• Open Access

AtCDKA;1 silencing in Arabidopsis thaliana reduces reproduction of sedentary plant-parasitic nematodes

Authors

  • Elke Van de Cappelle,

    1. Molecular Biotechnology Department, Faculty of Bioscience Engineering, Ghent University, Coupure links 653, B-9000 Ghent, Belgium
    Search for more papers by this author
  • Eva Plovie,

    1. Molecular Biotechnology Department, Faculty of Bioscience Engineering, Ghent University, Coupure links 653, B-9000 Ghent, Belgium
    Search for more papers by this author
    • Present address: Cardiovascular Research Center, Massachusetts General Hospital, Harvard Medical School, Charlestown, MA 02129, USA

  • Tina Kyndt,

    1. Molecular Biotechnology Department, Faculty of Bioscience Engineering, Ghent University, Coupure links 653, B-9000 Ghent, Belgium
    Search for more papers by this author
  • Wim Grunewald,

    1. Molecular Biotechnology Department, Faculty of Bioscience Engineering, Ghent University, Coupure links 653, B-9000 Ghent, Belgium
    2. VIB Department of Plant Systems Biology, Ghent University, Technologiepark 927, 9052 Ghent, Belgium
    Search for more papers by this author
  • Bernard Cannoot,

    1. Molecular Biotechnology Department, Faculty of Bioscience Engineering, Ghent University, Coupure links 653, B-9000 Ghent, Belgium
    Search for more papers by this author
  • Godelieve Gheysen

    Corresponding author
    1. Molecular Biotechnology Department, Faculty of Bioscience Engineering, Ghent University, Coupure links 653, B-9000 Ghent, Belgium
    Search for more papers by this author

* Correspondence (fax +3292646219; e-mail godelieve.gheysen@ugent.be)

Summary

The activity of the Arabidopsis thaliana cyclin-dependent kinase AtCDKA;1 is important throughout G1/S and G2/M transitions and guarantees the progression of the cell cycle. Inhibitor studies have shown that activation of the cell cycle is important for the development of nematode feeding sites. The aim of this study was to silence the expression of the AtCDKA;1 gene in nematode feeding sites to interfere with their development. Therefore, sense and antisense constructs were made for the AtCDKA;1 gene and fused to a nematode-inducible promoter which was activated in nematode feeding sites at an earlier time point than AtCDKA;1. Two transgenic A. thaliana lines (S266 and S306) containing inverted repeats of the AtCDKA;1 gene and with reduced AtCDKA;1 expression in seedlings and galls were analysed in more detail. When the lines were infected with the root-knot nematode Meloidogyne incognita, significantly fewer galls and egg masses developed on the roots of the transgenic than wild-type plants. Infection of the AtCDKA;1-silenced lines with Heterodera schachtii resulted in significantly fewer cysts compared with controls. The S266 and S306 lines showed no phenotypic aberrations in root morphology, and analysis at different time points after infection demonstrated that the number of penetrating nematodes was the same, but fewer nematodes developed to maturity in the silenced lines. In conclusion, our results demonstrate that silencing of CDKA;1 can be used as a strategy to produce transgenic plants less susceptible to plant-parasitic nematodes.

Introduction

Sedentary endoparasites, such as the root-knot nematode Meloidogyne incognita and the cyst nematode Heterodera schachtii, establish permanent feeding sites (galls and syncytia, respectively) in root tissue, thereby damaging plants by consuming large amounts of energy and nutrients from the roots. These complex multinucleated nematode feeding sites (NFSs) are metabolically highly active and are adapted to serve as a nutrient source for the nematode. Giant cells, induced by root-knot nematodes, are generated through sequential mitoses without cytokinesis (acytokinetic mitosis) possibly followed at later stages by endoreduplication. The root cells surrounding the giant cells swell, divide and give rise to a root knot or gall (Jones, 1981; Bleve-Zacheo and Melillo, 1997). Giant cells also have small vacuoles and a proliferation of smooth endoplasmic reticulum, ribosomes, mitochondria and plastids. In syncytia, formed by cyst nematodes, no mitosis can be observed. However, the enlargement of the nuclei implies that DNA synthesis occurs during and after the incorporation of new cells into the syncytium by dissolving cell walls, leading to multinucleated cells (Endo, 1964).

