for PA and caged in linen mesh supported by a bamboo stick. For the noninfested control plants, two trifoliate leaves of similar developmental stage were caged without aphids. Three biological replicates were set up for each aphid-infested or noninfested control. Each biological replicate included pooled samples from three plants. Samples were collected at 6 and 12 h after aphid infestation for both control and aphid-infested plants. After de-assembling the cage, aphids were brushed off using a fine paint brush; the noninfested control received a similar extent of brushing. Petioles were separated from leaflets at harvest. Petioles from the two trifoliate leaves per plant were combined and used for gene transcription analysis. All the samples were frozen in liquid nitrogen and stored at −80°C until RNA isolation.

Methyl jasmonate (MeJA) treatment and mechanical wounding

Treatment of plants with MeJA for gene expression studies was the same as described previously in Gao et al. (2007a). Two-week-old plants of M. truncatula cv A17 and Jester were treated with 10 μM MeJA. Control plants were mock treated with 5 μl ethanol l−1 of air. Ethanol was the solvent for MeJA. Petioles of 10 control or treated plants were harvested at 3 and 12 h after treatment, pooled, frozen in liquid nitrogen and stored at −80°C until RNA extraction. Three biological replicates of control and MeJA treatments were conducted.

For the wounding experiment, we used the approach described by Moran & Thompson (2001). Petioles and leaflets from two fully expanded trifoliate leaves of individual plants (age c. 3–4 wk) were simultaneously and gently punctured using a bundle of sewing needles, approx. 30–50 punctures per petiole or leaflet. Samples were collected at 1 and 6 h after wounding for both control and wounded plants. Petioles were separated from leaflets at harvest. Petioles from two trifoliate leaves per plant were combined and used for gene transcription analysis. All the samples were frozen in liquid nitrogen and stored at −80°C until RNA isolation.

RNA isolation and cDNA synthesis

Total RNA isolation was performed using the Purescript RNA isolation kit (Gentra Systems, Minneapolis, MN, USA) and followed by treatment with RNase-free DNase (Ambion Inc., Houston, TX, USA). RNA integrity was checked using electrophoresis on a 1.5% (v/w) agarose gel before and after DNase treatment. The absence of genomic DNA contamination was confirmed using real-time qRT-PCR on nonreverse-transcribed RNA employing a primer pair designed to amplify a 107-bp intron sequence of the housekeeping gene Ubiquitin (TC102473; 5′-TCCTCTAAGGTTTAATGAACCGG-3′; R, 5′-GAAAGACACAGCCAAGTTGCAC-3′). The first-strand cDNA was synthesized using oligo-dT12–18 (Qiagen, Hilden, Germany) and SuperScript III reverse transcriptase (Invitrogen GmbH, Karlsruhe, Germany). The efficiency of cDNA synthesis was examined using qRT-PCR to amplify cDNA fragments in the 5′ and 3′ regions of the Ubiquitin gene. Only cDNA samples that showed similar threshold cycle (CT) values (22 ± 0.75) and that had a 3′ : 5′ ratio of 1–1.5 (using the formula 2(CT5′–CT3′)) were considered for further analysis of TF transcripts.

Primer set of M. truncatula TFs

A total of 752 pairs of primers representing all annotated M. truncatula TFs available at the time of this study was included in the profiling experiment. The identification of the putative TFs and the design and verification of the primers have been described in Udvardi et al. (2007) and Kakar et al. (2008). This set of putative TFs with their corresponding BAC and TC accession numbers, as well as the sequences of the primer pairs used for amplification, is listed in Supporting Information Table S1.

qRT-PCR conditions and analysis

High-throughput profiling of the 752 TFs in M. truncatula cv A17 vs Jester, with or without BGA infestation, was conducted at the Max Planck Institute of Molecular Plant Physiology, Golm, Germany. PCRs were performed in an optical 384-well plate with an ABI PRISM® 7900HT sequence detection system (Applied Biosystems, Foster City, CA, USA) using SYBR® Green (Invitrogen, Carlsbad, CA, USA) to monitor dsDNA synthesis. The reactions and conditions of PCR were the same as in Kakar et al. (2008) and Verdier et al. (2008).

Data were analysed using sds 2.0 software (Applied Biosystems). All amplification plots were analysed with a threshold of 0.3 to obtain CT values. PCR efficiency (E) was estimated using LinRegPCR software (http://LinRegPCR.nl) with data obtained from the exponential phase of each individual amplification plot and the equation (1 + E) = 10slope (Ramakers et al., 2003). Primers with R2 < 0.99 and E < 0.6 were excluded from further analysis.

Selection of reference genes

Each primer plate contained eight housekeeping genes: Ubiquitin (TC102473), Actin2 (TC107326), GAPDH (Glyceraldehyde 3-Phosphate Dehydrogenase), PTB (Polypyrimidine Tract-Binding), UBC (Ubiquitin-conjugating enzyme E2), EF1α (Elongation factor 1 alpha), Tubulin and Helicase. The suitability of the housekeeping genes for transcript normalization was determined using geNORM software v3.4 (Vandesompele et al., 2002). The four genes with the greatest transcript stability values were GAPDH (MtC00030_GC; E = 0.96 ± 0.018), Actin2 (TC107326; AC137836_27.5; E = 0.85 ± 0.018), Ubiquitin (TC102473; AC137828–19.4; E = 0.87 ± 0.015) and Helicase (CB892427; E = 0.88 ± 0.017). The primer sequences of the eight housekeeping genes have been published in Kakar et al. (2008).

