Nematology 00 (2015) 1-17 brill.com/nemy

The plant parasite Pratylenchus coffeaecarries a minimal nematode genome

Mark BURKE 1,2,3, Elizabeth H. SCHOLL 4, David McK. BIRD 1,4, Jennifer E. SCHAFF 5,Steve COLEMAN 2, Randy CROWELL 6, Stephen DIENER 3, Oksana GORDON 6, Steven GRAHAM 3,

Xinguo WANG 6, Eric WINDHAM 3, Garron M. WRIGHT 3 and Charles H. OPPERMAN 4,∗1 Bioinformatics Research Center, NC State University, Box 7614, Raleigh, NC 27695-7614, USA

2 David H. Murdock Research Institute, General Administration, 150 Research Campus Drive,Kannapolis, NC 28081, USA

3 David H. Murdock Research Institute, Information Resources & Bioinformatics, 150 Research Campus Drive,Kannapolis, NC 28081, USA

4 Plant Nematode Genetics Group, Department of Plant Pathology, NC State University, Box 7253,Raleigh, NC 27695-7253, USA

5 Genomic Sciences Laboratory, NC State University, Box 7614, Raleigh, NC 27695-7614, USA6 David H. Murdock Research Institute, Genomics Sequencing Laboratory, 150 Research Campus Drive,

Kannapolis, NC 28081, USA

Received: 30 October 2014; revised: 10 April 2015Accepted for publication: 11 April 2015

Summary – Here we report the genome sequence of the lesion nematode, Pratylenchus coffeae, a significant pest of banana andother staple crops in tropical and sub-tropical regions worldwide. Initial analysis of the 19.67 Mb genome reveals 6712 proteinencoding genes, the smallest number found in a metazoan, although sufficient to make a nematode. Significantly, no developmentalor physiological pathways are obviously missing when compared to the model free-living nematode Caenorhabditis elegans, whichpossesses approximately 21 000 genes. The highly streamlined P. coffeae genome may reveal a remarkable functional plasticity innematode genomes and may also indicate evolutionary routes to increased specialisation in other nematode genera. In addition, the P.coffeae genome may begin to reveal the core set of genes necessary to make a multicellular animal. Nematodes exhibit striking diversityin the niches they occupy, and the sequence of P. coffeae is a tool to begin to unravel the mechanisms that enable the extraordinarysuccess of this phylum as both free-living and parasitic forms. Unlike the sedentary endoparasitic root-knot nematodes (Meloidogynespp.), P. coffeae is a root-lesion nematode that does not establish a feeding site within the root. Because the P. coffeae nematode genomeencodes fewer than half the number of genes found in the genomes of root-knot nematodes, comparative analysis to determine genesP. coffeae does not carry may help to define development of more sophisticated forms of nematode-plant interactions. The P. coffeaegenome sequence may help to define timelines related to evolution of parasitism amongst nematodes. The genome of P. coffeae is asignificant new tool to understand not only nematode evolution but animal biology in general.

Keywords – Caenorhabditis elegans, lesion nematode, Meloidogyne hapla, migratory endoparasitic nematodes, nematode evolution,protein encoding genes.

Collectively, plant-parasitic nematodes cause over US$150 × 109 annual crop losses worldwide. Although themajority of damage is caused by sedentary endoparasiticforms, migratory endoparasites cause extensive losses ina number of critical staple crops. The problem in thesub-tropics and tropics is particularly severe, and manydeveloping nations are seriously impacted in both food

∗ Corresponding author, e-mail: warthog@ncsu.edu

security and economics by plant-parasitic nematodes.Pratylenchus coffeae (lesion nematode) is found world-wide, though distributed primarily in tropical and sub-tropical regions (Gowen et al., 2005). Root-lesion nema-todes, Pratylenchus spp., are among the most economi-cally damaging plant-parasitic nematodes and are foundin a wide variety of crops, yet basic research on these

© Koninklijke Brill NV, Leiden, 2015 DOI 10.1163/15685411-00002901

M. Burke et al.

has been largely neglected in favour of the sedentary en-doparasites, including the root-knot and cyst nematodes.Pratylenchus spp. attack a wide variety of crop speciesand causes substantial damage to the cortical tissue, re-sulting in decay and development of ‘lesions’, which areavenues for fungal and bacterial infection. A devastatingpathogen of banana, plantain, vegetables and citrus, thehost range of P. coffeae includes more than 250 differentplant species. It is often found in combination with otherphytophagous nematodes, including the burrowing nema-tode, Radopholus similis, the spiral nematode, Helicoty-lenchus multicinctus, and the root-knot nematodes, Me-loidogyne spp. Pratylenchus coffeae probably originatedin the Pacific Rim/Southeast Asia region but is now dis-tributed worldwide. It was most likely transported to newregions on banana planting material (corms) during colo-nial expansion, and it remains the most important nema-tode pest of both diploid and triploid bananas in its cen-tre of origin. It is also a significant problem in Centraland South America on Cavendish cultivars (AAA), andin Africa. In Ghana, P. coffeae is reported to cause upto 60% production loss of plantains (AAB) and it is alsoa significant problem in South Africa. In other Africanbanana-growing regions its distribution is more localised,suggesting it may be a more recent introduction in theseareas.

Pratylenchus coffeae is a migratory endoparasite of theroot cortex and corm of banana, plantain and other Musaspecies. All life stages and both sexes of P. coffeae invadeand feed in the root and corm, and reproduction and eggdeposition occur over a 20-30-day period at 25-30°C. At19.6 Mb, P. coffeae has been reported to have the smallestgenome size yet reported for a multicellular animal,as determined by flow cytometry (Leroy et al., 2003).Infection by P. coffeae causes extensive necrosis of thecortex and corms, which results in lesions and breakingof roots; P. coffeae does not penetrate the stele (vascularcylinder) of the root. The lesion nematode enters the plantby puncturing the cell walls of the root and inserting itshead into the cell. Once the cell wall has been penetrated,the nematode injects the cell with saliva and subsequentlyingests the contents of the cell. Nematode activity withinthe root cortex causes reddish-brown to black, elongatedlesions that are readily seen when the roots are split open.The spread of the resulting lesions are often exacerbatedby the invasion of other rot-causing organisms, with rotsextending 2 cm or more into the corms, causing thecondition known as “blackhead” (Gowen et al., 2005).Root systems are weakened or destroyed, causing lack of

vigour and poor fruiting in infested plants. Such plantsare readily blown over, or “toppled”, and the roots areexposed (Gowen et al., 2005). When lesion nematodesare controlled, plant density is higher, bunch weight andfrequency of fruiting are increased and the need forreplanting reduced. Yield increases of 30-60% are oftenrecorded.