Many plant genes have been identified that are differently expressed in feeding sites than in normal root cells (for a review, see Gheysen and Fenoll, 2002). Of the different classes of genes that are up-regulated in NFSs, cell cycle genes have been studied extensively (Niebel et al., 1996; De Almeida-Engler et al., 1999; Goverse et al., 2000). The expression patterns of two cyclin-dependent kinases (AtCDKA;1 and AtCDKB1;1) and two mitotic cyclins (AtCycA2;1 and AtCycB1;1) have been analysed by promoter β-glucuronidase (GUS) fusions and in situ hybridization in Arabidopsis thaliana (Niebel et al., 1996; De Almeida-Engler et al., 1999). Within the first hours of feeding by root-knot nematodes and cyst nematodes, these four genes are strongly expressed in feeding cells before any cell divisions are visible. This suggests a rapid activation of the cell cycle, and is most probably induced by an initial stimulus from the nematode.

Experiments with cell cycle blockers have confirmed the importance of the cell cycle in feeding site induction and development (De Almeida-Engler et al., 1999). Using hydroxyurea (a cytostatic drug acting as an inhibitor of DNA synthesis by blocking the G1/S transition), early giant cell and syncytium development were completely inhibited in A. thaliana. Application at later stages had no effect, suggesting an important role for genome multiplication in the early stages of feeding site formation. Application of oryzalin (a drug that inhibits plant microtubule polymerization and arrests cells in the early M phase) in the early stages inhibited the development of root-knot and cyst nematodes. These data indicate that mitotic activity is required for correct feeding cell development (De Almeida-Engler et al., 1999).

As the cell cycle is reactivated in NFSs and, consequently, many cell cycle genes are up-regulated, the specific silencing of these genes in NFSs may be an effective approach to interfere with nematode development. In this article, we focus on AtCDKA;1. Constitutive expression of a dominant negative mutation of the key cell cycle regulator AtCDKA;1 is lethal in Arabidopsis (Hemerly et al., 1995). However, local silencing of AtCDKA;1 early in the development of feeding sites may be a successful approach to obtain a lower susceptibility to sedentary plant-parasitic nematodes. Therefore, a nematode-inducible promoter specific for the feeding site is required so as not to interfere with normal plant development. In this study, the AtWRKY23 promoter was chosen because of its high expression in NFSs compared with other plant tissues.

Results

Gene silencing is efficient in NFSs

Silencing is not equally efficient in different tissues (Andika et al., 2005; G. Marjanac, De Buck en Depicker, VIB Department of Plant Systems Biology, Ghent University, Ghent, Belgium, unpubl. data). Therefore, the efficiency of double-stranded RNA (dsRNA) silencing in NFSs was determined using the GUS reporter gene. The AtWRKY23 promoter (Att0001 in Barthels et al., 1997) was chosen as it is strongly up-regulated in NFSs (Barthels et al., 1997) and shows lower expression in other root and leaf tissues (W. Grunewald et al., unpubl. data). Transgenic lines with AtWRKY23::GUS supertransformed with 35S::GUS hairpin constructs were obtained (G. Marjanac, De Buck en Depicker, unpubl. data) and analysed for GUS expression after nematode infection. The efficiency of GUS suppression in NFSs was evaluated in two pAtWRKY23::GUS/p35S::hpGUS lines (3 and 5). The pAtWRKY23::GUS line was used as a control (Table 1). The pAtWRKY23::GUS line showed GUS activity in NFSs as expected. Very weak expression was observed in only 15.8% of NFSs of the pAtWRKY23::GUS/p35S::hpGUS_3 line (Table 1), whereas the NFSs of the pAtWRKY23::GUS/p35S::hpGUS_5 line were all GUS negative (Figure 1), indicating efficient silencing of GUS in NFSs. These observations reveal that gene expression (using GUS as a reporter) can be successfully silenced in NFSs.

Table 1.  Number of syncytia (Heterodera schachtii) and galls (Meloidogyne incognita) positive or negative for β-glucuronidase (GUS) expression
LineH. schachtiiM. incognitaTotal number of NFSsNFSs GUS+ (%)
GUS+GUS–GUS+GUS–
  1. NFS, nematode feeding site.

pAtWRKY23::GUS85  084 117099.4
pAtWRKY23::GUS/ p35S::hpGUS_322104 01313915.8
pAtWRKY23::GUS/ p35S::hpGUS_5 0 83 0 3 86 0.0
Figure 1.

