CCS52 and DEL1 genes are key components of the endocycle in nematode-induced feeding sites


  • Janice de Almeida Engler,

    Corresponding author
    1. Institut National de la Recherche Agronomique, UMR 1355 ISA/Centre National de la Recherche Scientifique, UMR 7254 ISA/Université de Nice-Sophia Antipolis, UMR ISA, 400 route des Chappes, Sophia-Antipolis, France
      (e-mails or
    Search for more papers by this author
  • Tina Kyndt,

    1. Department of Molecular Biotechnology, Ghent University, Coupure Links 653, B-9000 Ghent, Belgium
    Search for more papers by this author
  • Paulo Vieira,

    1. Institut National de la Recherche Agronomique, UMR 1355 ISA/Centre National de la Recherche Scientifique, UMR 7254 ISA/Université de Nice-Sophia Antipolis, UMR ISA, 400 route des Chappes, Sophia-Antipolis, France
    Search for more papers by this author
  • Elke Van Cappelle,

    1. Department of Molecular Biotechnology, Ghent University, Coupure Links 653, B-9000 Ghent, Belgium
    Search for more papers by this author
  • Veronique Boudolf,

    1. Department of Plant Systems Biology, Vlaams Instituut voor Biotechnologie, B-9052 Ghent, Belgium
    2. Department of Plant Biotechnology and Bioinformatics, Ghent University, B-9052 Ghent, Belgium
    Search for more papers by this author
  • Vanesa Sanchez,

    1. Institut National de la Recherche Agronomique, UMR 1355 ISA/Centre National de la Recherche Scientifique, UMR 7254 ISA/Université de Nice-Sophia Antipolis, UMR ISA, 400 route des Chappes, Sophia-Antipolis, France
    Search for more papers by this author
  • Carolina Escobar,

    1. Unidad de Genómica, Centro Nacional de Biotecnología-CSIC, Campus Universidad Autónoma de Madrid, 28049 Madrid, Spain
    Search for more papers by this author
  • Lieven De Veylder,

    1. Department of Plant Systems Biology, Vlaams Instituut voor Biotechnologie, B-9052 Ghent, Belgium
    2. Department of Plant Biotechnology and Bioinformatics, Ghent University, B-9052 Ghent, Belgium
    Search for more papers by this author
  • Gilbert Engler,

    1. Institut National de la Recherche Agronomique, UMR 1355 ISA/Centre National de la Recherche Scientifique, UMR 7254 ISA/Université de Nice-Sophia Antipolis, UMR ISA, 400 route des Chappes, Sophia-Antipolis, France
    Search for more papers by this author
  • Pierre Abad,

    1. Institut National de la Recherche Agronomique, UMR 1355 ISA/Centre National de la Recherche Scientifique, UMR 7254 ISA/Université de Nice-Sophia Antipolis, UMR ISA, 400 route des Chappes, Sophia-Antipolis, France
    Search for more papers by this author
  • Godelieve Gheysen

    Corresponding author
    1. Department of Molecular Biotechnology, Ghent University, Coupure Links 653, B-9000 Ghent, Belgium
      (e-mails or
    Search for more papers by this author

(e-mails or


The establishment of galls and syncytia as feeding sites induced by root-knot and cyst nematodes, respectively, involves a progressive increase in nuclear and cellular size. Here we describe the functional characterization of endocycle activators CCS52A, CCS52B and a repressor of the endocycle, DEL1, during two types of nematode feeding site development in Arabidopsis thaliana. In situ hybridization analysis showed that expression of CCS52A1 and CCS52B was strongly induced in galls and syncytia and DEL1 was stably but weakly expressed throughout feeding site development. Down-regulation and over-expression of CCS52 and DEL1 in Arabidopsis drastically affected giant cell and syncytium growth, resulting in restrained nematode development, illustrating the need for mitotic activity and endo-reduplication for feeding site maturation. Exploiting the mechanism of endo-reduplication may be envisaged as a strategy to control plant-parasitic nematodes.


Plants interact with a wide range of organisms, often leading to various pathologies. Among these are obligate sedentary endoparasites such as nematodes that establish intimate interactions with a diverse range of plant hosts. Root-knot nematodes (Meloidogyne spp.) are capable of inducing nematode feeding sites within the vascular tissue, containing 5–7 hypertrophied giant cells with multiple abnormally enlarged nuclei. Multinucleate giant cells result from numerous mitotic events in the absence of cytokinesis, and become highly polyploid, possibly by successive endo-reduplication cycles (Sijmons et al., 1994; de Almeida Engler et al., 1999). Cells surrounding young giant cells actively divide to form a multi-layered swelling resulting in a gall. In contrast, cyst nematodes (Heterodera spp.) induce a multinucleate feeding site by stimulating a single vascular cell to become a syncytium through incorporation of adjacent cells via extensive cell-wall dissolution and protoplast fusion (Sijmons et al., 1994; Grundler et al., 1998). Even though galls and syncytia follow a different developmental program, formation of these nematode feeding sites has several features in common. In Arabidopsis thaliana, both types of feeding cells are initiated and localized in the root vascular tissue adjacent to xylem elements. Both contain a dense cytoplasm and numerous large nuclei, undergo organelle proliferation and show elaborate ingrowths of peripheral cell walls (Hussey and Grundler, 1998; Mitchum et al., 2004). Both giant cells and syncytia are polynucleate feeding cells: giant cells via acytokinetic mitoses and syncytia by cell fusion. Due to the presence of enlarged nuclei, it is assumed that additional DNA replication cycles may play a primary role in the establishment of functional feeding cells. Polyploidy in giant cells was first suggested in the 1960s (Owens and Novotny Specht, 1964; Dropkin, 1965), but the mechanism by which it is derived has not been elucidated so far. Previous work has reported that the nuclear DNA content in giant cells increases significantly due to the increased nuclear size and chromosome number (Wiggers et al., 1990; Starr, 1993), possibly leading to drastic feeding cell expansion. Intense DNA synthesis in galls and syncytia also strongly indicated the presence of additional replication cycles in nematode feeding cells (Rubinstein and Owens, 1964; Rohde and McClure, 1975; de Almeida Engler et al., 1999). The endocycle is a variant of the eukaryotic cell cycle in which successive S phases follow each other without intervening mitosis or cell division (De Veylder et al., 2011). The DNA content of the cell is doubled with every new round of DNA replication, resulting in formation of cells with DNA ploidy levels of 4C, 8C, 16C or higher. The endocycle may be induced during biological processes such as cell differentiation, cell expansion, metabolic activity and stress. Plant endoploidy is typically observed in differentiated, outsized cells (e.g. Arabidopsis trichomes), endosperm and fruit (Chevalier, 2007; Larkin et al., 2007; Sabelli et al., 2007).