Damage to the root system results in stunting ofthe plant, decreased bunch weight, lengthening of theproduction cycle and toppling or uprooting of the plant.Indirectly, the reduced root growth results in reduced soilorganic matter and leaching of nutrients and increasederosion, furthering the decline in production.

In contrast to the migratory endoparasitic nematodes,the sedentary endoparasites (including the root-knot ne-matodes, Meloidogyne spp.) form complex feeding siteswithin the root system of the host plant. Infective Meloi-dogyne spp. juveniles are free in the soil and are func-tionally analogous to dauer larvae in Caenorhabditis ele-gans. They penetrate the root, preferentially in the zone ofelongation or at the site of a lateral root emergence, andmigrate intercellularly into the vascular cylinder, causinglittle or no wound response or damage to the plant in con-trast to migratory endoparasitic nematodes. Once in thevascular cylinder, the nematode makes a commitment toestablish a feeding site. Although the basis for this deci-sion is unknown, the events that immediately ensue arecentral to the host-parasite interaction, and involve dra-matic changes both in plant and nematode, leading to gi-ant cell induction and gall formation. The migration phasewithin the root is accompanied by extensive secretion ofproteins by the infecting nematode and this is also thecase for the lesion nematodes. Nematodes have a num-ber of secretory systems, and secretions play numerousand central roles in host-parasite interactions (Smant etal., 1998; Mitchum et al., 2013; Quentin et al., 2013). Allplant-parasitic nematodes have an extensible stylet con-nected to a muscular pharynx with three or five associatedgland cells. Various enzymatic functions for the secretionshave been proposed, and initial cloning and sequencingof genes encoding gland proteins have permitted the na-ture of the secretion products to be discerned with con-fidence. Mature Meloidogyne spp. females release hun-dreds of eggs into a proteinaceous matrix on the surface ofthe root. Following a first moult in the egg, motile second-stage infective juveniles hatch in the soil and typically re-infect the same plant.

2 Nematology

Genome of Pratylenchus coffeae

In this paper we report the genome sequence of Praty-lenchus coffeae and compare it with information on thegenomes of C. elegans and Meloidogyne spp.

Materials and methods

We sequenced the genome of P. coffeae using Roche454 technology from genomic DNA obtained from cul-tures collected from a banana plantation maintained onbanana at CORBANA, Costa Rica, and from cultured P.coffeae maintained on carrot discs at INRA-Martinique.We generated 0.615 Gb of useable sequence from WGSand two sets of paired-end libraries (3 kb and 8 kb), equat-ing to 31-fold coverage of the 19.6 Mb genome. We as-sembled the reads using Newbler and annotated using avariety of platforms. Transposable elements, non-codingRNAs and the protein-coding gene set were inferred us-ing a combination of predictive modelling and homology-based approaches. Orthology and synteny analyses wereconducted using established methods. We utilised pre-viously sequenced transcriptome data sequenced frommixed-stage P. coffeae populations to assemble contigsand used these data to aid gene predictions and explorekey genes associated with parasitism, behaviour, repro-duction and development. All proteins predicted from thegene set were annotated using databases for conservedprotein domains, gene ontology annotations and modelnematodes (C. elegans and Meloidogyne hapla). Essen-tiality predictions were conducted using established andin-house methods.

REPEAT ANALYSIS

RepeatMasker (v3.2.8) (Smit et al., 2004) was usedto calculate the distribution and abundance of repetitivesequences and the best match selected from overlappingmatches in RepeatMasker output to avoid calculating thesame region multiple times. Transposable elements (TEs)in the assembly were identified by de novo identificationof repeat families in the assembly based on signaturesand manual curation performed to determine potentialcoding regions on intact TEs. Our analysis determinedthat the majority of unannotated elements contained noopen reading frames longer than 25-30 bp, nor did theyhave significant matches to repetitive elements in thepublic databases.

PROTEIN CODING GENE PREDICTION

GlimmerHMM was used for first round ab initiogene prediction. A gff file containing C. elegans genepredictions was parsed for exon start and end positionsfor each gene. The exon positions and genome assemblywere used as a training set for GlimmerHMM (Haas et al.,2008). The trained data were then used to scan the entireassembly for predicted genes, resulting in 7533 predictedgenes.

In the second round of gene prediction, a referencedataset of 12 560 P. coffeae ESTs (>100 bp) was manuallycurated from EST clusters (Haegeman et al., 2011a) andpredictions of highly conserved genes were used for geneprediction training. Of these, 1975 had matches to theNCBI NR database (<e-25) and C. elegans WormBaserelease 221 (wormbase.org) and were used to train the ab-initio gene predictor Augustus (Stanke & Morgenstern,2005), resulting in 4931 predicted genes.

GeneDetective was used to compare P. coffeae to theC. elegans predicted protein database (WB release 221),resulting in 6268 unique matches. GeneDetective was alsoused to compare P. coffeae to the M. hapla predictedprotein database (HapPep 3.0), resulting in 5685 uniquematches.

We combined predictions for Glimmer and Augustuswith homology matches to WormPep, HapPep and NCBINR manually to resolve unique gene regions within the P.coffeae genome, resulting in the predicted protein datasetof 6712.

FUNCTIONAL ANNOTATION

Initial functional annotation was performed using In-terProScan to search against the InterPro protein fam-ily database, which included PROSITE, PRINTS, Pfam,ProDom, SMART, TIGRFAMs, PIR SuperFamily andSUPERFAMILY (Zdobnov & Apweiler, 2001; Batemanet al., 2004). A Pfam search (version 24.0) was also per-formed independently for the P. coffeae genome. GeneOntology annotation was derived using Blast2GO soft-ware (Goti et al., 2008) based on the BLAST (Altschul etal., 1990) match against NCBI non-redundant (NR) pro-teins with an E-value cutoff of 1e-10 and InterProScanresults.