β-Glucuronidase (GUS) expression in nematode feeding sites (NFSs) under the control of different promoters, and the effect of hairpin silencing on this expression. Nematodes inside the roots were stained pink using acid fuchsin. Top: silencing of GUS by a hairpin construct driven by the AtWRKY23 promoter. GUS-positive syncytium (a) of pAtWRKY23::GUS plants at 4 days after inoculation (DAI). Weak (b) and no (c) GUS staining in syncytium at 4 DAI of pAtWRKY23::GUS/hpGUS. Middle: very early activation of the AtWRKY23 promoter in NFSs. The pAtWRKY23 promoter (d, e, f) drives strong GUS expression in syncytia 14 h after inoculation (HAI) (d) and in giant cells at 22 HAI (e). (f) pAtWRKY23::GUS syncytium at 32 HAI. Bottom: up-regulation of AtCDKA;1 in NFSs. No GUS activity above background driven by the AtCDKA;1 promoter is observed at 24 HAI in NFSs (g), but weak activity is observed from 32 HAI onwards in syncytia (h) and giant cells (i). Bar, 100 µm.

The AtWRKY23 promoter is active in NFSs before the AtCDKA;1 promoter

Cell cycle genes are activated very early during NFS formation. To have a strong silencing impact on the cell cycle, a putative gene silencing construct should be active before or at least simultaneously with the target cell cycle gene. As the AtWRKY23 promoter may be a valuable candidate for the high and specific expression of transgenic constructs in NFSs, pAtWRKY23::GUS and pCDKA;1::GUS seedlings were analysed to compare the time of promoter activation during infection with the cyst nematode H. schachtii or the root-knot nematode M. incognita. At 14, 22, 24, 32, 40 and 48 h after inoculation (HAI), a GUS assay was performed on the roots. This experiment was repeated three times. It was found that, during M. incognita infection, it took at least 22 h for GUS staining driven by the AtWRKY23 promoter to be visible in giant cells. However, blue-stained feeding sites of H. schachtii were found as soon as 14 HAI. By contrast, weak GUS activity driven by the AtCDKA;1 promoter was observed in the vascular cylinder only from 32 HAI onwards (Figure 1).

AtCDKA;1 silencing in A. thaliana reduces nematode infection

To analyse whether the silencing of AtCDKA;1 could interfere with the infection process of plant-parasitic nematodes, A. thaliana was transformed with sense and/or antisense AtCDKA;1 constructs driven by the AtWRKY23 promoter (Figure 2). Gene silencing is generally correlated with the presence of multiple homologous copies organized as inverted repeat structures (De Buck et al., 2001) that can produce dsRNA to trigger gene silencing. In addition to transformation with a sense or antisense construct, co-transformation was also performed to enrich for inverted repeat structures. As a result of the activity of the AtWRKY23 promoter, which is strongest in but not limited to the feeding site (Karimi, 1999; W. Grunewald et al., unpubl. data), transgenic lines with silenced AtCDKA;1 may have an aberrant phenotype. Indeed, several of the generated transgenic lines showed an abnormal root morphology. More particularly, they formed fewer lateral roots than did Col-0 control plants (data not shown). These lines were excluded from further analysis, and only those lines (40 in total) for which no obvious morphological difference could be observed relative to wild-type Col-0 plants were used for infection experiments with the cyst nematode H. schachtii and the root-knot nematode M. incognita. Of the 40 lines transformed with the AtCDKA;1 construct and infected with nematodes, 18 showed a significantly smaller number of infections in one or more tests. Two of the lines, S266 and S306, that repeatedly showed a significantly smaller number of cysts (H. schachtii) and egg masses (M. incognita) (Figure 3) were chosen for a more detailed investigation. Both had a simple 3 : 1 segregation pattern, but Southern analysis revealed that both lines contained multiple linked T-DNAs. These complex integration patterns can cause aberrant RNAs and dsRNA, which can trigger gene silencing (reviewed by Baulcombe, 2004). To investigate whether the decrease in nematode reproduction could be inherited by the next generation, the homozygous F3 lines S266IV and S306I were infected with M. incognita. Both S266IV and S306I plants showed strongly decreased levels of gall formation and egg mass development relative to untransformed plants (P < 0.01) (Figure 4a). Similar results were obtained with H. schachtii infections (Figure 4b), and demonstrate that silencing of CDKA;1 is an efficient strategy to repress plant-parasitic nematode reproduction.

