Haemonchus contortus P-Glycoproteins Interact with Host Eosinophil Granules: A Novel Insight into the Role of ABC Transporters in Host-Parasite Interaction

Eosinophils are one of the major mammalian effector cells encountered by helminths during infection. In the present study, we investigated the effects of eosinophil granule exposure on the sheep parasitic nematode Haemonchus contortus as a model. H. contortus eggs exposed to eosinophil granule products showed increased rhodamine 123 efflux and this effect was not due to loss of egg integrity. Rh123 is known to be a specific P-glycoprotein (Pgp) substrate and led to the hypothesis that in addition to their critical role in xenobiotic resistance, helminth ABC transporters such as Pgp may also be involved in the detoxification of host cytotoxic products. We showed by quantitative RT-PCR that, among nine different H. contortus Pgp genes, Hco-pgp-3, Hco-pgp-9.2, Hco-pgp-11 and, Hco-pgp-16 were specifically up-regulated in parasitic life stages suggesting a potential involvement of these Pgps in the detoxification of eosinophil granule products. Using exsheathed L3 larvae that mimic the first life stage in contact with the host, we demonstrated that eosinophil granules induced a dose dependent overexpression of Hco-pgp-3 and the closely related Hco-pgp-16. Taken together, our results provide the first evidence that a subset of helminth Pgps interact with, and could be involved in the detoxification of, host products. This opens the way for further studies aiming to explore the role of helminth Pgps in the host-parasite interaction, including evasion of the host immune response.


Introduction
Soil transmitted helminths have a major impact on both human and animal health. These parasites are responsible for increased morbidity, mortality and premature birth [1][2][3][4]. While the host immune response against parasitic helminths is highly complex, granulocytes are known to be involved in both the initiation and effector immune response phases [5,6]. These responses are associated with a proliferation of T-helper 2 lymphocytes, plasma cells, eosinophil cells, basophils and mast cells [7,8]. Among these granulocytes, eosinophils constitute a homogeneous population of cells that are recruited from bone marrow into the blood and then host tissues where they survive for several days or even weeks [6]. In vitro studies have demonstrated that granulocytes interact with pathogenic helminths causing major damage to the parasites [9]. During degranulation, eosinophils release cationic proteins with significant cytotoxic activity such as the major basic protein, eosinophil peroxidase and eosinophil-derived neurotoxin [10]. In this respect, it is expected that successful establishment of the parasite in vivo might require efficient detoxification mechanisms to defend against the host immune response products. To date, these mechanisms remain largely unknown in helminths.
Pgps are membrane pumps involved in the active transport of many biological and xenobiotic substances using the energy provided by ATP hydrolysis. They are of particular interest here since vertebrate Pgps have been shown to be involved in the transport of immune cell products [11][12][13][14][15][16]. For example, inhibition of Pgp activity in natural killer cells has been correlated with a decrease in their cytotoxicity [17]. We have a precedent from research into resistance to the anthelmintic drug ivermectin, to expect that Pgps from the parasite may play a similar role to those in the host and so provide protection by the transport of host granulocyte products [18][19][20].
The sheep parasitic nematode Haemonchus contortus represents one of the most pathogenic species impacting the livestock industry [21]. Exposure of infective H. contortus larvae to host eosinophils in vitro reduces their ability to establish a subsequent infection in vivo, making this species an appropriate model to investigate the interaction beetween eosinophils and the parasite [22]. We have previously developed an egg based assay where fluorescent rhodamine-123 transport out of H. contortus eggs is mediated by Pgp [23,24]. In this study, we observed that eosinophil granule proteins increase this Rh123 transport, suggesting a role for Pgp in protection of the parasite. In an effort to identify the molecular mechanisms involved, we cloned sequences for several different Pgp genes from H. contortus. Quantitative PCR was used to investigate the expression kinetics of Pgp genes in different developmental stages in order to identify those that could potentially be involved in host product detoxification in vivo. We also examined changes in Pgp gene expression induced by exposure to eosinophil granules in vitro.
Here, we report for the first time, that several H. contortus Pgps show increased expression associated specifically with the parasitic stages and that two of these show a specific increase in expression following exposure to host granulocyte products.

