A molecular analysis of desiccation tolerance mechanisms in the anhydrobiotic nematode Panagrolaimus superbus using expressed sequenced tags

EST assembly

Of the 100 unigenes with unique hits to the NCBI nr database, 38 returned BLAST hits only to bacterial taxa. Bacterial sequences had been screened and removed from the dataset during the assembly process using stringent matching criteria (BLASTN with an e-value cut-off of 1e^-50), thus some of these 38 sequences may correspond to residual contaminant sequences. Others may represent sequences from bacterial associates of P. superbus, or horizontally transferred sequences, e.g. one unigene sequence (PSC01785) had highest similarity to an ankyrin gene from a Wolbachia endosymbiont from Culex quinquefasciatus[28]. The remaining 62 sequences returned hits to eukaryote taxa and these included several hits to plant LEA (late embryogenesis abundant) sequences, important in plant desiccation tolerance. PSC01785 had best similarity to an exo-β-1,3-glucanase, a member of the glycosyl hydrolase family 5 (GHF5), from the Stramenopile Oomycete Phythophthora infestans[29]. β-1,3-glucans are a major structural component of fungal cell walls and, as a microbivore, P. superbus may also use GH5 glucanases to incorporate soil fungi into its diet. Endo-β-1,3-glucanase GHF5 genes have been reported from the fungal feeding nematodes Bursaphelenchus xylophilus, B. mucronatus[30] and Aphelenchus avenae[31], with horizontal gene transfer (HGT) from bacteria being proposed to explain their evolutionary origin. Endo-β-1,4-glucanase (cellulase) genes, also belonging to GHF5, have been reported from several species of plant parasitic nematodes of the order Tylenchida [32], with HGT from bacteria also proposed as their likely source [33]. The GHF5 cellulases recently reported from the free living nematode Pristionchus[34] are most similar to sequences from two slime mould species (Amoebozoa). Thus the P. superbus sequence reported here represents another possible HGT event (this time from the eukaryote Stramenopile lineage) for nematode GHF5 genes.

Large EST datasets have been generated for nematode parasites, however with the exception of the model nematodes C. elegans, other Caenorhabditis species and Pristionchus pacificus, there are relatively few EST datasets for free-living nematodes. Metagenomic analyses of nematode ESTs [35, 36] show that nematode gene-space is substantially under-sampled. BLAST analyses of the EST cluster consensus sequences from 37 species represented in NemPep3 showed that 34-70% of the sequences from each nematode species were unique to that species [35]. Data from Nembase4 (derived from 54 parasites and 8 free living species) show that 61.8% of the predicted proteins had no GO, EC or KEGG annotations [37]. Thus the finding that 51.7% of the P. superbus unigenes correspond to novel sequences is consistent with previous studies and is a reflection of the diversity of nematode gene space.

Abundantly expressed transcripts

Ten abundantly expressed sequences were associated with reproductive function, eight corresponding to major sperm protein genes (MSPs) and two to vitellogenin genes. In total we detected 32 P. superbus MSP unigenes and 7 vitellogenin unigenes. In C. elegans MSPs are encoded by a multigene family comprising more than 50 genes [47, 48]. Nematode MSPs are small, basic proteins required for the amoeboid movement of sperm. A family of six genes vit-1 to vit-6 encode C. elegans vitellogenin [49, 50], a major yolk component which is expressed exclusively in the adult hermaphrodite intestine from which it is secreted into the pseudocoelomic space and taken up by oocytes [51]. Four structural genes were abundantly expressed in P. superbus: an actin family member (homolog of C. elegans act-2); a gene encoding a core histone protein required for chromatin assembly and chromosome function [52] and genes encoding two proteins associated with the 60S ribosomal subunit. Two abundantly expressed contigs were associated with lipid metabolism. PSC00187 encodes a homolog of C. elegans HEH-1 and human NPC2/He1, a cholesterol-binding protein whose deficiency in humans causes Niemann-Pick type C2 disease involving retention of cholesterol in lysosome [53, 54]. Transcripts for the mitochondrially encoded cytochrome c oxidase subunit 1, essential for oxidative phosphorylation and ATP synthesis, were also highly expressed in this mixed stage P. superbus library.