Figure 2.

Schematic representations of the AtCDKA;1 constructs used: (a) sense construct of AtCDKA;1 driven by the AtWRKY23 promoter, with a kanamycin resistance gene (nptII); (b) antisense construct of AtCDKA;1 driven by the AtWRKY23 promoter, with a hygromycin resistance marker (hpt).

Figure 3.

Infection tests with Meloidogyne incognita and Heterodera schachtii on the S266 and S306 F2 lines. (a) Plates were infected with 100 M. incognita second stage juveniles and, 6 weeks after infection, the numbers of egg masses were counted and compared with that of the control line. Six replica plates were counted and error bars indicate the standard error of each group. Significantly different groups are indicated by different letters. (b) Plates were infected with 100 H. schachtii second stage juveniles and, 6 weeks after infection, the numbers of cysts were counted and compared with that of the control line. Six replica plates were counted and error bars indicate the standard error of each group. Significantly different groups are indicated by different letters.

Figure 4.

Infection tests with Meloidogyne incognita and Heterodera schachtii on the homozygous S266IV and S306I lines. (a) Plates were infected with 100 M. incognita second stage juveniles and, 6 weeks after infection, the numbers of egg masses were counted and compared with that of the control line. Six replica plates were counted and error bars indicate the standard error of each group. Significantly different groups (Tukey) are indicated by different letters. (b) Plates were infected with 100 H. schachtii second stage juveniles and, 6 weeks after infection, the numbers of cysts were counted and compared with that of the control line. Six replica plates were counted and error bars indicate the standard error of each group. Significantly different groups are indicated by different letters.

To confirm that the smaller number of mature nematodes in the roots was caused by the inhibition of nematode development, and not related to a smaller number of nematodes penetrating the roots, H. schachtii infection was analysed at different time points. At 2 days after inoculation (DAI), no difference was observed between the different lines: equal numbers of stage 2 juveniles were present in the roots (data not shown). At 9 DAI, a significantly different ratio of developing J3/J4 to vermiform J2 was found in the transgenic lines relative to the controls (Figure 5).

Figure 5.

Ratio of developed (J3/J4) to undeveloped (J2) nematodes at 9 days after inoculation (DAI) with Heterodera schachtii on S266IV and S306I compared with control lines. Plates were infected with 100 second stage juveniles and, at 9 DAI, the numbers of J2 and J3/4 in the roots were counted and compared with that of the control line. Six replica plates were counted and error bars indicate the standard error of each group. Significantly different groups are indicated by different letters.

Expression analysis of AtCDKA;1

Early in the infection process, the AtWRKY23 promoter is strongly up-regulated during initial feeding site formation. However, expression is not restricted to feeding sites, but occurs also in the meristem and vascular tissue (Karimi, 1999; W. Grunewald et al., unpubl. data). This suggests that AtCDKA;1 silencing regulated by the AtWRKY23 promoter should be detectable in uninfected plant tissues by the sensitive quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR) method. In Figure 6, the relative concentrations of AtCDKA;1 mRNA in the transgenic lines to mRNA in the wild-type seedlings are shown. Both transgenic lines (S266IV and S306I) showed reduced expression of AtCDKA;1 in whole seedlings, rosettes and root tissue (data only shown for seedlings). As the silencing effect of the construct was visible in the total plant RNA extract, the silencing of AtCDKA;1 at the initial infection sites is most probably even stronger. Therefore, the expression of the AtCDKA;1 sense construct driven by the AtWRKY23 promoter was analysed more specifically in root galls at 6 DAI. The expression levels of AtCDKA;1 in galls of homozygous S266IV and S306I F3 plants were 44% and 27%, respectively, relative to AtCDKA;1 expression in the galls of wild-type plants (Figure 6).

Figure 6.