Assignments to conserved positions in metabolic andregulatory pathways were performed using Blast2GOsoftware based on the KEGG annotation resource (Kane-hisa & Goto, 2000). KEGG genes and KO term annota-

Vol. 00(0), 2015 3

M. Burke et al.

tions were assigned based on similarity searches with a1e-5 E-value cut-off.

ANNOTATION OF CARBOHYDRATE ACTIVE ENZYMES

(CAZYMES)

We annotated putative carbohydrate active enzymes(CAZymes) by association between CAZyme familiesand Pfam domains with an E-value threshold of 0.01and a bitscore threshold of 55. The annotation wassupplemented and confirmed manually using BLASTsearch similarities and protein length matches. Putativefunctions of the proteins were predicted by similarity toknown protein modules and presence of catalytic sitesusing BLASTP search against NCBI’s Conserved DomainDatabase service and InterProScan (http://www.ebi.ac.uk/Tools/InterProScan).

EFFECTOR CANDIDATES INVOLVED IN HOST-PARASITE

INTERACTION

BLASTP was used to search for P. coffeae homologuesof effectors from M. hapla, M. incognita, Heteroderaglycines, Globodera pallida and Bursaphelenchus xylo-philus. An E-value cutoff of 1e-5 was used to identifysignificant matches. In addition, candidate effectors weresought from the P. coffeae protein set using a bioinfor-matic approach. Secreted proteins were identified as thosehaving a potential signal peptide at their N-terminus pre-dicted by SignalP 3.0 (Petersen et al., 2011) and no trans-membrane domain within the mature peptide as predictedusing TMHMM 2.0 (Krogh et al., 2001).

RNAI PATHWAY

A total of 78 proteins known to be involved in coreaspects of the C. elegans RNAi pathway were identifiedfrom the literature and WormMart. Protein sequences, in-cluding known isoforms, were downloaded from Worm-Base (release WS221) and used as queries in TBLASTNand BLASTP searches against the predicted protein andcontig databases. Positive BLAST hits (with a bitscore �40 and an E-value � 0.01) were translated in all sixreading frames, and analysed for domain structure byBLASTP (through NCBI’s Conserved Domain Databaseservice) and InterProScan. The appropriate reading framein each case was then subjected to reciprocal TBLASTNand BLASTP against the C. elegans non-redundant nu-cleotide and protein databases on the NCBI BLASTserver (http://www.ncbi.nlm.nih.gov/BLAST), using de-fault settings. The identity of the top-scoring reciprocal

BLAST hit was accepted as identity of the relevant pri-mary hit, as long as that identity was also supported bydomain structure analysis.

DAUER-RELATED GENES

Protein sequences known to be involved in dauer for-mation and maintenance were retrieved from WormBaseand used as search strings in a series of tBLASTn andBLASTP searches against P. coffeae genome and proteinsequences. An E-value cutoff of 1e-10 was used to iden-tify significant matches.

CHEMOSENSORY BEHAVIOUR-RELATED GENES

We utilised genes identified in C. elegans knownto be involved in chemosensory behaviours to identifypotential orthologues in the P. coffeae genome. BLASTPand InterProScan were used to search for P. coffeaeorthologues of these proteins. All primary BLASTP hitsreturning with an E-value � 0.0001 and coverage ratio �0.7 were analysed for identity and domain structure byBLASTP (NCBI’s Conserved Domain Database service),WormBase (WS221) and TMHMM 2.0.

DATA AVAILABILITY

Raw sequence data has been deposited at NCBI (avail-able 25 March 2015) under BioProject ID PRJNA276478and accession numbers SRS857861, SRX891472,SRX892877 and SRX892880. Project informationis available online at http://www.ncbi.nlm.nih.gov/bioproject/?term=PRJNA276478.

Results and discussion

The P. coffeae genome sequence assembled into19.67 Mb on 5821 contigs, corresponding to the sizepreviously determined by flow cytometry (Leroy et al.,2003). Significantly, gene-finding algorithms predictedonly 6712 genes, the smallest gene complement yet iden-tified in a metazoan (Opperman et al., 2008; Rodelspergeret al., 2013). At 342 genes Mb−1, gene density in P. cof-feae is high compared to any previously sequenced nema-todes. The 38.1% G+C content is in the typical range fornematodes and prevalence of repetitive DNA is very low,with less than 1% total interspersed repeats (Table S1 inthe Supplementary Information). Despite a 3% polymor-

4 Nematology

Genome of Pratylenchus coffeae

Table 1. Genome comparisons among sequenced nematode species.

Species Size (Mb) Contigs N50 (Mb) G+C% Gene number Protein length(aa)

Exons pergene

Gene density

Caenorhabditis elegans 100.3 6 chr 17.5 35.4 20 517 340 5 249Pristionchus pacificus 153.2 18 083 1.2 42.8 24 217 241 8 140Brugia malayi 89.3 27 210 0.04 30.1 21 332 193 3 222Ascaris suum 265.3 29 831 0.4 37.9 18 449 234 5 70Trichinella spiralis 58.5 6863 6.4 33.9 16 380 192 4 280Meloidogyne incognita 82.1 9538 0.1 31.4 20 332 249 5 223M. hapla 54.0 3452 0.3 27.4 14 721 250 4 270Bursaphelenchus xylophilus 73.1 5527 1.0 40.3 18 074 263 4 242Pratylenchus coffeae 19.7 5821 0.01 38.1 6712 243 2 342Radopholus similis 63.2 8176 0.02 39.9 11 315 246 4 172Globodera pallida 124.7 6873∗ 0.122 36.7 16 419 361 6 132

∗ Scaffolds, contigs > 30 000.

Fig. 1. The 20 most common protein domains found in Pratylenchus coffeae based on an HMM search of Pfam, as compared to thenumber of occurrences of each domain found in Caenorhabditis elegans and Meloidogyne hapla. ∗Top 20 in C. elegans; ∗∗top 20 in M.hapla.

phism rate, our genome assembly accounts for approxi-mately 98% coverage of the genome (Table 1). One re-markable feature of the P. coffeae genome is that genesaverage only two exons per gene compared to four or morein other sequenced nematode species. This may account,in part, for the small genome size.