Twenty one of the abundant sequences listed in Table 2 are novel. These novel unigenes correspond to 641 ESTs, representing 8.4% of the total EST dataset. Data on the predicted physico-chemical parameters of the putative proteins encoded by these 21 unigenes are presented in Additional file 1. Thirteen (65%) of these novel unigenes encode a signal peptide indicative of a secreted protein. The association between sequence novelty and likely secretion has been noted previously in the parasitic nemataode Nippostrongylus brasiliensis[55]. Three of the putative novel proteins are predicted to be natively unfolded over 80-100% of their primary sequence. The P. superbus dataset contains a total of 2,059 novel unigenes. Further analysis of these novel sequences is presented in a later section.

Functional annotation

In order to identify candidate stress-related genes which may have a role in anhydrobiosis the annot8r program [56] was used to assign KEGG (Kyoto Encyclopaedia of Genes and Genomes) pathway annotations [57] and Gene Ontology (GO) terms [58] to the P. superbus unigenes.

Assignments to metabolic pathways using KEGG

One thousand six hundred and eighty four KEGG orthology assignments were inferred by searching for P. superbus unigenes that have homologs among the default set of manually curated eukaryotic genes in the KEGG database (which contains 26 genomes); similarly 1,412 KEGG assignments specific to the C. elegans genome were also inferred (Table 3). KEGG pathways associated with metabolism had the highest representation, with a large number of the P. superbus sequences associated with the 'carbohydrate metabolism', 'energy metabolism', 'lipid metabolism' and 'amino acid metabolism' pathways. In the environmental information processing category, 'signal transduction' was highly represented. Other highly represented pathways were found in the genetic information processing category including 'translation' and 'folding, sorting and degradation' and a large number of sequences had KEGG assignments to the human neurodegenerative disease sub-category. Many neurodegenerative diseases are associated with the dysfunction or overload of the protection systems responsible for repairing or degrading damaged proteins and macromolecules [59–61]. Cells exposed to severe water stress experience serious damage to their macromolecules and membranes; proteins lose their structures, become unfolded and aggregate. Thus anhydrobiotic organisms are adapted to survive cellular dehydration by deploying efficient cellular protection and repair systems (Figure 2) and it is likely that some gene products that have roles in anhydrobiotic protection in nematodes may also have human homologs which are required for neural survival. For example the molecular chaperone DJ-1, which is associated with familial Parkinson's disease [62, 63], is also upregulated in response to desiccation stress in the anhydrobiotic nematode Aphelenchus avenae[64]; and AAvLEA1, a natively unfolded late embryogeneis (LEA) protein which is upregulated in response to desiccation stress in A. avenae[65], has been shown in vitro to protect complex mixtures of proteins from aggregation [66].

KEGG Pathway Category	KEGG mapping of P. superbus transcripts to biochemical pathways
	Eukaryotes	C. elegans
1. Metabolism	474	438
1.1 Carbohydrate Metabolism	113	102
1.2 Energy Metabolism	70	68
1.3 Lipid Metabolism	61	47
1.4 Nucleotide Metabolism	36	33
1.5 Amino Acid Metabolism	76	71
1.6 Metabolism of Other Amino Acids	26	24
1.7 Glycan Biosynthesis and Metabolism	21	19
1.8 Metabolism of Cofactors and Vitamins	22	18
1.9 Biosynthesis of Polyketides and Terpenoids	7	9
1.10 Biosynthesis of Secondary Metabolites	13	14
1.11 Xenobiotics Biodegradation and Metabolism	29	33
2. Genetic Information Processing	294	272
2.1 Transcription	50	43
2.2 Translation	132	124
2.3 Folding, Sorting and Degradation	91	87
2.4 Replication and Repair	21	18
2.5 RNA Family	0	0
3. Environmental Information Processing	88	62
3.1 Membrane Transport	4	3
3.2 Signal Transduction	73	53
3.3 Signalling Molecules and Interaction	11	6
4. Cellular Processes	236	163
4.1 Transport and catabolism	91	73
4.2 Cell Motility	19	10
4.3 Cell Growth and Death	63	39
4.4. Cell Communication	63	41
5. Organismal Systems	250	181
5.1 Immune System	52	38
5.2 Endocrine System	60	49
5.3 Circulatory System	21	15
5.4 Digestive System	44	31
5.4 Excretory System	15	14
5.5 Nervous System	20	10
5.6 Sensory System	19	11
5.7 Development	10	5
5.8 Environmental Adaptation	9	8
6. Human Diseases	342	296
6.1 Cancers	63	39
6.2 Immune System Diseases	22	21
6.3 Neurodegenerative Diseases	134	130
6.4 Cardiovascular Diseases	36	33
6.5 Metabolic Diseases	2	3
6.6 Infectious Diseases	85	70
Total	1684	1412