Relative levels of AtCDKA;1 expression, determined by quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR), in homozygous silenced F3 lines compared with the wild-type. (a) Analysis of whole seedlings. (b) Analysis of galls dissected at 6 days after inoculation. Significantly different groups are indicated by different letters.

Discussion

A consistent and profound inhibition of expression of endogenous genes, as well as transgenes, by dsRNA has been demonstrated in plants (Smith et al., 2000; Wesley et al., 2001). However, it is not clear whether gene suppression can efficiently silence genes in all types of cells, especially in tissues containing meristematic cells (Andika et al., 2005; G. Marjanac, De Buck en Depicker, unpubl. data). To test whether dsRNA silencing is possible in NFSs, GUS-expressing plants supertransformed with a 35S promoter-driven hairpin GUS gene were infected with juveniles of M. incognita and H. schachtii and, at 4 DAI, GUS expression was analysed. The hairpin GUS construct was able to induce strong GUS silencing in NFSs at 4 DAI. This indicates that it should be possible to silence different genes in NFSs, and suggests that silencing of AtCDKA;1 and other nematode-responsive genes is a feasible strategy.

Strategy to silence AtCDKA;1 expression in NFSs

To reduce AtCDKA;1 expression, constructs were made by fusing sense or antisense AtCDKA;1 to the nematode-inducible AtWRKY23 promoter. The 35S promoter was not an option because of its constitutive expression throughout the plant. Moreover, despite the strong ‘constitutive’ nature of the 35S promoter, it has been reported that it is down-regulated in a large percentage of nematode-induced giant cells and syncytia in different plants (Goddijn et al., 1993; Urwin et al., 1997; Bertioli et al., 1999; Van Poucke et al., 2001). A construct with a dominant negative allele of AtCDKA;1 under the control of the 35S promoter was lethal when introduced into Arabidopsis plants (Hemerly et al., 1995), but the use of a specific At2S2 albumin promoter, which is active during late embryogenesis, to drive the expression of the dominant negative allele allowed the production of transgenic plants (Hemerly et al., 2000). The AtWRKY23 promoter is strongly up-regulated in NFSs and, although AtWRKY23 is also expressed in different plant tissues (Karimi, 1999; W. Grunewald et al., unpubl. data), its stronger and very early expression in NFSs makes the promoter a valuable tool to interfere with feeding site formation. Moreover, during both M. incognita and H. schachtii infection, the AtWRKY23 promoter was activated prior to the AtCDKA;1 promoter, which is a prerequisite for the inhibition of CDKA;1 during the early events of nematode infection. The earliest WRKY23-driven GUS expression was visible at 22 HAI and 14 HAI for M. incognita and H. schachtii, respectively, whereas the CDKA;1 promoter was activated at the earliest at 32 HAI. The difference in activation time of the WRKY23 promoter between M. incognita and H. schachtii can easily be explained by their different migratory behaviour. After penetration of H. schachtii juveniles into the roots by means of their robust stylet and with the help of secreted enzymes, they migrate intracellularly directly to the vascular cylinder. In contrast, M. incognita juveniles migrate through the intercellular space. Not able to pass the Casparian strips of the endodermis, they make a detour to the root tip (Sijmons et al., 1994). Reaching the root tip, they turn around and move into the vascular cylinder. Approximately 24 h after infection of A. thaliana, they stop migrating, induce a giant cell and transform into a parasitic juvenile (Von Mende, 1997).

AtCDKA;1 expression level and nematode infection

Several transgenic A. thaliana lines containing sense, antisense or sense/antisense constructs were generated. The homozygous progeny from two lines, S266 and S306, were investigated more intensively. These lines contain inverted repeat structures of the AtCDKA;1 gene, which can lead to dsRNA formation and the triggering of RNA silencing.

qRT-PCR on S266 and S306 seedlings showed a decrease in the AtCDKA;1 mRNA level relative to that of wild-type plants. An even larger decrease was observed in galls dissected at 6 DAI from these two transgenic lines.

Both lines (S266 and S306) showed normal root morphology, indicating that the remaining lower AtCDKA;1 levels were sufficient for normal root development. When the lines were infected with juveniles of M. incognita, significantly fewer galls developed relative to that in wild-type plants. A smaller number of nematodes fully developed into mature females to produce egg masses. This indicates that reduced levels of AtCDKA;1 interfere with normal feeding site development. Normal root growth in these lines is possibly linked to the lower efficiency of silencing in meristematic tissues.

