Molecular characterization and functional analysis of two new β-1,4-endoglucanase genes (Ha-eng-2, Ha-eng-3) from the cereal cyst nematode Heterodera avenae


  • H. Long,

    1. State Key Laboratory for Biology of Plant Diseases and Insect Pests, Institute of Plant Protection, Chinese Academy of Agricultural Sciences, Beijing
    2. Key Laboratory of Pests Comprehensive Governance for Tropical Crops, Ministry of Agriculture, Environmental and Plant Protection Institute, Chinese Academy of Tropical Agricultural Sciences, Danzhou, China
    Search for more papers by this author
  • D. Peng,

    Corresponding author
    • State Key Laboratory for Biology of Plant Diseases and Insect Pests, Institute of Plant Protection, Chinese Academy of Agricultural Sciences, Beijing
    Search for more papers by this author
  • W. Huang,

    1. State Key Laboratory for Biology of Plant Diseases and Insect Pests, Institute of Plant Protection, Chinese Academy of Agricultural Sciences, Beijing
    Search for more papers by this author
  • H. Peng,

    1. State Key Laboratory for Biology of Plant Diseases and Insect Pests, Institute of Plant Protection, Chinese Academy of Agricultural Sciences, Beijing
    Search for more papers by this author
  • G. Wang

    1. State Key Laboratory for Biology of Plant Diseases and Insect Pests, Institute of Plant Protection, Chinese Academy of Agricultural Sciences, Beijing
    Search for more papers by this author



Parasitism genes encoding secretory proteins expressed in the pharyngeal glands of plant-parasitic nematodes play a crucial role in nematode parasitism of plants. Two new β-1,4-endoglucanase genes (Ha-eng-2 and Ha-eng-3) expressed in the pharyngeal glands of the sedentary cyst nematode, Heterodera avenae, were cloned. Both of the predicted proteins have a putative signal peptide for secretion and a catalytic domain. Neither peptide linkers nor cellulose binding domains were present. In situ hybridization showed that the transcripts of Ha-eng-2 and Ha-eng-3 accumulated specifically in the two subventral gland cells of H. avenae. RT-PCR analysis confirmed that their transcriptions were strong in the preparasitic and early parasitic second-stage juveniles, and were undetectable at the late parasitic stages of the nematode. Cellulase activities of the recombinant proteins HA-ENG-2 and HA-ENG-3 were confirmed in vitro. Knocking down Ha-eng-2 using RNA interference reduced nematode infectivity by 40%. The results indicate that these β-1,4-endoglucanases can be secreted into plant tissues and play an important role in the wall degradation of plant cells during penetration and the migration of second-stage juveniles in host roots.


The cereal cyst nematode Heterodera avenae is considered as the most economically important plant-parasitic nematode of cereal crops worldwide, and the losses caused by H. avenae range from 30 to 100% (Bonfil et al., 2004). Its life cycle from egg to adult passes through four juvenile stages, in which only the second-stage juvenile (J2) and adult male are motile stages that are capable of migrating through the plant tissues. The J2 of H. avenae is also the infective stage that penetrates the root tip and migrates intracellularly through the cortex to the vascular cylinder, where it inserts its stylet into a selected parenchyma cell and induces its transformation into a feeding site (Jones, 1981; Heinrich et al., 1998). At the feeding site, the nematode becomes sedentary and undergoes three moults to the sexually mature adult stage while feeding (Lilley et al., 2005).

Plant cell walls are the main mechanical barrier to host penetration and migration by endoparasitic nematodes. To degrade cell walls, plant-parasitic nematodes secrete various types of cell wall-degrading enzymes and exert mechanical force by stylet thrusting. β-1,4-endoglucanases were the first cell wall-degrading enzymes to have been identified in two cyst nematode species more than a decade ago (Smant et al., 2000) and have subsequently been detected in different plant-parasitic nematodes (Hassan et al., 2010). Most of these identified β-1,4-endoglucanases (EC belong to the glycosyl hydrolase family 5 (GHF5) and show high similarity with bacterial cellulases (Mayer et al., 2011). However, the β-1, 4-endoglucanases of Bursaphelenchus xylophilus from the order Aphelenchida belong to GHF45 and have high similarities to fungal cellulases (Kikuchi et al., 2004). Horizontal gene transfer has been suggested as the origin of cell wall-degrading enzymes in plant parasitic nematodes (Jones et al., 2005).

