Transcriptome Profiles of the Protoscoleces of Echinococcus granulosus Reveal that Excretory-Secretory Products Are Essential to Metabolic Adaptation

Background Cystic hydatid disease (CHD) is caused by the larval stages of the cestode and affects humans and domestic animals worldwide. Protoscoleces (PSCs) are one component of the larval stages that can interact with both definitive and intermediate hosts. Previous genomic and transcriptomic data have provided an overall snapshot of the genomics of the growth and development of this parasite. However, our understanding of how PSCs subvert the immune response of hosts and maintains metabolic adaptation remains unclear. In this study, we used Roche 454 sequencing technology and in silico secretome analysis to explore the transcriptome profiles of the PSCs from E. granulosus and elucidate the potential functions of the excretory-secretory proteins (ESPs) released by the parasite. Methodology/Principal Findings A large number of nonredundant sequences as unigenes were generated (26,514), of which 22,910 (86.4%) were mapped to the newly published E. granulosus genome and 17,705 (66.8%) were distributed within the coding sequence (CDS) regions. Of the 2,280 ESPs predicted from the transcriptome, 138 ESPs were inferred to be involved in the metabolism of carbohydrates, while 124 ESPs were inferred to be involved in the metabolism of protein. Eleven ESPs were identified as intracellular enzymes that regulate glycolysis/gluconeogenesis (GL/GN) pathways, while a further 44 antigenic proteins, 25 molecular chaperones and four proteases were highly represented. Many proteins were also found to be significantly enriched in development-related signaling pathways, such as the TGF-β receptor pathways and insulin pathways. Conclusions/Significance This study provides valuable information on the metabolic adaptation of parasites to their hosts that can be used to aid the development of novel intervention targets for hydatid treatment and control.


Introduction
Cystic hydatid disease (CHD) is a serious parasitic zoonosis that is caused by the larval stages of Echinococcus granulosus, a cestode that poses a threat to public health as well as significant economic losses [1,2,3]. At present, more than 3 million people are infected with this parasite [4,5], and the prevalence reaches 10% in some areas [6,7]. The disease is difficult to control because appropriate diagnostic procedures are lacking and the available drugs are inefficient [8].
E. granulosus has a complex developmental cycle, involving eggs, oncospheres, protoscoleces (PSCs), and adult stages. Adult parasites live in the small intestine of dogs. After sexual maturation, numerous eggs are produced by the adult parasites and are then excreted with the dog feces. Infections occur in an intermediate host, when eggs containing larvae are ingested. Hydatid cysts (the larval stage or metacestode) develop in the internal organs (primarily in liver and lungs) of intermediate hosts.
The larval stages of E. granulosus are comprised of two layers of cyst wall: cyst fluid and PSCs [9].
As the only infectious form of the larval stages, PSCs can interact with both definitive and intermediate hosts. They mature into adult parasites when the hydatid cysts are ingested by the definitive host. They can also differentiate into new cysts when released into the body cavity of intermediate hosts upon cyst rupture [10]. Mouse models of CHD are often established via the intraperitoneal inoculation with PSCs, a method that has been widely applied to drug screening and vaccine development [11,12]. Overall, the PSC is an important infectious reagent that contributes to the transmission of CHD and also an excellent model system in which many aspects of the host-parasite interaction can be studied.
Understanding the elaborate immune evasion strategies and mechanisms of physiological adaptation of the PSCs is critical to ascertain effective intervention targets to control the prevalence of the parasite. In this study, we focus on the role of excretorysecretory products (ESPs) that are released by parasites, as these compounds are exposed directly to the immune system of the hosts and are engaged at the host-parasite interface [13]. The mechanism by which PSCs can subvert the immune environment via ESPs is the key to successful infection. Recently, we found that ESPs from adult E. granulosus could downregulate host immune responses by preventing dendritic cells (DC) from maturing, by impairing DC function and by inducing the generation of CD4 + CD25 + FoxP3 + T cells (unpublished data). Previous studies have shown that cystic fluids produced in the intermediate hosts can modulate DC differentiation and cytokine secretion [14], while antigen B released by the germinal cells of E. granulosus can direct immature DCs towards the maturation of a Th2 cell response [15]. Moreover, the ESPs from E. multilocularis larvae have been found to induce apoptosis and tolerogenic properties in DC in vitro [16]. To date, studies have focused primarily on the immune regulation of ESPs by the host, with little work undertaken to investigate the influence of ESPs on the physiological adaptation of parasites to their hosts. Interestingly, several intracellular proteins that were not previously thought to be exposed to the immune system of hosts have recently been identified in the ESPs of PSCs [9,17]. This finding suggests that parasite-derived ESPs are incorporated in the metabolites of the host [18,19].
Further investigations into the mechanisms of physiological adaptation of ESPs released by PSC have been hampered due to the paucity of information regarding ESPs. Although studies have utilized proteomics to identify the constituents of ESPs [9,[20][21][22], very few have been identified. This is largely because of interference from host proteins [20][21] and because of technical limitations of the methodologies used. In recent years, however, the combination of transcriptomics and proteomics has enabled the identification of an increasing number of parasitic proteins [23,24].
In this study, we used Roche 454 sequencing technology and in silico secretome analysis to explore the transcriptome profiles of E. granulosus PSCs and to elucidate the potential functions of the ESPs released by the parasite.