Compared to C. elegans, P. coffeae carries 33% of thegene complement and 50% of the gene space, furtheraccounting for its reduced genome size (Table 1). Thisis largely reflected by reduction in gene family mem-bers. In some cases, such as the oxidative phosphory-lation pathway, the complete pathway is observed in P.

Vol. 00(0), 2015 5

M. Burke et al.

coffeae (Fig. S1A in the Supplementary Information) butmost often where C. elegans may have multiple enzymesat a given point in a metabolic pathway (e.g., fatty acidmetabolism), P. coffeae generally has only one (Fig. S1Bin the Supplementary Information). This apparent lackof redundancy may reflect the niche each nematode oc-cupies. Caenorhabditis elegans is found in and arounddecaying organic matter in the soil, a complex environ-ment with many competing signals, whereas P. coffeaespends the majority of its life inside the relatively ‘noise-free’ plant root. Examining Pfam domains indicates P.coffeae is reduced in gene family numbers compared tothe more complex sedentary endoparasitic nematode M.hapla. Caenorhabditis elegans has larger numbers perfamily than either P. coffeae or M. hapla in most cases,although there are several rare exceptions and in a fewcases, the plant nematodes carry domains not found in C.elegans (Fig. 1).

A hallmark of metazoans is that they undergo embryo-genesis. In C. elegans, 2617 genes exhibit an embryo-lethal, RNAi phenotype. We obtained these genes fromWormMart and undertook a comparison to C. briggsae, P.

coffeae and M. hapla, revealing 1844 in common (Fig. 2).This is likely the core set of essential genes for nematodeembryogenesis. Taken as a percentage of total predictedgene numbers, these results indicate that approximately30% of the P. coffeae gene complement is dedicated to‘making’ the worm, compared with 10-15% in C. elegansand M. hapla. Of the 2617 genes screened, 533 embryoniclethal genes are present in C. elegans, C. briggsae and M.hapla but not in P. coffeae, which may indicate increaseddevelopmental complexity in the former genomes, or sim-ply reflect the vastly reduced gene number in P. coffeae.Additionally, there are 34 embryonic lethal genes found inC. elegans that are not present in any of the other three ne-matodes, suggesting expansion of gene number in C. ele-gans. Embryonic lethal genes reflect a variety of classes,many of which have unknown functions. The genes notfound in P. coffeae may reveal important developmentalcomparisons and may also indicate potential functions.

G-protein-coupled receptors (GPCRs) sense moleculesoutside the cell and activate signal transduction pathwaysand cellular responses to stimuli, and are involved in awide array of physiological processes. The GPCRs are the

Fig. 2. A comparison of Caenorhabditis elegans Embryonic Lethal protein sequences to C. briggsae, Meloidogyne hapla andPratylenchus coffeae. The majority of the genes can be found in all four genomes (1844 or 70.4%). There are 34 (1.2%) that arein C. elegans and not found in any of the other three genomes, including the closely related C. briggsae. Many of the genes not foundin P. coffeae are individual members of larger gene families.

6 Nematology

Genome of Pratylenchus coffeae

Fig. 3. Comparison of numbers of genes in a variety of gene families. Collagen, NHR and GPCR were chosen based on an interest incomparing the lifestyles of the different nematodes. Conversely, glyceraldehyde-3-phosphate dehydrogenase was chosen because it isbelieved to have no association with lifestyle or feeding, yet still represents a reduction in gene family size in Meloidogyne hapla andPratylenchus coffeae as compared to Caenorhabditis elegans.

largest gene family observed in C. elegans, which encodesmore than 1200, many with unknown function. By con-trast, M. hapla has 137 GPCRs (Opperman et al., 2008),whilst P. coffeae possesses only 55 (Fig. 3). Compari-son of key genes in the C. elegans olfactory system to P.coffeae reveals both conservation of the pathway and re-duced gene numbers. The gpa (G-protein subunit alpha)genes in C. elegans are involved in numerous chemosen-sory behaviours, including chemotaxis to water-solubleand volatile compounds (Bargmann, 2006), and at least15 gpa genes are known; M. hapla carries orthologues tothe majority of these genes. By contrast, P. coffeae car-ries only orthologues of gpa-3, gpa-5, gpa-12 and odr-3(Table 2). The downstream signal transduction and regu-latory genes in these pathways are largely conserved be-tween P. coffeae, M. hapla and C. elegans (Table 2), sug-gesting that P. coffeae possesses a basal olfactory sys-tem.

In contrast to fully conserved metabolic pathways,many developmental pathways in C. elegans are only par-tially conserved in the P. coffeae genome. Sex determina-tion is a key developmental event in all nematodes, yet P.

coffeae carries a limited number of proteins with signif-icant similarity to those involved in sex determination inC. elegans (Table 3). Of the major sex determining genesin C. elegans, only tra-3 and, the mog (masculinisationof germline) genes are highly conserved. Although someother downstream sex determination genes are also con-served, as a whole this pathway remains obscure in P. cof-feae. By contrast, many of the genes in other C. eleganspathways have clear orthologues in P. coffeae. For ex-ample, many genes involved in RNAi function of smallRNAs can unequivocally be found in P. coffeae, withthe sole exception of rde-4, which is not well conservedacross phylogenetic distances (Table 4). Not surprisingly,the RNAi phenomenon can be experimentally induced inP. coffeae (Joseph et al., 2012). Similarly, many genesinvolved in basic nematode development are well con-served reflecting their primary roles in generalised growthas opposed to response to specific environments. GeneOntology analysis and Pfam domain determinations sup-port these broad similarities in developmental processes(Figs 1, S2).

Vol. 00(0), 2015 7

M. Burke et al.

Table 2. Pratylenchus coffeae genes involved in chemosensorybehaviour with orthologues in Caenorhabditis elegans with E-values less than e-5.