In cyst and root-knot nematodes, cell wall-degrading enzymes are specifically produced in the nematode's subventral pharyngeal glands and injected into plant tissues through their stylet (Davis et al., 2004). Endoparasitic nematodes secreting β-1,4-endoglucanases are thought to hydrolyse the β-1,4-glycosidic bonds of cellulose in cell walls during their penetration and migration phases of the infection process in host plant tissues, because the cellulase activities of these enzymes are evident in vitro. Moreover, knockout of a β-1,4-endoglucanase gene with RNA interference significantly reduces the ability of Globodera rostochiensis to invade roots, indicating β-1,4-endoglucanase is important for plant parasitism of nematodes (Chen et al., 2005; Rehman et al., 2009).

Cloning parasitism genes encoding extracellular proteins secreted through the stylet into plant tissue is essential for understanding the molecular basis of nematode parasitism of plants (Davis et al., 2000; Williamson & Hussey, 2007). Although H. avenae causes serious yield losses of wheat worldwide, little is known about its pathogenic mechanisms, especially at the molecular level. Recently, Long et al. (2012) reported the identification of a β-1, 4-endoglucanase gene Ha-eng-1a expressed in the pharyngeal gland cells from this nematode. It consists of a signal peptide, a catalytic domain, and a cellulose-binding domain (CBD). The current study reports the cloning and characterization of two new β-1,4-endoglucanases without a CBD in H. avenae. Additionally, the RNA interference (RNAi) technique was used to study the significance of this β-1,4-endoglucanase in plant parasitism of H. avenae.

Materials and methods


Heterodera avenae was cultured on glasshouse-grown wheat Triticum aestivum cv. Wenmai 19 as described by Ferris et al. (1989), except the temperature of the seedling cultures was maintained at 16°C for the first week and then 22°C for the remaining growth period. The newly formed, dark brown cysts were incubated at 4°C for 6 weeks then at 16°C to stimulate the hatching of preparasitic second-stage juveniles (J2s). Wheat roots infected with H. avenae J2s were harvested 5, 10, 20 and 30 days post-inoculation (dpi), and different parasitic stages of H. avenae were collected by root blending and sieving (De Boer et al., 1999). Adult white females were directly hand-picked from root surfaces under a dissecting microscope at 40 dpi.

Nucleic acid extraction and synthesis

Genomic DNA was isolated from J2s (c. 100 000) of H. avenae and wheat leaves, as described by Ray et al. (1994) and Dellaporta (1993), respectively. Total RNA was isolated from different stages of H. avenae (c. 1000) with TRIzol (Invitrogen) according to the manufacturer's instructions. The mRNA was purified with the Oligotex mRNA Mini Kit (QIAGEN) from total RNA following the manufacturer's instructions, and then converted to first-strand cDNA by reverse transcription using the SuperScript III First-Strand Synthesis System for RT-PCR kit (Invitrogen) according to the manufacturer's instructions.

Isolation of cDNA and genomic DNA clones

The cellulase genes were cloned as part of an expressed sequence tag (EST) project carried out using a cDNA library derived from J2s of H. avenae (D. Peng, unpublished data). A 544 bp and a 487 bp cDNA fragment (clone 01C03 and 02D09, respectively) encoding β-1,4-endoglucanase genes were identified during a blast search of the sequences from the random sequencing of clones generated in this project. To obtain the full-length cDNA of clone 01C03, 5′ rapid amplification of cDNA ends (5′ RACE) was performed using 5′-RACE-Ready cDNA that was prepared from J2s of H. avenae and primers E2R-1 and E2R-2 (Table 1) according to the manufacturer's instructions (Invitrogen). A similar approach was used to amplify the full-length cDNA of clone 02D09 using the gene-specific primers E3R-1 and E3R-2 (Table 1).