Ethics statement
This study was performed in strict accordance with the recommendations provided in the Guide for the Care and Use of Laboratory

Sample collection
Hydatid cysts were collected from the livers of a naturally infected sheep in a slaughterhouse in Qinghai, China. Cyst fluids containing PSCs were sucked out of the cysts using a sterile syringe. After natural sedimentation for 10 min, PSCs were carefully collected from the sediment of cyst fluids and washed 10 times with saline solution. We then added 2 mL of Trizol reagent (Invitrogen, USA) to the well-washed PSCs. After continuous mixing with a pipette, the PSCs were stored at 280uC prior to use.

Genotyping the PSCs
Genomic DNA from the PSCs was extracted using the DNeasy tissue kit (Qiagen, Hilden, Germany) and used as a template for a polymerase chain reaction (PCR) [25]. The following two primer pairs were used to amplify the mitochondrial genes of Echinococcus species: cytochrome coxidase subunit 1 (cox1) gene (F: 59-TTGAATTTGCCACGTTTGAATGC-39; and R: 59-GAACC-TAACGACATAACATAATGA-39) and cytochrome b (cytb) gene (F: 59-GTCAGATGTCTTATTGGGCTGC-39; R: 59-TCT-GGGTGACACCCACCTAAATA-39). Each 25-mL reaction mixture contained 1 mL of template DNA, 12.5 mL Premix TaqH mix (TaKaRa Biomedicals, Tokyo, Japan), l mL of 10 mM of each primer, and 9.5 mL nuclease-free water. The procedure of PCR amplification consisted of 94uC for 1 min, 30 cycles of 94uC for 30 s, 56uC for 30 s, and 72uC for 1 min, followed by 72uC for 10 min, with a final holding step at 4uC. The PCR products were directly sequenced with a Dye Terminator Cycle Sequencing Kit (Amersham Biosciences, Tokyo, Japan) and ABI 3730 DNA Analyzer (Applied Biosystems, Foster City, USA). cDNA library preparation, Roche 454 sequencing and sequence assembly The total RNA was extracted from the PSCs in TRIzol reagent, and RNA quality was performed by gel electrophoresis with a 2100 BioAnalyzer (Agilent Technology, Santa Clara, USA). The sequencing protocol followed that described in Liao et al. [26], and was carried out at the Shanghai OE Biotech Company. cDNA was synthesized using 2 mg of total RNA with the SMART cDNA synthesis kit (Clontech Laboratories, Mountain View, USA) according to the manufacturer's instructions. The cDNA library

