Exosomes from Plasmodium yoelii-Infected Reticulocytes Protect Mice from Lethal Infections

Exosomes are 30–100-nm membrane vesicles of endocytic origin that are released after the fusion of multivesicular bodies (MVBs) with the plasma membrane. While initial studies suggested that the role of exosomes was limited to the removal of proteins during the maturation of reticulocytes to erythrocytes, recent studies indicate that they are produced by different types of cells and are involved in promoting inter-cellular communication and antigen presentation. Here, we describe the isolation and characterization of exosomes from peripheral blood of BALB/c mice infected with the reticulocyte-prone non-lethal Plasmodium yoelii 17X strain. Importantly, proteomic analysis revealed the presence of parasite proteins in these vesicles. Moreover, immunization of mice with purified exosomes elicited IgG antibodies capable of recognizing P. yoelii-infected red blood cells. Furthermore, lethal challenge of immunized mice with the normocyte-prone lethal P. yoelii 17XL strain caused a significant attenuation in the course of parasitaemia, increased survival time, and altered the cell tropism to reticulocytes. These results were obtained also when the exosomes were isolated from a P. yoelii-infected reticulocyte culture indicating that reticulocyte-derived exosomes carry antigens and are involved in immune modulation. Moreover, inclusion of CpG ODN 1826 in exosome immunizations elicited IgG2a and IgG2b antibodies and promoted survival, clearance of parasites and subsequent sterile protection of 83% of the animals challenged with P. yoelli 17XL. To our knowledge, this is the first report of immune responses elicited by exosomes derived from reticulocytes opening new avenues for the modulation of anti-malaria responses.


Introduction
Exosomes are 30-100-nm membrane vesicles formed by endocytosis of segments of the plasma membrane. The internalized segment generates multivesicular bodies (MVBs) containing small vesicles that are released as exosomes following fusion of the MVBs with the plasma membrane [1]. Initially described as a reticulocyte cargo-disposal mechanism for maturation to erythrocytes [2], exosomes are now known to be also secreted by many different types of cells, including dendritic cells, macrophages, B cells, and tumor cells [1,3,4]. Pioneering studies with exosomes secreted by Epstein-Barr virus (EBV)-transformed B-cells demonstrated stimulation of T-cell in an antigen-specific manner [5]. Since, the role of exosomes in antigen presentation and immune modulation has been amply demonstrated in different infections including Salmonella, Mycobacterium, and Toxoplasma where exosomes released by infected cells contain microbial proteins [1,6]. The role of exosomes in antigen presentation in malaria, however, has not been described.
In addition to exosomes, other vesicles termed microparticles (MPs) circulate in blood [7]. MPs should not be confounded with exosomes as MPs originate by budding or shedding from the plasma membrane as opposed to fusion of the MVBs with the plasma membrane. Moreover, MPs are heterogeneous in shape and bigger in size (100-1000 nm) and present different protein composition [8,9]. Of note, MPs have been described associated with malaria pathology both in human and rodent models [10]. Indeed, production of MPs by parasitized red blood cells with implications in malaria immune responses and inflammation has been recently described [11,12].
Exosomes were originally described in reticulocytes where they allow remodeling of the plasma membrane in the maturation to erythrocytes by eliminating specific proteins [2]. Remarkably, reticulocytes are the cells preferentially, if not exclusively, invaded by different malaria parasites such as Plasmodium vivax [13] and the reticulocyte-prone non-lethal Plasmodium yoelii 17X strain [14]. We thus hypothesized that reticulocyte-derived exosomes (rex) in such infections, in addition to their role as cargo-disposable machinery, should contain parasite proteins involved in immune modulation. Here, we describe the presence of parasite proteins in exosomes obtained from experimental infections of BALB/c mice infected with P. yoelii 17X (exPy). Moreover, we show that immunization of mice with exPy results in modulation of the immune response to P. yoelii infection. In addition, when combined with cytosinephosphate-guanosine oligodeoxynucleotides (CpG-ODN), immunization of mice with rexPy elicited IgG2a and IgG2b antibodies and protected mice from a lethal challenge. This is the first report of a role in immune modulation of reticulocyte-derived exosomes opening new avenues in the modulation of anti-malarial responses.