Isolation of sheep eosinophils and granule product purification
We combined the Percoll gradient method with flow cytometry assays in order to select an intact eosinophil population from the blood of infected sheep ( figure 1A). This approach allowed us to obtain a cell population containing approximately 95% eosinophils (figure 1B). Purity of the eosinophils was further checked by Giemsa May-Grunewald staining (figure 1C). These purified cells enabled us to extract eosinophil granule proteins by ultracentrifugation for subsequent experiments, including functional analysis in vitro and modulation of Pgp mRNA expression. Purified granule proteins were analyzed by SDS-PAGE highlighting approximately 10 distinct bands, ranging from 11 to 70 kDa. In contrast, total eosinophil lysate used as a control for granule purification efficiency contained several additional proteins ranging from 11 to over 100 kDa ( figure 1D).
We have previously reported that Pgp activity can be analyzed in H. contortus eggs using a rhodamine (Rh123) efflux assay [23,24]. In the present work, Rh123 efflux increased proportionally with exposure to granule proteins (figure 2). This relationship was established up to a protein concentration of 1250 mg/ml, beyond which a plateau of 75% stimulation was reached (P,0.0069). The loss of Rh123 from the eggs was not due to a loss of egg integrity since FITC was excluded from both untreated eggs and those exposed to eosinophil granule products, whereas FITC was easily able to enter eggs damaged by freezing ( figure S1). This result provided the first functional evidence that parasite Pgp activity can be modulated by host eosinophil granules products.

Identification of Pgp transcripts in the adult stage of H. contortus
In the present work 3' RACE-PCR experiments allowed the identification of 8 partial cDNA sequences corresponding to the 39 end from distinct H. contortus Pgp transcripts. All these cDNA sequences were then further extended in their respective 59 end by RT-PCR and among the eight candidates, four complete coding cDNA sequences have been obtained. Sequence descriptions including size and closest homologs in C. elegans and other nematode species are summarized in table 1. When the resulting complete, or partial, sequences matched unambiguously with a single orthologous gene in C. elegans, the corresponding H. contortus Pgp cDNA sequences were named following Beech et al [25]. No clear ortholog of Hco-pgp-16 could be identified in C. elegans and so the numerical series of Pgp genes was extended. Interestingly, we have identified three distinct H. contortus Pgp sequences for which the closest homolog in C. elegans was Cel-pgp-9 (i.e. Hco-pgp-9.1, Hcopgp-9.2 and Hco-pgp-9.3). Phylogenetic analyses of the complete sequences obtained in H. contortus were performed including homologous sequences in the closest species (Caenorhabditis elegans, (Cel), Caenorhabditis briggsae (Cbr), Ascaris suum (Asu), Parascaris equorum (Peq)). A split-decomposition network analysis revealed the phylogenetic proximity of the Hco-pgp-3 sequence with the Cel-pgp-3/Cel-pgp-4 group (figure 3). Hco-pgp-9.1 andCel-pgp-9 are clearly orthologous genes. Hco-pgp-16 clustered with Peq-pgp-16, Asu-mrp-3 and Cbr-CBG12369, suggesting the Ascaris sequence has been misidentified. Even though the Hco-pgp-16 sequence has no ortholog in C. elegans, it clustered within the Cel-pgp-3 and Celpgp-4 group. An alignment of the Hco-PGP-16 deduced aminoacid sequence with Cel-PGP-4 and Cel-PGP-3 revealed some amino acid similarity ranging from 66 to 67% respectively (figure S2) A subset of Pgp genes is specifically up-regulated during the parasitic phase of H. contortus life cycle In order to provide information