Introduction

Potato cyst nematodes (PCN), Globodera rostochiensis and G. pallida, are considered the most economically important plant-parasitic nematodes of potato plants and can reduce yields by 70% (Brown and Sykes 1983; Greco et al. 1982). Both species are present world-wide in over 50 potato-growing countries (Baldwin and Mundo-Ocampo 1991) and are listed as quarantine organisms in Europe. Infested fields should not be used for potato seed production for several years and appropriate control measures must be applied. Eradication of PCN from the infested field is difficult to achieve. The application of chemical nematicides has a limited effect on PCN, and raises concerns by consumers and has a potential negative environmental impact. Due to the long viability of dormant cysts, very long crop rotations should be used, which is economically impracticable (Trudgill et al. 2003). Natural resistance plays an important role in PCN control as the most efficient and environmentally sound approach for their management. However, resistance can be overcome by selection for resistance-breaking types in a heterogeneous nematode population that display polymorphic pathogenicity factors. PCNs are classified into pathotypes on the basis of their ability to multiply on different Solanum clones and potato cultivars. The International Pathotype Scheme introduced by Kort et al. (1977) recognizes, among the European populations, five pathotypes within G. rostochiensis (Ro1–Ro5) and three within G. pallida (Pa1–3), while the Latin American Scheme (Canto-Saenz and de Scurrah 1977) recognizes among the South American populations of these two species four and seven pathotypes, respectively. While investigations conducted using neutral markers have never allowed a clear identification of the different pathotypes in PCN (Folkertsma et al. 1996), no such investigations have been yet carried out using selected markers like nematode pathogenicity genes.

Globodera species as most cyst nematodes reproduce by obligate outcrossing. While G. pallida and G. rostochiensis parasitize potato, tomato and egg-plant (but not pepper and tobacco), G. tabacum is usually found on tobacco but is also able to develop on tomato, egg-plant, sweet pepper and non tuber solanaceous weeds. It is believed that G. tacabum consists in fact of three to four subspecies (Marché et al. 2001) which vary in their capacity to develop on different Nicotiana species. Globodera “mexicana” was found on non tuber solanaceous weeds, it can also parasitize tomato and egg-plant but is unable to develop on potato and tobacco. When plant-parasitic nematodes invade their hosts, they need to overcome the plant’s cell-wall barrier. In order to degrade this complex structure of pectin and cellulose fibres, plant-parasitic nematodes produce a mixture of lytic enzymes in their oesophageal glands (β-1,4 endoglucanases, pectate lyase, expansin, see Davis et al. 2009 for a review) and secrete them through the stylet into the plant. The pectate lyase enzymes degrade one of the cell-wall components, the unesterified polygalacturonate (pectate). The presence of pectate lyase enzymes in animals was discovered for the first time in G. rostochiensis by Popeijus et al. (2000). Since then pectate lyases were characterized from other plant-parasitic nematodes belonging to the genera Meloidogyne (Doyle and Lambert 2002; Huang et al. 2005), Heterodera (de Boer et al. 2002; Vanholme et al. 2007) and Bursaphelenchus (Kikuchi et al. 2006). The sequences already identified in PCN indicate the presence of two types of pectate lyases showing only 28% identity and 43% similarity and that are encoded most likely by paralogous genes designated as pel-1 and pel-2 (Kudla et al. 2007). The role of PEL-2 in the infection and parasitism of plants by PPN was demonstrated in the close relative of the PCN, the beet cyst nematode H. schachtii, by silencing of pel-2 in preparasitic juveniles, which resulted in significantly reduced infection efficiency (Vanholme et al. 2007). Transient expression in planta was successfully used as an alternative to enzymatic assay to also demonstrate the functional role of G. rostochiensis PEL-2 in plant parasitism (Kudla et al. 2007).

While the studies performed in the last decade identified pectate lyase as parasitism factors in different species of nematodes, there is little information on the variability of these genes within species. More, phylogenetic inferences conducted on nematodes PEL proteins suggest that several horizontal transfers have occurred independently from different bacterial origins which have led to distinct and highly diverse lineages in this gene family (Kudla et al. 2007; Vanholme et al. 2007). Within these lineages, it is unknown whether the sequences found in nematodes with common or different host ranges have evolved by convergent evolution or common ancestry.