Table 1. Primers used in this study
Primer namePrimer sequence (5′ – 3′)

Gene-specific primer pairs E2F-1 and E2F-2, and E3F-1 and E3F-2 (Table 1), flanking full-length open reading frames of Ha-eng-2 and Ha-eng-3 cDNA respectively, were designed and used to amplify the corresponding genomic sequences by long distance (LD) PCR from H. avenae genomic DNA. PCR products obtained were cloned into pGEM-T Easy vector (Promega) and sequenced.

Sequence analysis and phylogenetics

Initial sequence comparisons were conducted using blastx and blastp searching of protein databases in NCBI ( Prediction of a signal peptide for secretion and the cleavage site was performed using signalP v. 4.0 (Petersen et al., 2011). Molecular mass was predicted by protein machine ( The phylogenetic tree was generated based on the neighbour-joining method using mega v. 4 (Tamura et al., 1996). Bootstrap values were calculated from 1000 replicates.

Southern blot hybridization

Genomic DNA of H. avenae was completely digested overnight at 37°C with EcoRI or HindIII (New England BioLabs). Genomic DNA from wheat was used as the control. Digestion products were separated on a 0·8% agarose gel and blotted onto a Hybond-N+ membrane (Roche Diagnostics) using standard protocols (Sambrook et al., 1989). A DIG-labelled Ha-eng-3 probe was generated by PCR (Roche) on preparasitic J2 first-strand cDNA templates using the primers E3S1 and E3S2 (Table 1). The blot was hybridized at 42°C and detected following the instructions of the DIG High Prime DNA Labelling and Detection Starter Kit I (Roche).

Whole-mount in situ hybridization

DIG-labelled ssDNA probes were generated from the cloned cDNA fragment. Sense probes against Ha-eng-2 were prepared with the gene-specific primer E2H-1, and primer E2H-2 was used for the antisense probes. Probes against Ha-eng-3 were generated with primers E3H-1 (sense) and E3H-2 (antisense). In situ hybridization was performed according to De Boer et al. (1998) with minor modifications. Freshly hatched J2s were fixed in 3% paraformaldehyde at 4°C for 16 h, which was followed by an additional incubation at room temperature for 5 h, and hybridization was then performed overnight at 47°C.

RT-PCR analysis

Transcriptions of Ha-eng-2 and Ha-eng-3 in different developmental stages of H. avenae were determined by reverse transcription (RT)-PCR. Equal amounts of mRNA (200 ng) from each sample were used for first-strand cDNA synthesis. The resulting cDNA fragments were amplified using the primers as in the above in situ hybridizations (Table 1). Actin genes of H. avenae were amplified from each sample as a positive control using the primers ActinF and ActinR (Table 1). RT-PCR on J2s soaked in dsRNA was performed with the gene-specific primers E2i-1 and E2i-2 that were designed to not hybridize within the amplified region employed to generate dsRNA. PCR products were separated with 1·5% agarose gel electrophoresis and stained with ethidium bromide.

Heterologous expression and enzyme activity assay

Coding sequences of Ha-eng-2 and Ha-eng-3 (without predicted signal peptides) were subcloned into the expression vector pET-32a (Novagen) using primer pairs E2-SacI and E2-NotI, and E3-SacI and E3-NotI, respectively (Table 1). The recombinant plasmids were introduced into Escherichia coli strain BL21 (DE3) for expression (TaKaRa Biotech). Fusion proteins were determined on SDS-polyacrylamide gel electrophoresis (SDS-PAGE) with Coomassie brilliant blue staining. For the enzymatic assay, cell lysates of transformed E. coli were spotted on 1·5% agar, 0·5% carboxymethylcellulose (CMC; Sigma) plates. After incubation overnight at 37°C, the plates were stained with 0·1% Congo red (Sigma) and washed for 15 min with 1 m NaCl. Protein extracts from non-induced bacteria were used as a negative control.