on H. contortus Pgps that could potentially be involved in the host-parasite interaction, we investigated their expression kinetics through different developmental stages of the parasite. H. contortus presents a simple life cycle, including a free living and a parasitic phase within a single host. Eggs released in host faeces develop through three successive larval stages (L1 to L3) on the pasture. The transition to parasitism occurs when the host ingests infective third stage larvae (L3) while grazing. Then, larvae exsheath (xL3) and develop through a fourth larval stage (L4) into mature adult worms into the host abomasum.
The expression level of 9 different Hco-pgp mRNA (including Hco-pgp-2) was investigated in eggs, L3, L4 and adult male stages using real-time PCR. Note that adult females were not included as they contain eggs that could interfere with Pgp expression analysis.
All the Pgp genes were found to be expressed in the four different developmental stages studied in this work. However, for some Pgps striking differences of expression level have been observed between the free-living and the parasitic stages (figure 4). For example, Hco-pgp-3, Hco-pgp-9.2, Hco-pgp-11 and Hco-pgp-16 transcripts were found to be significantly more abundant in parasitic life stages (L4 and/or adult male) in comparison with free-living stages (eggs and/or L3) with an increased expression ranging from 7.5-fold (Hco-pgp-16) up to 250-folds (Hco-pgp-11). Within the parasitic stages, Hco-pgp-3 was found to be equally expressed in L4 and adult whereas Hco-pgp-11 and Hco-pgp-16 mRNA were found to be more abundant in adult stage than in L4. Among the three H. contortus Pgp-9 homologues it is noteworthy that only Hco-pgp-9.2 was found to be temporarily up regulated in the L4 stage in contrast with Hco-pgp-9.1 and Hco-pgp-9.3 that were found to be more abundantly expressed in free living stages. Hcopgp-3, Hco-pgp-9.2, Hco-pgp-11 and Hco-pgp-16 represent candidates of major interest for their potential involvement in host product detoxification in vivo even though we cannot rule out that other Pgps may also play an important role in the host parasite interaction. These four Pgps were selected for further characterization to investigate their potential interaction with eosinophil granule products.
Eosinophil granules induce the specific overexpression of Hco-pgp-3 and Hco-pgp-16 genes in exsheathed L3 larvae The Rh123 transport assay data suggest that eosinophil granule products may interact directly with Pgp transporters in H. contortus eggs. Previous studies of bacteria, mammals and nematodes report that Pgp substrates can specifically induce overexpression of their corresponding Pgp gene [26][27][28][29]. Taking advantage of this property, we examined which H. contortus Pgp genes respond specifically to host granule products in vivo. Expression of Hco-pgp-3, Hco-pgp-9.2, Hco-pgp-11 and Hco-pgp-16 was monitored by qRT-PCR in artificially exsheathed L3 H. contortus larvae that mimic the first parasitic stage in contact with eosinophil granules in vivo.  Larvae were exposed for 24 h to increasing concentrations of granule products ranging from 300 to 2 000 mg/ml. Comparisons between the stimulated and non-stimulated exsheathed L3 (control) showed a significant (P,0.01, n = 6) and dose dependent increase of mRNA expression levels of Hco-pgp-3 and Hco-pgp-16 in stimulated larvae, whereas expression of Hco-pgp-9.2 and Hco-pgp-11 remained unchanged ( figure 5). This result shows that eosinophil granule products contain compounds that specifically modulate the expression of Hco-pgp-3 and Hco-pgp-16 and suggests these as candidate Pgps that specifically protect H. contortus from granule products in vivo.

