Phytophthora infestans Has a Plethora of Phospholipase D Enzymes Including a Subclass That Has Extracellular Activity

In eukaryotes phospholipase D (PLD) is involved in many cellular processes. Currently little is known about PLDs in oomycetes. Here we report that the oomycete plant pathogen Phytophthora infestans has a large repertoire of PLDs divided over six subfamilies: PXPH-PLD, PXTM-PLD, TM-PLD, PLD-likes, and type A and B sPLD-likes. Since the latter have signal peptides we developed a method using metabolically labelled phospholipids to monitor if P. infestans secretes PLD. In extracellular medium of ten P. infestans strains PLD activity was detected as demonstrated by the production of phosphatidic acid and the PLD specific marker phosphatidylalcohol.


Introduction
In eukaryotes, phospholipid-based signalling and metabolism play important roles in numerous cellular processes. Phospholipids are ubiquitous components of all cell membranes and, as second messengers, they act as modulators of many cellular functions. Conversion of phospholipids is accomplished by enzymes that are well conserved across eukaryotes and play crucial roles in cellular regulation, metabolism, stress responses and phospholipid biosynthesis. Important classes of phospholipid modifying enzymes are phospholipid kinases, phospholipid phosphatases and phospholipases. The latter are a heterogeneous group of enzymes whose classification is based on their catalytic activity and substrate specificity.
Phospholipase D (PLD) catalyzes the hydrolysis of structural phospholipids at their terminal phosphoesteric bond, leading to the production of a hydrophilic constituent and phosphatidic acid (PA). The latter has emerged as a significant lipid mediator in many cellular processes. PA mediated signal transduction includes modulation of receptor signalling, cytoskeleton rearrangement, secretion, and vesicle trafficking during endocytosis or exocytosis [1]. All eukaryotes have PXPH-PLDs that are composed of the N-terminally located phosphoinositide binding domains Pleckstrin homology (PH) and PHOX homology (PX). These precede the catalytic site and regulatory motifs. Unique to plants are the C2-PLDs that contain the N-terminally located calcium/lipid binding domain C2. These two PLD subfamilies have a catalytic site with two highly conserved motifs each consisting of HxKxxxxD (hereafter HKD1 and HKD2). Bacteria have PLDs that harbour the catalytic site and regulatory motifs but lack phosphoinositide binding domains. Also ''non-HKD-PLDs'' have been identified, enzymes with PLD activity but lacking characteristic motifs [2]. Recently, a novel rice (Oryza sativa) PLD was described that contains the two HKD motifs as well as a N-terminal signal peptide (SP) [3].
Annotation of the genomes of Phytophthora sojae and Phytophthora ramorum revealed that Phytophthora has a variable set of PLD genes. Phytophthora lacks C2-PLD genes but possesses one PXPH-PLD gene. In addition, there are seventeen PLD genes that represent novel subfamilies including a subfamily with SP sequences that encodes potentially secreted PLDs [4]. In this study we performed an in depth analysis of the PLD subfamilies in Phytophthora infestans, the causal agent of potato late blight and the most notorious Phytophthora species. All Phytophthora PLD subfamilies were detected but the subfamily of PLDs with SPs (sPLD-likes) appears to be comprised of two unrelated subfamilies, sPLD-like-A and sPLDlike-B. Moreover, we developed a method that makes use of reconstituted plant membrane vesicles for analysing the presence of PLD activity in extracellular medium of P. infestans and provide evidence that P. infestans indeed secretes PLD. This finding points to a putative role for PLD in pathogenicity.