Results
Peripheral blood exosomes from P. yoelii 17X infections contain parasite proteins To determine whether exosomes derived from malaria-infected mice contain parasite proteins, we purified exosomes from peripheral blood of BALB/c mice infected with P. yoelii 17X at approximately two weeks post-infection (p.i.), when reticulocytosis reached 60-90%. Electron microscopy (EM) analysis revealed a homogeneous population of vesicles whose cup-shape ( Figure 1A) and size (median of 56.8 nm diameter) ( Figure 1B) were consistent with previous descriptions of exosomes [15]. To better characterize these vesicles, they were coated on latex beads, stained with a panel of FITC-labeled or PE-labeled antibodies and analyzed by flow cytometry. Of note, vesicles isolated both by sucrose cushion ( Figure 1C) as well as by ultracentrifugation/ filtration ( Figure 1D) display similar staining profiles for the proteins analyzed. CD107a (Lamp1) was found in vesicles isolated from blood of infected mice indicating that they were originated from an internal compartment and were not plasma membrane fragments. The presence of Lamp1 together with no Exosomes isolated from blood of infected mice were coated on latex beads, labeled with antibodies and detected by FACS analysis. (C) Histograms of representative marker of exosomes (Lamp1) and marker enriched in reticulocyte-derived exosomes (Tfrc) isolated by ultracentrifugation and sucrosegradient purification. (D) Representative staining markers of exosomes (Lamp1), reticulocyte-derived exosomes (Tfrc, Itga4) and exosomes derived from T-cells (CD3f), APCs (MHC II, CD86), and platelets (CD41) performed on exosomes purified by ultracentrifugation and filtration. Histogram for the CD40l (microparticles) and CD133 (membrane particle) are also shown. Controls (tinted histograms) were performed with isotype controls or directly coupled secondary Abs. doi:10.1371/journal.pone.0026588.g001 presence of the microvesicle marker CD40l or membrane particle marker CD133 [8] indicates that our preparation of vesicles are mainly exosomes ( Figure 1D). To determine whether the exosomes can come from different origins, we analyzed the presence of cell-specific molecules. Exosomes from infected blood showed the presence of CD71 (Tfrc) and Itga4 (integrin a4), molecules that were previously described in reticulocyte-derived exosomes [16]. Moreover, low staining for Antigen presenting cell (APC) markers class II MHC and CD86 and the platelet-specific marker CD41 (GpIIb) [17,18] were also found on exosomes obtained from plasma. In addition, we could not find staining for the T-cell marker CD3f [17,19], further indicating that vesicles obtained from these BALB/c-P. yoelii 17X infections are exosomes mostly originating from reticulocytes.