Discussion
Helminth Pgps have been widely studied for their involvement in xenobiotic transport and their association with drug resistance. Their implication in other critical functions such as the hostparasite interaction remained to be explored. We found that eosinophil granule products specifically affected the transport of Rh123 that has previously been shown to be mediated by Pgp transporters. Eosinophils are known to be key effectors during helminth infection and we hypothesized that parasitic helminth Pgp could play a role in detoxification of eosinophil granule products. Eosinophil granule products are mainly composed of four highly basic proteins that have been shown to be toxic for helminths. These include eosinophil peroxidase (EPO), major  Table S3. doi:10.1371/journal.pone.0087802.g003 basic protein (MBP), eosinophil cationic protein (ECP), and eosinophil-derived neurotoxin (EDN) [30,31]. Eosinophil granule proteins are stored in a crystalloid structure that can be purified by ultracentrifugation [32]. However purification of an eosinophil cell population from sheep blood remains challenging. For example Terefe et al. [22] reported that a single Percoll gradient (1.090 g/ ml) allowed the purification of sheep eosinophils with a purity ranging from 43 to 63%. In the present work, we optimized this approach using two gradient Percoll densities in combination with cell sorting that resulted in a cell population containing 95% eosinophils. Thus, we could hypothesize that most of the protein collected after eosinophil lysis and ultracentrifugation would correspond to granules products. Separation of purified sheep eosinophil granule protein with SDS-PAGE revealed a set of proteins ranging from 11 to 20 kDa. It is tempting to hypothesize that at least some of them could correspond to the major cationic proteins (MBP, ECP and EDN) presenting similar molecular weight as reported for human eosinophil granule proteins (i.e. 13, 16 and 18 kDa respectively) [33]. These granule proteins were then used in a Rh123 efflux assay to assess their potential role as modulators of Pgp activity. Using H. contortus eggs, we showed that eosinophil granule proteins are able to specifically interact with the parasite Pgp. This result provides functional evidence that granule proteins could represent potential substrates for the parasite Pgp in vivo. Interestingly, the relationship between granule protein concentration and the resulting Rh123 efflux is similar to those observed with a classical Pgp substrates such as colchicine [34] suggesting a strong interaction between H. contortus Pgps and granule proteins. Therefore this result provided a robust argument to further investigate this interaction at the molecular level.
In order to identify H. contortus Pgp genes potentially involved in granule protein transport in vivo our first objective was to explore their expression in a parasitic stage (adult male) of H. contortus. In contrast with mammals, nematodes possess a large diversity of Pgp genes. The model nematode Caenorhabditis elegans that is closely related to H. contortus encodes up to 14 Pgp genes (pgp-1 to pgp-14) [35]. However, in H. contortus, only a single complete Pgp sequence (Hco-pgp-2) has been reported to date [36]. In a recent study, Williamson et al developed quantitative PCR for several of these, corresponding to partial transcript sequences identified in the infective larval stage (L3), providing a basis for H. contortus Pgp expression analysis [37]. In the present work, H. contortus Pgp transcripts have been further investigated in the adult stage using an RT-PCR approach. We have also extended transcript sequence information including 3'UTR regions in order optimize transcript specific primer design for subsequent quantitative PCR experiments.
Analyses of the H. contortus Pgp sequences found a high degree of conservation with their respective homologs in C. elegans, with the striking exception of Hco-pgp-16, for which the closest homolog found in Genbank corresponds to the recently reported pgp-16 sequence from the distantly related horse parasite Parascaris equorum. Despite this lack of a clear ortholog within the free living species C. elegans, the closest homolog in H. contortus was the Hco-pgp-3 sequence. It is noteworthy that the complete cDNA sequence corresponding to Hco-pgp-3 gene identified in the present work perfectly overlaps with two distinct partial cDNA sequences previously designated as Hco-pgp-3 (HM635768) and Hco-pgp-4 (HM635766) stressing the need to obtain complete cDNAsequence before assigning a definitive nomenclature of H. contortus genes. Interestingly, we were able to identify three potential homologs of C. elegans Pgp-9 that mostly differ in their respective 39UTR region (data not shown). The origin and functional implication of such potential duplication events remain to be explored.
For parasitic nematodes, the transition to parasitism implies some major adaptations such as evasion of the host immune response and changes in nutrition, metabolism and growth [38,39]. In order to identify H. contortus Pgp genes potentially involved in host-parasite interaction and more specifically those involved in eosinophil granule product detoxification, we have compared the expression level of H. contortus Pgps in free living stages (eggs and L3) and parasitical stages (L4 and adult). Our results showed that Hco-pgp-9.2, Hco-pgp-11, Hco-pgp-3 and Hco-pgp-16 mRNAs are each overexpressed in L4 and/or adult in comparison with to free living stages.
Homologs of pgp-9 and pgp-11 have been associated with anthelminthic resistance in the parasitic nematodes Teladorsagia circumcincta and Parascaris equorum, respectively [40,41]. Overexpression of these genes in parasitic stages of H. contortus raises the question of potential functions during the interaction between host and parasite. The closest homolog of Hco-pgp-3 identified in databank is C. elegans pgp-3, which is involved in defense against natural toxins produced by plants and bacteria [42,43]. Interestingly, our phylogenetic analysis revealed that pgp-16 orthologs of H. contortus, Ascaris suum and Parascaris equorum are clustered within the same group as Cel-pgp-3, highlighting the importance of this Pgp group for future studies.
It is tempting to speculate that the specific over expression of Hco-pgp-3, Hco-pgp-9.2, Hco-pgp-11 and Hco-pgp-16 in parasitic stages might reflect their potential activation during the host/ parasite interaction. It has been reported that Pgp substrates can induce over expression of their cognate Pgp genes in nematodes [28,29] and we were able to show that, Hco-pgp-3 and Hco-pgp-16, but not Hco-pgp-9.2 or Hco-pgp-11, were specifically induced in exsheathed L3 (xL3) larvae exposed to eosinophil granule protein in a dose dependent manner. The specific induction of Hco-pgp-3 and the closely related Hco-pgp-16 represent an attractive and promising result suggesting that a subset of H. contortus Pgps might be involved in detoxification of host immune cell products. This lays the basis for future work aiming to decipher in more detail the mode of action of fractionated eosinophil proteins and other host products on parasitic helminth Pgps.
In conclusion, the study of nematode Pgps has, until now, been mainly based on their involvement in the detoxification of drugs in the rather artificial setting of anthelmintic treatment. The present work demonstrates for the first time that interactions occur between parasitic helminth Pgps and host cell products and provides some new insight about potential molecular mechanisms involved in overcoming the host immune response. This in turn may provide novel potential targets for the future control of H. contortus and other helminths of medical and veterinary importance.