In this study, we have selected one of the cell-wall degrading enzymes, PEL-2, and investigated the molecular variability of this parasitism factor at different scales: (1) within a wide range of PCN populations including South American populations, (2) between different PCN pathotypes and (3) between species of Globodera showing different host range. Though we were unable to identify polymorphisms in this gene supporting the pathotype definition, correlation was detected between pel-2 polymorphism and the ability of some Globodera species to develop on different plant hosts.

Materials and Methods

Nematode Populations

Nematode populations from different geographical locations and belonging to different pathotypes and subspecies were selected to represent variability within the species G. pallida, G. “mexicana”, G. rostochiensis and G. tabacum (Table 1).

Table 1 Nematode populations and designations of clones
Full size table

DNA and RNA Extraction

Globodera rostochiensis DNA was extracted from 10 cysts of each of the nine populations. The Wizard genomic DNA purification kit (Promega) was used following slightly modified manufacturer’s instructions. Extracted DNA was diluted in 20 μl of DNA rehydration solution. G. pallida, G. “mexicana” and G. tabacum DNA was extracted from one cyst from each population that was previously rehydrated at least overnight in distilled water. The eggs of the cyst were crushed and then suspended in proteinase K buffer (10 mM Tris pH 8.8, 1 mM EDTA, 1% Nonidet P-40 and 100 μg/ml proteinase K). The samples were placed at −80°C for 15 min, incubated for 1 h at 60°C and then for 15 min at 95°C. The tubes were centrifuged and 5 μl of the supernatant was used to amplify the pel-2 gene.

Total RNA was extracted from J2s with the RNeasy Mini kit (Qiagen) according to the manufacturer’s instructions. The RNAs were treated with DNase I (10 U, Ambion) for 15 min at 37°C and then purified by phenol/chloroform/isoamyl alcohol (25/24/1) extraction. The RNAs were precipitated with 3 M sodium acetate and absolute ethanol, and stored at −80°C. The reverse transcription was performed using the “SuperScript™ III One-Step RT-PCR System with Platinum® Taq High Fidelity” kit (Invitrogen). 3′raceadaptater primer (10 μM) and dNTP (10 mM) were added to each RNA sample, heated at 65°C for 5 min and placed immediately on ice. A mix of 4 μl of 5× PCR buffer, 2 μl of DTT (0.1 M), 2 μl of MgCl2 (25 mM) and 1 μl of RNAsin was heated at 42°C and added to each sample. The reactions were heated at 42°C during 2 min before the addition of 1 μl of Superscript III per reaction. The final incubation step was performed at 42°C for 1 h.

PCR Amplification

Two sets of primers for amplifying the major part of pel-2 gene in selected species were designed. Primers pel2-F (attcgttctttgcttctttt) and pel2-R (ttcttatcaccaatcgatca) based on previously determined Gr-pel-2 mRNA sequence (accession AY094613) were developed using Primer3 software (Rozen and Skaletsky 2000) and yielded PCR products in G. pallida and G. rostochiensis. PCR amplification in G. rostochiensis was performed in a total volume of 50 μl using 2 μl of extracted DNA, 100 μM of each dNTP, 0.1 μM of each primer, 3.0 mM MgCl2, 2 U of GoTaq polymerase (Promega) and the corresponding buffer. The PCR amplification was carried out in a Gene Amp PCR system 2700 thermal cycler (Applied Biosystems) or a gradient thermal cycler MJ Research PTC200 DNA Engine Gradient (GMI). The PCR protocol started with initial denaturation at 94°C for 3 min, followed by 35 cycles of denaturation at 94°C for 45 s, annealing at 51°C for 1 min and extension at 72°C for 2 min. The reaction was terminated by 7 min final extension at 72°C. Amplicons were visualised on 1% agarose gel. For G. pallida, G. “mexicana” and G. tabacum, the pel-2 gene was amplified in a final volume of 25 μl containing 1× Taq buffer, 2.5 mM MgCl2, 160 μM of each dNTP, 0.5 μM of each primer and 1 U of Goflexi Taq polymerase (Promega). The PCR program consisted of 5 min at 96°C, 35 cycles with 1 min at 96°C, 1 min at 58°C, 1 min at 72°C, and a final elongation step at 72°C during 7 min.

