Transcriptome of Aphanomyces euteiches : New Oomycete Putative Pathogenicity Factors and Metabolic Pathways Authors: Elodie Gaulin, Mohammed-Amine Madoui, Arnaud Bottin, Christophe Jacquet, Catherine Mathé, Arnaud Couloux, Patrick Wincker, Bernard Dumas Published: 2008, PLOS ONE DOI: 10.1371/journal.pone.0001723 ABSTRACT Aphanomyces euteiches is an oomycete pathogen that causes seedling blight and root rot of legumes, such as alfalfa and pea. The genus Aphanomyces is phylogenically distinct from well-studied oomycetes such as Phytophthora sp., and contains species pathogenic on plants and aquatic animals. To provide the first foray into gene diversity of A. euteiches , two cDNA libraries were constructed using mRNA extracted from mycelium grown in an artificial liquid medium or in contact to plant roots. A unigene set of 7,977 sequences was obtained from 18,864 high-quality expressed sequenced tags (ESTs) and characterized for potential functions. Comparisons with oomycete proteomes revealed major differences between the gene content of A. euteiches and those of Phytophthora species, leading to the identification of biosynthetic pathways absent in Phytophthora , of new putative pathogenicity genes and of expansion of gene families encoding extracellular proteins, notably different classes of proteases. Among the genes specific of A. euteiches are members of a new family of extracellular proteins putatively involved in adhesion, containing up to four protein domains similar to fungal cellulose binding domains. Comparison of A. euteiches sequences with proteomes of fully sequenced eukaryotic pathogens, including fungi, apicomplexa and trypanosomatids, allowed the identification of A. euteiches genes with close orthologs in these microorganisms but absent in other oomycetes sequenced so far, notably transporters and non-ribosomal peptide synthetases, and suggests the presence of a defense mechanism against oxidative stress which was initially characterized in the pathogenic trypanosomatids. INTRODUCTION The legume root rot disease caused by the oomycete pathogen Aphanomyces euteiches is one major yield reducing factor in legume production including pea and alfalfa throughout the world [1] – [3] . A. euteiches can infect plants at any age. The first symptoms on pea may be discerned on roots which are softened and water-soaked. The pathogen spreads rapidly through the cortical tissue and the fine branches of feeding rootlets are destroyed. Then the epicotyls become dark and eventually collapse. Leaves progressively turn yellow, starting at the bottom of the shoot. In severe cases, the plants collapse and die before forming any pod. Thick-walled oospores are resting stages and constitute the primary inoculum source in soil [4] . After germination, primary zoospores are released at the apex of sporangia. These bi-flagellate motile elements encyst at the root surface, germinate and infect the host root. After a few days, the pathogen can reproduce sexually in the host and form new oospores. These oospores can remain dormant in soil and organic debris for many years [5] . Effective chemical controls for Aphanomyces root rot of legumes are not available. Crop rotation and bioassay methods to assess the inoculum potential in the soil remain the only effective method to avoid disease. Using recurrent selection-based strategies, pea germplasms with partial resistance or tolerance were obtained [6] . Among Oomycetes, the genus Aphanomyces belongs to the Saprolegniales and comprises species that are destructive on plants, crustaceans and fishes such as A. astaci and A. invadans. Virulent genotypes of A. astaci were introduced in Europe giving rise to disease outbreaks among its crustacean host population, the freshwater crayfish [7] . This parasite is now considered as one of the 100 world's worst invasive species (Global Invasive Species Database, http// www.issg.org/database ). For the last 30 years, the epizootic ulcerative syndrome (EUS) a disease caused by A. invadans , has affected wild and farmed fish in Asia and Indo-Pacific regions and has been recently detected in the eastern seaboard of the US [8] . Aphanomyces genus also includes significant plant pathogens such as A. euteiches and A. cochlioides