Ethics statement
All animal experiments were approved by the Regional Centre-Limousin Ethics Committee (CL2006-012) and conducted under a license issued by the Directorate of veterinary services in Tours; France with the accreditation number: B-37-175-3.

Purification of eosinophils
Eosinophils from the peripheral blood of Haemonchus contortus infected sheep (21 days post-infection) were purified by discontinuous density Percoll gradients in combination with flow cytometry sorting. Briefly, 600 ml of blood were collected and mixed immediately with 50 ml of a solution containing ethylene diamine tetra-acetic acid (EDTA) at a concentration of 40 mg per ml in order to avoid blood coagulation. In order to lyse the erythrocyte cells, 150 ml of blood were mixed with 275 ml of water with stirring. After 20 s, cell lysis was stopped by adding 275 ml of 2,5X PBS solution. The cells were then centrifuged at 400 g for 5 min and the pellet washed 3 times in 50 ml of 1X PBS (0.14 M NaC1, 8 mM Na2HP04, 3 mM KCI, l.5 mM KH2P04. pH 7.4). A multiple discontinuous density gradient (1.040, 1.055, 1.070 and 1.080 g/ml) was prepared. The Percoll solutions were poured in a 15 ml centrifuge tube starting with the solution of the lowest density (1.040 g/ml) and finishing with the solution of the highest density (1.080 g/ml). The cell suspension was carefully layered on top of the gradient and centrifuged at 400 g for 15 min. The purified cells were collected at the 1.070/1.080 g/ml interface and washed with a 1X PBS. Isolated cells were further purified by flow cytometry (MoFloH, DakoCytomation, 4850 Innovation Drive, Fort Collins, CO) using Forward Scatter (FS) and Size Scatter (SS) parameters. The SS is a 90 degree light scatter that reflects the granularity of the cells and the forward angle light scatter (FS) is proportional to the cell size. The purity of the cell population was further checked by Giemsa May-Grunwald (Reactifs RAL H , Martillac, France) (GMG) staining according to the manufacture's recommendation.

Isolation of granules
The granules were isolated by ultracentrifugation according to the methods described by Slifman et al. [44]. In brief, the eosinophil cells were centrifuged at 600 g for 10 min and the resulting pellet was washed in 5 ml of 1X PBS. Eosinophils were then lysed in 0.25 M sucrose by repeated passage through an 18 gauge needle attached to a syringe. Eosinophil lysate was centrifuged at 600 g for 10 min in order to separate unbroken cells and granules. Granules were then pelleted by ultracentrifugation at 10 000 g for 10 min. The quantity of granule proteins was determined using a nanodrop spectrophotometer (Thermo Scientific, Waltham, Massachusetts, USA) and store at 280uC until used.

Denaturing Sodium Dodecyl-sulfate Polyacrylamide Gel Electrophoresis (SDS-PAGE)
Eosinophil lysate and purified eosinophil granules proteins were separated on SDS-PAGE acrylamide gels and stained with Coomassie blue [45]. Both samples were suspended in denaturation buffer and heated at 100uC for 5 min and loaded onto 10% polyacrylamide gels. Proteins were visualized after gel staining in a 0.25% Coomassie blue solution.

Parasites
For all experiments, the ''Weybridge'' Haemonchus contortus isolate was used. Sheep were infected with 6 000 infective third stage larvae (L3). Eggs were collected 28 days post infection from host faeces as previously described [46]. The L3 larvae were harvested from coprocultures and subsequently artificially exsheathed (xL3) following exposure to 0.1% hypochlorite solution for 10 min. Fourth stage larvae (L4) and adult male nematodes (A) were collected from the abomasal mucosa of sheep necropsied after 7 or 25 days post-infection respectively.