Since the first primer set resulted in weak or no amplification in most of the G. tabacum populations, a second internal degenerate primer pair was designed for G. tabacum. Primers Gt pel2-F (WWScKtYYtttttgccaatt) and Gt pel2-R (cYgaaatgatgCaRtttttg) were used to generate PCR products in the G. tabacum populations tested. Compared to the previous pel2 primer pair, Gt pel2-F and Gt pel2-R are located, respectively, 15 bp downstream pel2-F and 118 bp upstream pel2-R.

RT-PCR Amplification

RT-PCRs were conducted using two sets of primers. Primers pel2-F and pel2-R were used to amplify the pel-2 gene and primers pel1-Fwd4 (atgctttttgttatcatttcaa) and pel1-Rev3 (tagttgacaattttaaYWgcc) were used to amplify the pel-1 gene. PCR amplifications were performed in a total volume of 25 μl using 2 μl of the reverse transcriptase, 200 μM of each dNTP, 0.4 μM of each primer, 3.0 mM MgCl2, 2 U of GoTaq polymerase (Promega) and the corresponding buffer. The PCR protocol started with initial denaturation at 95°C for 1 min, followed by 30 cycles of denaturation at 95°C for 30 s, annealing at 56°C for 50 s and extension at 72°C for 1 min. The reaction was terminated by 5 min final extension at 72°C. Amplicons were visualised on 1% agarose gels.

Cloning and Isolation of Plasmid DNA

PCR products were cloned into the pGEM®-T Easy vector (Promega) and transformed into competent cells of E. coli JM109 (Promega) according to the manufacturer’s instructions. Plasmid DNA was isolated by JetQuick Plasmid Miniprep (Genomed) according to the manufacturer’s instructions.

Sequence Analysis and Evolutionary Analysis

The isolated plasmids were sequenced using universal primers SP6 and T7 by Macrogen Inc. (Korea). DNA sequences were assembled, further analysed and aligned using the computer software BioEdit v. 7.0.5.2 (Hall 1999). The alignments were manually corrected when necessary. Similarity to known sequences was determined by NCBI BLAST server (Altschul et al. 1997). Introns were determined by alignment with the corresponding mRNA, excised, and DNA sequences were translated into predicted protein sequences by BioEdit. This sequence alignment was subsequently imported in DnaSP 5.10 (Librado and Rozas 2009) for sequence analysis. Using, respectively, a window size and step size of 50 bp and 20 bp length, we looked for spatially varying selection that will inflate nucleotide diversity overall samples (πT) when compared to intra-population nucleotide diversity (πS). Neutrality tests (Tajima’s D and McDonald Kreitman’s test) were also computed in DnaSP. Genetic distances of DNA were estimated by Kimura 2 parameter model which takes into account transition versus transversion bias (higher rates of transitions than transversions). Phylogenetic trees were constructed using the maximum parsimony, neighbour-joining and minimum evolution methods implemented in Mega 4 (Tamura et al. 2007). Choice of the model and model parameters used to construct the trees were carried out using Modeltest 3.8 (Posada 2006).

Selective pressures were evaluated using the ratio of nonsynonymous to synonymous substitution rates per site (ω = K a/K s) using methods accessible via the DataMonkey web server (Kosakovsky Pond and Frost 2005; Scheffler et al. 2006; Delport et al. 2010).

Three-dimensional protein models were constructed with Swiss-Model Workspace (Arnold et al. 2006), while protein structures were compared by superimposing pdb files by TopMatch (Sippl and Wiederstein 2008; Sippl 2008).

Results

Structure of the pel-2 Gene in Globodera

Sequences of pel-2 gene were determined for G. pallida, G. “mexicana”, G. rostochiensis and G. tabacum clones representing variability within and between the species, by analysing populations from different geographical locations. Determined sequences were deposited at Genbank: 35 for G. pallida and G. “mexicana” (accessions GQ426889–GQ426923), 20 for G. rostochiensis (FJ810102–FJ810121) and 23 for G. tabacum (GQ426924–GQ426946). Determined DNA sequences of 1.028–1.136 bp long for G. pallida, G. “mexicana” and G. rostochiensis represent the major part of pel-2 gene, where the 5′-end is missing 5 aa of the signal peptide in comparison with previously predicted Gr-PEL-2 (AAM21970) (Kudla et al. 2007). Sequences determined for G. tabacum cover 943 – 950 bp of pel-2 gene where the 5′-end is missing 10 aa of the signal peptide and the 3′-end of the DNA sequences is missing the last 5 aa of the Gr-PEL-2 and the UTS.