Author Summary
The successful infection establishment of parasites depends on their ability to combat their host's immune system while maintaining metabolic adaptation to their hosts. The mechanisms of these processes are not well understood. We used the protoscoleces (PSCs) of E. granulosus as a model system to study this complex host-parasite interaction by investigating the role of excretory-secretory proteins (ESPs) in the physiological adaptation of the parasite. Using Roche 454 sequencing technology and in silico secretome analysis, we predicted 2280 ESPs and analyzed their biological functions. Our analysis of the bioinformatic data suggested that ESPs are integral to the metabolism of carbohydrates and proteins within the parasite and/or hosts. We also found that ESPs are involved in mediating the immune responses of hosts and function within key development-related signaling pathways. We found 11 intracellular enzymes, 25 molecular chaperones and four proteases that were highly represented in the ESPs, in addition to 44 antigenic proteins that showed promise as candidates for vaccine or serodiagnostic development purposes. These findings provide valuable information on the mechanisms of metabolic adaptation in parasites that will aid the development of novel hydatid treatment and control targets.
was constructed using a GS-FLX Titanium General Library Preparation Kit (Roche, Branford, USA) without normalization [27], and then sequenced using a half run on the Roche 454 GS-FLX Titanium platform. The modules built-in Newbler 2.5.3 (a de novo sequence assembly software, Roche, USA) was used to remove low quality sequences and assemble the remaining sequences. Briefly, the quality score trimming filter trims back from the 39 end of reads and was based on estimated quality scores (not the final quality scores) derived from an internal calibrated signal histogram. The error rate in a sliding window (default size of 40 bp) was calculated from the estimated quality scores and multiplied by an empirical scaling factor (default of 1.1). The window was moved leftwards until the estimated error rate in the window was ,1.0% (by default). If the resulting read was less than 40 bp (default), the read was discarded and not counted (numTrimmedTooShortQuality metric). After removing low quality sequences and sequencing adaptors, the remaining sequencing reads were assembled using the Newbler 2.5.3 with the 'extend low depth overlaps' parameter. All of the ESTs from the Roche 454 were used to run the final assembly. The resulting isotig consensus sequences and singletons were referred as 'unigenes' in the following study.

Bioinformatic analyses of transcriptomic sequence data
The software SOAP2 was used to map the raw sequence reads to the nonredundant sequence data [28]. Briefly, raw reads were aligned to the assembled, nonredundant transcriptomic data, to ensure that each read was mapped to a unique transcript. Reads mapped to more than one transcript were randomly assigned to one unique transcript, to ensure that they were recorded only once. Reads per kilobase per million reads (RPKM), the evaluation index of relative assessment of transcript abundance, was calculated using the standard formula [29].
After conceptual translation from the predicted coding domains of individual transcriptomic sequences, the functions of the potential proteins were predicted using InterProScan [35], employing the default parameters. According to their homology with conserved domains and with protein families, proteins inferred for E. granulosus PSC (EgPSC) were assigned to three gene ontology (GO) categories, including molecular function, cellular component and biological process [36]. The pathway analysis of inferred proteins was carried out using the KEGG (Kyoto Encyclopedia of Genes and Genomes) database [37].

In silico secretome analysis
Excretory-secretory proteins (ESPs) were predicted according to the methods described by Garg and Ranganathan [38,39]. Briefly, the secretory proteins were predicted utilizing the following five tools: ESTScan 3.0.3 [40] to translate the unigenes into putative proteins; SecretomeP 1.0 [41] for non-classical secreted proteins; SignalP 4.1 [42] for classical secreted proteins; TargetP 1.1 [43] for trimming mitochondrial proteins; and TMHMM 2.0 [44] for trimming transmembrane proteins. The predicted proteins with no transmembrane helices were thought to be ESPs.
All potential ESPs were blasted with known ESP sequences from E. granulosus (including nucleotide and protein sequences [9,7,[20][21][22] and our unpublished data) to validate the in silico secretome analysis. They were then annotated against GO, KEGG, Reactome (http://www.reactome.org/ReactomeGWT/entrypoint.htm1) and Panther (http://www.patherdb.org/) databases to identify functional groups and pathway annotations. Enrichment of KEGG pathways for genes with significant expression was calculated utilizing a classical hypergeometric distribution statistical comparison of the query gene list against all predicted E. granulosus genes. Caenorhaditis elegans pathways were used as a reference. Calculated P-values were subjected to FDR correction, with p,0.05 taken as the threshold for significance.

Genotyping of E. granulosus PSCs
The genotype of E. granulosus PSCs used in this study was sheep G1, as the PCR fragment amplified from cytb gene showed the highest identity (99%) to the E. granulosus G1 genotype referenced in GenBank (accession AF297617, S1 Figure). This was consistent with the fact that sheep G1 strain is the most common strain worldwide [60].