Putative anhdyrobiotic and stress response genes constitutively expressed in unstressed P. superbus

A comparison the P. superbus EST unigene dataset with EST datasets from other anhydrobiotic nematodes

The 3,982 P. superbus unigenes were compared to EST unigene datasets from three other species of anhydrobiotic nematodes [31, 102, 149] to identify putative homologous protein families which may reveal some of the core anhydrobiotic processes shared by these nematodes. Plectus murrayi is an Antarctic soil nematode adapted to survive desiccation and freezing [102]. Aphelenchus avenae is a slow desiccation strategist soil dwelling fungiverous nematode. Ditylenchus africanus is an endoparasite of plants with peanut as its primary host. It migrates to the pods and seeds of the ground nut and can survive in an anhydrobiotic state in the seeds [149]. The phylogenetic relationships of these nematodes are indicated in Figure 1.

The combined dataset from the four anhydrobiotic nematodes comprised 10,791 unigenes. All against all BLASTP analyses of the predicted peptide sequences for these unigenes, followed by their classification into putative homologous groups using the TRIBE Markov clustering algorithm as implemented in the MCL software package [150], has identified 7,063 unigene families, where 6,308 consist of singletons. The distribution of these unigene families across the four nematode species is summarized in Figure 3(a). A total of 67 unigene families contain transcripts from all four anhydrobiotic nematodes. While our analysis is based on an incomplete coverage of the transcriptomes of all four nematodes, these 67 families provide a first indication of subsets of genes common to the four species, some of which may be involved in anhydrobiotic processes. The annotation of the 67 unigene families (inferred by BLAST), together with a list of their component P. superbus contigs is presented in Additional file 5. These families include representatives of several of the anhydrobiotic and stress response proteins discussed in the previous section. Among these are protein kinases and HMG proteins; glutathione S-transferase; sHSP, HSP70; HSP90; peptidyl-prolyl cis-trans isomerase; several components of the UPS system and RIC1, a poorly characterized family which encodes plasma membrane proteins that are expressed in response to high salt or low temperature conditions in plants [144]. Members of the nematode specific transthyretin-related (ttr) family [151] are also included among the 67 unigenes. The function of ttr genes remains elusive; so far the function of just one nematode ttr gene product has been discovered (TTR-52 mediates the recognition and engulfment of apoptotic cells in C. elegans[152]). Figure 3(b) shows the distribution of homologs of the P. superbus putative stress response genes from Table 4 across the four anhydrobiotic nematode species. These BLAST annotations are presented in full in Additional file 3. Since this analysis is based on partial transcriptomes of the four nematodes the results need to be interpreted conservatively, however the data show that these four anhydrobiotic nematodes express a great diversity of stress responsive genes. Surprisingly none of the 13 LEA unigenes were common to all four nematode datsets, and 8 LEA sequences were found only in P. superbus. This may indicate that constitutive expression of LEA transcripts is higher in P. superbus than in the three other anhydrobiotic nematodes. When more complete coverage of the transcriptomes of anhydrobiotic nematodes and other anhydrobiotic animals becomes available comparative transcriptomic analyses will be a powerful tool for the identification of candidate genes and processes required for successful anhydrobiotic survival.

Analysis of novel ESTs

Of the 3,982 unigenes in our dataset 2,059 (51.7%) have no significant similarity to any sequences in the Genbank or NemPep databases. The Prot4EST algorithm [46] was used to translate these novel unigenes into putative peptides. Analysis of the physical properties of these putative peptides reveals that 149 of them are predicted to lack a fixed tertiary structure (100% intrinsically disordered), while an additional 296 peptides are predicted to be 50-99% disordered. Intrinsically disordered proteins (IDPs) are hydrophilic, being characterized by a high proportion of polar and charged amino acids and low sequence complexity; they also have a low content of the hydrophobic amino acids which would normally form the core of a folded globular protein [153]. These physical features also occur in LEA proteins. A plot the hydropathy [111] of P. superbus putative novel peptides and the 13 predicted P. superbus LEA proteins against their predicted degree of disorder (determined using the IUPred program [110]) shows that there are 225 P. superbus peptides with a GRAVY (Grand Average of Hydropathy) value of ≤ -1 and > 50% disordered (Figure 4).