Three exons and two introns were determined in the partial sequences of pel-2 genes by aligning the genomic sequences with corresponding cDNA. The introns follow the GT-AG rule for cis-splicing and vary in their size (Fig. 1). All the introns were located in the same positions in all of the sequences with intron 1 inserted in phase 2 and intron 2 inserted in phase 0 in a protosplice site corresponding to a KTG/KK sequence.

Fig. 1
figure 1

Comparison of the genomic pel-2 DNA sequences from different Globodera species showing differences in intron sizes among sequences of the same species and between species. Sizes are indicated in nucleotides, non shaded parts represent introns and shaded parts exons. Predominant sizes observed at least in half of the clones of each species sequence dataset are indicated in bold, exponents indicated after each intron sizes represent the number of populations in each species where this variant was observed. Gtab G. tabacum, Gros G. rostochiensis and Gpal G. pallida

Full size image

After excision of introns, DNA sequences were translated into 247 aa long protein sequences (237 in G. tabacum). Short deletions or point mutations led to premature stop codons in one, two and six of the predicted protein sequences in G. rostochiensis, G. pallida and G. tabacum, respectively and were therefore presumed to be pseudogenes. Additionally, the only two sequences obtained for G. “mexicana” were also presumed to be pseudogenes. The putative pseudogenes were omitted from the further analyses of variability to avoid any bias they could introduce into results. Two DNA sequences of one G. rostochiensis population were identical (clones B7i and B7k), while all the rest of determined DNA sequences were different from each other. BLASTP analysis showed high similarity to the previously described Gr-PEL-2 (AAM21970; Kudla et al. 2007) sequence with E values between 2e−134–8e−137, 2e−121–4e−124, e−108–2e−110, for G. rostochiensis, G. pallida and G. tabacum sequences, respectively. Best BLASTP hits with non nematode organisms were obtained in all cases with bacterial (Streptomyces sp.) pectate lyases. Due to the synonymous mutations 18, 31 and 17 diverse DNA sequences were translated into 12, 21 and 14 diverse protein sequences of G. rostochiensis, G. pallida and G. tabacum, respectively.

Variability of the pel-2 Gene in Globodera

Variability among determined aligned DNA sequences can be observed through the sequences as point mutations affecting 373 positions (236 transitions and 137 transversions), short (up to 21 bp long) insertions or deletions and one longer (41 bp) indel in the first pel-2 intron of G. rostochiensis (supplementary material, Figure S1). All together 89 indels were detected in our sequence dataset, all located in the non coding parts of the gene (58 in intron 1 and 31 in intron 2). Entropy plots of intra-species DNA variability (Figure S1) revealed a relatively constant variability along the stretch of DNA sequences in all species, with somewhat higher variability in the first intron.

The polymorphism analysis of the coding region within each species is summarized in Table 2. The three investigated Globodera species showed from 5.5 to 9.2% of polymorphic sites with slightly more variation observed in G. pallida compared to the two other Globodera species. The highest variability was found at the third base of the codons, comprising over one half of variable sites. However, 32–42% of the pel-2 mutations occurred in the first and second positions of the codons that globally correspond to non-synonymous mutations. In overall, the polymorphic sites observed in our dataset were equally distributed between synonymous and replacement mutations.

Table 2 Polymorphisms of the 741 (712 in G. tabacum) bp long coding region of pel-2
Full size table

To determine whether positive selection pressure was acting in these data, we carried out evolutionary analyses using two neutrality tests (Tajima’D and McDonald Kreitman) implemented in DnaSP. None of these neutrality tests provided significant results (Table 2). Results of the McDonald and Kreitman’s test showed an excess of radical fixation while only the G. tabacum versus G. rostochiensis comparison showed an excess of radical polymorphism which is consistent with the slightly increase of replacements versus synonymous mutations observed in these two species compared to G. pallida. Complementary evolutionary analyses based on the estimation of the dN/dS ratio were conducted using the PARRIS method available through the DataMonkey web server. Again, no significant sign of positive selection was detected since the M2 model for positive selection (LnL = 2808.88) does not appear better adapted to the dataset than the M1 model for neutrality (LnL = 2808.88). Site-specific methods like SLAC, IFEL and REL revealed 7–18 sites (7 by REL, 12 by SLAC and 18 by IFEL) significantly (posterior probability >95%) under negative (purifying) selection with 7 sites found by all three methods (T30, T96, G129, Q139, K174, Q206 and S247; numbering according to Gr-pel-2 complete cds). No site under positive selection (posterior probability >95%) was found common by all three methods.