Nevertheless, it is questionable whether strong silencing can be demonstrated by analysing dissected galls from CDKA;1-silenced plants. If silencing is sufficiently strong, no galls will be formed, and, indeed, fewer galls were observed in the silenced lines. In the galls that are present, silencing is probably not sufficiently strong in these locations to inhibit NFS formation, and so little silencing will be seen.

As cell cycle inhibitor experiments have also shown that cell cycle activation is important for syncytium development, the transgenic lines were also tested with H. schachtii infection. Cyst development at 6 weeks after inoculation and J3/J4 development at 9 DAI were significantly lower in the transgenic lines relative to the controls, but the total number of juveniles at 2 and 9 DAI was similar in all lines.

This demonstrates that both AtCDKA;1-silenced lines are equally well invaded by nematodes as the controls, but that nematodes are inhibited in their development, most probably because of inadequate feeding site development.

These experiments confirm that the cell cycle is important for the establishment of feeding sites, and that silencing an important gene involved in feeding site formation could be a strategy to obtain resistant plants.

Experimental procedures

Transgenic lines

The following lines were obtained from VIB Department of Plant Systems Biology, Ghent University, Ghent, Belgium (G. Marjanac, De Buck en Depicker, unpubl. data): pAtWRKY23::GUS/p35S::hpGUS_3 and pAtWRKY23::GUS/p35S::hpGUS_5. These lines contain a GUS gene expressed from the AtWRKY23 promoter and a hairpin GUS construct expressed from the 35S promoter. The lines were infected with nematodes and subjected to a GUS assay at 4 DAI.

The Att0001 promoter, which is induced very early after infection by root-knot and cyst nematodes (Barthels et al., 1997), was identified by screening of a collection of Arabidopsis lines tagged using root-knot nematode infection, and was later identified as the promoter of the AtWRK23 gene in reverse orientation (Barthels et al., 1997; W. Grunewald et al., unpubl. data).

The 3.95-kb AtWRKY23 promoter was cut from pARM1-aIII (Karimi, 1999) by SmaI and ligated into the pZERO vector (Invitrogen, Carlsbad, CA, USA), which was opened by the EcoRV restriction enzyme.

AtCDKA;1 cDNA was isolated from the pCDC2aAT vector (Ferreira et al., 1991) and combined with the AtWRKY23 promoter. The expression cassette was inserted in sense orientation into the pGSV4/3′nos vector (Hérouart et al., 1993), containing a kanamycin marker gene. The hygromycin marker containing vector pNE/3′nos (De Buck et al., 1999) was used for insertion of the expression cassette in antisense orientation (Figure 2).

All cloning steps were performed according to the protocols described by Sambrook et al. (1989), and the constructs were confirmed by restriction fragment analysis and sequencing.

The binary vectors were transferred to Agrobacterium tumefaciens strain C58C1(pMP90) (Koncz and Schell, 1986) by three-parental mating (Ditta et al., 1980) with Escherichia coli HB101 containing the pRK2013 helper plasmid.

Arabidopsis thaliana (Col-0) plants were transformed by one or two binary vectors by means of the floral dip method (Clough and Bent, 1998). pAtWRKY23::CDKA;1S transgenic plants were obtained on kanamycin (50 mg/L) medium, pAtWRKY23::CDKA;1AS plants on hygromycin (10 mg/L) medium and pAtWRKY23::CDKA;1S-AS plants on both media. The in vitro growth conditions were set under a 16-h light/8-h dark photoperiod at 21 °C on germination medium (Valvekens et al., 1988).