To compare data from different PCR runs and different cDNA samples, CT values were normalized against the geometric mean of these four reference genes, whose transcript levels were most stable across the biological samples and runs. The average of the geometric mean of these four genes for all 24 samples and a total of 48 plate runs was CT = 21.23 ± 0.47. For normalization, the mean reference gene CT value was subtracted from the CT value of the TF gene of interest, yielding a ΔCT value. The expression level (Exp) of each TF gene was calculated using the formula E(−ΔCT).

Three biological replicates were analysed for each of the eight treatments: aphid susceptible A17 and resistant Jester with or without BGA infestation for 6 or 12 h. Individual Student’s t-tests were used to determine significant effects of aphid infestation on Exp of individual TF genes (infested A17 or Jester compared with its noninfested control). In addition, a ratio of transcript levels was calculated for each TF gene by dividing the mean Exp of infested A17 or Jester plants by the mean Exp of noninfested plants.

Follow-up qRT-PCR analysis of selected TF genes

Further analysis of selected genes in the same pair of M. truncatula cultivars with infestation by PA, MeJA treatment or wounding, as well as in other M. truncatula genetic backgrounds with BGA infestation, was conducted at CSIRO Plant Industry Division, Floreat, Western Australia. The qRT-PCRs and conditions were the same as described previously in Gao et al. (2007a). Relative gene expression was derived using 2−ΔCT, where ΔCT represents the CT of the gene of interest minus the CT of Tubulin (E = 0.98 ± 0.016) or Ubiquitin (E value as above) in the case of cv Cyprus and cv Caliph, which have significantly different mRNA levels of Tubulin gene (L-L. Gao, unpublished). Where required, the significance of differences between relative gene expressions was analysed by two-way ANOVA (genotypes by treatments) and compared by the least-significant difference test at a 5% significance level using GenStat 6.2 (Lawes Agricultural Trust, Rothamsted, UK).

Results

TF expression profiles during susceptible and resistant BGA interactions

The expression of 752 putative TFs was examined in the petiole tissue of BGA-infested and noninfested M. truncatula plants. Previous studies have shown high levels of aphid-responsive gene expression in the petioles of resistant M. truncatula (Gao et al., 2007a). Of the 752 TF genes investigated, 162 TFs (22%) were not expressed at a detectable level in petioles under any of the conditions examined (Table S1). The most likely explanation is that these TFs are expressed elsewhere, but not in petioles. The transcription profiles of the remaining TF genes were first compared between the pair of closely related resistant (Jester) and susceptible (A17) M. truncatula lines without aphid infestation. With the exception of one gene, CCHC (Zn) (AC130807_30.241), whose transcription level appeared to be constitutively higher in A17 than in Jester, the expression levels of the rest of the TF genes did not differ significantly (P > 0.05) between the two lines at the 6- and 12-h time points (Table S1).

Among the 590 (78%) genes for which transcripts were detected, a total of 82 TF genes from 30 TF gene families exhibited differential expression between aphid-infested and noninfested plants (P ≤ 0.05 and fold change ≤ 0.5 for downregulation or ≥ 2 for upregulation) in either the susceptible or resistant interaction (Fig. 1, Table 1). Most of these expression differences (63/82) were identified in the resistant interaction, whereas only 19 TF genes were differentially expressed in the susceptible interaction. Interestingly, only one TF gene, C2H2 (Zn) (AC135797_13.2), showed differential expression in both interactions, but with a stronger induction in the resistant interaction (Table 1).

Figure 1.

 Venn diagrams of the numbers of transcription factor genes induced or repressed on bluegreen aphid infestation in Medicago truncatula susceptible A17 and resistant Jester with ratios of ≤ 0.5 (repressed) or ≥ 2 (induced) and P ≤ 0.05 after aphid infestation for 6 h (a) or 12 h (b) based on three biological replicates. A complete list of genes, including the ratios and P values, is shown in Table 1. *Number of genes induced or repressed at 6 h which were also induced or repressed at 12 h. †Genes were overlapping between groups.

Table 1.   Transcription factor (TF) genes responding to bluegreen aphid infestation in Medicago truncatula susceptible A17 and resistant Jester
TF familyGene accessionSusceptible A17Resistant Jester
6 h12 h6 h12 h
P valueRatioP valueRatioP valueRatioP valueRatio
  1. Values shown indicate the P values from t-test and the ratio of the average relative expression in infested to noninfested control plants for each gene. Data were from three biological replicates.

  2. Only genes with P ≤ 0.05 and fold change ≤ 0.5 (for repressed, underlined) or ≥ 2 (for induced, in bold) are considered.