Roche 454 transcriptome sequencing and reads assembly
A total of 330,188 raw reads (mean length = 411.8 bp) were generated. The data is stored in Sequence Read Archive (SRA, No. SRP040541). After trimming to remove adaptors, low quality reads and polyN tail sequences, 329,927 clean reads remained (mean length = 400.3 bp; Table 1). Clean reads were assembled and produced about 26,514 unigenes ranging in size form 150-3,357 bp (mean = 501.5 bp). These included 4,175 isotigs ranging in size from 154 to 3,357 bp and 22,339 singletons of 150 to 1,710 bp. Approximately 84% of the isotigs were.500 bp, while most singletons (85.97%) were between 300 and 800 bp in size ( Table 1, S2 Figure). The numbers of EgPSCs unigenes matching known sequences are listed in Table 1. In summary, 26,514 unigenes were inferred from our transcriptome. The large majority of these (17,861, 67.4%) exhibited the highest level of homology to proteins in E. multilocularis, followed by proteins from E. granulosus (17,732; 66.9%), Caenorhabditis elegans (8,946; 33.7%) and S. mansoni (2,159; 17.5%). Moreover, 22,910 (86.4%) contigs were mapped to the E. granulosus genome and 17,705 (66.8%) of these were distributed within the coding sequence (CDS) region, which suggested that our results were reliable.

Potential secretome database
PSCs are an important, infectious component of the larval stages of E. granulosus that can interact with both definitive and intermediate hosts [10]. The adaptive mechanisms that facilitate this interaction between host and parasite is of great interest to our understanding of the transmission of this widespread disease. Preliminary investigations suggest that parasites secrete certain molecules to assist in host tissue colonization [13]. We therefore focused on the components of ESPs released by PSCs and their potential roles in the physiological adaptation to their hosts and/or themselves.
Of the 26,514 unigenes identified, 19,576 were translated into proteins by ESTScan, 437 proteins were predicted to be classical secreted proteins using SignalP, while 592 were predicted to be non-classical secreted proteins according to SecretomeP. The classical and non-classical proteins were then analyzed using TargetP software for mitochondrial proteins, which resulted in the removal of 25 proteins. A further 123 transmembrane proteins were removed from the secretory protein dataset by TMHMM. In total, we obtained 881 ESPs using the four tools. A further 1,399 proteins that showed a high degree of similarity to experimentally identified secreted proteins were added by the Blast program. Thus, a total of 2,280 proteins were finally predicted as secretory proteins (Table 2).
To validate the in silico secretome analysis, we compiled a list of all experimentally identified ESP sequences of E. granulosus from the NCBI database and from previous studies (47 nucleotides and 77 proteins) [9,17,[20][21][22], and then blasted the putative ESP sequences with the known ESP sequences (see Table S2). Ninetyone proteins were successfully mapped to the known ES proteins, of which 18 shared 100% identity and 33 shared 95%-99% identity. In addition, most known ESPs from other parasites [62] were matched successfully to those identified in our study. More importantly, domains in ESPs of Teladorsagia circumcincta (including metridin-like ShK toxin, lectin, proteinase inhibitor I29, and allergen V5/Tpx-1) were also found in the ESPs of EgPSC, which strengthens the concept that parasites employ universal ESPs to mediate parasite-host interplay [55]. Overall, these data suggest that the ESPs of EgPSCs identified in this study were reliable.
To date, there have been five proteomic studies regarding E. granulosus that have identified just 157 ESPs among them [9,17,[20][21][22]. In this study, approximately 500 ESP domains were found, including known proteins (Table S3), a result that significantly expands the known ES components of EgPSCs. For example, WD40 repeats [63,64], G-protein-coupled receptor (GPCR) [65] and Cadherin [66] all presented novel ESPs that were involved in parasite development-related processes. Recent studies using genome-wide and transcriptome data provide comprehensive information about the growth and development of E. granulosus [31,67]. The results of this study extend this
Of the 2,280 putative ESPs, only 1,406 were mapped to known functions (Table S3). These proteins included not only many common and abundant 'house-keeping proteins' (e.g., ribosome proteins, cytochrome subunit proteins, and enzymes involved in carbohydrate and protein metabolism), but also some rare but interesting proteins (e.g., putative receptor and antigenic proteins). This highlights the important roles of ESPs in parasite survival and development within hostile host environments. Below, we characterize these potential ESPs in greater detail.