RNAi soaking and infection assay

RNAi soaking was developed based on that described by Urwin et al. (2002) and Chen et al. (2005) with minor adaptations. Fluorescein isothiocyanate (FITC) uptake was first analysed by soaking. Freshly hatched J2s of H. avenae (c. 5000) were soaked in a solution (1 mg mL−1 FITC, 3 mm spermidine, 50 mm octopamine and 0·05% gelatin, adjusted with 0·25 × M9 buffer) for 36 h at 16°C on a rotator. FITC uptake was examined under a fluorescence microscope. For knocking down Ha-eng-2, dsRNA against Ha-eng-2 was synthesized with the Hiscribe T7 In Vitro Transcription Kit according to the manufacturer's instructions (New England BioLabs) using the primers T7E2-F and T7E2-R. Approximately 8000 freshly hatched J2s were incubated in the dsRNA solution (2 mg mL−1 dsRNA, 3 mm spermidine, 50 mm octopamine and 0·05% gelatin, adjusted with 0·25 × M9 buffer) for 36 h at 16°C. Control nematodes were soaked in solutions without dsRNA or with dsRNA targeted against gfp. For each reaction, c. 2000 J2s were used for RT-PCR, and the remaining nematodes (c. 6000) were used for infection assays. The soaking experiments with dsRNA were repeated in triplicate.

For the nematode infection assay, T. aestivum cv. Wenmai 19 surface-sterilized seeds were germinated in Petri dishes on 1% water agar plates for 2 days in the dark at room temperature. The germinated seeds were then sown in 50 mL Falcon tubes containing 30 mL of sterilized sandy soil (Seah et al., 1998). Three tubes were used with three plants per tube for each treatment. After 6 days, young plants were inoculated with soaked nematodes at a density of c. 300 J2s per tube. Roots were harvested at 10 dpi for acid fuchsin staining (Bird, 1983), and the nematodes within roots were counted under a microscope. The infection assays were repeated in triplicate and the results were analysed using Duncan's multiple range test.


Heterodera avenae cellulase cDNAs

Full-length cDNA sequences of the two clones 01C03 and 02D09 obtained by 5′ RACE were designated as Ha-eng-2 (JN861115) and Ha-eng-3 (JN861117), respectively. Ha-eng-2 cDNA comprises a total of 1215 bp (excluding the polyA tail), including a 999 bp open reading frame (ORF). The 5′-untranslated region (UTR) is 45 bp before the ATG initiation codon and the 3′-UTR is 171 bp long containing a typical polyadenylation signal (AATAAA) 11 bp upstream of the polyA tail. The ORF encodes a deduced protein of 333 amino acids with a calculated molecular weight of 36708 Da and a pI of 8·55. A signal peptide of 20 amino acids with a cleavage site between Ser20 and Leu21 was predicted by signalP at the N-terminus of the deduced protein.

The complete Ha-eng-3 cDNA is 1184 bp in length (excluding the polyA tail) and contains a 972 bp ORF encoding 324 amino acids. The 5′-UTR is 38 bp and the 3′-UTR is 174 bp long containing a polyadenylation signal (AATAAA) 12 bp upstream of the polyA tail. The molecular weight and pI of the deduced protein were calculated as 35675 Da and 9·16, respectively. A predicted secretion signal sequence terminated immediately upstream of a protease cleavage site between amino acids Ser20 and Leu21. Neither Ha-eng-2 nor Ha-eng-3 transcripts contained a consensus spliced leader sequence at the 5′ end.