The plant endocycle is controlled by diverse factors or gene products, and functional analyses of mutants and transgenic plants with aberrant levels of endo-reduplication have led to identification of key regulators of the endocycle. Examples are the two CCS52 classes identified in plants: CCS52A (Cdh1/Fzr/Srw1-type), which is also found in yeast and animals; and CCS52B, which is presumed to be plant-specific (Tarayre et al., 2004). In Arabidopsis, the CCS52A class is represented by two family members (CCS52A1 and CCS52A2), whereas there is only one CCS52B gene (Fülöp et al., 2005). CCS52 is a cell-cycle switch protein that behaves as an adaptor protein that is important for activation of the anaphase-promoting complex/cyclosome (APC/C) and is involved in conversion of mitotic cycles into endocycles. Another example of an endocycle regulator is the plant homologue of the archaeal DNA topoisomerase VI that is required for successful progression of the endo-reduplication cycle in Arabidopsis (Sugimoto-Shirasu et al., 2002, 2005). RHL1 encodes the ROOT HAIRLESS 1 protein and forms a multiprotein complex with plant topoisomerase VI (Sugimoto-Shirasu et al., 2002). Sugimoto-Shirasu et al. (2002) also suggested that topoisomerase VI is required to resolve entangled chromosomes during endocycles above 8C.

In contrast to CCS52 and RHL1, E2Fe/DEL1 (hereafter referred to as DEL1) is an inhibitor of endo-reduplication and preserves the mitotic state of proliferating cells by suppressing transcription of genes required to enter the endocycle (Vlieghe et al., 2005; Lammens et al., 2008). DEL (DP-E2F-like) genes encode atypical E2F-like proteins designated E2F7/E2F8 in mammals. Arabidopsis has three DEL genes (DEL1, DEL2 and DEL3). Loss of DEL1 function results in augmented ploidy levels, while ectopic expression of DEL1 results in decreased endo-reduplication levels.

Plant biotrophic interactors consistently establish specialized interaction sites where nutrient exchange occurs. Augmented plant nuclear DNA ploidy has been described during numerous interactions, including fungal and bacterial symbionts and parasitic fungi and nematodes (Wildermuth, 2010). Here, we address the occurrence of endo-reduplication triggered by two types of nematodes (gall-forming root-knot nematodes and syncytium-forming cyst nematodes) during establishment of their feeding sites, and investigate components potentially participating in this process. The DEL1 and CCS52 gene family and RHL1 may be essential components of the endocycle machinery active in nematode feeding sites in host roots. Therefore, we generated knockdown and over-expression lines of three CCS52 genes and analysed the effect in nematode-infected roots. We also examined comparable lines for DEL1 and RHL1, which have been described as key regulators of the endocycle (Sugimoto-Shirasu et al., 2005; Vlieghe et al., 2005). Data reported suggest that increased DNA ploidy levels in nematode feeding sites are linked to endocycle activation. Collectively, the results support the conclusion that mitotic and DNA replication processes cannot operate independently in favour of a successful nematode feeding site and nematode maturation.


In situ localization of CCS52 and DEL1 transcripts

Sketches of a gall induced by Meloidogyne incognita (Figure 1a) and a syncytium induced by Heterodera schachtii (Figure 1b) are presented to provide an overview of both nematode feeding sites. Sections of galls induced by M. incognita and syncytia induced by H. schachtii were hybridized in situ to localize transcripts of three cell-cycle genes (CCS52A1, CCS52B and DEL1) reported to be involved in endocycle control (Cebolla et al., 1999; Vlieghe et al., 2005). All three genes are expressed in the root vascular tissue where nematodes induce feeding sites (Figure 1i for DEL1) (de Almeida Engler et al., 2009; Plant Systems Explorer Image Database at CCS52A1 is highly expressed in giant cells (Figure 1c) and syncytia (Figure 1d), and less so in neighbouring cells. Transcript levels of CCS52B are high in galls (Figure 1e,e′) and syncytia (Figure 1f,f′). CCS52B expression is patchy in syncytia and is particularly strong in the proliferating xylem elements in both types of feeding sites. However, DEL1 is weakly expressed in giant cells (Figure 1g,g′) and syncytia (Figure 1h).

Figure 1.