Species and Pathotype Specific Polymorphisms

Protein sequence variability was observed within and between the species at, respectively, 62 and 51 variable amino acid sites. Among the 51 inter species variable sites, only 8, 9 and 15 appeared as conserved between all determined clones of G. rostochiensis, G. pallida and G. tabacum, respectively, making them potentially interesting sites to develop tools for species-specific identification. One of such position is position 114 (numbering according to Gr-PEL-2 protein) which was found to be conserved within all the clones of a given species but different in each species (K114, H114 and Y114 encoded by AAA, CAC and TAC in G. tabacum, G. rostochiensis and G. pallida, respectively). Additionally, 46 variable nucleotides were found to distinguish between PCN and TCN (tobacco cyst nematode, G. tabacum). Among these, 17 have no impact on the protein sequence (synonymous change) but 29 will affect the protein sequence in 24 different codons (K/T24, A/E33, K/I36, S/V38, R/H61, R/K83, E/D117, S/A120, G/N132, K/H134, K/R149, K/V168, N/K171, A/V176, Q/E178, I/L192, T/N208, S/C216, V/E217, K/Q225, P/K228, I/V235, S/A238, G/A241).

Among all the intra-species variability observed no clear correlation can be established between these polymorphisms and PCN pathotype classifications or TCN subspecies. In order to get insights on the putative usefulness of pel-2 to account for these differences, we took advantage of working with different populations in each species to look for spatially varying selection along the coding sequence. Comparison of the nucleotide diversity overall samples with the intra-population nucleotide diversity revealed different patterns of variation among the three investigated species (Fig. 2). No inflation of total nucleotide diversity compared to within population nucleotide diversity was observed in TCN suggesting that PEL-2 is not accountable for differences between the TCN subspecies. However, some inflation was observed for G. pallida in the COOH end terminal part of the protein, suggesting that PEL-2 could be accountable for differences between PCN pathotypes.

Fig. 2
figure 2

Difference between nucleotide diversity computed overall populations (πT) and within populations (πS) in various populations of G. tabacum, G. rostochiensis and G. pallida. Peaks are indicative of spatially varying selection

Full size image

Three-Dimensional Modelling to Assess the Functional Impact of the Amino Acid Polymorphisms Observed

Seven residues were previously determined as the active site and calcium binding site in Gr-PEL2: E57, D82, E102, D130, K131, K153, R156 (Kudla et al. 2007). Five of these sites (E42, D115, K116, K138, R141) appeared conserved in all our determined sequences while D67 (aspartic acid) was substituted with V (valine) and G (glycine) in two G. tabacum clones (GTT1-26, GTA1-11, respectively) and E87 (glutamic acid) was substituted with (G) glycine in G. rostochiensis clone B7a. Furthermore, most of the PL3 family members have two sets of closely spaced cysteine residues to support two disulphide bonds, present also in the majority of our dataset (C86, C90, C158 and C161). However, in two sequences (GPS8-42 and GTS1-21) one cysteine was replaced with arginine (R) affecting the second disulphide bond. In order to get insight in the functional impact of such polymorphisms, three-dimensional protein structure modelling was carried out on these variants and models were built to pinpoint differences between predicted PEL-2 proteins of G. pallida, G. tabacum and G. rostochiensis. Automated fold recognition technique identified the pectate lyase PEL15 from Bacillus sp. strain KSM-P15 (PDB accession 1ee6) as the best fitting template for all determined PEL-2 sequences and pectate lyase PELI from Erwinia chrysanthemi (PDB accession 3b8yA) as an alternative template in two clones GPS4-27, GTA2-37). All predicted three-dimensional protein models determined in this study have similar three-dimensional structure. No structural difference was observed between the variants at the active site positions 67 and 87. Interspecies comparison of the predicted tertiary structure of Globodera PEL-2 isozymes revealed only minor differences (Fig. 3). None of these differences in proteins folding comprises the seven aa residues involved in catalytic activity and calcium binding suggesting similar enzymatic activity of all forms.