  3. *Genes chosen for further study. As a result of the changing nature of the tentative consensus annotation, members of a gene family have not been annotated in detail; however, for the purpose of this article, gene family members which were studied in detail have been numbered to differentiate individual genes.

ABI3-VP1AC148404_7.110.080.510.572.280.691.370.040.46
ABI3-VP1AC143338_28.10.441.600.030.420.450.250.371.55
AP2/EREBPAC144591_8.2910.581.220.270.740.441.280.020.40
AP2/EREBPAC148348_6.1810.173.260.810.870.053.100.791.21
AP2/EREBPAC148348_6.20.860.800.4231.720.013.970.120.01
AP2/EREBPAC151527_13.10.421.220.490.860.002.920.550.82
ARFAC144483_9.10.681.120.411.350.430.780.010.42
ARFAC140023_13.2510.590.900.051.930.560.700.002.20
AUX/IAAAC148405_5.50.590.840.080.690.721.090.050.36
bHLH-1*AC148775_2.10.850.910.215.090.0032.030.048.76
bHLHAC130800_18.20.330.080.150.490.830.880.023.09
bHLH-2*AC123547_15.50.911.010.392.050.004.680.042.20
bHLHAC148762_23.20.161.420.230.520.150.620.020.23
bHLHAC144431_26.20.770.850.000.320.980.990.010.10
BTB/POZAC144517_7.80.531.220.271.280.360.730.030.41
BTB/POZAC136503_46.20.180.470.800.900.140.070.000.21
bZIPAC136504_23.10.381.160.060.700.070.830.010.23
bZIPAC144805_7.10.291.680.020.250.940.960.250.22
bZIPAC140545_26.20.191.940.750.890.010.540.000.31
bZIPAC140034_1.20.251.380.831.090.140.450.000.25
bZIPAC146721_7.10.571.100.390.550.901.030.020.31
C2C2 (Zn)AC136450_7.91.001.000.710.830.570.800.030.25
C2H2 (Zn)AC137995_31.10.581.200.710.890.012.650.004.65
C2H2 (Zn)AC149210_2.60.971.030.581.660.004.690.581.17
C2H2 (Zn)AC125368_6.20.342.450.420.110.471.650.033.08
C2H2 (Zn)AC135797_13.20.002.710.272.940.0016.050.014.62
C2H2 (Zn)AC139344_17.20.621.370.580.850.510.560.010.35
C2H2 (Zn)AC140671_11.10.281.520.111.830.004.170.004.62
C2H2 (Zn)AC146683_9.10.601.260.530.330.022.080.061.91
C2H2 (Zn)AC119418_5.80.400.110.810.760.140.080.040.21
C2H2 (Zn)AC148485_10.20.521.550.050.140.131.620.480.71
C3HAC151527_13.50.490.620.152.110.910.960.030.31
CCHC (Zn)AC126778_8.1210.740.960.541.500.013.170.900.96
CCHC (Zn)AC144538_22.10.681.140.730.750.012.300.241.61
CCHC (Zn)AC140551_7.80.100.030.429.000.541.320.040.30
CCHC (Zn)CT573029_22.20.401.520.357.210.0022.750.0210.79
CCHC (Zn)AC121233_16.1710.461.130.010.390.791.060.590.81
GRASAC152185_1.3210.521.110.000.480.400.830.050.63
GRASAC144539_7.90.891.120.890.870.830.870.030.45
GRFAC141113_5.10.951.050.010.350.052.760.731.14
HD familyAC136974_11.10.731.140.330.600.590.850.030.36
HD familyAC149494_4.10.900.950.041.850.420.850.012.93
HD familyAC141864_5.10.731.220.960.980.651.210.050.38
HD-likeAC124958_13.20.601.550.030.180.160.460.650.78
HD-likeAC147009_5.30.511.380.412.170.022.490.521.25
HD-likeAC148915_2.80.561.200.040.440.650.890.720.85
HD-likeAC149803_7.20.240.310.020.280.491.220.580.90
HSFAC146585_18.20.271.070.301.580.021.530.012.27
HTHAC146743_11.2310.391.250.020.490.621.040.180.57
HTHAC135797_13.30.540.810.070.490.231.790.012.33
HTHAC138453_8.70.242.110.980.990.170.500.010.31
LIMAC125482_8.1710.501.140.300.740.410.850.050.45
LRRAC146720_15.20.550.560.630.400.0112.670.210.27
MADSAC144517_41.20.081.340.540.790.601.120.040.44
MADSAC135316_15.20.281.360.320.130.020.190.240.64
MYB/HD-likeAC126006_21.20.251.200.030.450.950.980.080.35
MYB/HD-likeAC135465_23.10.931.010.651.140.740.880.022.01
MYB/HD-likeAC140030_6.20.921.030.720.890.670.850.030.26
MYB/HD-likeAC136840_24.20.061.620.060.560.270.760.050.18
MYB/HD-likeAC149803_7.30.381.260.110.780.590.820.010.37
MYB/HD-likeAC146842_9.50.424.940.010.120.900.900.390.11
MYB/HD-likeAC149642_2.70.381.280.010.460.220.650.060.33
MYB/HD-likeAC149493_7.60.351.530.341.280.005.010.152.48
MYB/HD-likeAC124961_18.10.023.590.661.370.500.690.390.24
MYB/HD-likeAC124961_19.20.670.560.690.560.390.050.042.16
MYB/HD-likeAC136955_16.20.130.720.040.480.970.980.481.80
NACAC126789_19.50.731.170.951.020.220.740.000.43
NACAC131026_11.10.413.700.860.900.010.460.200.47
NACAC141112_7.20.960.980.071.540.143.870.003.54
p53-likeAC133863_33.30.052.400.590.340.820.850.870.84
PHDAC155099_1.20.680.800.020.480.250.530.730.86
PHDAC144541_12.80.042.230.460.670.791.180.350.64
RCC1AC139708_3.20.271.430.501.210.013.170.035.68
RRAC121233_16.10.041.710.340.890.420.850.010.50
SBPAC151915_8.20.831.060.170.770.050.760.020.46
WRKYAC122726_21.1110.900.980.810.910.012.950.054.81
WRKY-1*AC151481_14.10.215.430.452.110.049.240.0028.04
WRKYAC150799_4.10.811.070.411.860.263.210.013.57
WRKY-2*AC137827_18.1610.061.730.422.220.0133.100.0029.22
ZF DHHCAC146747_7.1010.141.360.140.640.042.610.181.37