In situ localization of CCS52A1, CCS52B and DEL1 transcripts in Meloidogyne incognita-induced galls, Heterodera schachtii-induced syncytia and non-infected roots.
(a, b) Sketches of a mature gall induced by M. incognita (a) and a mature syncytium induced by H. schachtii (b). The epidermis is shown in green, the cortex in yellow, the endodermis in orange, neighbouring cells in blue and feeding cells in white. Root-knot and cyst nematodes induce feeding sites and exploit the cell-cycle machinery in different ways. Within the first type of feeding site (galls), the giant feeding cells undergo mitotic events resulting in multiple nuclei, which subsequently undergo ectopic DNA replication. The second type of feeding site (syncytia) becomes multinucleate via neighbouring cell fusion, and incorporated nuclei undergo ectopic DNA replication. Mitosis in syncytia is only important prior to incorporation of neighbouring cells to the feeding site. Nuclei of giant cells undergo mitosis and possibly endo-reduplication, whereas the syncytial feeding cell possibly undergoes only endo-reduplication.
(c, d) CCS52A1 expression in a gall at 7 DAI (days after inoculation) (c), and in a syncytium at 7 DAI (d).
(e, e′, f, f′) CCS52B expression in a gall at 7 DAI (e, e′) and in a syncytium at 7 DAI (f, f′).
(g, g′, h) DEL1 expression in a gall at 7 DAI (g, g′) and in a syncytium at 7 DAI (h).
(i) mRNA localization of DEL1 in an Arabidopsis root.
Longitudinal tissue sections were hybridized with 35S-labelled antisense RNA probes. Hybridization signals are visible as white dots under dark-field optics and black dots under bright-field optics. NC, neighbouring cells; asterisks, giant cells; S, syncytium; VT, vascular tissue; n, nematode. Scale bars = 50 μm.

Expression analysis and ploidy levels in CCS52A and CCS52B over-expressing and knockdown lines

Three independent lines ectopically expressing CCS52B (35Spro:CCS52B1III, 35Spro:CCS52B2III and 35Spro:CCS52B5III) were analysed by quantitative RT-PCR and showed a high expression level compared to the control (Figure S1a). However, decreased CCS52A1 and CCS52A2 expression was detected in RNAi knockdown lines expressing gene-specific RNAs (hp lines) for both CCS52A genes, but significantly less for CCS52A1 (Figure S1b). Low CCS52B expression was observed in the hpCCS52B line (Figure S1b). The hpCCS52AB line was designed to knockdown both CCS52A and CCS52B genes. Due to the varying expression levels of the three CCS52 genes in all hairpin lines analysed, lower ploidy levels were used for the choice of line analysed in this study (Figure S1c). Flow cytometry of knockdown lines (hpCCS52A, hpCCS52B and hpCCS52AB) showed a decreased ploidy in cotyledon cells (Figure S1c). Roots of 21-day-old plants of the CCS52B over-expressing lines were also analysed by flow cytometry, revealing that increased CCS52B expression resulted in more nuclei with a DNA content up to 32C and 64C and much fewer nuclei with 2C and 4C (Figure S1d). The highest ploidy levels were observed in line 35Spro:CCS52B5III (hereafter referred to as CCS52BOE), which was subsequently used for morphological analysis. The main root apical meristems of CCS52BOE seedlings were slightly bulged compared to wild-type roots, and leaf trichomes often contained four or five branches compared to the three branches observed in wild-type leaves (Figure S2a,b). Root sections of the CCS52BOE line displayed enlargement of cells of the root apex and elongation zone, causing a misalignment compared to wild-type (Figure S2c,e). Disturbed morphology of root cells was also observed in the CCS52A2OE line (Figure S2d).

Ectopic expression and down-regulation of CCS52 genes interfere with giant cell and syncytium development

Analysis of roots in lines in which the CCS52 genes were knocked down or over-expressed showed that the root architecture was similar to wild-type plants. Root-knot or cyst nematode infection on the CCS52BOE line was far more efficient than in wild-type roots (Figure S3). Feeding sites caused by both nematodes were often present over the entire root length, preventing accurate counting of induced galls (Figure S3a) or syncytia (Figure S3b). This excessive infection of Arabidopsis roots was less evident in the CCS52A2OE line. Nevertheless, galls (compare WT in Figure 2a with Figure 2c) and syncytia (compare WT in Figure 2h with Figure 2j) induced in both CCS52BOE and CCS52A2OE lines showed premature maturation compared to wild-type feeding sites. Fast-maturing giant cells (Figure 2c,e,f) and to a lesser extent syncytia (Figure 2j) contained larger nuclei compared to wild-type infected plants. Fewer nuclei containing large chromocenters (compare WT in Figure S4b with Figure S4a), as well as fewer neighbouring cells, were apparent in the CCS52BOE and CCS52A2OE galls. Fewer dividing neighbouring cells were also observed in syncytia (Figure 2j). As galls and syncytia matured (21 days after inoculation, DAI), fewer expanded and abnormally shaped giant cells (compare WT in Figure 2b with Figure 2d) and syncytia (compare WT in Figure 2i with Figure 2k) were observed in CCS52BOE or CCS52A2OE lines. In addition, a decrease in nematode size, indicative of delay in their development, was perceived in these gall sections. Giant cells in the CCS52A2OE line stopped expanding and showed a retracted cytoplasm surrounded by enhanced wall invaginations and stubs (Figure 2g). Enlarged giant cell nuclei in the CCS52BOE line as well as the CCS52A2OE line showed bulky chromocenters, suggesting the occurrence of endo-reduplication (Figure 2f,g,g′ and Figure S4a compared to WT in Figure S4b′). Measurements of individual nuclei in giant cells of CCS52BOE and CCS52A2OE lines at different gall stages confirmed the increased nuclear size observed in giant cells (Figure S5).

Figure 2.

 Morphology of Meloidogyne incognita-induced galls and Heterodera schachtii-induced syncytia in Arabidopsis roots over-expressing CCS52.
(a, b) Galls at 3 DAI (a) and at 21 DAI (b) in wild-type infected seedlings.
(c, d) Galls at 3 DAI (c) in plants over-expressing CCS52B, showing abnormally enlarged nuclei (arrowheads) in giant cells, and at 21 DAI (d) showing aberrant giant cells and nematodes with delayed development.
(e, f, g, g′) Galls in plants over-expressing CCS52A2 at 3 DAI (e), at 7 DAI (f) showing large nuclei, and at 21 DAI (g, g′) showing giant cells containing nuclei with bulky chromocenters (black arrows in g′) and lacking cytoplasm.
(h, i) Syncytia in wild-type seedlings at 3 DAI (h) and 21 DAI (i).
(j, k) Syncytia in plants over-expressing CCS52B at 3 DAI (j) and at 21 DAI (k) showing enlarged NCs and vacuolated syncytia.
Bright-field images of sections stained with toluidine blue. Asterisks, giant cells; n, nematode; NC, neighbouring cells; S, syncytium. Scale bars = 50 μm.