Genomic clones

The corresponding genomic sequences of Ha-eng-2 (JN861114) and Ha-eng-3 (JN861116) are 2460 bp and 1919 bp long (both from ATG to the stop codon), respectively. Both Ha-eng-2 and Ha-eng-3 contain seven introns and share the same intron positions at the protein level. However, the introns in Ha-eng-2 were larger than those in Ha-eng-3 (44–610 bp vs 43–286 bp, respectively). Compared to other plant-parasitic nematode endoglucanases, Ha-eng-2 and Ha-eng-3 share the same intron numbers and locations with Hg-eng-2, Hg-eng-3, Hg-eng-4 (endoglucanase genes from H. glycines) and Gr-eng-2 (an endoglucanase gene from Globodera rostochiensis) (data not shown).

A Southern blot containing genomic DNA from H. avenae was hybridized with a DIG-labelled DNA probe generated from Ha-eng-3 cDNA. The probe hybridized to multiple fragments in both the EcoRI and HindIII digests of H. avenae genomic DNA, and the sizes of the hybridization bands ranged from 0·5 to 15 kb (Fig. 1). No hybridization signal was detected with genomic DNA of wheat (Fig. 1). Neither the genomic coding region nor the cDNA contained EcoRI or HindIII sites; thus Ha-eng-3 could be classified into a small multigene family.

Figure 1.

Southern blot analysis of Ha-eng-3. The blot contains genomic DNAs from Heterodera avenae (N) and wheat leaves (W) digested with HindIII (H) and EcoRI (E) and hybridized with a Ha-eng-3 cDNA probe. DNA size markers are indicated in kb.

Homology search and phylogenetic analysis

blastp searches revealed that the mature protein sequences of HA-ENG-2 and HA-ENG-3 showed significant similarity with those of family 5-glycosyl hydrolases. The highest similarities for HA-ENG-2 and HA-ENG-3 were with β-1,4-endoglucanases from H. glycines (AAC15708, 7e−152 and AAC48321, 9e−167, respectively), and the highest non-nematode similarities for both HA-ENG-2 and HA-ENG-3 were with Cellulophaga algicola (bacteria; ADV50035, e−85 and 5e−93, respectively). A conserved domain search showed that HA-ENG-2 and HA-ENG-3 comprise a single catalytic domain protein (from 37 to 288 aa and 37 to 283 aa for HA-ENG-2 and HA-ENG-3, respectively) which is recognized as GHF5 cellulase. The GHF5 signature and two active site glutamic acid residues were identified in HA-ENG-2 and HA-ENG-3. Neither peptide linkers nor cellulose binding domains were discerned in HA-ENG-2 and HA-ENG-3.

HA-ENG-2 and HA-ENG-3 share 80·2% identity in amino acid sequences. The catalytic domain sequences of HA-ENG-2 and HA-ENG-3 were highly conserved, being 81·6% identical to each other. The identities between catalytic domains of HA-ENG-2 and HA-ENG-1 (β-1,4-endoglucanases from H. avenae, FJ839965) and HA-ENG-3 and HA-ENG-1 were 62·1 and 69·1%, respectively. A phylogenetic tree was generated from an alignment of H. avenae sequences with GHF5 β-1,4-endoglucanases from cyst nematodes. One β-1,4-endoglucanase (Mi-eng-1, AF323087) from Meloidogyne incognita was used as the out-group. In the resulting tree, the β-1,4-endoglucanase sequences from H. avenae clustered together with the majority of cyst nematode β-1, 4-endoglucanases in a large group, with the other cyst nematode β-1, 4-endoglucanases clustering into one separate clade (Fig. 2).

Figure 2.

Phylogenetic tree of the cyst nematode endoglucanase genes. The bootstrap values are indicated at each node and were calculated from 1000 repetitions. Deduced protein sequences are from Heterodera glycines Hg-eng-1 (AF052733), Hg-eng-2 (AF052734), Hg-eng-3 (AF056048), Hg-eng-4 (AY325809), Hg-eng-5 (AY336935) and Hg-eng-6 (AY163572), Globodera rostochiensis Gr-eng-1 (AF056110), Gr-eng-2 (AF056111), Gr-eng-3 (AF408154) and Gr-eng-4 (AF408157), Heterodera schachtii Hs-eng-1 (AJ299386) and Hs-eng-2 (AJ299387), Globodera tabacum solanacearum Gts-eng-1 (AF182392) and Gts-eng-2 (AF182393). Meloidogyne javanica Mi-eng-1 (AF323087) is included in the analysis to root the resulting tree.