Next, we determined the proteome of exosomes by liquid chromatography-tandem mass spectrometry (LC-MS/MS). Re-sults confirmed that exosomes purified from plasma of BALB/c mice infected with P. yoelii 17X contain many of the proteins most frequently found in exosomes such as Hsp8/Hsp90, Actb, Eno1, Anxa2, Ywhaz, Msn, Cfl1 and Actg1 [18] (Tables S1,S2). Moreover, a large number of proteins previously described in rex (Hsc70, CD47, aquoporin1 and Tfrc) [20][21][22] were also identified in our proteomic analysis. In contrast, no markers of MPs (CD40l) or apoptotic vesicles (Histones) [23] were found in the 100,0006g pellet. Remarkably, proteomic analysis revealed the presence of several P. yoelii antigens in exosomes (Table 1, S2). These parasite antigens included serine-repeat antigens, merozoite surface proteins 1 and 9, enzymes, proteases, heat-shock proteins, and hypothetical proteins. Altogether, these results showed that peripheral blood exosomes obtained from infections of BALB/c mice with P. yoelii 17X contain parasite proteins likely derived from infected reticulocytes. Immunization with peripheral blood exosomes from P. yoelii 17X infections elicits parasite-specific immune humoral responses and modulates the course of infection upon lethal challenge To investigate whether peripheral blood exosomes derived from Plasmodium-infected mice could induce parasite-specific immune responses, we immunized mice with exosomes obtained from mice infected with the non-lethal P. yoelii 17X strain (exPy) and noninfected animals (exC) at days 0 and 20. Twenty days after the second immunization, sera were tested by immunofluorescence (IFA) for the presence of anti-P. yoelii antibodies. As shown in Figure 2A, immunizations with exPy elicited IgG antibodies capable of specifically recognizing infected RBCs. To further evaluate these immune responses, mice immunized with exC and exPy were challenged twenty days later with 5610 5 -10 6 P. yoelii 17XL lethal parasites. Daily examination of the animals and blood smears by Giemsa staining revealed an attenuation in the course of parasitaemia ( Figure 2B) and higher survival times ( Figure 2C) only in mice immunized with exPy. In addition, these mice had a significant increase of reticulocytosis (from 1-2% up to 65%) as well as a change in the parasite's cell tropism to reticulocytes ( Figure 2D). This change in cell tropism does not appear to be simply the result of an increase in reticulocytosis as the percentage of circulating normocytes did not fall below 45%; yet, close to 90% of infected cells were reticulocytes. More importantly, these results demonstrated that peripheral blood exosomes in this rodent malaria model induced immune responses which modulated the course of infection.
Reticulocyte-derived exosomes from infections are responsible for immune modulation, reticulocytosis and change of cell-tropism Exosomes obtained from the blood of BALB/c mice infected with P. yoelii 17X are likely to be originated from reticulocytes; yet, exosomes from other cells capable of antigen presentation, such as dendritic cells could contribute to the final pool of exosomes. To exclude this possibility, isolated reticulocytes from blood of P. yoeliiinfected mice were cultured for 24 hours and exosomes purified from supernatants by ultracentrifugation and filtration. Analysis of vesicles released by these cells by EM confirmed their purity and Figure 2. Exosomes from blood of P. yoelii 17X-infected mice modulate the course of P. yoelii 17XL infection. (A) Humoral IgG response after exosomal immunization of BALB/c mice. Bright-field and fluorescence image projections of immune sera recognizing 17XL-pRBCs. The images correspond to sera of representative mice. One non-immunized (NI) mouse, one mouse immunized with exosomes from control animals (exC) and two mice immunized with exosomes from mice infected with the P. yoelii17X non-lethal strain (exPy). Staining for mouse IgG (green) and DNA (blue) is shown. Scale bar represents 5 mm. (B) Time-course parasitaemia (mean6SD), (C) survival curve, (D) reticulocytosis (right axis, dotted line) and percentages of infected reticulocytes (left axis, continuous line) after 5610 5 -10 6 P. yoelii 17XL infections of groups of BALB/c mice previously immunized with exosomes from control animals (exC) (n = 6) and exosomes from mice infected with the P. yoelii 17X non-lethal strain (exPy) (n = 6). Non-immunized (NI) mice (n = 4) were untreated. Data correspond to 3 independent experiments. (B) P. yoelii 17X mean peak level of parasitemia is 50.3%61.6% for NI, 51.77%61.02% for exC and 39.14%62.17% for exPy. The reduction was evaluated by analysis of variance and was statistically significant (P,0.001). (C) Differences in the survival curves between NI and exPy (P,0.05) are statistically significant (Log-rank (Mantel-Cox Test)). doi:10.1371/journal.pone.0026588.g002 revealed a size and morphology compatible with exosomes ( Figure 3A). Moreover, flow cytometry analysis of rex bound to aldehyde/sulfate beads revealed the expression in the surface of Lamp1, Tfrc, and Itga4 markers but not of MHC II, CD3f or CD41, confirming their origin from reticulocytes ( Figure 3B). In vitro-purified rex from infections (rexPy) or from anemic mice (rexC) [21] were then used to immunize new groups of mice at 20day intervals with two i.v. doses of 5 mg of rexPy or rexC. Nonimmunized mice (NI) were untreated. Twenty days after the second immunization, all mice were infected with 5610 5 P. yoelii 17XL lethal parasites. Noticeably, we observed the same results as those observed in immunizations with exosomes from peripheral blood, namely higher survival times, increase in reticulocytosis (from 1-2% up to 60%) and infection preference for reticulocytes in mice immunized with exPy ( Figure S1). Thus, rex from infections contain parasite proteins and are involved in modulation of immune responses upon infection.