GUS histochemical assay and fuchsin staining

Plants were pre-treated for 30 min with 90% acetone at 4 °C to permeabilize the tissue. The tissue was then washed three times on a shaker for 5 min in phosphate buffer (0.05 m NaPO4, pH 7.2) at room temperature. To avoid dissimilarity of substrate infiltration [charged ferri/ferrocyanide ions infiltrate more slowly in the cytoplasm than in 5-bromo-4-chloro-3-indoyl-β-d-glucuronide (X-gluc)], the samples were equilibrated in GUS pre-incubation solution [phosphate buffer + 0.5 mm K-ferrocyanide (K4[Fe(CN)6]·3H2O) + 0.5–3 mm K-ferricyanide (K3[Fe(CN)6])] at 37 °C. The tissue was incubated in freshly prepared and pre-warmed GUS assay solution (pre-incubation solution + 2 mm X-gluc) for the appropriate time (10 min to overnight) at 37 °C in the dark. The samples were washed three times for 10 min in phosphate buffer to stop the reaction. The tissue was fixed for 2 h in freshly prepared 2.5% glutaraldehyde at room temperature, and washed three times for 20 min with phosphate buffer. Analysis was performed with a DIC stereomicroscope (Nikon, eclipse TE2000, Kanagawa, Japan).

After GUS staining, the nematodes were visualized using acid fuchsin. The infected roots were fixed and stained for 5 h in a solution of equal parts of 95% ethanol and glacial acetic acid, containing 17.5 mg/L acid fuchsin. The root tissue was destained by soaking in a saturated solution of chloral hydrate (4.5 g/mL of H2O) for 16 h. After rinsing the roots with stained nematodes in tap water, they were stored in acidified glycerine (± five drops of 1.0 m HCl to 50 mL of regular glycerine).

Nematode stock

The cyst nematode H. schachtii was propagated on white mustard (Sinapsis alba) roots. The mustard seeds were sterilized (2 min in 70% ethanol and 12 min in 5% NaOCl and 0.1% Tween 20, and washed six times in sterile double-distilled H2O) and germinated on Knop medium (Sijmons et al., 1991). Three weeks later, the plants were inoculated with second stage juveniles (J2). Mature cysts (6–8 weeks old) were harvested and collected in sterile sieves soaked in 3 mm ZnCl2 to hatch at 21 °C.

The root-knot nematode M. incognita was cultivated on the roots of pea (Pisum sativa). The seeds were sterilized (soaked for 30 min in sterile double-distilled H2O, 5 min in 70% ethanol and 15 min in 5% NaOCl and 0.1% Tween 20, and washed six times in sterile double-distilled H2O) and germinated on KNOP medium (Sijmons et al., 1991). After 2–3 weeks, the root tips were infected with J2. About 5–7 weeks later, galls with egg masses were dissected and allowed to hatch in sieves in 5–8 mL of sterile water at 28 °C.

Infection tests

For each homozygous transgenic line and wild-type A. thaliana Columbia, around 40 seeds were collected in miracloth bags and sterilized for 2 min in 95% ethanol and 12 min in 5% NaOCl and 0.1% Tween 20, and washed six times in sterile double-distilled H2O. They were grown on 0.5 × Murashige and Skoog germination medium supplemented with the appropriate antibiotics. Two weeks later, the plants were transferred to KNOP medium (Sijmons et al., 1991) on 12 × 12 cm2 Petri dishes, resulting in five plants per plate and six replica plates per line. The Petri dishes were placed slightly tilted to promote unidirectional root growth. One week later, two root tips of each plant were inoculated with 100–250 M. incognita or H. schachtii J2 from the sterile nematode cultures. For the test shown in Figure 4b, the hatched juveniles were submitted to an extra surface sterilization, resulting in lower infection rates. Plants were grown in vitro at 21 °C under a 16-h light/8-h dark photoperiod. Six weeks after infection, the number of galls and egg masses (M. incognita) or cysts (H. schachtii) were counted on each plate and compared with the control lines. These infection experiments were performed twice with each nematode on the plants of the F2 and F3 generations. For analysis of nematode development, the number of penetrated J2 at 2 DAI and the number of J2/J3/J4 at 9 DAI were counted after inoculation with H. schachtii.

The data were analysed statistically by SPSS (SPSS Inc., Chicago, IL, USA) using the Kolmogorov–Smirnov test for normality, the Levene test for homogeneity of variance and Duncan's or Tukey's test to compare the means. Groups indicated by a different letter indicate a significant difference at P < 0.05.

RNA extraction

The roots and rosettes of ±40 uninfected plants (14 days old) were harvested and immediately deep frozen. The roots of 16-day-old plants were inoculated with J2 of M. incognita. Six days later, 80–100 galls per line were dissected. The dissected tissues were deep frozen at –80 °C.