Expression analysis

In situ hybridization was performed to localize the expression of Ha-eng-2 and Ha-eng-3 in nematodes. DIG-labelled antisense probes generated from Ha-eng-2 and Ha-eng-3 specifically hybridized with transcripts in the two subventral gland cells of the preparasitic J2 (Fig. 3a,c respectively). No hybridization was observed in H. avenae with the control sense cDNA probes of Ha-eng-2 and Ha-eng-3 (Fig. 3b,d respectively). The developmental expression patterns of Ha-eng-2 and Ha-eng-3 were also determined by RT-PCR employing the gene-specific primers E2H-1 and E2H-2 of Ha-eng-2 and E3H-1 and E3H-2 of Ha-eng-3. The transcripts of both Ha-eng-2 and Ha-eng-3 were strongly detected in two nematode stages (pre-J2s and J2, 5 dpi) and were weak in one nematode stage (J2, 10 dpi), but were undetectable in other parasitic stages (juveniles 20 and 30 dpi, adult females; Fig. 4). The successful amplification of an actin fragment indicates that cDNA was acquired from different developmental stages. In order to discriminate any genomic amplification, the same primers used in RT-PCR were also used for the amplification of the genomic DNA of H. avenae. Different amplification sizes demonstrate the absence of genomic DNA contamination in cDNA amplifications.

Figure 3.

Localization of Ha-eng-2 and Ha-eng-3 transcripts in the pharyngeal gland cells of Heterodera avenae revealed by in situ hybridization. Fixed nematode sections were incubated with Ha-eng-2 antisense (a) or sense (b), and Ha-eng-3 antisense (c) or sense (d) cDNA probes. Scale bar: 50 μm.

Figure 4.

Developmental expressions of Ha-eng-2 and Ha-eng-3 detected by PCR amplification (24 cycles) of cDNAs from different stages of Heterodera avenae. Lanes 1–7 indicate the size of Ha-eng-2 and Ha-eng-3 amplified from cDNA templates from: Lane 1, preparasitic second-stage juveniles (pre-J2s); lane 2, 5 days post-inoculation (dpi) juveniles; lane 3, 10 dpi juveniles; lane 4, 20 dpi juveniles; lane 5, 30 dpi juveniles; lane 6, adult females; and lane 7, genomic DNA templates. All cDNA templates were amplified with primers of the β-actin gene from H. avenae as a positive control.

To test whether the HA-ENG-2 and HA-ENG-3 proteins exhibited cellulase activity, the predicted ORFs were subcloned into the pET-32a vector and the expression in E. coli was induced. After staining, bands of c. 47 and 46 kDa, corresponding to recombinant HA-ENG-2 and HA-ENG-3, respectively, were detected in the total proteins by SDS-PAGE (Fig. 5a). These sizes are similar to the molecular mass predicted from the amino acid sequences of these recombinant proteins carrying tags and lacking the N-terminal signal. Cell lysates of transformed E. coli containing Ha-eng-2 and Ha-eng-3 constructs showed significant hydrolytic activity in a cup plate assay using carboxymethylcellulose as the substrate (Fig. 5b). The lysates of E. coli transformants that were not induced to express the protein (data not shown), or that harboured the empty plasmid vector, did not show such activity (Fig. 5b). Based on these results, it is confirmed that HA-ENG-2 and HA-ENG-3 encode functional cellulases.

Figure 5.