Immunization with rex combined with CpG-ODN elicits protective immune responses CpG oligodeoxynucleotide (CPG-ODN) is a potent immunostimulator proving beneficial in protection against rodent malaria [24]. We thus tested the immune enhancement effect of CpG-ODN on immunization with rex. To do so, BALB/c mice were immunized subcutaneously (s.c.) with 10 mg of rexC or rexPy combined with 10 mg CpG-ODN. Twenty days after the first immunization, mice were boosted s.c. with 5 mg of exosomes and forty days later sera from immunized mice were tested by IFA and ELISA for the presence of antibodies against P. yoelii. As shown in Figure 4, immunizations with rexPy, but not with rexC or nonimmunized mice, elicited IgG antibodies recognizing infected RBCs ( Figure 4A). In addition, ELISA analysis demonstrated that elicited antibodies were predominantly of the IgG2a and IgG2b isotypes with no detected differences in the levels of IgG3 and IgM ( Figure 4C).
Next, immunization with rexPy in combination with CpG-ODN was tested for the ability to protect mice against a lethal challenge with P. yoelii 17XL. Significantly, 83% of the mice immunized with rexPy survived to a lethal infection ( Figure 5A) showing concomitantly reticulocytosis ( Figure 5D) and a change of cell-tropism to reticulocytes ( Figure 5C). Of importance, mice which survived the primary infection showed a complete clearance of infection ( Figure 5B) and remained immunoprotected, showing no detectable parasitaemia by Giemsa smears, after re-challenges with P. yoelii 17XL at days 74, 116, and 156 post first infection.

Discussion
Reticulocyte-derived exosomes function as a cargo-disposal machinery of unwanted proteins and molecules in the maturation of reticulocytes to erythrocytes [2]. Noticeably, malaria parasites such as P. vivax and the P. yoelii 17X strain predominantly, if not exclusively, infect reticulocytes. Here, we describe for what we believe is the first time, that rex from experimental infection of BALB/c mice with the P. yoelii 17X strain contain parasite proteins and that in addition to their cargo-disposable role in differentiation to erythrocytes, rex have a role in modulating immune responses.
The function of rex has been less explored than the one of exosomes derived from antigen-presenting cells, B cells, or tumor cells. In fact, different studies have identified proteins associated with these vesicles [25] and this has been recently confirmed and expanded in a proteomic analysis of rex from rats [26]. Our proteomic data from exosomes isolated from blood of P. yoelii 17Xinfected mice revealed the presence of proteins previously described in rex such as Hsc70, CD47, aquoporin1 and Tfrc [20,22]. Moreover, we identified the Na + -independent neutral aminoacid transporter LAT-1 [27], a function that was originally described in rex [28]. Remarkably, our proteomic data also revealed the presence of 31 parasite proteins within or associated with exosomes obtained from reticulocyte-enriched blood of infected animals. These proteins include several antigens such as merozoite surface proteins 1 and 9, as well as blood stage surface antigens, a rhoptry-associated protein and a variant Yir protein. In addition, we identified 6 hypothetical proteins, 3 of them having GO annotations related to binding (PY06307), pathogenesis (PY06423), and immune response/antigen processing presentation (PY05295). Moreover, we found 9 enzymes related to proteolysis and metabolic processes. The function of these enzymes is yet to be determined but it is worth recalling that some of them seem to have moonlight functions in malaria [29][30][31][32]. In fact, we identified enolase (PY06644) which has been found in different locations suggesting non-glycolytic functions of this enzyme [33] and which upon immunization as a recombinant protein partially protects mice against P. yoelii [33]. While proteomic analysis of rexPy, per se, remains to be determined, this data unequivocally demonstrates the presence of malaria proteins in exosomes derived from peripheral blood in this reticulocyte-prone rodent malaria model.