Knockdown of CCS52A, CCS52B or CCS52AB genes also had a profound effect on gall and syncytium development (Figure 3). Decreased expression of either CCS52 gene resulted in galls with small giant cells (compare WT in Figure 3a with Figure 3b–f) and undersized syncytia (compare WT in Figure 3g with Figure 3h–l) containing little cytoplasm and showing a delay in nematode development. DAPI staining revealed that, in contrast to the amoeboid-shaped nuclei in wild-type giant cells (Figure S4c), nuclei in the CCS52B knockdown line were smaller, elongated and often clustered (Figure S4d). These changes in nuclear morphology were, to a lesser extent, also observed in syncytia nuclei (compare WT in Figure S4e with Figure S4f). Measurements of giant cell nuclei with reduced levels of CCS52AB confirmed our observations on gall sections (Figure S5).

Figure 3.

 Effect of CCS52 knockdown on Meloidogyne incognita-induced galls and Heterodera schachtii-induced syncytia.
(a) Gall in a wild-type root at 7 DAI.
(b) Gall in the CCS52A knockdown line at 7 DAI.
(c) Gall in the CCS52B knockdown line at 7 DAI showing small giant cells containing few cytoplasm.
(d) Gall in the CCS52AB knockdown line at 7 DAI showing small giant cells with cell-wall stubs (arrowhead).
(e) Gall in the CCS52B knockdown line at 21 DAI showing disordered gall development.
(f) Gall in the CCS52AB knockdown line at 21 DAI showing undersized galls and delayed nematode development.
(g) Syncytium in a wild-type root at 7 DAI.
(h, i) Syncytia in the knockdown CCS52B line at 7 DAI (h) and 21 DAI (i).
(j) Syncytium in the knockdown CCS52A line at 7 DAI.
(k, l) Syncytia in the knockdown CCS52AB line at 7 DAI (k) and 21 DAI (l) showing malformed feeding sites.
Bright-field images of sections stained with toluidine blue. Asterisks, giant cells; n, nematode; NC, neighbouring cells; LRM, lateral root meristem; S, syncytium. Scale bars = 50 μm.

Down-regulation of CCS52B resulted in aberrant syncytia containing irregularly dividing neighbouring cells. Syncytia in this line sometimes incorporated cortical and epidermal cells (Figure 3h). Mature syncytia contained small nuclei, and feeding site expansion was clearly affected (compare WT in Figure 3g with Figure 3i). Analysis of CCS52A and CCS52AB knockdown lines revealed that syncytia contained less cytoplasm and were malformed (Figure 3j–l).

Nematode infection tests of CCS52A, CCS52B and CCS52AB knockdown lines revealed that root-knot and cyst nematodes were able to penetrate and to induce galls or syncytia, but fewer juveniles were able to mature compared to wild-type infected plants (Figure 4a,b).

Figure 4.

 Infection tests by Meloidogyne incognita and Heterodera schachtii in CCS52 knockdown Arabidopsis lines.
The numbers of egg masses (a) and cysts (b) are decreased compared to wild-type. Asterisks indicate values that are significantly different from wild-type at < 0.05 (Student’s t test).

Ectopic expression and knockout of DEL1 affects mitosis and the development of galls and syncytia

To analyse the involvement of DEL1 in gall and syncytium development, over-expression (hereafter referred to as DEL1OE) and knockout (del1-1) plants were investigated (Figure 5). When infected with root-knot nematodes, ectopic DEL1 expression resulted in galls containing small-sized giant cells that attempted to divide (compare WT in Figure 5a with Figure 5b,c), with highly proliferating neighbouring cells (Figure 5b). Mature galls contained small malformed giant cells and profuse cell-wall invaginations and stubs (Figure 5d). Loss-of-function analysis for DEL1 in the del1-1 line infected with root-knot nematodes revealed malformed giant cells containing little cytoplasm (Figure 5e). Morphological analyses of syncytia also demonstrated that syncytium development was significantly affected. Ectopic DEL1 expression also resulted in malformed syncytia (compare WT in Figure 5f with Figure 5g). These feeding sites were small, contained homogeneous cytoplasm lacking the typical small vacuoles, and neighbouring cells failed to fuse properly (Figure 5g). Loss-of-function analysis of DEL1 in the del1-1 line infected with cyst nematodes showed abnormally formed syncytia, smaller than wild-type, which failed to expand and had fewer neighbouring cells (compare WT in Figure 5f with Figure 5h). Nuclear size analysis of giant cells in wild-type (Figure S4g) and DEL1OE roots revealed that nuclei were smaller in the transgenic line (Figures S4h and S5).

Figure 5.

 Morphology, and promoter activity of CDKB1;1, in Meloidogyne incognita-induced galls and Heterodera schachtii-induced syncytia in DEL1 over-expressing and knockout lines.
(a) Gall in a wild-type root at 7 DAI.
(b–d) Gall in a plant over-expressing DEL1 at 7 DAI (b) showing ectopic cell division around small giant cells, at 14 DAI (c) showing a giant cell that attempted to divide (arrow), and at 21 DAI (d) containing giant cells comprising cell-wall stubs (black arrows) and multiple nuclei (red arrows).
(e) Gall of the del1-1 line at 14 DAI showing malformed giant cells containing little cytoplasm.
(f) Syncytium in a wild-type root at 21 DAI.
(g, h) Syncytia in a plant over-expressing DEL1 at 7 DAI (g) and a magnification of the feeding site (inset), and at 14 DAI (h) showing disturbed feeding site expansion.
(i–k) CDKB1;1pro:GUS expression in a wild-type gall at 14 DAI (i), in a gall of the DEL1 over-expressing line at 14 DAI (j), and in a gall of the del1-1 line at 14 DAI (k).
(l, m) CDKB1;1pro:GUS expression in a wild-type syncytium at 14 DAI (l), and in a syncytium of the DEL1 over-expressing line at 14 DAI (m).
(n, o) CDKB1;1pro:GUS expression in a syncytium of a del1-1 mutant background line at 14 DAI (n) and 21 DAI (o).
(a–h) Bright-field images of galls and syncytia sections stained with toluidine blue. (i–o) GUS expression visualized under dark-field conditions Asterisks, giant cells; NC, neighbouring cells; n, nematode; S, syncytium. Scale bars = 50 μm.