causing the most serious seedling disease of sugar beet in terms of yield loss, persistence of the oospores in soil, and difficulty of control [9] , [10] . By contrast to this diversity of hosts, some oomycete genera, such as the Peronosporale Phytophthora is composed of species which are only pathogenic to plants. Such diversity within the oomycetes could reflect different evolutionary histories and possibly different mechanisms of infection between Saprolegniales and Peronosporales [11] . Until recently, oomycetes genomics focussed only on Phytophthora spp. with the completion of genome sequencing projects and large scale identification of expressed sequences [12] – [17] . These data, combined to functional studies, led to the identification of new factors of pathogenicity such as the large family of RxLR effectors which are thought to manipulate the plant cell upon their translocation into the host cytoplasm [18] . However little is known about the biology and pathology of other oomycetes, notably those belonging to the Saprolegniales. To fill this gap, we have undertaken a large scale sequencing of A. euteiches cDNAs. This microorganism is of interest since it is able to interact with the model legume Medicago truncatula, offering the opportunity to engage genetic approaches for deciphering this interaction [9] . Here we report on the characterization of an EST collection comprising 18,684 EST assembled into 7,977 unique sequences from two different cDNA libraries. A systematic comparative strategy was engaged to identify specific A. euteiches gene sets absent in Phytophthora species. This work greatly expands our knowledge of this group of destructive pathogens and points to the large evolutionary divergence within the oomycete lineage. MATERIALS AND METHODS Aphanomyces euteiches (ATCC201684) was maintained on Corn Meal Agar medium (8.5 g/l) at 28°C in the dark. Synthetic medium [19] supplemented with pectin from apple (10 g/l) was used to perform growth assay on pectin. CDNA LIBRARIES CONSTRUCTION Two unidirectional cDNA libraries from A. euteiches were prepared from mycelium grown on medium containing yeast extract/glucose (library MYC) and from mycelium in interaction with Medicago truncatula roots (library INT) [20] . A total of 18,684 expressed sequence tags (9,224 from library MYC and 9,460 from library INT) were submitted to the EBI databank for accession number assignation and were assembled into 7,977 unigenes [20] . SEQUENCE ANALYSIS AND DATABASE SEARCHES Sequences describe here could be downloaded from the Aphanomyces database AphanoDB ( [19] , http://www.polebio.scsv.ups-tlse.fr/aphano/ ) and cDNA clones are available at the CNRGV ( http://cnrgv.toulouse.inra.fr/ENG/ ). Data searches and analyses have been conducted on AphanoDB using local tools. For detection of RxLR effectors, all open reading frames (ORFs) encoding at least 100 amino acids (delimited by two stops) were extracted. This led to 136,617 putative ORFs. Signal peptides were predicted with SignalP version 3.0 [21] , with a hidden Markov model probability cutoff of 0.8 and 10,131 candidate secreted proteins were obtained. Three independent strategies were used to look for the RxLR motif over all these sequences: - Search for the regular expression RxLR-x(1,40)-[ED][ED][KR] (with the fuzzpro program from EMBOSS suite [22] ) in the first 100 residues downstream of the signal peptide cleavage site. This led to the identification of 5 sequences, among which only two were retained as being in the same frame as the BLASTX matches. - Search with HMM profiles performed with the hmmsearch program from the HMMer 2.2 package [23] with two alternative profiles, RdEER [24] and cropped [18] . From the 2 sequences obtained with this method none was in the frame in agreement with BLASTX matches. - Search for the regular expression RxLx[EDQ]-x(1,40)-[ED][ED][KR] (with the fuzzpro program from EMBOSS suite) in the first 100 residues downstream of the signal peptide cleavage site. The RxLx[EDQ] motif was the original signal required for secretion identified in Plasmodium falciparum [25] : 2 out of the 3 resulting sequences were discarded because of inconsistency of the strand relatively to similarity searches. Multiple alignments