Peripheral blood contains other type of vesicles termed microparticles (MPs) which have been described in malaria infections both in humans and animal models. Thus, MPs from platelets bind to erythrocytes transferring platelet antigens which are postulated to facilitate cytoadherence to brain tissues leading to cerebral malaria [34]. In addition, MPs have also been recently described in human falciparum and vivax infections [12,35] and have been associated with acute vivax malaria symptoms [35]. Moreover, in rodent models, MPs derived from malaria infected RBCs have been recently shown to lead to inflammation through macrophage activation [11]. Noticeably, if shedding of these vesicles occurs during P. yoelii 17X infections, they did not contribute to our results since MPs were discarded during our exosome preparation (pellet after 15,0006g centrifugation, see methods). Moreover, flow cytometry revealed no significant amounts of CD40l, a MP marker, in our vesicle preparations. In addition, EM analysis allowed us to check for purity of our isolated exosomes thus discarding any preparation containing contamination with particles that differ from the cup-shape and size reported for exosomes.
Immunizations of BALB/c mice with exosomes derived from peripheral blood or from rex purified from in vitro cultures from infections, elicited IgG antibodies capable of recognizing P. yoeliiinfected RBCs, induced reticulocytosis and changed the celltropism to reticulocytes of the normocyte-prone lethal P. yoelii 17XL strain upon infection. Moreover, when combined with CpG-ODN, immunizations with rexPy obtained from reticulocyte in vitro culture elicited the same responses as above in addition to conferring complete and long lasting protection in close to 85% of the mice tested. Of note, antibody responses in this lethal strain of P. yoelii are mainly suppressed during infections [36]; yet, these immunizations elicited IgG2a antibodies capable of recognizing P. yoelii-infected RBCs. This humoral response is likely associated with protection since it has been described that passive transfer of the IgG2a fraction of hyperimmune plasma can modulate parasitaemias of BALB/c mice infected with P. yoelii 17XL [37]. Remarkably, reticulocytosis, change of cell-tropism and protection mediated by IgG2a antibodies were reported upon immunizations of BALB/c mice with the P. yoelii 235-kDA antigen [38] and with the P. yoelii MSP8 protein [39]. Moreover, in this latter study, genome-wide global transcriptional analysis revealed the presence of up-regulated genes coding for parasite antigens also identified in our proteomic analysis such as MSP1, MSP9, rhopthry-associated proteins and blood stage surface antigens (pAgI and pAg3) [39]. Protection in all these reports seems depending on elicitation of IgG2a antibodies that putatively block entrance into mature red blood cells and which, upon infections with the lethal strain, would favor the survival of parasites entering reticulocytes. In turn, it is tempting to speculate that infected reticulocytes should release rex containing parasite proteins which act in inter-cellular communication and antigen presentation, thus facilitating reticulocytosis and the mounting of a protective and long-lasting immune response. In the absence of supporting experimental evidence, this remains to be demonstrated.
To our knowledge, the results presented here provide the first description of exosomes, not MPs, containing malaria proteins. Moreover, our study also shows that, when combined with CpG-ODN, rex from infections have the capacity of elicitation of protective and long-lasting immune responses against lethal infections in this rodent malaria model. While further experimentation is required to determine the exact role of rex during infections, rex represent a new player in immune response to malaria that could also be used as a vaccine tool against infections.

Ethics statement
All the animal studies were performed at the animal facilities of Hospital Clinic in Barcelona in accordance with guidelines and protocols approved by the Ethics Committee for Animal Experimentation of the University of Barcelona CEEA-UB.