Garay-Arroyo et al.[154] proposed that LEA proteins are contained within a larger group of proteins called 'hydrophilins' that accumulate in response to osmotic stress in prokaryotes and eukaryotes. The characteristics that define this group are a glycine content of greater than 6% and hydropathy index of less than -1. Our dataset contains 170 novel putative peptides that meet these criteria (Figure 5). The P. superbus unigenes predicted to encode LEA proteins (Table 5) were identified on the basis of BLAST searches. Analysis of their physical properties reveals that all of these putative LEA proteins are hydrophilic, having GRAVY values ranging from -0.62 to -1.42; six are predicted to be 100% unstructured, a further three are largely (77-99%) unstructured, but three putative LEA unigenes: PSC01853, PSC00514 and PSC00416 are predicted to be only partly disordered (9, 10 and 18% disorder, respectively). Eleven of the 13 putative LEA sequences also have a glycine content of greater than 6%.

Many IDPs proteins function by molecular recognition: either by transient, or permanent binding to a structured partner molecule [155]. However the functions of some IDPs depend directly on the extended random coil conformation of the disordered state--the so-called entropic chain effect [156]. Entropic chain effects are likely to be central to many of the functions of LEA proteins. The elongated, natively unfolded conformation of LEA proteins may help to form a "molecular shield" [66], preventing protein aggregation and denaturation. These hydrophilic, proteins also have the capacity to bind and retain water molecules and, at later stages of dehydration process, an abundance of charged amino acids may enable some LEA proteins to replace water at the hydrogen bonding sites of dehydrated proteins. Although LEA proteins are natively unfolded when fully hydrated, some LEA proteins, including AavLEA1 [157], have been shown to develop secondary structure as they become desiccated [158], leading to the suggestion that some LEA proteins might function as intracellular space-filler molecules which prevent the collapse of cells as they become desiccated [105]. The combined group of putative hydrophilic proteins identified in Figures 3 and 4 contains 294 individual novel sequences. These sequences represent an important group of candidate anhydrobiotic genes that merit further investigation.

Expression of putative stress related genes upon desiccation

P. superbus is capable of surviving exposure to a dry atmosphere in desiccation chambers containing silica gel without the need for prior preincubation to mild desiccation stress [15]. However it likely that in its natural habitat P. superbus would experience more gradual change from a condition in which its cells and tissues are fully hydrated to one of extreme dehydration. In addition, intrinsic behavioural (coiling/clumping) responses or morphological adaptations (such as surface lipids [14] or possibly SXP/RAL-2 cuticular proteins) may slow the rate of water loss in P. superbus and allow time for inducible molecular protection mechanisms to be put in place. We used qPCR to investigate the inducible expression of several unigenes that represent homologues of stress related genes in other organisms. The expression of five of the 19 genes tested was upregulated in P. superbus following exposure to 98% RH for 12 h (Figure 6).

Three antioxidant genes gpx (glutathione peroxidase), dj-1 and prx (which encodes a 1-Cys peroxiredoxin) were upregulated in response to desiccation stress in P. superbus. ROS accumulation is triggered by cellular dehydration and our qPCR data show the importance of enzymatic antioxidant defense systems during the induction of anhydrobiosis. Glutathione peroxidases (GPx) catalyse the reduction of H₂O₂ and GPx have been previously found to be upregulated in A. avenae and in P. murrayi in response to desiccation [64, 102]. DJ-1 is a multifunctional protein associated with familial Parkinson's disease [62, 63]. One of its proposed functions is a redox-dependent molecular chaperone activity [63] and a role for DJ-1 as an atypical peroxiredoxin-like peroxidase in inactivating mitochondrial H₂O₂ has also been proposed [159]. We show that the expression of dj-1 is upregulated 9-fold in P. superbus in response to desiccation stress. This gene is also upregulated in response to desiccation stress in the anhydrobiotic nematode Aphelenchus avenae[64]. Peroxiredoxins (Prx) comprise two classes: 1-Cys Prx and 2-Cys Prx, based on the number of cysteinyl residues directly involved in catalysis [92]. Animal Prx sequences comprise 3 clades [160]: clades A and B contain 2-Cys Prx, while 1-Cys Prx occur in clade C which also contains plant 1-Cys Prx sequences [160]. In plants 1-Cys Prx are seed-specific [161]: they accumulate during seed maturation and their expression declines during germination, an expression pattern is also characteristic of many lea genes. A seed-specific 1-Cys Prx, was found to be abundantly expressed during desiccation in the leaves of the resurrection plant Xerophyta viscosa[162] and transcripts encoding a 1-Cys Prx are also upregulated during rehydration of the anhydrobiotic moss Tortula ruralis[163]. Here we show that a 1-Cys Prx is upregulated in response to desiccation in P. superbus, revealing a further parallel between the desiccation tolerance mechanisms of anhydrobiotic nematodes and plants.