Bioinformatic analysis
Putative PLD genes in the P. infestans genome database of strain T30-4 (http://www.broad.mit.edu/annotation/genome/phytophthora_ infestans) were identified by several methods including automatic annotation, BLAST searches with the P. sojae and P. ramorum PLD gene models [4], and with representative PLD sequences from NCBI GenBank. Putative PLD gene models were further analysed based on characteristic conserved motifs and manually corrected when needed. Multiple alignments were made in ClustallW2. Phylograms were constructed by Mega4.1 using the Minimum Evolution method with a bootstrap test based on 5000 replicates, and the Poisson correction method [5]. Signal and anchor peptides were detected via SignalP 3.0 (http://www.cbs.dtu.dk/services/SignalP/) using default parameters. Putative transmembrane domains were analysed as described before [6]. Searches for additional proteins domains were performed via publically available databases. For determining synteny between Phytophthora spp. gene models present in regions flanking PLD genes were extracted and used for reciprocal BLAST analysis. Pairs of genes sharing best reciprocal BLAST hits were assigned as orthologues. Intergenic lengths were determined by taking the distance between the PLD gene and the neighbouring gene model thereby ignoring mobile elements.
Phytophthora infestans culture condition and sampling P. infestans strains were routinely grown at 18uC in the dark on V8 media or on Rye agar medium supplemented with 2% sucrose [7]. Extracellular medium was obtained by flooding full grown plates with V8 medium. After overnight incubation samples were taken and immediately centrifuged for 2 min at 10,000 g. The supernatant was collected and filtered over 0.2 mM filters to remove cell debris.
Plant cell suspension culturing and radio-labelling of phospholipids Cell suspensions of Nicotiana tabacum were grown as described [8]. Metabolically labelled 32 P-labelled phospholipids were obtained by overnight incubation of 1-2 ml cell suspension with  100 mCi carrier-free 32 PO 4 32 (GE Healthcare, Diegem, Belgium). The labelled cells were divided into 200 ml aliquots in 2-ml Eppendorf vials and labelling was terminated by the addition of 20 ml perchloric acid (50%, v/v). After brief vortexing, the samples were frozen in liquid nitrogen. After 5 min, 450 ml CHCl 3 :MeOH:HCl (50:100:1, v/v) was added and the mixture was briefly mixed and snap-frozen in liquid nitrogen. During thawing the tubes were shaken vigorously and thereafter snapfrozen again to improve lipid recovery. After subsequent thawing, 450 ml CHCl 3 and 200 ml 0.9% (w/v) NaCl was added to induce a two-phase system. Tubes were then vigorously shaken for 15-30 min and centrifuged for 2 min. The lower-phase was transferred to a fresh tube containing 450 ml CHCl 3 :MeOH:1 M HCl (3:48:47, v/v), mixed for 1 min and centrifuged for 2 min. Subsequently the upper-phase was removed and the lower organic phase was dried under a stream of N 2 (g) and the isolated lipids were dissolved in 20 ml CHCl 3 and stored at 220uC until usage.

Phospholipase D assay
The metabolically labelled phospholipids were dried under N 2 (g) and resuspended into vesicles in 200 ml 10 mM Tris buffer (pH 6.4) by sonication (VWR ultrasonic cleaner). Aliquots (10 ml) were supplemented with 5 ml buffer containing 40% propanol. The assay was initiated by adding 85 ml medium (control) or medium in which P. infestans was cultured (extracellular medium), followed by a 1 hour incubation at room-temperature. The reaction was stopped by adding 375 ml CHCl 3 :MeOH:HCl (50:100:1 [v/v]), and processed as described [8].

TLC analysis
Lipid samples were separated by thin layer chromatography on Merck silica 60 TLC plates (Darmstadt, Germany) using the ethyl acetate system as described [9]. Radiolabelled phospholipids were visualized by phosphoimaging (Storm, Molecular Dynamics; Sunnyvale, CA, USA).