Of the 19 genes with altered transcript levels in the susceptible interaction, four genes from four separate gene families exhibited upregulation at 6 h, albeit with only a c. 2.3–3.6-fold increase in aphid-infested plants compared with noninfested control plants. No upregulation of TF genes caused by aphid infestation was observed in the susceptible interaction at the 12-h time point. The remaining 15 genes from 11 different gene families showed downregulation and, in every case, not until 12 h after aphid infestation, with the HD-like family (3/15) and MYB/HD-like family (4/15) of TFs most represented among this class of differentially expressed genes.

By contrast, in the resistant interaction, 22 genes from 13 families were upregulated at 6 h, and only two genes were downregulated in response to BGA infestation. At 12 h, 20 genes from 11 families were upregulated and 29 genes from 18 families were downregulated (Table 1). Overall at 6 or 12 h, 32 genes from 17 families were upregulated in the resistant interaction following aphid infestation with a relatively high proportion (16/32) of these belonging to the AP2/EREBP, bHLH, C2H2 (Zn) and WRKY families. For some gene members, the transcript levels were consistently higher (up to c. 33-fold) in the aphid-infested than noninfested control plant at both time points, for example bHLH gene members (bHLH-1, AC148775_2.1; bHLH-2, AC123547_15.5) and WRKY gene members (WRKY-1, AC151481_4.1; WRKY-2, AC137827_18.161). We decided to focus intially on these four TFs, as the bHLH gene family appeared to be specifically overpresented in this M. truncatula–BGA resistant interaction, and members of the WRKY gene family have been found to be key regulators of other biotic stress responses. The results of the detailed studies for these four genes are shown in Figs 2–7 and Table 2.

Table 2.   Summary of the transcript changes of the two bHLH and two WRKY genes in the Medicago truncatula lines in response to bluegreen and pea aphid infestations and to methyl jasmonate (MeJA) treatment or mechanical wounding
Genes1Time point2Bluegreen aphidMeJAMechanical woundingPea aphid
A17 (S)Jester (R)Cyprus (S)Caliph (R)Borung (S)Mogul (R)A17 (S)Jester (R)A17 (S)Jester (R)A17 (S)Jester (R)
  1. 1bHLH-1 (AC148775_2.1); bHLH-2 (AC123547_15.5); WRKY-1 (AC151481_4.1); WRKY-2 (AC137827_18.161).

  2. 26 h and 12 h for aphid infestation; 3 h and 12 h for MeJA treatment; 1 h and 6 h for mechanical wounding (see the Results section for details of the choices of time points).

  3. S, bluegreen aphid susceptible.

  4. R, bluegreen aphid resistant.

  5. −, no significant change in transcription level in the plant at the defined time point with the specific treatment when compared with its nontreated control.

  6. +, gene transcription level was significantly higher in the plant at the defined time point with the specific treatment compared with its nontreated control (P < 0.05).

  7. *Significantly higher transcription level in aphid-resistant plants compared with susceptible plants at the defined time point with the specific treatment (P < 0.05).

bHLH-1First+++*+++
Second+++++
bHLH-2First+++
Second+++++++*
WRKY-1First+++*+
Second+++++*
WRKY-2First++++++++*
Second++++++++*

Close association between AKR-mediated aphid resistance and the induction of the bHLH and WRKY gene members