were conducted using the program CLUSTALW [26] and visualized with BOXSHADE ( http://www.ch.embnet.org/software/BOX_form.html ). CLUSTALW results were submitted to the WebLogo server [27] ( http://weblogo.berkeley.edu ). - - Search for the regular expression RxLR-x(1,40)-[ED][ED][KR] (with the fuzzpro program from EMBOSS suite [22] ) in the first 100 residues downstream of the signal peptide cleavage site. - This led to the identification of 5 sequences, among which only two were retained as being in the same frame as the BLASTX matches. - - Search with HMM profiles performed with the hmmsearch program from the HMMer 2.2 package [23] with two alternative profiles, RdEER [24] and cropped [18] . From the 2 sequences obtained with this method none was in the frame in agreement with BLASTX matches. - - Search for the regular expression RxLx[EDQ]-x(1,40)-[ED][ED][KR] (with the fuzzpro program from EMBOSS suite) in the first 100 residues downstream of the signal peptide cleavage site. The RxLx[EDQ] motif was the original signal required for secretion identified in Plasmodium falciparum [25] : 2 out of the 3 resulting sequences were discarded because of inconsistency of the strand relatively to similarity searches. RESULTS Two unidirectionnal cDNA libraries were constructed using mRNAs isolated from mycelium of A. euteiches cultivated either in vitro in a liquid medium (MYC library) or grown in contact to M. truncatula roots (INT library). A total of 18,684 high-quality sequences (9,224 for the MYC library and 9,460 for the INT library) corresponding to the 5′ end of cDNA inserts were acquired. A 7,977 unigene set was assembled from EST data. Among the unigenes, 2,843 consensus sequences were assembled from multiple sequence reads and 5,134 were singlets. The 7,977 unigene set was annotated by comparison to the NCBI non-redundant (nr) protein database (5-17-2007 Version) using BLAST analyses [28] . Proteomes of seven fully sequenced organisms were added to the analysis. These include the proteomes from two oomycetes ( P. sojae and P. ramorum ), the diatom Thalassiosira pseudonana , the pathogenic fungus Nectria haematococca , the model plant Arabidopsis thaliana and the apicomplexa parasites Toxoplasma gondii and Plasmodium falciparum . Overall, about 70% of the sequences showed homology to previously described genes in the NCBI non-redundant protein database using a BLASTX E value cut-off of 10 −5 ( Figure 1 ). As expected, a large proportion of sequences (80%, E value cut-off of <10 −5 ) had significant similarities to Phytophthora predicted proteins. Interestingly, about the same fraction (50%, E value cut-off of 10 −5 ) of A. euteiches sequences showed homology to a plant, diatom, apicomplexa or fungal sequence whereas these organisms are distantly related ( Figure 1 ). The unigene set was then annotated by comparisons with the Pfam database of protein domains [29] . 45% of the sequences showed homology with a protein domain with an E value<10 −5 . All the data were stored in a public database called AphanoDB ( www.polebio.scsv.ups-tlse.fr/aphano/ ) which is described elsewhere [20] . Consensus and singlet sequences were named with a unique identifier with the prefix ‘Ae’ for the species, 3 digits for the number of EST included in the contig, the 2 digits ‘AL’ for the strain and 5 digits as unique number. Unigene sequences were blasted using the BLASTX software to the nr database and to the proteome of P. sojae (Ps), A. thaliana (At), N. haematococca (Nh), T. pseudonana (Tp) and T. gondii (Tg). For each class of e-value the percentage of unigenes showing homology is indicated. GENES HIGHLY REPRESENTED IN THE INTERACTION LIBRARY Determination of the number of sequences specific of each library provided a good indication that the two growth conditions used for library construction allowed the expression of divergent transcriptomes. More than 40% of the sequences appeared to be specific for one library ( Table 1 ). To identify unigenes with a statistical difference in transcript abundance, we used a method described in [30] based upon the frequencies which genes occur in each library and false discovery rate control for multiple test corrections [31] . Eight genes abundantly represented ( p -value