CDKB1;1 is a G2/M specific marker, and its therefore expression reflects cells that potentially present mitotic activity. Control wild-type CDKB1;1pro:GUS plants (Figure 5i) as well as CDKB1;1pro:GUS × DEL1OE plants (Figure 5j) showed strong promoter activity in galls. In contrast, CDKB1;1 promoter activity decreased in giant cells of del1-1 crossed with the CDKB1;1pro:GUS line, confirming the reduced mitotic activity in galls (Figure 5k).

Expression analysis of the mitotic marker CDKB1;1pro:GUS in wild-type syncytia confirmed strong CDKB1;1 promoter activity during early developmental stages (Figure 5l) with progressive decrease as they matured (de Almeida Engler et al., 1999). A strong CDKB1;1 promoter activity was observed in syncytia over-expressing DEL1 (Figure 5m), but it was down-regulated in the del1-1 line (Figure 5n,o).

Root-knot and cyst nematode infection tests performed on the DEL1OE line revealed decreased nematode reproduction (Figure 6).

Figure 6.

 Infection tests by Meloidogyne incognita and Heterodera schachtii in the DEL1 over-expressing line.
The numbers of egg masses (a) and cysts (b) decreased in the DEL1 over-expressing line compared to wild-type. Asterisks indicate values that are significantly different from wild-type at < 0.05 (Student’s t test).

Knockout of RHL1 causes minute gall formation and blocks cyst nematode infection

Analysis of rhl1 knockdown mutants showed that nematodes were able to penetrate, migrate and induce minute multinucleate giant cells that did not expand further (Figure 7a–c). Only a few neighbouring cells divided asymmetrically, giving rise to tiny galls. Infection at 7 DAI only proceeded until 14 DAI because rhl1 seedlings only lived for 3 weeks. Although cyst nematodes attempted to penetrate the roots of rhl1 mutants, they did not induce syncytia.

Figure 7.

 Morphology of Meloidogyne incognita-induced galls in roots of the rhl1 knockout line.
(a, b) Longitudinal sections (a) and cross-sections (b) of minute galls in infected rhl1 roots at 7 DAI. Bright-field images of gall sections stained with toluidine blue.
(c) Nuclei of a gall in infected rhl1 roots at 7 DAI. Fluorescence image of DAPI-stained nuclei.
Asterisks, giant cells; n, nematode. Scale bars = 50 μm.


Previous reports have highlighted polyploidy in nematode feeding cells by the presence of multiple large nuclei with a high chromosome number compared to normal root cells (Endo, 1971a,b; Starr, 1993; de Almeida Engler et al., 2004). Endopolyploidy may be the result of (i) multinucleate cells originating from acytokinetic mitosis, (ii) endo-reduplication, (iii) nuclear fusion or (iv) endomitosis (Chevalier et al., 2011). Our analysis, combined with functional predictions for genes involved in altering ploidy levels, provides insights into the components of the endocycle present in galls induced by root-knot nematodes and syncytia induced by cyst nematodes.

In almost all lines analysed here (CCS52A2OE, CCS52BOE, hpCCS52A, hpCCS52B, hpCCS52AB, DEL1OE, del1-1 and rhl1), nematodes were able to penetrate, migrate and induce galls or syncytia in Arabidopsis roots. Only lack of RHL1 completely hindered syncytium induction. In contrast, nematode feeding site expansion was affected in all lines analysed. Inhibition of the endocycle by CCS52 knockdown or DEL1 over-expression did not hamper mitosis but prevented formation of large polyploid nuclei, resulting in smaller feeding cells. Conversely, both inhibition and stimulation of endo-reduplication affected nematode feeding site development. Reproduction was considerably lower for all lines tested. Our current analysis, combined with data obtained from previous studies, supports the idea that the mitotic cycle or endocycle alone is not sufficient to support giant cell and syncytium development, and that a balanced interplay between both processes is a prerequisite for nematode feeding site maturation.

CCS52 and DEL1 are expressed differentially in nematode feeding sites

mRNA in situ hybridization showed that high transcript levels of CCS52A and CCS52B, encoding APC/C activators, are found during gall and syncytium development. Previous studies have shown high transcript levels of MtCCS52 at early stages (6 DAI) and decreased expression at later stages (60 DAI) after inoculation of Medicago truncatula with the root-knot nematode Meloidogyne incognita (Koltai et al., 2001). Determination of the promoter activity of CCS52A in 7-day-old galls reinforced these findings (Favery et al., 2002). Based on their known function, CCS52 genes possibly promote endo-reduplication in maturing giant cells and syncytia through degradation of mitotic cyclins by the APC/C complex (Kondorosi and Kondorosi, 2004). Our previous work has shown that transcriptional activity of CYCB1;1 is high during early giant cell and syncytium development but down-regulated at later stages (de Almeida Engler et al., 1999). Therefore, the APC/C pathway, which is responsible for ubiquitin-dependent proteolysis of mitotic activators, may be triggered in giant cells for mitotic control and to allow endo-reduplication. A mitotic block is even more obvious in syncytia, in which mitosis was never observed except in proliferating neighbouring cells that are later incorporated into the syncytium. DEL1 expression is not induced in tissues undergoing endo-reduplication (Vlieghe et al., 2005) and DEL1 is weakly expressed in galls and syncytia, possibly allowing endocycle onset in both types of feeding sites.