The material was ground with a Retsch MM301 machine (Haan, Germany) for 35 s at a frequency of 25 s−1. The metal grinding ball was removed and 1 mL of Trizol (Invitrogen) was added. The mixture was vortexed and incubated for 5 min at room temperature; 200 µL of chloroform was added, followed by a 3-min incubation at room temperature. After centrifugation for 15 min at 7 000 g at 4 °C, the upper phase was removed to a fresh tube and 500 µL of isopropyl alcohol was added, mixed and incubated for 10 min at room temperature. After centrifugation for 10 min at 7 000 g at 4 °C, the supernatant was removed and the RNA pellet was washed with 75% ethanol and dissolved in 50 µL of diethyl pyrocarbonate (DEPC)–H2O.

The residual DNA was removed by treating 1 µL of total RNA (5 µg) with 2 µL of DNase (1 U/µL) (Fermentas, St. Leon-Rot, Germany), 1 µL of RNase inhibitor (1 U/µL) (Invitrogen) and 2 µL of DNase buffer (10×) in a total volume of 20 µL, followed by incubation at 37 °C for 30 min. After the addition of 2 µL of ethylenediaminetetraacetic acid (EDTA) (25 mm), the enzymes were inactivated for 10 min at 65 °C; 2 µg of RNA was transcribed in first-strand cDNA via the reverse transcriptase kit of Invitrogen.

qRT-PCR

The reaction mix was prepared by Robot Robotics4 (Corbett Research, Cambridgeshire, UK) in 100-µL tubes fitted into the 72-well GeneDisc of the Rotor-Gene (Corbett Research). Each 25-µL reaction mix contained the following: 2 µL of cDNA, 1 µL of forward primer (70 nm), 1 µL of reverse primer (70 nm), 12.5 µL of Absolute™ QPCR SYBR® Green Mix (ABgene, Epsom, UK) and 8.5 µL of double-distilled H2O. The formation of primer dimers was opposed by the use of a hot start PCR of 15 min at 95 °C. This was followed by 40 cycles of 20 s at 95 °C, 30 s at 58 °C and 20 s at 72 °C, and fluorescence measurement at gain 6. PCR was completed with a melting step between 72 and 95 °C with 1 °C temperature increments for 5 s each step. The primers used for amplification of the AtCDKA;1 fragment were 5′-CCTCTTGAAAGAAATGCAGCACAGC-3′ and 5′-GAGAAATCAGGAGTAGAATCCATGTGC-3′ and, for the AtACT2 fragment, 5′-GGCTCCTCTTAACCCAAAGGC-3′ and 5′-CACACCATCACCAGAATCCAGC-3′.

The qRT-PCR data were analysed by the relative quantification method inline image, with the threshold level set at the lowest level of the ‘sweet spot’ zone of the curve in the exponential phase. To locate this point, the data were analysed several times using the inline image method with the threshold set at different levels. These levels were chosen in the exponential phase, located at least 10 times above the baseline and at 10, 20, 25, 30, 35, 40, 45, 50 and 60% of the average of the maximum fluorescence from the different samples. The zone in which the inline image results show maximum similarity is the ‘sweet spot’ zone. The fraction of fluorescence originating from the target gene is expressed in a characteristic and narrow melting peak at the melting temperature (Tm) of the amplicon. At lower temperatures, no notable peaks originating from primer dimers or from aspecific products were visible. The Ct values of both the samples of interest and the calibrator (wild-type plant tissue) were normalized to that of the housekeeping gene AtACT2. The measurements were repeated twice for galls at 6 DAI, roots and leaves, and thrice for complete plants.

Acknowledgements

This research was supported by a grant from Ghent University (BOF-01G00805), a PhD grant to Elke Van de Cappelle (BOF-01J11602) and a PhD grant to Eva Plovie from the Institute for the Promotion of Innovation through Science and Technology in Flanders (IWT-Vlaanderen). We are very grateful to Janice de Almeida Engler for discussions and to Gordana Marjanac, Sylvie de Buck and Ann Depicker (VIB Department of Plant Systems Biology, Ghent University) for providing the GUS lines supertransformed with the GUS hairpin construct.

Ancillary