Mice and parasites
Female BALB/c mice, 6 to 8 weeks of age, were used throughout the study. The P. yoelii 17XL (lethal) and 17X (nonlethal) cloned lines were originally obtained from MR4 (http:// www.mr4.org/). Infections were initiated by intraperitoneal (i.p.) injection of 10 5 -10 6 pRBCs from the tail blood of donor mice at 5-10% parasitaemia. Parasitaemia was monitored by Giemsa staining of blood smears calculating the percentage of pRBCs over total RBCs in three optical fields of approximately 300 RBCs.
Reticulocytosis was induced by phlebotomy as previously described [21]. Blood was collected by cardiac puncture in anesthetized mice. Brilliant Cresyl Blue (BCB)-Giemsa staining was used to detect reticulocytes [40]. Briefly, five ml of mouse blood was mixed with an equal volume of 1% BCB in 0.65% sodium chloride and incubated at room temperature (RT) for 20 min. After incubation, smears were prepared and left to air dry. After fixation with methanol, the smears were counterstained with Giemsa. Percentage of reticulocytes over total RBCs was counted in different optical fields with, at least, 1000 RBCs.

Purification of reticulocytes
Reticulocytes were obtained from mice blood collected in EDTA. We have used blood of BALB/c anemic mice and of mice infected with P. yoelii 17X strain at 20-30% parasitaemia. To isolate the reticulocytes, blood was centrifuged at 10006g for 20 min. After two washes in phosphate-buffered saline (PBS), reticulocytes were purified by layering them on top of a Percoll/ NaCl density gradient as previously described [21,41]. Briefly, 5 ml of 1.096 g/mL of density Percoll solution was placed in 15 mL tubes and 2 mL of 1.058 g/mL Percoll were layered over. Thereafter, 2 mL of blood diluted in PBS to a final hematocrit of 50% were layered on top of each tube. Tubes were centrifuged at 2506g for 30 minutes at 4uC. Reticulocytes were collected from the interface of the two Percoll layers and washed twice with PBS. We routinely obtained 763610 7 reticulocytes per uninfected mice and 3.660.6610 8 reticulocytes per P. yoelii 17X-infected mice. Purified reticulocytes were cultured for 24 h at 37uC in DMEM, supplemented with 5 mM glutamine, 5 mM adenosine, 5 mM inosine, 5% fetal calf serum, 50 U/mL penicillin, and 50 mg/mL streptomycin at 1-3% hematocrit. To remove exogenous exosomes in the culture medium, the fetal calf serum was precentrifuged (100,0006g overnight) (o.n.).

Exosomes
Exosomes were purified from the plasma of uninfected mice or mice infected with P. yoelii 17X by sequential centrifugations at 5006g for 30 min, 15,0006g for 45 min and 100,0006g for 2 h at 4uC (adapted from [17]). The final pellet was resuspended in PBS, filtered through a 0.22-mm membrane, and centrifuged at 100,0006g for 2 h at 4uC. The pellets were resuspended in PBS, and the protein content was determined by Bradford assay. Usually, we are able to obtain the equivalent of 5-15 mg of proteins in exosomes per ml of blood (7.862.4/uninfected mice, 12.761.9/P. yoelii 17X mice). Routinely, exosomes purity was assessed by ME analysis.
Exosomes from supernatant of culture were obtained with the same differential centrifugation method.
For some applications, exosomes were loaded in a 30% sucrose cushion (adapted from [15]). Briefly, exosome pellet from first centrifugation at 100,0006g was resuspended in 6 ml PBS and loaded on top of 1 ml sucrose/D 2 O solution (20 mM Tris, 30% sucrose, pH 7.4). Samples were centrifuged in a swinging bucket rotor at 100,0006g for 75 min. Approximately 0.75 ml of the cushion containing exosomes was collected, diluted with PBS and centrifuged at 100,0006g for 90 min. Washed exosomes were then resuspended in PBS and protein content was measured using Bradford assay.