Receptor Pratylenchus coffeae contig

G-protein associatedgpa-3 5406gpa-5 2518gpa-13 1300odr-3 5406gcy-35 66gcy-36 66

Signal transductiondaf-11 3188daf-21 5230ocr-2 1147odr-1 4191osm-9 1147tax-2 371tax-4 1373

Regulatorsegl-4 5718goa-1 5406grk-2 2709kin-29 3385let-60 3145osm-9 1147tax-6 1401ttx-4 1958

The ability to form dauers is broadly conserved acrossthe Nematoda. Genetic analysis in C. elegans identified 32genes as dauer-affecting (daf ) (Hu, 2007); P. coffeae car-ries strong orthologues of 18 C. elegans daf genes, andweak orthologues of two more (Table 5). Pratylenchuscoffeae carries a robust orthologue of the daf-12 gene en-coding the dafachronic acid (DA) receptor, further sup-porting the deeply conserved central role of DA-daf-12signalling in nematode development. The molecular iden-tities of those genes not found in P. coffeae appear relatedto specific developmental cues connected to life style, anddemonstrate that although the basic mechanical aspects ofdevelopment are conserved, response to environment inparasitic vs free-living nematodes is substantially differ-ent.

It has been hypothesised that horizontal gene transfer(HGT) has played a significant role in the evolution offeeding and pathogenesis in plant-parasitic nematodes(Scholl et al., 2003; Danchin et al., 2010), and many ofthese genes encode nematode effectors (Mitchum et al.,2013; Quentin et al., 2013). HGT was first reported for

Table 3. Caenorhabditis elegans sex determination gene ortho-logues in Pratylenchus coffeae and Meloidogyne hapla.

C. elegans gene P. coffeae M. hapla

sex-1 e-16 e-25fox-1 NF NFxol-1 NF NFsdc-1 NF e-10sdc-2 NF NFsdc-3 NF NFher-1 NF NFtra-1 e-9 e-32tra-2 NF NFtra-3 e-30 e-63laf-1 e-115 e-167gld-1 e-55 e-71fog-1 e-8 e-14fog-2 NF NFfog-3 NF e-7mag-1 NF e-59mab-3 NF e-27fem-1 e-7 e-14fem-2 NF e-8fem-3 NF NFfbf-1 e-13 e-13fbf-2 e-13 e-14nos-3 NF NFmog-1 e-166 0mog-4 0 0mog-5 0 0mog-6 (cyn-4) e-16 e-143

NF = not found.

cyst nematode cellulase genes (Smant et al., 1998), whichwere demonstrated to be involved in nematode migrationthrough the root, feeding, and potential modification ofcell walls during feeding site development (Haegeman etal., 2011a; Mitchum et al., 2013). There are six cellulasegenes in M. hapla, arising from a combination of HGTand gene duplication. By contrast, P. coffeae carries onlytwo cellulase genes (Table 6). A comparison of effectorenzymes between P. coffeae and Meloidogyne spp. revealsthat P. coffeae possesses a similar suite of types but ingreatly reduced numbers. It has recently been reportedthat cellulase gene HGT is an ancient event that occurredin a progenitor species leading to the Pratylenchidae (fromwhich Meloidogyne spp. evolved), and that additionalcellulase genes in species of root-knot and cyst nematodesarose from duplication events (Rybarczyk-Mydlowska etal., 2012). Our data support this hypothesis and suggestthat acquisition of cellulase genes by HGT was a crucial

8 Nematology

Genome of Pratylenchus coffeae

Table 4. RNAi gene orthologues in Pratylenchus coffeae andMeloidogyne hapla with E-values less than e-5.

Caenorhabditis elegans Pratylenchuscoffeae

Meloidogynehapla

drh-1 + +drh-2 (pseudo)drh-3 + +smg-2 + +smg-5smg-6 +rde-2rde-4zfp-1mut-7mut-16dcr-1 + +rrf-1 + +rrf-2 + +rrf-3 + +ego-1 + +rde-1 + +SAGO-1 +SAGO-2 + +PPW-1 + +PPW-2 + +C04F12.1 + +hrd-1 + +WAGO-2 + +WAGO-4 + +M03D4.7WAGO-1 + +T22B3.2 + +WAGO-10 +hpo-24 + +WAGO-11 + +WAGO-5 + +ZK218.8 +eri-1eri-3eri-5eri-6/7 + +gfl-1 + +sid-1sid-2rsd-2rsd-3 + +rsd-6drsh-1 + +alg-1 + +alg-2 + +alg-4 + +PRG-1 + +

Table 4. (Continued.)

Caenorhabditis elegans Pratylenchuscoffeae

Meloidogynehapla

PRG-2 + +ERGO-1 + +vig-1tsn-1 +pash-1xpo-1 + +xpo-2 + +xpo-3csr-1 + +NRDE-3 + +ain-1ain-2adr-1adr-2lin-15bsomi-1xrn-1 +xrn-2 +cid-1 + +ekl-1 +ekl-4 + +ekl-5ekl-6mes-2 + +mes-3rha-1 + +

early step in the evolution of phytophagous feeding innematodes.

While little is known about effector molecules secretedfrom the less specialised migratory endoparasitic species,comparative analysis of effectors between sedentary andmigratory endoparasitic forms may shed light on evolu-tion of molecules necessary for both successful infectionand formation of specialised feeding sites (Haegeman etal., 2011b). Comparative analysis revealed effector candi-dates that are present in the sedentary endoparasites (root-knot and cyst nematodes) but not in P. coffeae, suggest-ing these putative effectors may be intriguing candidatesfor specialisation and complex feeding site formation (Ta-ble 7).

This observation can be extended to a number of othergene families thought to have arisen in plant-parasitic ne-matodes by a combination of HGT and duplication. Thereare 22 pectate lyases (PL), many with similarity to bac-terial PLs, which have undergone substantial duplicationand relocation within the M. hapla genome. By contrast,

Vol. 00(0), 2015 9

M. Burke et al.

Table 5. Dauer pathway gene orthologues in Pratylenchus coffeae, Bursaphelenchus xylophilus and Meloidogyne hapla.