To further examine the link between AKR-mediated aphid resistance and the induction of the two bHLH gene members (bHLH-1 and bHLH-2) and two WRKY gene members (WRKY-1 and WRKY-2), the transcript levels of these genes were measured following BGA infestation in two additional genotypes, A17AKR+ and JesterAKR−, which were specifically developed for molecular genetic analysis of BGA resistance (Gao et al., 2007a). Line A17AKR+ contains AKR and only differs from the susceptible line A17 by c. 1%, whereas JesterAKR− no longer contains AKR and only differs from the resistant line Jester by c. 1%. A17 and Jester were also included to allow comparison of the results with those of the TF profiling experiment described above. A17AKR+ has a degree of resistance to BGA similar to that of Jester, and JesterAKR− and A17 have a similar degree of susceptibility to BGA (Gao et al., 2007a). As shown in Fig. 2, the relative transcript levels of the four TF genes were significantly higher in BGA-infested plants containing the AKR gene (A17AKR+, Jester) compared with noninfested plants at both 6 and 12 h, consistent with the results described above and shown in Table 1. There was no significant (P > 0.05) difference between transcript levels for these genes in BGA-infested susceptible A17 and JesterAKR− compared with noninfested controls at both 6 and 12 h. Similarly, no significant difference in transcript levels for any of the four genes was observed between BGA-infested resistant Jester and A17AKR+. These results indicate that the activation of these bHLH and WRKY gene members by BGA is highly correlated with the presence of the AKR resistance gene.

Figure 2.

 Transcript levels of members of the bHLH and WRKY gene families in petioles of Medicago truncatula A17, A17AKR+, Jester and JesterAKR− at 6 and 12 h following bluegreen aphid infestation (hatched bars) or without infestation (open bars). (a) bHLH-1; (b) bHLH-2; (c) WRKY-1; (d) WRKY-2. The relative transcript abundance of each gene was normalized to a Tubulin gene, and calculated using the formula 2−ΔCT, where ΔCT represents the threshold cycle value (CT) of the gene of interest minus the CT value of Tubulin. Values are the mean and standard error of three biological replicates. *Significantly higher transcription level in aphid-infested plants compared with noninfested control plants of the same M. truncatula line at the defined time point based on one-way ANOVA (P < 0.05).

Induction of bHLH and WRKY genes in other genetic backgrounds containing BGA resistance

To further investigate the links between BGA resistance and the activation of the bHLH and WRKY genes, aphid-induced expression of these four TF genes was measured in two other BGA-resistant lines of M. truncatula, both genetically quite distinct from A17/Jester, and their respective, susceptible, closely related partners (resistant line Caliph vs susceptible line Cyprus, and resistant line Mogul vs susceptible line Borung). The transcript levels for the WRKY and bHLH genes increased specifically in the resistant line Caliph at 12 h following BGA infestation, with the exception of bHLH-1, whose transcript level also increased in both the BGA-infested susceptible line Cyprus and resistant line Caliph at 6 h (Fig. 3). Increases in the transcript levels of all four genes were also observed in the resistant line Mogul following BGA infestation, but not in its susceptible, closely related partner Borung. However, these genes were only induced significantly early at 6 h following BGA infestation and were no longer induced at 12 h (Fig. 4).

Figure 3.

 Transcript levels of members of the bHLH and WRKY gene families in petioles of Medicago truncatula susceptible Cyprus and resistant Caliph at 6 and 12 h following bluegreen aphid infestation (hatched bars) or without infestation (open bars). (a) bHLH-1; (b) bHLH-2; (c) WRKY-1; (d) WRKY-2. The relative transcript abundance of each gene was normalized to a Ubiquitin gene, and calculated using the formula 2−ΔCT, where ΔCT represents the threshold cycle value (CT) of the gene of interest minus the CT value of Ubiquitin. Values are the mean and standard error of three biological replicates. *Significantly higher transcription level in aphid-infested plants compared with noninfested control plants of the same M. truncatula line at the defined time point based on one-way ANOVA (P < 0.05).

Figure 4.

 Transcript levels of members of the bHLH and WRKY gene families in petioles of Medicago truncatula susceptible Borung and resistant Mogul at 6 and 12 h following bluegreen aphid infestation (hatched bars) or without infestation (open bars). (a) bHLH-1; (b) bHLH-2; (c) WRKY-1; (d) WRKY-2. The relative transcript abundance was calculated as described for Fig. 2 and values are the mean and standard error of three biological replicates. *Significantly higher transcription level in aphid-infested plants compared with noninfested control plants of the same M. truncatula line at the defined time point based on one-way ANOVA (P < 0.05).

Differential responses of bHLH and WRKY genes to MeJA treatment

Our previous study showed that the genes involved in the octadecanoid pathway were exclusively or predominantly induced in the resistant line Jester following BGA infestation (Gao et al., 2007a). This raises the question as to whether the TFs are responding indirectly to aphid-induced JA or its analogues in Jester, rather than responding directly to aphid feeding. To address this question, the transcript levels of these four genes were measured in both A17 and Jester in response to MeJA treatment at 3 and 12 h, an early and late response to MeJA treatment. With the exception of WRKY-2, whose transcript levels increased in both A17 and Jester plants following MeJA treatment at 12 h, the transcript levels of the three other TF genes were not significantly (P > 0.05) different between MeJA-treated and untreated control A17 or Jester plants at either time point (Fig. 5).

Figure 5.