Metabolism of carbohydrates for parasite energy and nutrition
The interaction of pathogens with mammalian hosts leads to a variety of physiology responses that drive the adaptation of the interacting partners to their new environments and conditions [19]. The ESPs released by parasites might be important actors in this process of adaptation, because they are involved in the metabolism of carbohydrates [68]. We identified a total of 122 domains (summarized in Table S4), of which, 32 proteins were identified to have a higher level of expression in the parasite (Table 4).
E. granulosus has evolved an optimal strategy to gain energy and nutrition from its host using ESPs (Fig. 1). Firstly, the parasite can regulate glycolysis (GL). We identified nine enzymes associated with GL, including the rate-limiting enzymes PFK1 and pyruvate kinase. Through GL, non-essential amino acids (e.g., glutamine, aspartic acid, arginine, proline, histidine, alanine, tyrosine and cysteine), fatty acids, adenine and hypoxanthine nucleotides, as well as pyrimidine, could be synthesized to support parasite development and growth. Alternatively, glucose and other carbohydrates could be synthesized via gluconeogenesis (GN) when alternative carbon sources (e.g., glucogenic amino acids, lactate, and glycerol) were available. In addition to the reversible enzymatic GL steps, several reactions are essential in the GN pathway from pyruvate via oxaloacetate to glucose: the reactions catalyzed by pyruvate carboxylase, phosphoenolpyuvate carboxykinase (PEPCK), fructose-1, 6-bisphosphatase, and glucose-6phosphatase leading to oxaloacetate, phosphoenolpyruvate (PEP), fructose-6-phosphate, and glucose. Finally, tricarboxylic acid (TCA) enzymes, such as aconitate hydratase, succinate dehydrogenase complex, malate dehydrogenase, were identified in the TCA cycle. Other enzymes involved in carbohydrate metabolism are shown in Table 4.
Certain enzymes have been recognized to play key roles in the development of parasites. Phosphoglucose isomerase (PGI), one of glycolytic enzymes, has been found to stimulate parasite growth and the formation of novel blood vessels nearby the developing metacestode [69]. Vaccinating mice with recombinant PGI increases their resistance towards a secondary infection challenge [69]. Similarly, PEPCK is a novel egg antigen of S. mansoni [70] and an abundant protein in adult parasites that is related to numerous metabolic pathways (e.g., endocrine function, excretion and carbohydrate metabolism [22].
To date, only five ESPs have been identified to participate in this metabolic process [17]. The results of this study support the role of these proteins in metabolic adaptation to their hosts and, more importantly, demonstrate that many more ESPs may be used by E. granulosus to regulate carbohydrate metabolism. Further work is required to identify these additional ESPs and establish their functions.

Control of parasite homeostasis
Following infection with E. granulosus, the intermediate host produces a significant immune response that affects the growth and development of parasites [71,72], while the parasites initiate effective evasion mechanisms to counteract adverse host environments.
In this study, we found that 36 ESP domains were molecular chaperones (Table S5), and identified a further 25 proteins that were present with high levels of abundance (Table 4), including several novel molecules (heat shock proteins, HSP90 and HSP40, universal stress protein [Usp], calreticulin, calcineurin B, GrpE in the HSP60 family and Gp96). HSP90 was the most strongly expressed of all the molecular chaperons (Fig. 2), suggesting it is Others one of the key molecules in mediating parasite development. This is supported by the fact that nitration of HSP90 is known to induce cell death [73], and HSP90 has been used as a drug target in protozoa intervention [74]. Previous studies have also shown that UspA and Usp8 are associated with stress resistance and growth in bacterial species [75]. ESPs might disrupt the expression of intracellular 70 protein in the host immune cells, while the parasite itself might release HSP70 to prevent damage from those same cells [76]. These molecular chaperone-like proteins may be released to regulate the stress responses that arise in the extremely harsh intestinal environments of definitive hosts (e.g., numerous highly active proteases, variable pH levels). E. granulosus may secrete proteases or inhibitors to digest host proteins, or to protect itself from digestion by endogenous or hostderived proteinases. In this study, 39 proteases, including serine, aspartic, metallo-and cysteine proteinases, and five inhibitors, were inferred among the set of ESPs (see Table S6). Several of these (serine, cysteine, and the proteinase inhibitors) are likely to be important targets for parasite intervention and control [77][78][79]. However, only three proteases and two protease inhibitors were strongly expressed in the set of ESPs (Table 4). More sensitive technologies will therefore be required to identify other proteases that were expressed at lower levels of abundance.
In contrast, the action of antioxidant enzymes is a key component of parasite survival during infection. In this study, seven ESPs were identified as antioxidant enzymes, including glutathione transferase, peroxiredoxin, thioredoxin, Cu 2 + /Zn 2 + superoxide dismutase, and neuronal nitric oxide synthase protein inhibitor. These molecules might be utilized by the parasite to detoxify the reactive oxygen species produced by the host environments [80].