Results
Sixteen genes in P. infestans encode proteins with phospholipase D hallmarks The genome sequence of P. infestans harbours eighteen genes encoding a PLD, all of which can be classified in the subfamilies identified previously in P. sojae and P. ramorum (Table 1). PXPH-PLD, PXTM-PLD and TM-PLD are single copy genes. Three genes encode members of the PLD-like subfamily and the remaining twelve encode putatively secreted PLDs (sPLD-likes) based on the presence of a predicted SP. Alignments revealed that eleven sPLD-likes share sustained protein similarity (.50% overall identity) whereas one (sPLD-like-1) lacks significant identity with any of the others (15% ; Table S1). sPLD-like-1 was therefore designated as a type A (sPLD-like-A) and the others as type B (sPLD-like-B). Two of the latter, sPLD-like-4 and -11, lack continuous open reading frames (ORFs) and are considered as pseudogenes. In retrospect, also P. sojae and P. ramorum have one sPLD-like-A ortholog ( Table 1).
The PLD proteins were analyzed for characteristic conserved motifs ( Table 2). The two catalytic motifs HKD1 and HKD2 are conserved in P. infestans PXTM-PLD, PXPH-PLD and sPLD-like-A. In TM-PLD, the PLD-likes and sPLD-like-Bs one HKD motif is modified. In TM-PLD HKD2 is changed in HND, as in the P. sojae and P. ramorum TM-PLDs. The PLD-likes and sPLD-likes lack the aspartate residue (D) of HKD1 at the expected position.
Additional motifs in PLDs have been recognized as regulatory or binding domains. In PXPH-PLD both the PX and PH domain are readily conserved. PXTM-PLD contains a PX domain and five transmembrane domains [6]. A remnant of a PH domain (E-value 1.64E +03 ) was detected between transmembrane domains   Table 1 and [4]) were included in the analysis. On the right, the vertical lines represent scaffolds containing one or more (s)PLD-like genes (rectangles). The scaffold numbers are shown above or below each line. On scaffolds with two or more (s)PLD-like genes the number in each rectangle indicates the distance in kb between the start of this gene and the start of the first PLD-like gene (marked by 0) on the scaffold. (s)PLD-likes that cluster in the phylogram have the same color. Dotted lines connecting the scaffolds indicate that the genomic regions are syntenic. doi:10.1371/journal.pone.0017767.g002 two and three. As described previously PXTM-PLD has a ''PIP 2binding domain'', a ''PC-binding site'' (IYIENQFF motif) and a Ga-protein binding motif DRY/RVYVVV between HKD1 and HKD2 [6]. In human HsPLD2 and Arabidopsis AtPLDd, the PIP 2binding domain was recently identified as a tubulin binding region and the IYIENQFF and the DRY/VYVVV motifs appear to act as an actin binding fragment [10,11,12]. P. infestans PXPH-PLD and PXTM-PLD encompass all these domains. Both, the PIP 2 -binding domain and the IYIENQFF motif are present in PLD-likes and sPLD-like-Bs. Notably, in these cases the C-terminal hydrophobic region of the DRY/RVYVVV motif is conserved but the DRY motif is not found. In PXTM-PLD and plant PLDs it is replaced by EKF [6]. In TM-PLD an EPF is located at the expected position but the hydrophobic region is lacking. sPLD-like-1 solely consists of the catalytic site with the two HKD motifs ( Table 2). Downstream of HKD2, PXPH-PLDs and C2-PLDs contain a so-called IGSANIN motif that is supposed to play a role in membrane attachment. The motif is fully conserved in PXPH-PLD whereas in PXTM-PLD, the PLD-likes and sPLD-like-Bs a slightly modified motif is found. In TM-PLD the motif is barely Figure 3. Phylogram of Phytophthora sPLD-like-A orthologs and their homologs in various organisms. The consensus minimal evolution tree constructed from the amino acid sequences is shown. For protein sequence identifiers see Table 3. doi:10.1371/journal.pone.0017767.g003 recognizable; only two out of seven amino acids are conserved. sPLD-like-A lacks the motif but instead has a ''VGSANMD'' motif that is located directly downstream of HKD1 (Table 2).