 Transcript levels of members of the bHLH and WRKY gene families in Medicago truncatula A17 and Jester at 3 and 12 h following methyl jasmonate (MeJA) treatment (hatched bars) or without treatment (open bars). (a) bHLH-1; (b) bHLH-2; (c) WRKY-1; (d) WRKY-2. The relative transcript abundance of each gene was normalized to a Tubulin gene, and calculated using the formula 2−ΔCT, where ΔCT represents the threshold cycle value (CT) of the gene of interest minus the CT value of Tubulin. Values are the mean and standard error of three biological replicates. *Significantly higher transcription level in aphid-infested plants compared with uninfested control plants of the same M. truncatula line at the defined time point based on one-way ANOVA (P < 0.05).

Responses of bHLH and WRKY genes to mechanical wounding

Mechanical wounding is also able to cause the induction of the octadecanoid pathway, and so the transcript levels of the four aphid-responsive TF genes were also measured in both A17 and Jester, 1 or 6 h following wounding. Based on our observation using the electronic penetration graph technique (Tjallingii, 1978), BGAs usually begin phloem ingestion 4–5 h after being caged on the plants. Therefore, the 1- and 6-h time points after wounding were approximately equivalent to 6 and 12 h after the infestation of aphids. Plant responses to mechanical wounding have been well characterized in tomato and Arabidopsis, in which the induction of JA-related genes is typical on mechanical wounding (Schilmiller & Howe, 2005; Abraham et al., 2009). To establish the wounding responses in M. truncatula, the transcription levels of the genes involved in JA biosynthesis and JA-responsive genes, such as genes encoding lipoxygenase, proteinase inhibitor and vegetative storage protein (Gao et al., 2007a), were analysed following wounding treatment (see the Materials and Methods section). A significantly higher transcription level of these genes was observed at 6 h in both A17 and Jester plants on wounding treatment compared with the nontreated control plants (Fig. S1). The induction of these JA-related genes suggests the effectiveness of wounding treatment, although the change in JA-related gene transcription was not seen at the early 1-h time point after wounding (results not shown).

With the exception of WRKY-1, whose transcript levels did not change significantly (P > 0.05) in either genotype and at either time point on wounding treatment, the transcript levels of the three other TF genes were significantly (P < 0.05) higher in wounded than in untreated control plants (Fig. 6). Although bHLH-1 was significantly induced in both A17 and Jester at both 1 and 6 h, with higher transcript levels at 1 than 6 h, for bHLH-2, the transcript level was higher only at 6 h. For WRKY-2, the increase occurred at both 1 and 6 h, and the transcript level was higher at 6 than 1 h after treatment (Fig. 6d). For all four genes, there was no significant (P > 0.05) difference between A17 and Jester with or without treatment at either 1- or 6-h time points.

Figure 6.

 Transcript levels of members of the bHLH and WRKY gene families in Medicago truncatula A17 and Jester at 1 and 6 h following mechanical wounding treatment (hatched bars) or without treatment (open bars). (a) bHLH-1; (b) bHLH-2; (c) WRKY-1; (d) WRKY-2. The relative transcript abundance was calculated as described for Fig. 2 and values are the mean and standard error of three biological replicates. *Significantly higher transcription level in aphid-infested plants compared with noninfested control plants of the same M. truncatula line at the defined time point based on one-way ANOVA (P < 0.05).

Induction of bHLH and WRKY gene members by PA

Previously, we have identified resistance to PA in the M. truncatula line Jester (Gao et al., 2008). The biology of resistance to the two aphid species shared similarity, with resistance in both cases occurring at the level of the phloem, requiring an intact plant and involving a combination of antixenosis, antibiosis and plant tolerance (Gao et al., 2008). Despite these obvious similarities, resistance to these two aphid species is controlled by two independent resistance genes in Jester (Guo et al., 2009). To examine whether the TF genes were activated in response to PA, the transcript levels of the two bHLH genes and two WRKY genes were measured in A17 and Jester after PA infestation.

The expression patterns of the four TF genes after PA infestation showed more variation than was observed after BGA infestation (Fig. 7). The transcript levels of the two bHLH genes increased in the resistant line Jester following PA infestation. For bHLH-1, the increase occurred at 12 h and for bHLH-2 at both 6 and 12 h (Fig. 7a,b). By contrast, the transcript levels of the two WRKY genes increased in both susceptible and resistant plants following PA infestation (Fig. 7c,d). However, for WRKY-1, the transcript levels increased at both 6 and 12 h in Jester, whereas, in A17, an increase in the transcript level was only observed at 12 h (Fig. 7c). For both WRKY genes, although the transcript levels increased in both susceptible and resistant plants following PA infestation, they were significantly (P < 0.05) higher in the infested resistant Jester than the susceptible A17 plants, when compared at each time point (Fig. 7c,d).

Figure 7.

 Transcript levels of members of the bHLH and WRKY gene families in petioles of Medicago truncatula susceptible A17 and resistant Jester at 6 and 12 h following pea aphid infestation (hatched bars) or without infestation (open bars). (a) bHLH-1; (b) bHLH-2; (c) WRKY-1; (d) WRKY-2. The relative transcript abundance was calculated as described for Fig. 2 and the values are the mean and standard error of three biological replicates. *Significantly higher transcription level in aphid-infested plants compared with noninfested control plants of the same M. truncatula line at the defined time point based on one-way ANOVA (P < 0.05).