Direct regulation of host immunological responses
In previous experiments we demonstrated that following infection with EgPSCs the microenvironment of the murine peripheral immune system undergoes several changes. These included T cell activation and the accumulation of immunosuppressive cells, such as myeloid-derived suppressor cells (MDSC) and CD4 + CD25 + FoxP3 + T cells (Treg) [71]. Such alterations might occur via the action of ESPs as many ESPs have been found to redirect host immune responses [13,17]. In this study, we found several ESPs that contribute to immune regulation following infection (Table 4). Tegument protein (Teg) is known to induce a biased Th2 cell immune response related to chronic infection [81], while 14-3-3 proteins are associated with resistance to the immune responses mediated by local cells [82]. In addition, the antigen B (AgB) family are important in immune evasion because the antigen is secreted at variable amounts [83], and have also been demonstrated to direct immature DC maturation towards a preferential Th2 immune response [15].
Notably, cysteine proteinases have been reported to inhibit Th1 immune response via the induction of IL-4, which is the main cytokine responsible for Th2 differentiation [84]. HSP70 has been shown to stimulate both of types of response in CHD patients [85]. Also, the intraperitoneal injection of calreticulin (CRT) significantly influences Th1/Th2 balance [86]. Hence, these proteins might be novel immunoregulatory molecules that contribute to immune evasion.

Signaling pathways
We found that EgPSC possesses many signaling pathways such as P13K-Akt, mitogen-activated protein kinase (MAPK), Wnt, calcium, HIF-1, insulin, estrogen and chemokine signaling (Table  S1). However, in the putative set of ESPs, only G-protein, calcium, Table 3. Cont. Caenorhaditis elegans pathways were used as a reference. The ESP corresponding to each pathway can be found in Table S9 (Table S7), which indicated their importance in parasite-host interactions and physiological processes. Notably, we found that G-protein-coupled receptors (GPCRs), TGF-b and insulin signaling pathways might closely associate with the development of EgPSCs. For example, GPCRs can activate the G-proteins located within the cell. They work cooperatively to deliver varied signals, which in turn regulate various physiological processes [87]. However, the exact function of G-protein signaling in parasites remain unclear.
Studies have shown that TGF-b and insulin signaling pathways in C. elegans can trigger an 'alternative' developmental pathway, and can regulate and transit the environmental stresses on the first larval stage of the parasite [88,89]. In particular, the disruption of both signaling pathways leads to arrested development in this species [90,91]. Indeed, the TGF-b pathway is speculated to regulate developmental events in parasitic nematodes [92], as molecules involved in the TGF-b pathway have been found in several parasitic nematodes including Brugia pahangi, Brugia malayi and Parastrongyloides trichosuri [93][94][95]. The role of TGFsignaling in E. granulosus development and growth warrants further investigation. A recent study revealed that host insulin acts as a stimulant for parasite development within the host liver and that E. multilocularis senses the hormones of hosts through an evolutionary-conserved insulin signaling pathway, which demonstrates the importance of insulin signaling for parasite survival [96].