Phylogenetic relationships of PLD genes
To analyse the relationship between Phytophthora PLDs and those of other organisms, PLD sequences were retrieved from genome databases. For an overall comparison, P. infestans PLDs were clustered with the 16 PLDs encoded in the rice genome. Rice has a putatively secreted PLD (OsPLDQ) and hence, has three PLD subfamilies [3].P. infestans PXPH-PLD and PXTM-PLD cluster with rice PXPH-PLDs. Most sPLD-likes and PLD-likes cluster in one large clade separated from the rice C2-PLDs. The sPLD-like-A of P. infestans clusters with OsPLDw. TM-PLD is an outlier that is absent in rice (Figure 1).
Further study revealed that PXPH-PLDs from plants and mammals are the closest homologs of P. infestans PXTM-PLD and PXPH-PLD (not shown). For TM-PLD, low similarity was found only with three P. infestans proteins with a DUF803 domain (PITG_05621, PITG_09535 and PITG_12469). Since the DUF803 domain corresponds to the transmembrane domains an additional analysis was performed for the C-terminal part harbouring the catalytic site. With an E-value of 6e 26 a PLD from Phaeobacter gallaeciensis (ZP_02144336) was the closest homolog.
Evolution of (s)PLD-like genes in Phytophthora The large number of PLD-like and sPLD-likes genes in Phytophthora could point to gene duplication events in a common ancestor. To investigate this we constructed a phylogram of all (s)PLD-likes in the three Phytophthora species (Figure 2). The PLDlikes with the exception of Pi-PLD-like-2 group in one clade and Pi-PLD-like-1 and -3 each have their own ortholog in P. sojae and P. ramorum. The bulk of the tree represents the sPLD-like-B family and remarkably includes Pi-PLD-like-2 that groups with three Pi-sPLD-like-B genes, i.e. Pi-sPLD-like-9, -11 and -12, in a distinct P. infestans specific subclade. Another P. infestans specific cluster is observed in the subclade comprising Pi-sPLD-like-2, -3, -4 and -5 which all four group with the same ortholog in P. sojae and P. ramorum. Other subclades lack a P. infestans ortholog. This nonuniform distribution is also evident in Table 1 in which the closest homologs of the P. infestans (s)PLD-like genes are listed based on BlastP. Several sPLD-like-Bs share the same P. sojae or P. ramorum protein as closest homolog rather than having unique sets of three orthologs representing one copy in each of the species. This suggests that some sPLD-like-Bs in P. infestans evolved directionally so that the paralogs are more related to each other than to orthologs. Despite this close relationship the paralogs still show significant differences at identity level (Table S1).  Figure 2). Five genes are located on scaffold_1, including two sPLD-like-B genes that are 12 kb apart, and two PLD-like genes, 21 kb apart. Three sPLD-like-B genes are located on scaffold_18 spanning 127 kb. Also the two sPLD-like-B genes on scaffold_10 are far apart (135 kb). Scaffold_27 and scaffold_181 have two sPLD-like-B genes whereas sPLD-like-A is located on scaffold_63. When compared to P. sojae and P. ramorum, the distribution of PLD genes within P. infestans seems more dispersed. For example, in P. sojae, six sPLD-like-B genes are located within an 18 kb region and similarly, in P. ramorum, five within 15 kb [4].
To examine the extent of synteny between species, genomic regions flanking each PLD gene were analysed. The average intergenic length was 1.7 kb, 2.4 kb and 6.7 kb, respectively, for P. ramorum, P. sojae and P. infestans. The regions carrying sPLD-like-1 ( Figure 2) and the single copy genes PXPH-PLD, PXTM-PLD and TM-PLD (not shown) showed conserved synteny. The majority of the PLD-like and sPLD-like genes in P. infestans are surrounded by genes that co-localize with PLD-like and sPLD-like genes in the other two species although genes are often rearranged (Figure 2). Scaffold_181 of P. infestans might have been the result of a duplication event involving Scaffold_1. For both PLD-like-2 and sPLD-like-9, some of the surrounding genes were found on P. sojae scaffold_64 and P. ramorum scaffold_112 suggesting another duplication event. Alternatively, scaffold_228 which is only 184 kb in size, might in reality be part of scaffold_18. For genes flanking sPLD-like-11 and -12 homologs were found in the other two species but not in regions harbouring sPLD-like-B genes. Based on the phylogeny and the synteny it seems plausible that the two P. infestans specific sPLD-like-B clades arose by gene duplication after emergence of the species infestans. Moreover, since P. sojae and P. ramorum are more at the base of the Phytophthora phylogeny than P. infestans the absence of P. infestans orthologs in some subclades is probably due to gene loss.

sPLD-like-A represents an independent peak in PLD evolution
Phytophthora sPLD-like-A is quite divergent from other PLDs. The closest relatives are found in plants while more distant sPLDlike-A sequences are widespread in metazoan, viruses and prokaryotes. This suggests they are all descendants of an ancient gene ( Figure 3; Table 3). In contrast, the sPLD-likes-Bs and PLD-likes are most analogous to PLD-like sequences derived from actinomycetes ( Figure 4). Representative sPLD-like-A homologs were further analysed for potential characteristics (Table 3) and some clade representatives were aligned ( Figure 5). It was noted that the prediction of the SP for Phytophthora orthologs is indecisive and hardly discriminates between a SP or a signal anchor (SA). Signal anchors are found in proteins that are inserted but not cleaved in the ER. sPLD-like-A homologs from plants and Dictyostelium have an obvious SP with an HMM score higher than 0.98. In contrast, a SA was predicted for many other sPLD-like-As, including all mammalian derived homologs (HMM score .0.98). Some, including those of Caenorhabditis elegans, Drosophila erecta and virus derived sequences, have neither a SP nor a SA (Table 3).
All sPLD-likes-As have a correct HKD1 except for Physcomitrella patens. For HKD2, a substitution is observed in nearly all mammalian and in the Dictyostelium homolog. The ''VGSANMD'' motif adjacent to the HKD1 motif is well conserved among all. Phytophthora sPLD-like-As are highly homologous (.75%) but with sPLD-like-As from other organisms the homology is much lower ( Figure 5). The conserved regions mainly cover the HKD1 and HKD2 motifs and surrounding regions.