Caenorhabditis elegans gene Pratylenchus coffeae Meloidogyne hapla Bursaphelenchus xylophilus

daf-1 e-63 e-23 Ydaf-2 e-60 e-35 Ydaf-3 e-26 e-15 Ydaf-4 e-24 e-26 Ydaf-5 NF NF NFdaf-6 e-130 e-73 NFdaf-7 e-5 e-7 Ydaf-8 e-13 e-8 Ydaf-9 e-30 e-27 Ydaf-10 (osm-4) e-137 e-34 Ydaf-11 e-65 e-45 Ydaf-12 e-54 e-14 Ydaf-14 e-6 NF Ydaf-15 e-121 e-34 Ydaf-16 (daf-17) e-13 e-36 Ydaf-18 e-72 e-34 Ydaf-19 (daf-24) NF e-17 NFdaf-20 e-54 e-51 NFdaf-21 e-0 e-149 NFdaf-22 NF e-140 NFdaf-23 (age-1) e-124 e-72 Ydaf-25 e-30 e-33 NFdaf-37 NF e-9 NFdaf-38 e-27 e-53 NF

NF = not found.

Table 6. Comparison of CAZyme encoding genes in Meloidogyne hapla, M. incognita, Pratylenchus coffeae, Pristionchus pacificus,Caenorhabditis elegans and Brugia malayi.

Species Cellulose/Xylan Pectin/Pectate Arabinose Total

GH5_2 GH5_8 GH28 PL3 GH43

Meloidogyne hapla 6 1 2 22 2 33M. incognita 21 6 2 30 2 61Pratylenchus coffeae 1 2 1 3 2 9Pristionchus pacificus 7 0 0 0 0 7Caenorhabditis elegans 0 0 0 0 0 0Brugia malayi 0 0 0 0 0 0

P. coffeae carries only three PL genes (Table 6). Phylo-genetic analysis shows two of the three P. coffeae PLs asdivergent from the main group of M. hapla PLs, suggest-ing an HGT event followed by gene duplication. The thirdP. coffeae PL lies within a clade of M. hapla PLs, indi-cating continued evolutionary pressure on the gene sincespeciation from the common ancestor to the two nema-todes (Fig. 4). This observation can be extended to other

gene families thought to be important in feeding site de-velopment by M. hapla. We hypothesise that duplicationand relocation have played an essential role in the evolu-tion of sedentary nematode parasitism and specialisation.

The P. coffeae genomic sequence reveals an animalgenome with approximately half the predicted genes ofother species thus far characterised, including the single-celled amoeba, Naegleria gruberi. Although possessing a

10 Nematology

Genome of Pratylenchus coffeae

Table 7. Potential sedentary endoparasitic nematode-specific effector molecules identified by comparison to the Pratylenchus coffeaegenome.

Accession number Description

AF531164 Meloidogyne incognita putative pharyngeal gland cell secretory protein 5 (msp5)AF531165 M. incognita putative pharyngeal gland cell secretory protein 6 (msp6)AF531166 M. incognita putative pharyngeal gland cell secretory protein 7 (msp7)AF531168 M. incognita putative pharyngeal gland cell secretory protein 8 (msp8)AF531169 M. incognita putative pharyngeal gland cell secretory protein 9 (msp9)AY134431 M. incognita putative pharyngeal gland cell secretory protein 12 (msp12)AY134432 M. incognita putative pharyngeal gland cell secretory protein 13 (msp13)AY026357 Heterodera glycines pectinase precursorAY134433 M. incognita putative pharyngeal gland cell secretory protein 14 (msp14)AY134434 M. incognita putative pharyngeal gland cell secretory protein 15 (msp15)AY134436 M. incognita putative pharyngeal gland cell secretory protein 17 (msp17)AY134437 M. incognita putative pharyngeal gland cell secretory protein 18 (msp18)AY134438 M. incognita putative pharyngeal gland cell secretory protein 19 (msp19)AY134439 M. incognita putative pharyngeal gland cell secretory protein 20 (msp20)AY134442 M. incognita putative pharyngeal gland cell secretory protein 23 (msp23)AY134443 M. incognita putative pharyngeal gland cell secretory protein 24 (msp24)AY134444 M. incognita putative pharyngeal gland cell secretory protein 25 (msp25)AY142117 M. incognita putative pharyngeal gland cell secretory protein 34 (msp34)AY142118 M. incognita putative pharyngeal gland cell secretory protein 33 (msp33)AY142119 M. incognita putative pharyngeal gland cell secretory protein 35 (msp35)DQ841123 M. hapla parasitism protein 16D10AY134441 M. incognita putative pharyngeal gland cell secretory protein 22 (msp22)AY142120 M. incognita putative pharyngeal gland cell secretory protein 30 (msp30)AY134435 M. incognita putative pharyngeal gland cell secretory protein 16 (msp16)AF469058 H. glycines cellulose binding proteinAJ251758 Globodera rostochiensis partial mRNA for hypothetical protein (clone A41)

small genome, P. coffeae possesses genes that all nema-todes appear to utilise during growth and development.In addition, it contains a complement of genes specific toplant-feeding nematodes, although P. coffeae carries re-duced numbers in multi-gene families compared to themore specialised root-knot nematodes (e.g., M. hapla). In-dividual P. coffeae genes may provide multiple functions,roles played by neofunctionalised and subfunctionalisedgene families in other nematodes. Perhaps what is mostintriguing about the P. coffeae genome is what it does notcontain compared to the specialised root-knot nematodes;this provides tantalising clues about genes potentially in-volved in giant cell formation (Table 7). Because thesegenes are not found in other nematode species, they rep-resent robust candidates for genes underlying specialisedfunctions within the plant-nematode interaction. Further,the similarities in suites of genes specific to plant-parasiticnematodes indicate a basal core of functions related tothe phytophagous lifestyle. We speculate that P. coffeae

represents a more primitive species of plant-feeding ne-matode than root-knot nematodes, and that evolution ofspecialised feeding site formation was due to a combina-tion of gene family expansion and horizontal gene transferevents.