Fig. 3
figure 3

The predicted three-dimensional models of the pectate lyase 2 protein structure for G. pallida, G. tabacum and G. rostochiensis clones were constructed based on Bacillus sp. PEL15 pectate lyase (PDB 1ee6) and Erwinia chrysanthemi PELI pectate lyase (PDB accession 3b8yA). Representative clones of Gp-PEL2 (GPS4-27, GPS9-7), Gt-PEL2 (GTT5-41, GTA2-37) and Gr-PEL2 (B1 h, B9j and B10a) display slightly different tertiary structure of the isozymes with variable regions indicated by black arrows

Full size image

Phylogenetic Analysis and Putative Convergent Evolution of pel-2 in PCN Species

Analysis of the phylogenetic relationships among the pectate lyase sequences is presented in Fig. 4. The tree obtained from the exons of pel-2 sequences determined in this study clusters together sequences from the same species in monophyletic groups with a high bootstrap support, but present poor phylogenetic resolution at the intra-specific level even when the intron sequences are added to the dataset (data not shown). As expected from this low intra-species resolution, neither the PCN nor TCN populations form separate groups according to their pathotype or subspecies classification. The G. “mexicana” sequences clustered within the G. pallida with 100% support of the node. Surprisingly whatever the reconstruction method used (maximum parsimony, neighbour-joining or minimum evolution), the PCN, G. rostochiensis and G. pallida pel-2 sequences appeared more similar to each other than to G. tabacum.

Fig. 4
figure 4

Phylogenetic analysis of the Globodera pel-2 sequences. Here the dendrogram obtained using the Neighbor Joining reconstruction method from the exons of pel-2 sequences is shown, but similar species topology was observed using the Minimum Evolution or Maximum Parsimony reconstruction methods on either DNA or amino acid sequence dataset. Phylogenetic analysis were conducted using 1000 bootstrap and the Tamura-Nei model (gamma = 0.62) as defined by Modeltest

Full size image

In order to get insight in this unusual species topology, genetic distances of pairwise comparisons were estimated using the Kimura 2 parameters model. A very clear multimodal distribution was observed for the pel-2 gene corresponding to three separate normal distributions (Fig. 5). The first mode corresponds to the Globodera intra-species comparisons. The means of this first mode was very similar among all species considered (0.02 for G. pallida and 0.01 for both G. rostochiensis and G. tabacum). The second and third modes both correspond to comparisons at the inter species level. In contrast to the first mode, the mean value differs according to the species comparison studied. The smallest distance is observed among PCN species (mean = 0.08), while higher genetic distances were observed when comparing PCN to TCN (mean = 0.13), suggesting that evolution of pel-2 occurred at a different rate in TCN than in PCN.

Fig. 5
figure 5

Distribution of the Globodera pel-2 pairwise comparisons among classes of genetic distances calculated with the Kimura 2 parameters (K2P) model. Comparisons between intraspecific sequences are indicated by dots column, comparisons between sequences of G. pallida and G. rostochiensis in dash columns and comparisons between G. tabacum (tobacco cyst nematode) and the two other Globodera species (potato cyst nematode) in black columns

Full size image

pel-2 Expression in G. “mexicana”

As pel-2 sequences we obtained in G. “mexicana” were determined as pseudogenes following in silico analysis, transcription of this gene was tested in G. “mexicana” infective preparasitic juveniles (J2). It has been shown previously that numerous parasitic genes, including Gr-pel-2, are expressed in J2 larvae (Kudla et al. 2007), therefore only this developmental stage was tested. No amplification of pel-2 transcripts was observed in G. “mexicana” while a product of the expected size was obtained in G. pallida (Fig. 6). Lack of pel-2 expression in the juvenile stage of G. “mexicana” was confirmed. Nonetheless other members of the pectate lyase gene family are expressed in G. “mexicana”, as demonstrated by the successful transcript amplification of the pel-1 copy in both G. pallida and G. “mexicana” (Fig. 6).