Only one of the three lea sequences tested was upregulated, but this sequence (PSC01853) was upregulated 9.8 fold in response to desiccation stress. Of the four P. superbus hsp sequences assayed only one, an shsp sequence, was upregulated in response to desiccation. The genes encoding HSP70 and HSP90 are constitutively expressed in the Antarctic nematode Plectus murrayi and are not upregulated further by desiccation [164]. Similarly only 2 of 6 hsp70 paralogues show higher expression levels in diapausing eggs of the rotifer Brachionus plicatilis than in other metabolically active life stages [165]. sHSP are the major holding chaperones which prevent the irreversible aggregation of unfolding proteins [122, 123]. They have been shown to accumulate in anhydrobiotic encysted larvae of the brine shrimp Artemia franciscana[124] and they are also abundantly expressed in diapausing eggs of B. plicatilis[165]. These expression profiles for representatives of different HSP classes in anhydrobiotic animals from three different phyla suggest that sHSP proteins, in particular, have an important role in maintaining the integrity of the proteome during anhydrobiosis.

Nematode culture

Panagrolaimus superbus (strain DF5050) was obtained from Prof. Bjorn Sohlenius, Swedish Museum of Natural History, Stockholm. The nematodes were cultured at 20°C in the dark on nematode growth medium (NGM) plates containing a lawn streptomycin resistant E. coli strain HB101 obtained from the Caenorhabditis Genetics Center, University of Minnesota, USA. The NGM was supplemented with streptomycin sulfate (30 μg ml^-1).

cDNA library construction and EST generation

Total RNA was extracted from mixed stage unstressed worms using the TRIzol^® reagent (Invitrogen, Carlsbad, USA). The cDNA library was prepared using the SMART™ cDNA Library Construction Kit Long-Distance (LD) PCR protocol (Clontech, Mountain View, CA 94043 USA). Fifty ng of total RNA was used for the SMART cDNA synthesis and there were 25 PCR cycles in the LD PCR amplification step. The cDNAs were cloned into the pDNR-Lib vector (Clontech) and transformed into E. coli DH10B cells. A total of 15,360 recombinant E. coli were picked using a Q-Bot™ robot (Genetix, Hampshire BH25 5NN, UK) and transferred to 384 well microtitre plates containing freezing media [166] and chloramphenicol (30 μg ml^-1) and the plates were stored at -80°C. The cDNA inserts from individual transformants (n = 9,216) from the cDNA library were sequenced by the Sanger method at the Scottish Crop Research Institute, Dundee (4,224 clones) and at The GenePool, University of Edinburgh (4,992 clones).

Clustering and sequence analysis

The raw EST sequences were processed through the PartiGene pipeline [27], first using trace2dbEST which removes vector-derived sequences, poor quality sequences and ESTs shorter than 150 bp, followed by CLOBB [167], an iterative program which groups the sequences on the basis of BLAST similarity into clusters that are putatively derived from the same gene. Clusters containing more than one sequence were then assembled into consensus sequences using phrap (http://www.phrap.org). The partial transcriptome consists of these consensus sequences, along with those clusters that contain only one sequence (singletons). Potential bacterial contaminant sequences (28 contigs) and nematode rRNA genes (27 contigs) were identified using a BLASTN search of the P. superbus consensus sequences against the GenBank nucleotide database (nt), with an e-value cut-off of 1e^-50 to identify significant matches. Genes encoded on the mitochondrial genome were identified by BLASTN using 27 nematode mtDNA genomes from GenBank as queries against the P. superbus consensus sequences with an e-value cut-off of 1e^-10.

The consensus sequence for each unigene was translated using prot4EST [46]. Each unigene was then subjected to a BLASTP search (e-value cut-off of 1e^-4) against a non-redundant custom database containing sequences from a variety of sources: the GenBank NR database, Wormpep (version 224) (http://www.sanger.ac.uk/Projects/C_elegans/WORMBASE/current/wormpep.shtml) and an extended version of the Nempep4 database (http://www.nematodes.org/nembase4/), which we have named Nempep4+. Nempep4+ has been supplemented with sequences from the following nematodes: Plectus murrayi, Ditylenchus africanus, Aphelenchus avenae, Trihinella spiralis, Wuchereria bancrofti, Loa loa and Pristionchus pacificus. Putative genes were annotated using the annot8r algorithm [56]. This software tool assigns Gene Ontology (GO) terms, Enzyme Commission (EC) numbers [168] and Kyoto Encyclopaedia of Genes and Genomes (KEGG) pathway data [57] to EST sequences based on BLAST searches against annotated subsets of the EMBL UniProt database [169]. All BLAST results were parsed and the corresponding annotations were saved in a relational postgreSQL database (http://www.postgresql.org). A web interface where the annot8r annotations can be subjected to keyword queries and where output clusters can retrieved is available at http://www.nematodes.org/nembase4/species_info.php?species=PSC. Supplemental ortholog assignment and pathway mapping were carried out using the KAAS-KEGG Automatic Annotation Server [170].