Potential targets for diagnosis and vaccine development
CHD has a global distribution and causes high rates of morbidity and has a high socio-economic burden in several countries [97]. The Eg95 vaccine induces a high antibody titer in sheep and goats, which protects them against CHD [98]. However, due to antigenic variation caused by genotypic diversity [99], the common Eg95 vaccine does not bind the antibodies of all E. granulosus species, which limits its utility. We suggest that the ESPs of EgPSCs are an excellent alternative candidate for a vaccine, as they are easy to prepare and safer for human health. More importantly, the ESPs obtained by in vitro culture have shown a 92.07% protection rate against a high dose of egg infection in sheep (1,000 eggs per sheep) [100].
Using in silico secretome analysis, we identified 44 antigenic proteins present at high abundance in our set of ESPs (Table 4). Of these, elongation factor 1 alpha, antigen B8/1, myophilin, thioredoxin peroxidase, phosphoglycerate mutase, heat shock protein 90a and actin, were the most abundant. In addition, HSP70, enolase, 14-3-3, phosphate glucose isomerase, malate dehydrogenase, glutathione S-transferase were also present at high abundance in the set of ESPs (Table S8). These abundant proteins hold enormous potential as diagnostic markers or intervention targets. Indeed, malate dehydragenase (MDH) has been tested for the immunodiagnosis of E. granulosus, while thiredoxin peroxidase (TPx) has been used for the immunodiagnosis of human CHD [101]. Likewise, the 14-3-3 molecule has been demonstrated to be a candidate vaccine against E. granulosus in mice [12], while recombinant GST protein has been used in the diagnosis of echinococcosis [102].
Proteins that are present at lower levels of abundance might also be relevant as diagnostic markers or target molecules for vaccine development. In this study, these include antigen 5 (Ag5), calreticulin, calcineurin B, thioredoxin, phosphoglucomutase, fructose-bisphosphate aldolase and gp96 (Table S8). Many of these have already shown promise for serodiagnostic purposes. For example, Ag5 is a dominant immunogenic and diagnostic antigen of the E. granulosus metacestode in both adults and PSCs [22]. Similarly, calcineurin B has been previously identified as a candidate for a vaccine or drug target [103]. Surprisingly, the E. granulosus-specific protein domain antigen B (EgAgB) family, which are well known as diagnostic targets, were undetectable in this study. This result was consistent with previous observations that little or no AgB is secreted by in vitro cultured PSCs [17,104]. Previous studies have demonstrated that the germinal layer, but not the PSC, contributes to the primary secretion of AgB [17]. Thus, serological examination based on the AgB antibody would not be useful in early-stage PSC infection as only minute amounts of AgB antibody are produced at that time.
There are currently just two methods for the treatment of hydatid disease: surgery and the use of benzimidazole, both of which give unsatisfactory results. Hence, novel treatment compounds are urgently needed. In this study, we have identified several secretory drug targets for echinococcosis (Table 4, Table  S3), including GPCRs, threonine and tyrosine protein kinase and nuclear hormones, which have been the targets of successful new drug discoveries [65]. Insulin signaling [96], thyrotropin-releasing hormone receptor, pancreatic hormone-like or transforming growth factor-b (TFG-b) families have been linked to the larval developmental of E. multilocularis. Thus, interventions that utilize these molecules could also arrest parasite growth. In addition, GL enzymes could be drug targets for parasites that rely on the GL pathway for growth and development [22]. Finally, HSP90 has been used as a drug target in protozoa intervention programs [74].

Conclusions
The larval stages of E. granulosus are pathogenic to human, which therefore have become the research focus of CHD. Parkinson et al. [2012] first reported genes with features that reflect physiological adaptations of different parasite stages, including PSCs, and revealed abundant long non-protein coding transcripts, upregulated fermentative pathways, candidate apomucins and a set of platyhelminth-specific gene products, which greatly increased the quality and the quantity of the molecular information regarding E. granulosus [67]. The most newly published genome of the parasite also uncovered several key events of the parasites, including the species-specific genes AgB family, bile salt pathways and Cavb1 gene variation associated with praziquantel sensitivity [31]. Those studies have provided a molecular understanding of the growth and development of E. granulosus. In this study, we focused on the transcriptome of PSCs, which is the only infective component of the larval stages. We present novel and urgently needed information regarding the components of ESPs released by PSCs and their potential roles in the metabolic adaptation of parasites to their hosts. We suggest that intracellular ESPs are essential to the metabolism of carbohydrates within their hosts and that various molecular chaperones with a high level of expression may play a role in resisting harsh host environments. We also reveal a set of antigenic ESPs that show promise as candidates for vaccine development or in the development of serodiagnostic markers. Such findings will encourage more novel strategies for the treatment and control of CHD.
Although the coverage of the transcriptome data in this study was not deep as the genome-wide study [31,67], these findings are novel and hold importance for understanding the mechanisms of parasite metabolic adaptations within their hosts. Overall, this study adds supplementary knowledge regarding the genomics of E. granulosus, and deepens our understanding of host-parasite interactions.