Acknowledgements

M.B., E.H.S. and C.H.O. contributed equally to thiswork. We thank Jorge Gonsalez and Miguel E. Munoz(Dole, Costa Rica), and Patrick Quénéhervé (INRA, Mar-tinique, France) for samples and DNA of Pratylenchuscoffeae, Annelies Haegeman and Godelieve Gheysen(Ghent University, Ghent, Belgium) for preliminary ac-cess to the P. coffeae transcriptome, Brian R. Kerry (de-ceased) (Rothamsted Research, Harpenden, UK) for dis-cussions, Peter M. Digennaro for critical reading of themanuscript and Reenah L. Schaffer for proofreading andtechnical editing. Funding and support for this work was

Vol. 00(0), 2015 11

M. Burke et al.

Fig. 4. Phylogeny of three pectate lyase (PL) genes in Pratylenchus coffeae with 22 PL found in Meloidogyne hapla. Two of the threeP. coffeae PLs lie within a single clade basal to the M. hapla PLs, indicating a possible separate acquisition of the PL gene. The thirdP. coffeae PL lies within the clades containing the M. hapla PLs, and may indicate continuing evolutionary pressures on the gene afterspeciation of the common ancestor to the two nematodes. A Streptomyces coelicolor PL is used as the outgroup. Numbers indicateconfidence based on 1000 bootstrapped datasets.

provided by APC, Schneider Electric (West Kingston,RI, USA), the David H. Murdock Research Institute(Kannapolis, NC, USA) and the North Carolina Agricul-tural Research Service. C.H.O. and M.B. conceived thework and designed research and, together with E.H.S.and D.M.B., wrote the manuscript. J.E.S., R.C., O.G.and X.W. performed DNA preparation, library construc-tion and sequencing. C.H.O., E.H.S., S.D., S.G., E.W.,G.M.W. and M.B. performed computational and dataanalysis. All authors reviewed and commented on themanuscript.

References

Abad, P., Gouzy, J., Aury, J.-M., Castagnone-Sereno, P.,Danchin, E.G.J., Deleury, E., Perfus-Barbeoch, L., An-thouard, V., Artiguenave, F., Blok, V.C. et al. (2008). Genomesequence of the metazoan plant-parasitic nematode Meloido-gyne incognita. Nature Biotechnology 26, 909-915.

Altschul, S.F., Gish, W., Miller, W., Myers, E.W. & Lipman,D.J. (1990). Basic local alignment search tool. Journal ofMolecular Biology 215, 403-410.

Bargmann, C.L. (2006). Chemosensation in C. elegans. In:The C. elegans Research Community (Ed.). Wormbook.

12 Nematology

Genome of Pratylenchus coffeae

DOI:10.1895/wormbook.1.123.1, available online at http://www.wormbook.org.

Bird, D.McK., Williamson, V.M., Abad, P., McCarter, J.,Danchin, E., Castagnone-Serrano, P. & Opperman, C.H.(2009). Genome analysis of sedentary endoparasitic plant-parasitic nematodes. Annual Review of Phytopathology 47,333-351.

Danchin, E.G., Rosso, M.N., Vieira, P., de Almeida-Engler, J.,Coutinho, P.M., Henrissat, B. & Abad, P. (2010). Multiplelateral gene transfers and duplications have promoted plantparasitism ability in nematodes. Proceedings of the NationalAcademy of Science of the United States of America 107,17651-17656.

Duncan, L.W., Inserra, R.N., Thomas, W.K., Dunn, D., Mustika,I., Frisse, L.M., Mendes, M.L., Morris, K. & Kaplan, D.T.(1999). Molecular and morphological analysis of isolates ofPratylenchus coffeae and closely related species. Nematro-pica 29, 61-80.

Finn, R.D., Mistry, J., Tate, J., Coggill, P., Heger, A., Pollington,J.E., Gavin, O.L., Gunasekaran, P., Ceric, G., Forslund, K. etal. (2010). The Pfam protein families database. Nucleic AcidsResearch 38, D211-D222.

Fritz-Laylin, L.K., Prochnik, S.E., Ginger, M.L., Dacks, J.B.,Carpenter, M.L., Field, M.C., Kuo, A., Paredez, A., Chapman,J., Pham, J. et al. (2010). The genome of Naegleria gruberiilluminates early eukaryotic versatility. Cell 140, 631-642.

Gotz, S., Garcia-Gomez, J.M., Terol, J., Williams, T.D., Nagaraj,S.H., Nueda, M.J., Robles, M., Talón, M., Dopazo, J. &Conesa, A. (2008). High-throughput functional annotationand data mining with the Blast2GO suite. Nucleic AcidsResearch 36, 3420-3435.

Gowen, S.R., Quénéhervé, P. & Fogain, R. (2005). Nematodeparasites of bananas and plantains. In: Luc, M., Sikora, R.A.& Bridge, J. (Eds). Plant parasitic nematodes in subtropicaland tropical agriculture, 2nd edition. Wallingford, UK, CABInternational, pp. 611-645.

Haas, B.J., Salzberg, S.L., Zhu, W., Pertea, M., Allen, J.E.,Orvis, J., White, O., Buell, C.R. & Wortman, J.R. (2008).Automated eukaryotic gene structure annotation using EVi-denceModeler and the Program to Assemble Spliced Align-ments. Genome Biology 9, R7.

Haegeman, A., Joseph, S. & Gheysen, G. (2011a). Analysis ofthe transcriptome of the root lesion nematode Pratylenchuscoffeae generated by 454 sequencing technology. Molecularand Biochemical Parasitology 178, 7-14.

Haegeman, A., Mantelin, S., Jones, J.T. & Gheysen, G. (2011b).Functional roles of effectors of plant-parasitic nematodes.Gene 492, 18-31.

Hu, P.J. (2007). Dauer. In: The C. elegans Research Community(Ed.). Wormbook. DOI:10.1895/wormbook.1.144.1, availableonline at http://www.wormbook.org.

Joseph, S., Gheysen, G. & Subramaniam, K. (2012). RNAinterference in Pratylenchus coffeae: knock down of Pc-

pat-10 and Pc-unc-87 impedes migration. Molecular andBiochemical Parasitology 186, 51-59.

Kanehisa, M. & Goto, S. (2000). KEGG: Kyoto Encyclopediaof Genes and Genomes. Nucleic Acids Research 28, 27-30.

Kikuchi, T., Cotton, J.A., Dalzell, J.J., Hasegawa, K., Kan-zaki, N., McVeigh, P., Takanashi, T., Tsai, I.J., Assefa, S.A.,Cock, P.J. et al. (2011). Genomic insights into the originof parasitism in the emerging plant pathogen Bursaphelen-chus xylophilus. PLoS Pathogens 7, e1002219, DOI:10.1371/journal.ppat.1002219.