Fig. 6
figure 6

Transcript amplification of pel-2 and pel-1 by RT-PCR in G. “mexicana” and G. pallida. Expected amplicon sizes are 840 bp for pel-2 and 785 bp for pel-1. M DNA ladder (each 100 bp). Lanes 1–3 negative control, G. pallida cDNA and G. “mexicana” cDNA, respectively

Full size image

Discussion

In this study, we investigated the sequence variability of the parasitism gene, pel-2, among different populations and species of Globodera. Substantial variability was observed in the dataset which is consistent with host range preferences of the species investigated. The sequences obtained display the same intron/exon structure with introns inserted in the same position in all the sequences suggesting that orthologous rather than paralogous sequences were obtained during the study. Paralogous copies of other nematode cell-wall degrading enzymes, like cellulase genes, have large variations in intron positions (Gao et al. 2004) and other studies support the high conservation of the intron/exon gene structure among clades (Rokas et al. 1999; Wada et al. 2002), though there are contradictions to this rule (Goetze 2006). As already observed in Meloidogyne for the Mj-cbp-1 gene (Adam et al. 2008), intron length was found quite variable within species as well as within populations. Presence of large indels (>40 bp) contributed to intron variability in the case of this pectate lyase gene. The introns follow the GT-AG rule for cis-splicing but the protosplice site of these introns (KTG/KK) differs slightly from the common eukaryotic MAG/GT protosplice site (Qiu et al. 2004) suggesting a nascent spliceosome, a feature related either to the basal position of nematodes among the eukaryotes or to the acquisition of this gene through horizontal gene transfer from rhizobacteria as suggested from previously published phylogenetic analysis (Kudla et al. 2007; Vanholme et al. 2007).

Relatively large intra-species variability of the studied sequences was observed between clones of G. rostochiensis, G. pallida and G. tabacum. Despite a high percentage of non synonymous variation among the point mutations observed, no sign of positive selection was detected in our dataset while sites under negative selection were robustly predicted by different methods indicating that pel-2 have evolved in the Globodera genus under purifying rather than diversifying selection. Therefore, the variability observed among our sequences could be explained either as a large variability of a single gene or by the existence of a small pel-2 gene family with very similar gene structure. The presence of a small multigene family was already confirmed by hybridization for PEL-1 and PEL-2 in Meloidogyne and Heterodera (Doyle and Lambert 2002; Huang et al. 2005; Vanholme et al. 2007). Recently, the genome sequence of M. incognita revealed 30 pectate lyases from the PL3 family (Abad et al. 2008). However, from our genomic sequencing one cannot rule out that the different sequences obtained represent isozymes rather than allozymes. The ongoing genome sequencing projects on G. pallida (http://www.sanger.ac.uk/sequencing/Globodera/pallida/) and G. rostochiensis (Michael Rott, personal communication) should reveal the exact copy number of pel-2 gene present in these species and their sequence polymorphism.

Although three-dimensional models showed similar PEL-2 protein structure in all determined clones, suggesting similar protein function, 47 protein sequences were determined with variability in primary structure, including the active and Ca2+ binding sites. Moreover, the substitutions of negative and hydrophilic D67 with nonpolar and hydrophobic V and G as well as the substitution of negatively charged E87 with the nonpolar G could influence the activity of these particular variants. These modifications were observed in G. tabacum tabacum from Connecticut, USA, G. tabacum azteca from Mexico and in G. rostochiensis from Venezuela, all native American populations where the host range diversity is wide. Therefore, different predicted proteins could represent enzymatic variants with different levels of activity or slightly different catalytic potential for different substrates. However, their potentially different enzymatic activities were not tested in this study.

Similarity analysis does not cluster clones from the same pathotypes as distinct clades. Although the analysis was planned to reveal possible differences between PCN pathotypes, the sequence variability results of this pathogenicity gene do not support the actual classification of PCN pathotypes. We cannot rule out that no association between sequence variants and pathotypes was detected because of insufficient number of populations and clones of certain pathotypes in our study. Nonetheless, we should notice that it is difficult to obtain numerous populations representing the different pathotypes and/or to make the pathotype profiling for new populations which is a very time and nematode consuming task. Our results emphasize that the intra population variability of pel-2 can be as high as the inter population variability. If such a feature is not restricted to pel-2 but common to most of the parasitism genes it will be impossible to find in these genes any molecular support to the pathotype definition. In the cases of G. rostochiensis and G. tabacum, the lack of correlation observed can be explained by the high intra-population variability of pel-2. However, the picture is not as clear in the case of G. pallida. For this gene and in this species, the pattern of variation observed in the COOH end terminal part of the protein revealed an inflation of the inter-population diversity compared to the intra-population diversity, a pattern that will have been expected from the existence of pathotypes or from local adaptation among the investigated populations. No polymorphism in the COOH end terminal part of PEL-2 correlate to the actual G. pallida pathotype classification. However, other groupings can be observed and one can notice that nucleotide position A/G689 perfectly distinguishes between clade I (+Europe) G. pallida populations and Peruvian populations of clades II–V. Such partition among the G. pallida populations appears in accordance with some pathogenicity results (see below).