Extracellular PLD activity
The finding that PLDs in two of the subfamilies have putative signal peptides urged us to validate if P. infestans secretes PLDs. First, the expression of the various PLD genes was analysed using P. infestans EST depositories [13] and Nimblegen data [14]. ESTs were identified (Table 1) for PXPH-PLD, PXTM-PLD, TM-PLD, sPLD-like-1, sPLD-like-12 and PLD-like-1. Most ESTs are derived from mycelial libraries which is in agreement with the Nimblegen expression profiles ( Figure S1). The mycelial tissue was therefore tested as the source for extracellular PLD activity. A mix of metabolically labelled plant phospholipids was presented as substrate to cell free P. infestans extracellular medium. The reaction was performed in the presence of a primary alcohol. PLDs have the unique capability to transfer a phosphatidate intermediate to short-chain primary alcohols (acting as substitute for water) resulting in the formation of phosphatidylalcohols [15]. These transphosphatidylation products are easily detectable by thin layer chromatography (TLC) due to their metabolic stability and unique migration properties. Incubation of extracellular medium of P. infestans mycelial cultures with the metabolically labelled phospholipids in the presence of 2% propanol resulted in the production of both PA and phosphatidylpropanol (PPro) whereas in the control V8 medium no PLD activity was detected. Increase in PA and appearance of the non-pre-existing PPro was observed for all P. infestans strains tested ( Figure 6) and this points to a ubiquitous presence of extracellular PLD activity. PA and PPro levels varied per strain which probably reflects differences in growth rate, viability of the mycelium, or variation in the amount of secreted activity among strains. Transphosphatidylation was already observed at the lowest propanol concentration tested (0.1%; Figure 6, inset).
To exclude that release of intracellular PLDs from dead or dying cells causes the observed activity, V8 medium of a flooded plate was carefully removed, fresh medium was added and after overnight incubation the extracellular medium was again tested for PLD activity. In the ''fresh'' extracellular medium PLD activity appeared to be even higher than in the extracellular medium prior Figure 5. Phylogram of P. infestans PLD-likes and type B sPLD-likes and homologs in various organisms. The consensus minimal evolution tree constructed from the amino acid sequences is shown. For accession numbers of the P. infestans sequences see Table 1. Other sequences that were all selected by Blast searches are from: Arabidopsis (accession nr. NP_188194 for AtPLDa1, NP_175666 for AtPLDa2 and NP_197919 for AtPLDa3), Frankia alni (P_711983), Janibacter (ZP_00995229), Kineococcus radiotolerans (YP_001361943), Kribbella flavida (YP_003383471), Mycobacterium smegmatis (YP_884788), Oceanibulbus indolifex (ZP_02154842), Saccharopolyspora erythraea (YP_001106428), Streptomyces ambofaciens (CAJ89461), Streptomyces pristinaespiralis (YP_002197876) and Streptosporangium roseum (YP_003339462). doi:10.1371/journal.pone.0017767.g005 Figure 6. Profiles of plant phospholipids exposed to P. infestans extracellular medium reveals PLD activity. Metabotically labelled phospholipids were isolated from tobacco suspension cell culture, sonicated into vesicles, mixed with extracelllular medium collected from various P. infestans strains and incubated for 60 min in the presence of a primary alcohol (2% propanol). Lipids were (re)extracted, separated by TLC and visualized by phosphoimaging. As control, vesicles were treated with V8 medium. to the replacement, demonstrating that the extracellular PLD activity is due to active release of PLD by living cells (Figure S2).