Rhodamine 123 (Rh123) assay
The effect of eosinophil products on parasite Pgp activity was determined by changes in the accumulation of rhodamine (Rh123) in the H. contortus eggs. Approximately fifteen thousand eggs were suspended in deionized water (1 ml) and were centrifuged at 1 000 g for 1 min. The supernatant was discarded. Rh123 solution (2 ml of 1.5 mM Rh123 in deionized water) was added to the egg pellet and incubated for 5 min at 20uC in the dark. Eosinophil granule products were then added at four distinct final concentrations (300, 600, 1250 and 2 500 mg/ml of granule proteins) and the eggs were incubated for a further 10 min at 20uC before being centrifuged at 1 000 g for 1 min. The supernatant was discarded and the eggs were washed three times with 5 ml of ice cold deionized water. Eggs were kept in the dark for 60 min before analyses. Specific fluorescence was measured using a Quanta Master spectrofluorometer (PTI, NJ, USA) equipped with a 75 W xenon lamp (l excitation = 495 nm, l emission = 525 nm). Data were given in arbitrary units of green fluorescence corrected for the native green fluorescence of eggs and expressed as the percentage of fluorescence in control eggs. Three replicates per condition were analyzed. The Rh123 efflux without eosinophil granules was normalized to 0. The Hill equation was used to describe the dose-response relationships [47] by calculating the following equation: where Bottom is the baseline and Top is the maximum effect, EC50 the dose giving half the maximum effect, X the granule protein concentration (log). Bottom values were adjusted to 0, corresponding to the fluorescence in the control eggs.

Molecular Biology
In order to identify H. contortus Pgps, total RNA was extracted from 10 adult males. Frozen parasite pellets were homogenized in Trizol reagent (Invitrogen, Carlsbad, California, USA) and RNA was isolated according to the manufacturer's recommendations. RNA pellets were dissolved in 35 ml of RNA secure re-suspension solution (Ambion, Austin, Texas, USA). DNA was removed form RNA samples using RQ1 RNase-Free DNase (Promega, ). The RNA concentrations were measured using a nanodrop spectrophotometer (Thermo Scientific, Waltham, Massachusetts, USA) and adjusted to 100 ng/ml.
First strand cDNA synthesis was performed on 1 mg of total RNA using the oligo (dT) RACER primer (Invitrogen) and the superscript III reverse transcriptase (Invitrogen,Carlsbad, California, USA) according to the manufacturer's instructions. Using first strand cDNA prepared from adult H. contortus, the 39 ends of Pgp cDNAs were isolated by 39RACE PCR using the GeneRacer kit (Invitrogen, Carlsbad, California, USA) according to the manufacturer's recommendations. 3'RACE PCR amplicons were obtained with forward primers designed upon H. contortus supercontig sequences (available at http://www.sanger.ac.uk/cgi-bin/ blast/submitblast/h_contortus) In order to extend sequence in their 5' ends, a new set of PCR experiments was performed using either the SL1 primer (GGTTTAATTACCCAAGTTTGAG) or forward primers designed from supercontig sequences. The complete list of primers is provided in the table S1. All PCR reactions were performed with a programmable thermocycler (Biometra, Gottingen, Germany). 39 RACE experiments were carried out in a final volume of 25 ml containing 15 ng of firststrand cDNA, 0.5 units of LongAmp TM Taq polymerase (New England Biolabs, Ipswich, United Kingdom), 200 mM of each dNTP and 0.4 mM of each primer. The reaction mixture was denatured by heating to 94uC for 30 sec, followed by 32 cycles at 94uC for 30 sec, 54-60uC (depending on specific melting temperature of primers) for 30 sec, 72uC for 30-240 sec (depending on the amplicon size). A final extension step was performed at 72uC for 7 min. Amplification products were cloned in PGEM-T vector (Promega, charbonnieres, France) and sequenced by GATC biotech (Konstanz, Germany).

Sequence comparison and Phylogenetic analyses
Database searches were performed using the Haemonchus contortus blast server (http://www.sanger.ac.uk/cgi-bin/blast/submitblast/ h_contortus) and the BLAST Network Service (National Center for Biotechnology Information), using the BLASTX algorithm [48].
The nucleotide sequences (Genbank-accession numbers are provided in Table S3) were translated and amino acids sequences were aligned using MUSCLE [49]. The corresponding nucleotide alignments were obtained by concatenating codons using the REVTRANS server (http://www.cbs.dtu.dk/services/RevTrans/) [50]. The model of nucleotide substitution that best fitted the data was determined using jModeltest version 2.0 [51] on the Phylemon 2 server [52].
Phylogenetic relationships between Pgp sequences were defined for all the full length nucleotide sequences using the splitdecomposition method implemented in SPLITSTREE version 4.11.3 [53] using the general time reversible (GTR) scheme of substitution with a gamma distribution model of site rates variation (jModeltest results).