CCS52 knockdown or ectopic expression inhibit nematode development

Selective degradation of mitotic cyclins via the ubiquitin-mediated proteolytic pathway requires the APC/C, which is activated by CCS52 and CDC20 (Vinardell et al., 2003). Down-regulation of CCS52 leads to reduced endo-reduplication (Cebolla et al., 1999), resulting in lower ploidy levels in the plant. We observed that gall and syncytium development as well as nematode reproduction were markedly reduced in CCS52 knockdown lines as well as in plants showing ectopic expression of DEL1, CCS52A2 and CCS52B. Lack of CCS52 may prevent the induction of endo-reduplication cycles in galls and syncytia, affecting their size and morphology. Likewise, Vinardell et al. (2003) showed that antisense expression of CCS52A in Medicago truncatula affected nodule development and nitrogen fixation.

Although ectopic expression of CCS52 affected tissue organization at the root apical meristem, we observed higher nematode penetration rates than in wild-type Arabidopsis seedlings. Moreover, stimulating the endocycle induced additional galls and syncytia in roots, and early giant cell and syncytium expansion (3 DAI). Ectopic CCS52A2 and CCS52B expression prematurely induced the endocycle and hampered mitosis, resulting in giant cells that contained fewer but enlarged nuclei. This imbalance caused by inhibition of the mitotic phase and the endocycle resulted in giant cells with less cytoplasm. This rapid switch results in giant cells of reduced size containing fewer but enlarged nuclei, and decreased nematode growth and reproduction. These observations suggest that the endocycle cannot solely drive giant cell development. In the case of syncytia, ectopic CCS52A2 and CCS52B expression resulted in enlarged nuclei that clustered and occupied a large fraction of the undersized feeding site. Possibly, lack of an accurate balance of mitotic activity and endo-reduplication caused disorganization in neighbouring cells, disturbing syncytium expansion. Full feeding site expansion is prevented and nematode development is delayed and often arrested at the J2 stage, with few nematodes maturing into large pear-shaped females.

Ectopic expression and knockout of DEL1 compromises giant cell and syncytium development

DEL1 operates as a specific repressor of the endocycle, and its ectopic expression promotes cell division and inhibits cell growth, with decreased ploidy levels. Loss of DEL1 function results in increased ploidy (Vlieghe et al., 2005). Root-knot nematode infection of DEL1OE plants resulted in small, occasionally dividing, multinucleated giant cells that formed wall stubs. As giant cells in wild-type plants never divide and do not often show remnants of cell walls, these data strongly indicate that DEL1 over-expression drives giant cells into mitosis that occasionally leads to cytokinesis. The observations in the DEL1OE line confirm our previous hypothesis that multiple nuclei resulting from acytokinetic mitotic events are not sufficient to drive giant cell expansion (de Almeida Engler et al., 1999). Similar to the reduced giant cell size observed in this study, Vlieghe et al. (2005) observed that leaves were smaller in DEL1OE lines, suggesting a general inhibition of cell expansion in these lines. The presence of small nuclei in the DEL1OE line suggests blocked or decreased endo-reduplication, consistent with the fact that DEL1 represses transcription of endocycle genes (Vlieghe et al., 2005). DEL1 over-expression may prevent mitotic or post-mitotic giant cells from initiating the endocycle by suppressing transcription of genes required to start DNA replication. DEL1 over-expressing giant cells and syncytia showed a high proliferation of neighbouring cells caused by enhanced mitotic activity, possibly reflecting the previously observed accumulation of G1 cells that have exited the G2 phase of the mitotic cycle (Vlieghe et al., 2005). DEL1 over-expression also disrupted neighbouring cell morphology and organization in syncytia, limiting their incorporation and resulting in undersized feeding sites and reduced nematode reproduction.

The cytological observations are supported by the high CDKB1;1pro:GUS activity in galls and syncytia in DEL1OE, illustrating induction of the mitotic state in feeding cells. We showed that CDKB1;1 promoter activity in the del1-1 mutant decreased in galls and syncytia, suggesting a low mitotic status and resulting in small giant cells, possibly due to premature onset of the endocycle. Cell-cycle inhibition probably affected neighbouring cells of syncytia before their incorporation, resulting in smaller feeding sites.

DEL1 negatively controls endo-reduplication through direct transcriptional control of APC/C activity by repressing the CCS52A2 promoter (Lammens et al., 2008), thus preventing premature endocycle onset (De Veylder et al., 2011). The accumulation of CCS52A2 during late S and G2 phase in the del1-1 mutant may stimulate giant cells and syncytia to proceed from division to the endocycle.

Knockdown of RHL1 delays giant cell development

ROOT HAIRLESS genes encode subunits of topoisomerase VI with DNA-binding activity that participate in decatenation of duplicated chromosomes (Hartung et al., 2002; Sugimoto-Shirasu et al., 2002). The RHL1 protein plays a crucial role in ploidy-dependent cell growth in Arabidopsis, and has been associated with defects in the third round of the endocycle in hypocotyl cells (Sugimoto-Shirasu et al., 2002, 2005). We showed that root-knot nematodes were capable of infection and gall induction in the rhl1 mutant. Mitosis followed by two rounds of DNA replication appears sufficient to form polynucleate, polyploid, minute giant cells, but they were unable to develop further. Once endo-reduplication is triggered in giant cells, repeated endocycle rounds possibly require DNA decatenation and disentanglement enzymes. In contrast, cyst nematodes attempted root penetration but were unable to induce a syncytium in the rhl1 mutant, possibly because syncytium induction already depends on endo-reduplication as opposed to giant cell initiation.