Discussion

This article provides an overview of the TF genes expressed during the early stages of BGA infestation in M. truncatula susceptible and resistant plants. Eighty TF genes belonging to 30 TF families were responsive to BGA infestation (Table 1). This large number of TF genes may reflect the complexity of aphid defence regulation and a broad spectrum of transcriptional reprogramming in the plants during the early stages of aphid infestation. Some of these TFs may be key regulators of aphid defence, and these may have gone undetected if relatively low-sensitive microarray techniques had been used in this study. Indeed, several microarray studies, mostly involving compatible plant–aphid interactions, have failed to identify or have identified a much smaller number of TFs in response to aphid infestation, although some of these were taken at later time points than in our study (Moran et al., 2002; Levy et al., 2005; Park et al., 2006; Couldridge et al., 2007; van Oosten, 2007; Kusnierczyk et al., 2008; Li et al., 2008).

TF profiling revealed different patterns of TF gene expression in M. truncatula during susceptible and resistant interactions with BGA (Table 1). The lack of transcriptional changes of TFs during the initial infestation (at 6 h) and the subsequent suppression of TFs during aphid feeding (at 12 h) in susceptible A17 plants provide further evidence at the molecular level that aphids may be able to avoid or suppress plant responses during the early infestation process in susceptible interactions (Walling, 2008). By contrast, there were a significantly larger number of TF genes differentially expressed in resistant Jester plants than in susceptible A17 plants on BGA infestation (61 genes in Jester vs 19 in A17). A significantly larger number of genes were also seen to be differentially expressed in resistant than susceptible soybean cultivars (not closely related) when the transcription profiling of 18 000 soybean genes was analysed at the same time points (at 6 and 12 h) on soybean aphid infestation (Li et al., 2008). This trend was also observed in incompatible and compatible interactions of soybean with avirulent and virulent strains of Pseudomonas syringae and other plant–pathogen systems (Zou et al., 2005; Klink et al., 2007). This suggests that resistant plants are able to trigger rapid and massive transcriptional reprogramming in order to hamper further infestation by aphids, which may be similar to plant defence against microbial pathogens. In the M. truncatula–BGA interaction, the responses of TF genes in resistant plants were as early as 6 h and were before the induction of defence-related genes, which occurred later than 12 h following aphid infestation (Gao et al., 2007a). This suggests that these TFs are among the early aphid-responsive genes in resistant plants. Some of the upregulated TFs are likely to represent key regulators that mediate the network of gene expression required for successful aphid resistance.

Among the suppressed TF genes in the BGA susceptible interaction, members of the HD-like and MYB/HD-like families showed relatively high representation (7/15). Downregulation of a MYB gene was also seen in tobacco (Nicotiana attenuata) after infestation by a chewing insect Manduca sexta (Hui et al., 2003). MYB/HD-like TF genes have been shown to have highly diversified biological functions (Jin & Martin, 1999; Yanhui et al., 2006), although there is relatively little information available with regard to their function in plant defence. In Arabidopsis, MYB family members have been shown to be responsive to plant defence elicitors and bacterial/fungal pathogens, whereas overexpression of a MYB gene increased plant resistance to a chewing insect, the fall armyworm (Spodoptera frugiperda), and a knockout of AtMYB102 resulted in enhanced susceptibility to the white cabbage butterfly (Pieris rapae) (Johnson & Dowd, 2004; McGrath et al., 2005; De Vos et al., 2006; Libault et al., 2007). In tobacco, a MYB gene has been shown to play a part in defence against tobacco mosaic virus (Hui et al., 2003). Other studies have suggested that the MYB family is one of the major TF families that regulate genes related to photosynthesis and metabolites (Grotewold, 2005; Allan et al., 2008; Malitsky et al., 2008; Saibo et al., 2009). It remains to be determined whether the downregulation of the MYB/HD-like genes observed in BGA-infested susceptible A17 plants is part of the aphid’s manipulation of the plant’s defence response to establish a successful feeding site. Alternatively, the downregulation of TF genes may play a role in the reduction of the photosynthetic rate and the suppression of genes related to photosynthesis, as shown in microarray studies involving compatible plant–aphid interactions (Macedo et al., 2003; Voelckel et al., 2004; Zhu-Salzman et al., 2004; Qubbaj et al., 2005; Botha et al., 2006; Eulgem, 2006; Macedo et al., 2009).