Electron microscopy
Electron microscope analysis was performed with exosomes fixed in 2% paraformaldehyde (PFA) o.n. at 4uC. Fixed exosomes were deposited onto formvar grids for 20 min. Grids were then fixed in glutaraldehyde 1% for 5 min, washed in distilled water and negatively stained for 5 min with a solution of uranyl-oxalate (pH = 7) and for 10 min at 4uC with uranyl acetate(4%)-methyl cellulose (2%). After thoroughly drying, grids were observed with a JEOL 1010 transmission electron microscope. Micrographs were used to quantify the diameter of exosomes.

Flow cytometry
Five micrograms of purified exosomes were incubated with 10 ml of 4 mm-diameter aldehyde/sulfate beads (4-mm beads, Invitrogen) for 15 min at room temperature (RT). PBS was added to a final volume of 1 ml and exosome-beads were incubated with rotation overnight at 4uC. After adsorbing exosomes, beads were subjected to glycine (100 mM final for 30 min at RT), followed by the addition of 0.5% BSA in PBS to saturate any remaining free binding sites on the beads. Coated beads were incubated with different antibodies (some directly conjugated to fluorescent dyes) diluted in PBS/0.5% BSA 30 min at 4uC. After washing, samples were incubated as needed with fluorescent-labeled secondary antibodies diluted in PBS/0.5% BSA for 30 min at 4uC. Samples were analyzed by flow cytometry using a FACSCanto instrument. A minimum of 5610 5 single beads per sample was examined.

Proteomic analysis of exosomes
Exosomal preparations were dissolved in 8 M urea 0.4 M NH 4 HCO 3 , reduced with 5 mM dithiothreitol for 15 min at 50uC, alkylated with 10 mM iodoacetamide for 30 min, at room temperature, and the reaction was diluted to obtain a final concentration of 1 M urea. After digestion with trypsin for 16 h, samples were desalted using POROS R2 ziptips [42], redissolved in 0.1% formic acid (FA), loaded in a strong-cation exchange trap column (5 mL, Optimized Technologies) and eluted by injecting increasing concentrations of NaCl (0, 25, 50, 100, 200, and 500 mM NaCl in 0.5% FA/2% acetonitrile (ACN)) using the autosampler, which is part of the 1D Plus nanoHPLC system (Eksigent). Eluted peptides were trapped and washed (0.5% FA in 2% ACN) in a laboratory-made C18 column (1 cm, 75 mm, Phenomenex Luna C18, 5 mm); separation was achieved in a capillary C18 column (20 cm, 75 mm, Phenomenex Luna C18, 5 mm), using a linear acetonitrile (ACN) gradient of 5-40% B solvent over 100 min (solvent A: 5% ACN/0.1% FA, solvent B: 80% ACN/0.1% FA). Peptides were analyzed online on a LTQ XL/ETD (Thermo Fisher Scientific) mass spectrometer equipped with a Triversa Nanomate (Advion) nanospray source. Full scan MS spectra were collected from 400-1700 m/z range and the ten most intense ions were selected for fragmentation twice, before dynamically excluding for 1 min. MS/MS spectra were converted to DTA files and submitted to database search using Sequest. The search parameters were: (i) full tryptic digestion with 1 missed cleavage site allowed; (ii) 2 Da of peptide mass tolerance; and (iii) methionine oxidation and cysteine carbamidomethylation as variable and static modifications, respectively. The database (129,706 total sequences) was built with the forward and reverse sequences of P. yoelii v5.5 (http://plasmodb.org), in addition to possible murine (IPI mouse v3.58) and human (keratins derived from IPI human v3.58) contaminant sequences downloaded from http://www.ebi.ac.uk/IPI/IPIhelp.html. Data were filtered with DCn$0.05, peptide probability #0.05 and Xcorr $1.5, 2.4 and 3.2 for 1+, 2+, and 3+ charged peptides, respectively. After assembling the peptides into proteins and redundant proteins into protein groups, the dataset was further filtered with Xcorr sum (sum of Xcorr of unique peptides for each protein) $3.5 to obtain a false-discovery rate of 0.8% at protein level.