Concluding remarks

It has been believed for many years that root-knot and cyst nematodes push their feeding cells to undergo endo-reduplication due to their large nuclei. During infection on host plants, nematodes may profit from the endo-reduplication competence of root vascular tissue cells in the elongation and differentiation zone. This physiological status may allow their re-differentiation into giant cells or syncytia. Both types of feeding sites undergo a differentiation program involving several rounds of DNA replication accompanied by nuclear and cell expansion, resulting in large multinucleated and polyploid cells. Nuclear enlargement has also been observed in leaf cells before fungal penetration (Genre et al., 2008), and cell enlargement has been observed in mesophyll cells underlying fungal-infected epidermal cells (Chandran et al., 2010). Biotrophs undergo extensive body growth, imposing a substantial continued metabolic requirement on the plant host. Here we demonstrated that nematodes exploit key endocycle regulators to induce large feeding cells. The data presented, combined with our previous findings (de Almeida Engler et al., 1999) (Figure 8), strongly corroborate the idea that plant-parasitic nematodes reprogram the plant cell-cycle machinery, inducing mitotic cycles in giant cells and neighbouring cells that will be incorporated into a syncytium, followed by repeated endo-reduplication cycles within both types of feeding cells. Previous reports proposed that endo-reduplication is a general mechanism that is essential to sustain the great metabolic demands in favour of a wide range of plant–pathogen interactions (Wildermuth, 2010). Here we provide further evidence that gene amplification by endo-reduplication may be essential to maintain high metabolic activity in giant cells and syncytia to ensure maturation and reproduction of both types of nematodes.

Figure 8.

 Schematic representation of the cell cycle and endocycle within Meloidogyne incognita-induced giant cells and Heterodera schachtii-induced syncytia.
(a) (1) Scheme for a giant cell undergoing acytokinetic mitotic cycles producing several nuclei (violet circles). The mitotic cell cycle comprises the classic G1/S/G2/M phases and is characterized by high CDK/CYC activity as well as CYCA2 expression. Mitosis generates multiple giant cell nuclei that subsequently undergo several rounds of endo-reduplication (2) to generate nuclei with variable sizes and ploidy levels. At this stage, CDK/CYC activity decreases and mitotic cyclin expression is blocked, but CCS52A1, CCS52A2 and CCS52B expression remain high. Decreased DEL1 expression in giant cells may facilitate the switch from the cell cycle to the endocycle allowing CCS52A2 activity.
(b) (3) Scheme for a syncytium containing neighbouring cells that undergo a cytokinetic cell cycle consisting of S and M phases interrupted by two gap phases (G1 and G2). This mitotic activity results in daughter cells that fuse to the feeding site to generate a multinucleated syncytium. This syncytium further undergoes similar endocycles (4) as in giant cells.
The asterisk (*) represents involvement of the RHL1 protein in the endocycle of both galls and syncytia.

Experimental procedures

Constructs and plant transformation

Whole 14-day-old A. thaliana seedlings were harvested and immediately deep-frozen. Samples were ground and RNA was extracted using Trizol (Invitrogen, After DNaseI treatment and heat inactivation, reverse transcription was performed using 2 μg RNA at 42°C for 2 h in a total volume of 40 μl: 20 μl RNA, 8 μl first-strand buffer (5×) (Invitrogen), 1 μl Superscript II reverse transcriptase (Invitrogen), 2 μl dNTPs (10 mm), 4 μl dithiothreitol (0.1 m), 1 μl oligo(dT)25 primer (700 ng μl−1) and 4 μl ddH20. To generate the CCS52B over-expression construct, the complete cDNA coding sequence of CCS52B (AY099581) was amplified from cDNA of A. thaliana by three subsequent (nested) PCR reactions using the following primers (Table S1): EVPLO51-52 in the first PCR, EVPLO71-72 in the second PCR, and EVPLO53-54 in the third PCR. This coding sequence was inserted by Gateway cloning into the pK7WG2 vector (Karimi et al., 2002), containing the CaMV 35S promoter and polyadenylation signal. To generate CCS52 hairpin constructs, a 390 bp fragment of CCS52B was amplified from Arabidopsis cDNA by 3 nested PCRs, using primers EVPLO51-52, EVPLO71-72 and EVPLO73-74, respectively. As Arabidopsis contains two homologues of the CCS52A gene, a fragment with maximal homology between CCS52A1 and CCS52A2 was selected for gene silencing. This 361 bp fragment was amplified from Arabidopsis cDNA by two subsequent nested PCRs using primers EVPLO75-58 and EVPLO59-60, respectively. To silence all CCS52 genes, a chimeric fragment was constructed by blunt-end cloning of the 390 bp CCS52B and 360 bp CCS52A fragments (referred as CCS52AB). Three hairpin constructs (hpCCS52A, hpCCS52B and hpCCS52AB) were created based on these gene fragments by Gateway cloning in pK7GwiWG2(II) containing the CaMV 35S promoter (Karimi et al., 2002). Arabidopsis thaliana Col-0 plants were transformed via the floral-dip method (Clough and Bent, 1998). For all knockdown lines, expression of CCS52A1, CCS52A2 and CCS52B was evaluated in 14-day-old seedlings by quantitative RT-PCR, and transgenic lines with significant reduction in expression were selected for morphological analysis and infection tests. For analysis of ectopic expression of CCS52A2, the over-expression line generated by Lammens et al. (2008) was used. Seeds were germinated on MS germination medium containing appropriate antibiotics for mutant selection. All further analyses were performed on homozygous T2 lines.