Krogh, A., Larsson, B., von Heijne, G. & Sonnhammer, E.L.(2001). Predicting transmembrane protein topology with ahidden Markov model: Application to complete genomes.Journal of Molecular Biology 305, 567-580.

Leroy, S., Duperray, C. & Morand, S. (2003). Flow cytometryfor parasite nematode genome size measurement. Molecularand Biochemical Parasitology 128, 91-93.

Mitchum, M.G., Hussey, R.S., Baum, T.J., Wang, X., Elling,A.A., Wubben, M. & Davis, E.L. (2013). Nematode effectorproteins: an emerging paradigm of parasitism. New Phytolo-gist 199, 879-894.

Opperman, C.H., Bird, D.M., Williamson, V.M., Rokhsar, D.S.,Burke, M., Cohn, J., Cromer, J., Diener, S., Gajan, J., Graham,S. et al. (2008). Sequence and genetic map of Meloidogynehapla: a compact nematode genome for plant parasitism.Proceedings of the National Academy of Science of the UnitedStates of America 105, 14802-14807.

Petersen, T.N., Brunak, S., von Heinje, G. & Nielsen, H.(2011). SignalP 4.0: discriminating signal peptides fromtransmembrane regions. Nature Methods 8, 785-786.

Quentin, M., Abad, P. & Favery, B. (2013). Plant parasiticnematode effectors target host defense and nuclear functionsto establish feeding cells. Frontiers in Plant Science 4,1-7.

Rödelsperger, C., Streit, A. & Sommer, R.J. (2013). Struc-ture, function and evolution of the nematode genome. In:eLS. Chichester, UK, John Wiley & Sons. Available on-line at http://www.els.net.prox.lib.ncsu.edu (DOI:10.1002/9780470015902.a0024603).

Rybarczyk-Mydlowska, K., Maboreke, H.R., van Megen, H.,van den Elsen, S., Mooyman, P., Smant, G., Bakker, J. &Helder, J. (2012). Rather than by direct acquisition via lat-eral gene transfer, GHF5 cellulases were passed on from earlyPratylenchidae to root-knot and cyst nematodes. BMC Evolu-tionary Biology 12, 221, DOI:10.1186/1471-2148-12-221.

Scholl, E.H., Thorne, J.L., McCarter, J.P. & Bird, D.M. (2003).Horizontally transferred genes in plant-parasitic nematodes:a high-throughput genomic approach. Genome Biology 4,R39.

Smant, G., Stokkermans, J.P., Yan, Y., de Boer, J.M., Baum,T.J., Wang, X., Hussey, R.S., Gommers, F.J., Henrissat,B., Davis, E.L. et al. (1998). Endogenous cellulases inanimals: isolation of beta-1,4-endoglucanase genes from twospecies of plant-parasitic cyst nematodes. Proceedings of the

Vol. 00(0), 2015 13

M. Burke et al.

National Academy of Science of the United States of America95, 4906-4911.

Smit, A.F.A., Hubley, R. & Green, P. (2004). Repeat-Masker Open-3.0.1996-2004. Available online at http://www.repeatmasker.org.

Stanke, M. & Morgenstern, B. (2005). AUGUSTUS: a webserver for gene prediction in eukaryotes that allows user-defined constraints. Nucleic Acids Research 33, W465-W467.

van Megen, H., van den Elsen, S., Holterman, M., Karssen,G., Mooyman, P., Bongers, T., Holovachov, O., Bakker, J. &Helder, J. (2009). A phylogenetic tree of nematodes basedon about 1200 full length small subunit ribosomal DNAsequences. Nematology 11, 927-950.

Zdobnov, E.M. & Apweiler, R. (2001). InterProScan – anintegration platform for the signature-recognition methods inInterPro. Bioinformatics 17, 847-848.

Table S1. Repetitive DNA content in the Pratylenchus coffeae genome.

Number of elements∗ Length occupied (bp) Percentage of sequence

Retroelements 54 28 724 0.15%SINEs 0 0 0.00%Penelope 1 66 0.00%LINEs 2 316 0.00%

L2/CR1/Rex 1 250 0.00%LTR elements 52 28 408 0.14%

BEL/Pao 17 1521 0.01%Gypsy/DIRS1 35 26 887 0.14%

DNA transposons 16 1150 0.01%hobo-Activator 4 199 0.00%Tc1-IS630-Pogo 3 262 0.00%MuDR-IS905 4 341 0.00%Tourist/Harbinger 1 102 0.00%

Unclassified 11 1520 0.01%Total interspersed repeats 31 394 0.16%Small RNA 32 4807 0.02%Satellites: 2 156 0.00%Simple repeats 537 23 541 0.12%Low complexity 26 332 1 148 916 5.85%

File name: Pcoffeae_LargeContigs.fasta; sequences: 5821; total length: 19 625 775 bp (19 625 775 bp excl. N/X-runs); GC level:38.09%; bases masked: 1 208 308 bp (6.16%).∗ Most repeats fragmented by insertions or deletions have been counted as one element.

14 Nematology

Genome of Pratylenchus coffeaeA F

ig.

S1.

Keg

gdi

agra

ms

indi

catin

gpr

esen

ceof

enzy

mes

in:

A:

the

oxid

ativ

eph

osph

oryl

atio

npa

thw

ay;

and

B:

the

fatty

acid

met

abol

ism

path

way

.E

nzym

esin

the

path

way

high

light

edin

pink

are

pres

ent

inC

aeno

rhab

diti

sel

egan

sas

wel

las

Mel

oido

gyne

hapl

aan

dP

raty

lenc

hus

coffe

ae.

Tho

sein

yello

war

efo

und

only

inC

.el

egan

san

dM

.hap

la.E

nzym

eshi

ghlig

hted

with

gree

nar

efo

und

only

inC

.ele

gans

.

Vol. 00(0), 2015 15

M. Burke et al.B F

ig.S

1.(C

ontin

ued.

)

16 Nematology

Genome of Pratylenchus coffeae

Fig. S2. The most abundant Gene Ontology terms in the Pratylenchus coffeae dataset for the biological process, cellular component,and molecular function categories.

Vol. 00(0), 2015 17