Among the upregulated TF genes in the resistant interaction, four gene families, AP2/EREBP, bHLH, C2H2 (Zn) and WRKY, showed relatively high representation (16/32), and the members from these TF families were exclusively induced in the resistant aphid interaction. Although transcription profiling studies of TF genes involved in R gene-mediated resistance have not been reported, the overrepresentation of the bHLH family in the resistant M. truncatula–BGA interaction appears to be distinct from the transcription signatures of TFs associated with plant defence responses in other studies (McGrath et al., 2005; Libault et al., 2007; Naoumkina et al., 2008). Although the specific members from each TF family varied depending on the interaction, the overrepresentation of AP2/EREBP and WRKY families among the upregulated TF genes was commonly observed in Arabidopsis plants in response to a fungal elicitor (e.g. chitin), a fungal pathogen Alternaria brassicicola and the plant hormone MeJA, as well as in M. truncatula cell suspension in response to a yeast elicitor (McGrath et al., 2005; Libault et al., 2007; Naoumkina et al., 2008). The C2H2 (Zn) family was among the overrepresented TF families in Arabidopsis in response to chitin (Libault et al., 2007). Members of AP2/EREBP, WRKY and C2H2 (Zn) families have also been shown to regulate plant defence responses against other biotic and abiotic stresses (Singh et al., 2002; Gutterson & Reuber, 2004; Eulgem, 2005; Qu & Zhu, 2006; Eulgem & Somssich, 2007).

Further evidence for the involvement of the MtWRKY and MtbHLH gene families more generally in aphid resistance was obtained by looking at additional independent interactions. The higher and/or exclusive induction of these genes was consistently observed in two other BGA-resistant M. truncatula cvs Caliph and Mogul, although the timing and magnitude of the induction varied depending on the genetic background (Figs 3a,b, 4a,b). This suggests that the induction of these MtbHLH and MtWRKY genes is probably linked to the R gene-mediated resistance against BGA. Our results (Fig. 2) on the M. truncatula line A17AKR+ (contains AKR and only differs from the susceptible line A17 by c. 1%) and JesterAKR− (no longer contains AKR and only differs from the resistant line Jester by c. 1%) further support this hypothesis. Members of the WRKY and bHLH families have also been implicated in the R gene-mediated resistance in soybean against soybean aphid (Li et al., 2008). In Arabidopsis, AtWRKY70 is required for both basal defence and full R gene (RPP4)-mediated disease resistance against the oomycete pathogen, Hyaloperonospora parasitica (Knoth et al., 2007).

Earlier and/or higher induction of the four TF genes was also seen in response to infestation by a different aphid species, PA (Fig. 7). In Jester, resistance to PA is mediated by a different gene, APR, than that mediating resistance to BGA (Gao et al., 2008; Guo et al., 2009; L. G. Kamphius, S. Guo & K. B. Singh, unpublished). The combined results for BGA and PA support a general role for bHLH and WRKY TFs in the mediation of aphid resistance, at least in M. truncatula. It is interesting to note that PA infestation induces these TF genes but, unlike the BGA response in Jester, the interaction does not seem to induce JA signalling (Gao et al., 2008). This suggests that these MtWRKY and MtbHLH members may act near the top of the JA signalling cascade of R gene-mediated aphid resistance or, alternatively, may function independently of the JA pathway. Consistent with this hypothesis was the observation that three of the four TFs examined did not respond to JA treatment, and therefore may act upstream of the octadecanoid pathway previously associated with BGA resistance in Jester (Gao et al., 2007a).

The specific function of the MtWRKY and MtbHLH genes in aphid resistance is yet to be characterized. Currently, there is a lack of information on the function of bHLH genes in relation to plant defence, whereas the members of the WRKY gene family have been relatively well studied, mostly in Arabidopsis (Eulgem et al., 2000; Toledo-Ortiz et al., 2003; Eulgem, 2005, 2006; X. Li et al., 2006; Eulgem & Somssich, 2007). Based on sequence similarity, MtWRKY (WRKY-1) and MtWRKY (WRKY-2) fall into Arabidopsis WRKY groups III and IIc, respectively (Eulgem et al., 2000). Interestingly, the three soybean WRKY genes which were induced by soybean aphid in a resistant cultivar, but not in a susceptible cultivar, also fall into the Arabidopsis WRKY group IIc (Li et al., 2008). Rice WRKY89 is a member of AtWRKY group III and has been implicated in resistance to the phloem-feeding white-backed leafhopper (Wang et al., 2007). All of these are in different Arabidopsis WRKY groups than tobacco WRKY3 and WRKY6 (WRKY group I), the silencing of which leads to increased susceptibility of tobacco plants to the chewing insect Manduca sexta via the JA signalling pathway (Skibbe et al., 2008). Thus, our WRKY TF results are consistent with the hypothesis that sucking insect defence resembles more closely that of pathogens rather than chewing insects (Walling, 2008). The discovery of TF genes associated with aphid resistance is important because they are likely to orchestrate genome-wide changes in transcription that lead to full R gene-mediated aphid resistance. For this reason, they are important targets for optimizing plant protection to aphids in agriculture.

Acknowledgements

We thank Heping Han and Elaine Smith, CSIRO Plant Industry, for technical support and Drs Armin Schlereth and Thomas Ott, Max-Planck Institute of Molecular Plant Physiology, for valuable discussions on the data analysis. This work was supported, in part, by the Department of Education, Science and Training (DEST) in Australia to K. B. S. and by the EU-FP6 Grain Legumes Integrated Project (FOOD-CT-2004-506223) to M. K. V.. L. G. K. is the recipient of a CSIRO/OCE Post-Doctoral Fellowship.

Ancillary