Analysis of CCS52 expression by quantitative RT-PCR

Expression levels of the CCS52 gene family members in transgenic lines were evaluated by quantitative RT-PCR. The target region chosen for specific amplification of each gene was outside the regions used for silencing. The housekeeping gene ACT2 (actin) was used as a control for mRNA quality and quantity (Table S1). The reaction mixture was prepared by a Robotics4 robot (Corbett Research, in 100 μl tubes [2 μl cDNA, 1.75 pmol for each primer, 12.5 μl ABsolute™ QPCR SYBR® Green Mix (ABgene, in a total volume of 25 μl]. A first denaturation step (5 min at 95°C) was followed by 45 cycles at 95, 58 and 72°C, respectively. The PCR reaction was completed with a melting step between 72 and 95°C, with 1°C temperature increments for 5 sec. Primers ccs52A1fq3 and ccs52A1rq3 for CCS52A1 (At4g22910), ccs52A2fq2 and ccs52A2rq2 for CCS52A2 (At4g11920), and ccs52Bfq1and ccs52Brq2 for CCS52B (At5g13840) were used. ACT2 primers (actin2fq and actin2rq) were used for the internal standard. For quantitative expression analysis of CCS52A1, CCS52A2 and CCS52B, the inline image method was used (Livak and Schmittgen, 2001). Two biological replicates and two technical replicates were performed. All data are expressed relative to the amount of CCS52 expression in wild-type Arabidopsis Col-0 plants.

Arabidopsis growth conditions and root growth and feeding site analysis

Arabidopsis seeds were surface-sterilized, germinated and infected as described by de Almeida Engler et al. (1999). Plantlets were grown vertically, with a 16 h light/8 h dark photoperiod at 21°C/18°C, respectively. Galls and syncytia of roots infected with 100 surface-sterilized freshly hatched M. incognita or H. schachtii J2 were harvested for morphological analysis at 7, 14 and 21 DAI. Samples were immediately frozen in liquid nitrogen for molecular analysis or embedded for microscopic analyses.

mRNA in situ hybridization

Fixation, embedding and sectioning were performed for galls and syncytia dissected at 7, 14 and 21 DAI. Samples were fixed (2.5% glutaraldehyde) and incubated under rotation at 4°C for 48 h. Dissected nematode feeding sites were then dehydrated for 2 h in an increasing ethanol series (15–100% ethanol). Gene-specific sense and antisense probes for CCS52A, CCS52B and DEL1 were generated as previously described (de Almeida Engler et al., 2001). The in situ hybridization procedure was performed as described by de Almeida Engler et al. (2001). Sections were stained with 0.05% toluidine blue and images were taken using a digital Axiocam camera (Zeiss) using standard dark- and bright-field optics.

Flow cytometric ploidy analyses

Plant material from 21-day-old plants was harvested and frozen in liquid nitrogen, chopped with a razor blade in 300 μl cold Galbraith buffer (Galbraith et al., 1983), and filtered over a 30 μm mesh. Nuclei were stained with 1 μg ml−1 4,6-diamidino-2-phenylindole (DAPI). One thousand nuclei for each sample were analysed using a BRYTE HS flow cytometer (Bio-Rad, At least two measurements were performed per sample. Multiple comparisons between plant lines were determined using SPSS 12.0.0 [multivariate analysis accompanied by post hoc tests; significance level of 0.05 (IBM,].

Nematode infection tests

Three-week-old Arabidopsis seedlings were inoculated with 100 J2 of sterile-cultured H. schachtii or surface-sterilized M. incognita per plant. Data shown for each line represent means ± SD of 50 seedlings analysed in two independent biological repetitions. For nematode reproduction, cysts and egg masses per plant were counted 6–8 weeks after infection, and statistically compared with wild-type Arabidopsis. The transgenic lines were compared to wild-type plants using the F test for homogeneity of variance, and an independent sample t test to compare means (significance level of 0.05).

Histochemical GUS analyses

Plants harbouring the CDKB1;1 promoter fused to the GUS gene (CDKB1;1pro:GUS) (Boudolf et al., 2004) were crossed with the T-DNA insertion mutant del1-1 and the DEL1OE line (Vlieghe et al., 2005). Homozygous lines were used for further analysis. GUS assays were performed as described by de Almeida Engler et al. (1999). Nematode-infected roots of CDKB1;1pro:GUS, CDKB1;1pro:GUSxDEL1OE and CDKB1;1pro:GUSxdel1-1 lines were harvested 7, 14 and 21 DAI, and analysed for GUS activity. Embedded infected roots were sectioned (3 μm), placed on microscope slides and mounted in Depex (Sigma). Micrographs were taken using an AxioCam digital camera (Zeiss) under dark-field illumination.

Morphological analysis and nuclear measurements

Uninfected and nematode-infected roots were fixed, embedded and imaged as described by Clément et al. (2009). For observation of nuclei, sections were stained for 5 min with 1 μg ml−1 DAPI in Pipes buffer, and rinsed in distilled water. Slides were mounted in 90% ultrapure glycerol and analysed by epifluorescence microscopy. Measurements were obtained from more than 100 nuclei within more than 50 histological sections of giant cells of Arabidopsis wild-type, down-regulated lines for CCS52 genes (hpCCS52A and hpCCS52AB) and CCS52OE at 7, 14 and 21 DAI. As nuclei of giant cells are not perfectly round, the longer and smaller dimensions were measured and the mean was calculated.

Whole-mount in situ analysis of nuclei

Nuclear analysis of cleared nematode feeding sites was performed as described by Vieira et al. (2012). Propidium iodide-stained samples were mounted with Citifluor™ on a microscope slide.

Confocal microscopy

Cleared samples were analysed using a Zeiss LSM 510 META confocal microscope. Dye excitation was achieved using a 543 nm HeNe laser, and emission light was captured using a long-pass 560 nm emission filter. Z-stacks were generated using 1 μm thick optical sections, and images are presented as maximum-brightness projections.


We thank Lucette Pjarovsky, Maria Montero, Baptiste Gautier and Floriane Marsens for helping with mutant analysis, and Nathalie Bosselut for the artwork in Figures 1 and 8. We are grateful to Eva Kondorosi (Institut des Sciences du Végétal-CNRS, Gif-sur-Ivette, cedex France) for the DNA for construction of CCS52 probes, and to Keiko Sugimoto (RIKEN Plant Science Center Yokohama Institute, Yokohama City, Kanagawa, Japan) and Keith Roberts (University of East Anglia, Norwich Research Park, Norwich, UK) for the rhl1 knockout line.