Enzyme activity assay of recombinant proteins HA-ENG-2 and HA-ENG-3. (a) Sodium dodecyl sulphate polyacrylamide gel of the lysate of Escherichia coli cells expressing Ha-eng-2 and Ha-eng-3 from the plasmid vector pET-32a. M, molecular mass marker; lane 1, protein extract from induced E. coli cells containing Ha-eng-2 constructs; lane 2, protein extract from induced E. coli cells containing Ha-eng-3 constructs; lane 3, protein extract from non-induced E. coli cells containing Ha-eng-2 constructs; lane 4, protein extract from induced E. coli cells harbouring the empty pET-32a vector; lane 5, protein extract from non-induced E. coli cells containing Ha-eng-3 constructs. (b) Cellulase activity assay on carboxymethylcellulose agar plate test in (plate 1) and (plate 2), cell lysates of transformed E. coli containing Ha-eng-2 and Ha-eng-3 constructs, respectively. The hydrolytic activity in the substrate is visualized as a clear halo by a Congo red stain. No halo could be detected while applying the lysates of E. coli harbouring the empty plasmid (plate 3).

Influence of Ha-eng-2 RNAi on nematode infectivity

A fluorescent signal was used as a tracer to check the uptake ability of J2 of H. avenae. After 36 h of soaking in the FITC solution, the fluorescent dye could be clearly observed in the pharyngeal lumen, median bulb and intestine of the nematode (Fig. 6a). Semiquantitative RT-PCR experiments showed a decrease in the mRNA levels of Ha-eng-2 when the nematodes were soaked with dsRNA against Ha-eng-2, indicating the effective silencing of Ha-eng-2 (Fig. 6b). Infection studies showed that the treatment of J2 with Ha-eng-2 dsRNA significantly reduced the infection ability of nematodes (40·1%; = 0·001) compared to the control (Table 2). The proportion of nematodes inside the roots relative to those transferred to the roots was 22·0% (SD ± 5·7) for the nematodes treated with Ha-eng-2 dsRNA, whereas 36·7% (SD ± 2·8) of the nematodes from the gfp dsRNA treatment were found inside the roots 10 dpi (Table 2). Nematodes soaked with gfp dsRNA infected equally well as those soaked in the solution without dsRNA; the differences between the two controls were not significant (Table 2).

Figure 6.

Efficiency of RNA interference by soaking. (a) Fluorescence microscopy showing fluorescein isothiocyanate (FITC) uptake by preparasitic second-stage juveniles of Heterodera avenae. S, stylet; pl, pharyngeal lumen; m, median bulb. Scale bar: 20 μm. (b) Semiquantitative RT-PCR (24 cycles) showing the expression of β-actin (412 bp) and Ha-eng-2 (705 bp) in pre-J2s soaked in the soaking buffer without dsRNA (N), pre-J2s soaked in dsRNA against gfp (C) and pre-J2s soaked in dsRNA against Ha-eng-2 (T).

Table 2. Number of Heterodera avenae inside the roots 10 days after being inoculated with dsRNA and non-dsRNA treated nematodes
TreatmentNumber of nematodes used for inoculationNumber of penetrating nematodesPercentage of penetrating nematodesa
  1. a

    Values followed by different letters are significantly different (P = 0·05) using Duncan's multiple range test.

Non-dsRNA300116 ± 22·138·6 ± 7·4 a
gfp dsRNA300110 ± 8·536·7 ± 2·8 a
Ha-eng-2 dsRNA30066 ± 17·322·0 ± 5·7 b


This paper describes the characterization of Ha-eng-2 and Ha-eng-3 genes which encode the functional cellulases from H. avenae. The presence of the polyA tail at the 3′ end of the cDNAs and introns within the corresponding genomic DNA fragments, as well as the observation that the Ha-eng-3 probe specifically hybridized to H. avenae genomic DNA in Southern blot analysis, indicate that Ha-eng-2 and Ha-eng-3 are of nematode origin. Moreover, the deduced protein sequences HA-ENG-2 and HA-ENG-3 are more similar to prokaryotic β-1,4-endoglucanases than eukaryotic β-1,4-endoglucanases, which supports the hypothesis that nematode cell wall-degrading enzymes were acquired from bacteria through horizontal gene transfer (Jones et al., 2005).