Discussion
Eighteen PLD genes were identified in the genome of P. infestans and this correlates with findings in P. sojae and P. ramorum [4]. It clearly exceeds the number of PLD genes in yeasts, mammalians, and even plants. Also the number of PLD subfamilies in Phytophthora is staggering: six subfamilies including two distinct potentially secreted sPLD-like subfamilies, type A and B. Like many plant pathogens Phytophthora species have large secretomes [14], anticipating potential functions in pathogenicity. Our finding that P. infestans secretes PLD activity in the medium suggests that Phytophthora exploits PLDs to modify host tissues and as such, PLDs may function in pathogenicity.
Phytophthora PXPH-PLD and PXTM-PLD are closely related to canonical PXPH-PLDs as illustrated by conserved catalytic and regulatory domains with the exception of the transmembrane domains in PXTM-PLD. A preliminary functional analysis revealed that homozygous PXTM-PLD knock-out mutants show aberrant growth behaviour [16]. TM-PLD lacks homology with any other protein outside Phytophthora. Within Phytophthora species TM-PLD is highly conserved suggesting that this protein has a function in these organisms although not necessarily as a PLD.
The sPLD-like-Bs and PLD-likes have all regulatory domains typical for PLDs but lack the lipid binding domains and the DRY motif. Most possess an altered HKD1 motif lacking aspartate. Nevertheless, the motif might still be able to participate in a catalytic reaction. Attempts to express the genes in E. coli were unsuccessful. The transformation rate was low and in all viable transformants the constructs showed frameshifts or point mutations leading to inactive proteins (unpublished observations). The abundance of sPLD-like-B isoforms suggests that they have a prominent function.
In contrast to eleven sPLD-like-Bs, P. infestans has only one sPLD-like-A which, unlike sPLD-like-Bs, has an inconclusive prediction for its signal peptide. sPLD-like-A homologs exist outside oomycetes. The homology, however, is mostly restricted to the catalytic motifs and there is no uniform signal peptide prediction. The latter could point to different destinations of the various sPLD-like-As, in or outside the cell. Unlike the metazoan homologs, the plant homologs have an obvious signal peptide. This might be essential to cross cell walls and reach membranes in adjacent cells. Like plants, Phytophthora has cell walls so for Phytophthora sPLD-like-A one could argue in favour of a signal peptide for rather than a signal anchor. However, this needs further experimentation.
The large number of sPLD-likes suggests that Phytophthora secretes PLD activity. Since most sPLD-likes have a modified HKD motif, the substrate specificity is unknown and unanticipated activities could be encountered. Potential PLD activity monitoring problems, due to substrate specificity, phospholipid composition and ratio, alternative phospholipase and kinase activities and buffer preferences were circumvented by exploring a novel in vitro approach based on labelled plant phospholipids. We demonstrated that all tested P. infestans strains secrete PLD activity although the identity of the active enzyme remains to be established. The observed differences in PLD activity among strains probably reflects variations in growth rate, viability and the amount of enzyme(s) secreted. Analysis of similar samples by an alkaline TLC system revealed that PtdCho and PtdGro are the main substrates (data not shown). The PLD activity was capable to transphosphatidylate at low alcohol concentrations as expected for common PLDs [2]. This novel in vitro assay could be a first-class tool to dissect the biochemical characteristics of the PLD activity and to identify the enzyme(s) responsible.
Phytophthora is a hemibiotrophic plant pathogen and as such it is conceivable that secreted PLDs are utilized to function as generators of PA at the outside layer of host cell plasma membrane. PA has been characterized as a multifunctional phospholipid with direct or indirect impact on many cellular processes [1,17]. In mammalian cells, exogenous added PLD activates G-protein-coupled receptors thereby mimicking hormones and growth factors [1]. Adding exogenous PA to P. infestans zoospores triggers encystment [9] whereas exogenous PA addition to plant cell-suspension culture results in MAPK activation, actin cytoskeleton rearrangements, reactive oxygen species generation, chlorosis, cell death responses [1,8] and stimulation of secretion [17]. Ergo, if Phytophthora PLDs initiate PA production at the plant membrane several cellular processes may be activated that influence the infection process. Figure S1 Nimblegen microarray data for Phytophthora infestans PLD genes. Each bar represents the average of two individual hybridisations. Samples were taken from P. infestans mycelium (strain T30-4) on various agar media (Pea agar, V8 agar and RS agar) or from infected potato leaves, 2-5 days postinoculation (DPI). Nimblegen microarray data are available in GEO under accession number GSE14480 [14]. (TIF) Figure S2 PLD activity is continuously released into fresh medium. Metabolically labeled phospholipids were incubated for 60 min in the presence of 2% propanol with fresh control medium (c), extracellular medium obtained by flooding (1 e ) and refreshed extracellular medium (2 e ) of P. infestans strain 88069. Lipids were extracted, separated by TLC and visualized by phosphoimaging.

(TIF)
Table S1 Amino acid identity between P. infestans sPLD-likes. The comparison was performed using Vector NTI and the results are presented as percentage (%) identity. (DOC)