The definition of PCN into pathotypes has been questioned and disputed in the past Trudgill 1985; Mugniéry et al. 1989; Nijboer and Parlevliet 1990) and despite many attempts to differentiate PCN pathotypes with molecular (Folkertsma et al. 1994, 1996; Gonzalez et al. 1995; Blok et al. 1998; Bendezu and Evans 2001) or biochemical approaches (Fox and Atkinson 1984; Phillips et al. 1992; Fullaondo et al. 2001; da Cunha et al. 2006), bio-tests as a way of defining PCN pathotypes have not been replaced since the pathotyping scheme does not correlate perfectly with genetic variability observed in PCN. It is questionable what classification would then best represent nematode pathogenicity differences, since the new Council Directive 2007/33/EC strongly advises the use of resistant varieties to control PCN. Though, not supported by a high bootstrap value, the pel-2 phylogenetic analysis allows the distinction of Peruvian G. pallida populations from clade I (populations GPS3 and GPS4) that appear strongly related to the European populations from Peruvian populations of clades II–V. These results are in accordance with the determination of the origin of Western European populations that are all derived from clade I (Plantard et al. 2008). But it should be noticed that these results are also in accordance with the spectrum of efficiency of many potato resistances that when found efficient against European G. pallida populations are also efficient against Peruvian clade I populations but not against the other Peruvian clades (Mugniéry et al. 2007; Grenier et al. 2010). However, for the practical reasons in potato production, the resistant potato variety could be selected by bio-tests even without determination of pathotype or virulence group of the PCN population present in a certain area (Urek et al. 2008).

An interesting feature is that for the pectate lyase 2 gene and protein the trees present a topology different from the admitted species topology. G. rostochiensis and G. pallida pel-2 and PEL-2 sequences are more similar to each other than to G. tabacum. On the other hand, in the accepted phylogenetic tree of the cyst nematodes based on the ITS region of rRNA gene, G. tabacum clusters together with G. rostochiensis and forms a sister clade to G. pallida (Subbotin and Moens 2006). As both G. rostochiensis and G. pallida display the same host range but a different one from G. tabacum, which cannot parasitize potato plants, it is tempting to speculate that pel-2 gene polymorphism may form the basis of this adaptation. It is known that the pectic polysaccharides in plant cell walls are structurally diverse (Willats et al. 2001) and it has been noticed that potato and sugar beet have some particular modifications of the homogalacturonal, the widespread and abundant domain of plant pectins (Pauly and Scheller 2000).

To develop this hypothesis further, we investigated more closely the situation of the pel-2 gene in G. “mexicana” a close relative of G. pallida as shown in Fig. 4 but that is unable to develop on potato. None of the clones determined for G. “mexicana” was predicted to correspond to a functional copy as premature stop codons were observed in the translated sequences obtained. In order to corroborate the potential implication of pel-2 in the ability of Globodera species to develop on potato, we have tested pel-2 expression in G. “mexicana”. RT-PCR results showed absence of pel-2 transcript in this species but their presence in G. pallida suggesting that the reason for the inability of this particular species to develop on cultivated potato can also be related to this pectate lyase gene copy. Given the absence of a transformation system for plant-parasitic nematodes, RNAi appears as a useful tool for analysing the function of nematode genes. It would be expected if our hypothesis holds true that silencing the pel-2 copy in PCNs will affect the ability of these nematodes to develop on potato but not on other host plants like tomato and egg-plant which are common hosts for G. rostochiensis, G. pallida and G. “mexicana”. Careful design of such RNAi experiment is required to ensure that only one particular copy among the pectate lyase gene family will be impaired. A task that will be easier to achieve once the complete genomic arsenal will be described in G. pallida.