Immunizations and challenge
For immunizations, mice were injected intravenously with two doses of 5 mg of exosomes obtained by ultracentrifugation and filtration and resuspended in PBS at days 0 and 20. Nonimmunized mice were untreated. Endotoxin level of exosomes was ,0.005EU/mg (determined by kinetic chromogenic LAL assay, Genescript). Immunized animals were challenged twenty days after the second immunization with 5610 5 -10 6 P. yoelii 17XL parasites, and parasitaemia and reticulocytosis were followed daily using Giemsa-stained blood smears.
For CpG immunizations, mice were injected subcutaneouslly (s.c.) with 10 mg of exosomes and 10 mg CpG ODN-1826. Endotoxin level of exosomes was ,0.005EU/mg (determined by kinetic chromogenic LAL assay, Genescript). Twenty days after, mice were immunized with 5 mg of exosomes. Twenty days after the second immunization, mice were infected with 5610 5 P. yoelii 17XL. Surviving animals were re-infected three times with P. yoelii 17XL at days 74, 116 and 156 after the first infection. Parasitaemia was followed using Giemsa-stained blood smears.

Antibody responses
Humoral responses elicited by exosomes were studied by immunofluorescence and ELISA using mouse sera collected 20 days after immunizations and tested for parasite recognition. IFA assays were performed on P. yoelii 17XL-infected blood smears fixed with cold methanol for 2 min and air-dried before blocking with 5% BSA/PBS for 30 min at room temperature (RT). Slides were incubated with mouse sera diluted 1/10 in 0.5% BSA/PBS o.n. at 4uC followed by incubation for 1 h at RT. Reactive IgGs were detected using an anti-mouse IgG antibody conjugated to Alexa Fluor 488 (Invitrogen) at 1:200 dilution for 1 h at RT. After washing, nuclei were stained with DAPI (Invitrogen, 5 mg/mL) for 7 min at RT. Sera recognizing pRBCs were imaged for bright field and fluorescence using a Leica SP5 microscope fitted with an inverted 636oil objective. Sera from control non-immunized mice and mice immunized with exosomes from uninfected animals were used as negative controls.
Serum antibody levels were determined by an enzyme-linked immunosorbent assay (ELISA) against P. yoelii antigen, with pooled sera (n = 6 mice per group). MaxiSorb immunoplates (Nunc) were coated with 200 ng of parasite antigen, obtained as described [43], in carbonate-bicarbonate buffer pH 9.6, overnight at 4uC. Wells were blocked with 100 ml 5% BSA in coating buffer at 37uC for 3 h. After four washes with 0.05% Tween in PBS, two-fold serial dilutions of serum samples were added in 1% BSA in PBS. After incubation for 1 h at 37uC, wells were washed and incubated with peroxidase-conjugated anti-mouse IgG. For isotype analysis bound antibodies were detected with peroxidase conjugated anti-mouse (IgG1, IgG2a, IgG2b, IgG3, or IgM) antibodies (Abcam). After incubation for 1 h, wells were washed and incubated with 50 ml TMB substrate (Pierce) for 15 min. The reaction was stopped with 50 ml 1N sulfuric acid per plate well. Absorbance was measured at 450 nm. Figure S1 Immunization with exosomes from 17Xinfected reticulocytes modulated the course of P. yoelii 17XL infection. (A) Survival curve, (B) time-course parasitaemia (mean6SD), (C) percentages of infected reticulocytes and (D) reticulocytosis after 5610 5 P. yoelii 17XL infections of groups of BALB/c mice previously immunized intravenously (i.v.) with 5 mg exosomes from control reticulocytes (rexC) (n = 5) or 5 mg of exosomes from reticulocytes infected with the P. yoelii 17X nonlethal strain (rexPy) (n = 5). Non-immunized (NI) mice (n = 5) were untreated. Data correspond to 2 independent experiments. (A) Differences in the survival curves between NI and rexPy (P,0.05) are statistically significant (Log-rank (Mantel-Cox Test). (TIF)