Real-time PCR
Expression of H. contortus Pgps in different developmental stages was analyzed using total RNA extracted from 25 ml of egg pellet, 5 000 L3, 400 L4, 10 adult males respectively using Trizol reagent (Invitrogen). In order to investigate the eosinophil granule effects on Pgp expression modulation, total RNA was extracted from xL3 incubated in water (control) or with increasing concentrations of eosinophil granule products (ranging from 300 to 2 000 mg/ml of proteins) during 24 hours. One microgram of total RNA from different nematode stages was used to generate cDNA with the oligo (dT) RACER primer (Invitrogen) and the superscript III reverse transcriptase (Invitrogen) according to the manufacturer's instructions. Quantitative RT-PCR experiments were carried out with specific forward and reverse primers designed either in identified sequences. Primer sequences used for quantitative RT-PCR experiments are provided in table S2. PCR amplification products were run on an electrophoresis gel and sequenced to ensure the absence of non-specific amplification products and that the primers were suitable for qRT-PCR.
Real-time PCR experiments were performed by monitoring the increase of fluorescence of IQ SYBRH Green in real-time (Bio-Rad, Hercules, CA, USA) with the Rotor-Gene 3000 (Corbett Research, Sydney Australia). The relative gene expression fold changes were calculated by the Gene Expression Analysis software (Bio-Rad, Hercules, CA, USA) using H. contortus gapdh, actin, and btubulin as reference genes. Pgp mRNA expression of eggs or xL3 incubated in the absence of eosinophil granules was normalized to 1. The data are presented as fold changes in mean 6 SEM of mRNA expression. For the analysis of H. contortus P-glycoprotein mRNA expression in different nematode developmental stages, statistically significant differences between eggs and other stages were analyzed with one-way ANOVA test followed by Bonferroni's Multiple Comparison Test. Differences were considered significant when P,0.05. Statistically significant differences between xL3 in the presence and absence of eosinophil granules were analyzed with the non-parametric Man-Whitney test using Prism software (v5.02, GraphPad Software, San Diego, Ca, USA). Figure S1 Fluorescein isothiocyanate (FITC) uptake assay performed on H. contortus eggs following exposure to host granule products. In order to control H. contortus egg viability after exposure to host eosinophil granule products, an FITC uptake assay was performed on eggs incubated in PBS (A), eggs incubated with eosinophil granule products (B), eggs frozen at 280uC. The absence of FITC uptake in the eggs incubated with or without host eosinophil granule products (A and B) confirmed their viability. In contrast, strong fluorescence associated with FITC uptake was observed in eggs previously frozen at 280uC during one hour (C). These stained eggs represent a control for egg mortality. Bars: 50 mm. Method: Approximately one thousand H. contortus eggs were suspended in PBS 1X and were incubated 1 h at room temperature with host granule proteins at a final concentration of 2 500 mg/ml. As a control of viability or mortality, eggs were also incubated in PBS 1X at room temperature or frozen at 280uC during one hour respectively.

Supporting Information
Fluorescein isothiocyanate (FITC) solution (20 ml of 1 mg/ml FITC in PBS 1X) was added to the eggs and incubated for 1 h at 24uC in the dark. The eggs were then washed three times with 1 ml of ice cold deionized water and analyzed by fluorescence microscopy (DP50 camera, Olympus) using a filter band-pass with 450 to 480 nm of excitation and 515 nm of emission. (TIF) Figure S2 Alignment of Hco-PGP-16, Hco-PGP-3, Cel-PGP-3 and Cel-PGP-4 sequences. Hco-PGP-16, Hco-PGP-3, Cel-PGP-3 and Cel-PGP-4 sequences were aligned using the MUSCLE algorithm [49] and further processed using the GeneDoc program. Typical features of Pgp including Walker A and Walker B motifs, ABC transporter signatures and the 12 transmembrane domains are highlighted. Amino acids common to the four sequences are shaded in dark blue. Amino acids conserved between Hco-PGP-3, Cel-PGP-3, Cel-PGP-4 but not Hco-PGP-16 are shaded in light blue. Amino acids conserved between Hco-PGP-16, Cel-PGP-3, Cel-PGP-4 but not Hco-PGP-3 are shaded in red. (TIF)