Both HA-ENG-2 and HA-ENG-3 proteins comprise a single GHF5 catalytic domain, devoid of linker domain and cellulose-binding domain. To date, β-1,4-endoglucanases identified in the order Tylenchida consist of several protein domain structure variants, of which the largest one only includes a GHF5 catalytic domain. The nematode GHF5 endoglucanase genes must have been duplicated several times after the horizontal gene transfer event, with an occasional sequential loss of the linker and CBD (Ledger et al., 2006). Furthermore, the ancestral gene had already been duplicated in the early evolution of plant-parasitic nematodes and gained numerous introns. Thereafter, some introns were gained, lost or shifted in certain lineages, and additional duplication events took place frequently (Kyndt et al., 2008). HA-ENG-2 and HA-ENG-3 contained the conserved intron locations and clustered with major cyst nematode endoglucanases including HG-ENG-2 and HG-ENG-3 from H. glycines. Similarly, it is assumed that HA-ENG-2 and HA-ENG-3 may have lost the domains of cellulose binding and linker during subsequent duplications.

In situ hybridization revealed that both Ha-eng-2 and Ha-eng-3 transcripts specifically accumulated in the two subventral gland cells of H. avenae J2. For endoparasitic nematodes, the subventral gland cells function during the early parasitic process, i.e. plant invasion and migration of J2 through plant tissues, whereas the dorsal gland cell is involved in the induction and maintenance of the feeding site (Davis et al., 2000). The presence of a signal peptide in the predicted protein sequences of HA-ENG-2 and HA-ENG-3, the localization of Ha-eng-2 and Ha-eng-3 transcripts in the two subventral gland cells, as well as the expressions of Ha-eng-2 and Ha-eng-3, were only observed in motile infective J2s. The cellulase activity seen in vitro indicates that HA-ENG-2 and HA-ENG-3 proteins may be secreted into plant tissues by nematodes, degrading the plant cell wall and facilitating the penetration and migration of J2s in plant tissues.

To illustrate the importance of Ha-eng-2 in the parasitism of H. avenae, gene expression was knocked down using RNAi. In plant-parasitic nematodes, RNAi has been demonstrated to be effective in both sedentary and migratory endoparasites and used to analyse the function of pathogenicity factors (Maule et al., 2011). Knocking down the expression of Ha-eng-2 in preparastitic nematodes reduced the penetration rate of plants by more than 40%, which is unlikely to be attributed to chemical additives (e.g. spermidine, octopamine and gelatin) or the non-specific presence of dsRNA (gfp) as the controls contained both. The results confirm those of Chen et al. (2005) who reported that silencing two β-1,4-endoglucanase genes Gr-eng-1 and Gr-eng-2 of the potato cyst nematode Globodera rostochiensis by RNAi resulted in a significant reduction in infection. Knocking down Gr-eng-3 and Gr-eng-4 also led to a reduction of G. rostochiensis infectivity by 57% (Rehman et al., 2009). These findings are consistent with the general idea that β-1,4-endoglucanases are essential during the migration of nematodes in plant roots.

Heterodera avenae leads to significant economic yield losses of wheat worldwide, which cannot be readily managed. Application of RNAi technology may provide a new approach for its control. In contrast to other cyst nematodes (e.g. H. glycines, H. schachtii and G.  rostochiensis) and root knot nematodes, H. avenae is still poorly characterized at the molecular level. Parasitism genes from H. avenae have seldom been reported. Further studies are needed to identify more parasitism genes of H. avenae using a variety of approaches. Once the essential parasitism genes have been characterized, transgenic resistant hosts expressing the dsRNA that targets the genes may be developed, and the management of H. avenae may be reinforced by means of future biotechnology.


This research was supported by the National Key Basic Research Program of China (973 Program, 2013CB127502), an International Cooperation Research Grant of the Ministry of Science and Technology of China (2009DFB30230) and Nature Sciences Foundation of China (30921140411, 31171827). Professors Maurice Moens and Zhou Guanghe are thanked for their scientific advice and revising the manuscript.