To identify putative unigene families among four anhydrobiotic nematodes, EST consensus sequences were kindly provided by various groups: 1,387 Plectus murrayi sequences [102] from Dr. B. Adhikari; 2,596 Ditylenchus africanus sequences [149] from Dr. A. Haegeman and 2,700 A. avenae sequences [31] from Dr. N. Karim. All ESTs were translated into peptide sequences using prot4EST [46] and they were subjected to an all-vs-all BLASTP analysis to identify pairwise similarities. A graph representation of the homologous relationships among the unigenes was constructed, where each node is a unigene and an edge is drawn between any two nodes that have a BLASTP match. Each edge is weighted by -logE where E is the e-value of the alignment between two similar unigenes. E-values of 0 are transformed into 1e^-200, i.e. an edge weight of 200. This graph is then used by the MCL algorithm [150] as input, with an inflation parameter of 2.1, to classify the unigenes into putative families.

Translation and primary structure analysis of novel ESTs

Novel ESTs were translated into putative peptide sequences using prot4EST which incorporates the ESTScan2.0 [171] and DECODER [172] programs which use 'de novo' prediction methods for predicting the amino acid sequence of cDNA sequences with putative sequencing errors (particularly insertions and deletions). The glycine content and the Grand Average of Hydropathy (GRAVY) values of these putative peptides were determined using the ProtParam tool available at http://www.expasy.ch/tools/protparam.html[173]. The GRAVY value is calculated as the sum of hydropathy values of all the amino acids, divided by the number of residues in the sequence [111]. Predictions of the extent of intrinsically disordered regions within each putative peptide were determined using the IUPred program (using the long disorder prediction algorithm) [110] kindly provided by Dr. Zsuzsanna Dosztányi.

Real-time relative qPCR analysis of gene expression

A mixed population of nematodes was vacuum filtered onto 25 mm Supor^® Membrane Disc Filters at a concentration of 2,000 nematodes per filter. Five replicate filters were prepared for each treatment. The filters were placed in an 8 L glass desiccation chamber over a saturated solution of potassium dichromate (to generate an RH of 98% [174]) for 12 h at 20°C in the dark. Nematodes were then washed off the filters with distilled water and the nematodes from the five filters were pooled together. RNA was extracted using the TRIsure™ (BIO-38033, Bioline) method followed by treatment with DNAse I (Promega, M6101). Control nematodes were placed directly in to TRIsure and flash frozen with liquid N₂ without vacuum filtration.

Total RNA (1 μg per reaction) was converted to cDNA using the Roche Transcriptor First Strand cDNA Synthesis Kit (04 379 012 001). One μl of cDNA from the above reaction was used for each real time qPCR reaction. These reactions were carried out on a Roche LightCycler 480 thermocycler using Roche SYBR I Master 1 kit (04 707 516 001). Each qPCR reaction also contained 5 μl SYBR Master Mix, 0.002 pmole of each primer and 2 μl H₂0. Primers were designed to produce an amplicon of approximately 125 bp for each gene tested. (These primer sequences are presented in Additional file 6). Relative expression data were calculated with the LightCycler 480 Efficiency Method analysis software using the second derivative maximum option. The P. superbus ama-1 and rpl32 genes were used as a reference. Having established that the crossing point data were normally distributed and that the variance of the controls and treatment data were equal, two sample Student's t-tests were carried out to identify statistically significant differences in expression levels between the controls and the experimental treatments.

Availability of supporting data

These P. superbus EST sequences have been deposited in dbEST with the accession numbers GW405912-GW413517. The unigene sequences and annotations along with their constituent ESTs can be downloaded from the NEMBASE4 database [37] at http://www.nematodes.org/nembase4/species_info.php?species=PSC and the unigene annotations can also be subjected to keyword queries at the NEMBASE4 database.