Background

Schistosomiasis, caused by the parasitic blood fluke of the genus Schistosoma, remains a major global public health problem that affects more than 230 million people living in the endemic areas of 77 countries worldwide (more than 95% in Africa) (http://www.who.int/mediacentre/factsheets/fs115/en/index.html). Three major pathogenic schistosome species, Schistosoma mansoni, Schistosoma haematobium, and Schistosoma japonicum, are known to contribute to the disease loads. The treatment of schistosomiasis is based on the long-term mass treatment with the only available drug, praziquantel (PZQ). However, concerns about the problem of drug resistance have necessitated a search for alternative treatments [1–3]. The decoding of the genomes of three major pathogenic blood flukes has provided a valuable entity for the systematic identification of possible drug targets against the parasite [4–42]. As a result, we observed that the SjGIPC3 gene was ubiquitously expressed at different developmental stages, but in a stage-biased pattern. The expression level of the SjGIPC3 gene was lower in the cercarial stage than in any other stages in which it was detected, suggesting that the function of SjGIPC3 is mainly exerted in the stages within the host. In adult worms, the transcriptional level of SjGIPC3 was significantly higher in male than in female worms (Figure 4).

Figure 4
figure 4

Determination of relative expression levels of SjGIPC3 transcripts at different developmental stages by qRT-PCR. Transcriptional levels were calibrated according to the comparative 2−ΔΔCt method using the housekee** gene SjPSMD4 as an endogenous control, and were normalized relative to the cercarial stage. Error bars represent the standard deviations of the mean from the three replicates. C, cercariae; S, hepatic schistosomula; M, male adult worms; F, female adult worms; E, eggs.

Full size image

Determination of the ligand binding specificity of SjGIPC3

To investigate the binding properties of the PDZ domain of SjGIPC3, we screened an arbitrary peptide library using Y2H assays. When the library was screened against the SjGIPC3-PDZ domain, 37 positive clones were sequenced, each encoding a unique carboxy-terminus. Among them, 17 belong to Class I PDZ binding motifs, 16 belong to Class II PDZ binding motifs, and four are unclassified (Table 1). When the library was screened against the full-length SjGIPC3, 42 positive clones were obtained. Sequence analysis revealed that three pairs of clones were identical; thus a total of 39 unique carboxy-terminal sequences were obtained. Among them, 24 belong to Class I PDZ binding motifs, 12 belong to Class II PDZ binding motifs, and three are unclassified (Table 1).

Table 1 The unique ligand carboxyl termini obtained from Y2H assays using the SjGIPC3-PDZ domain or full-length SjGIPC3 as bait
Full size table

In the first assay using the SjGIPC3-PDZ domain as bait, the domain showed predominant preference for hydrophobic amino acids at the extreme carboxyl terminus (P0), especially Leu (37.8%), Ile (32.4%), or Val (18.9%). The P-1 of the carboxyl terminus slightly prioritizes polar amino acids, such as Ser (16.2%), Asn (10.8%), or Gln (10.8%), and acidic amino acids, such as Asp (13.5%) or Glu (13.5%). At P-2, the polar amino acids Ser (37.8%) and Thr (10.8%) and the aliphatic amino acids Ile (21.6%) and Leu (18.9%) were preferentially selected. At P-3, Ser (13.5%), Asp (18.9%), and Glu (24.3%) were preferred, along with the rare occurrence of other amino acids (Figure 5A). Based on the statistical analysis of unique carboxy-terminal sequences, a consensus-binding sequence can be deduced as -[S/D/E]-[S/T/I/L]-[S/N/Q/D/E]-[V/I/L]*.

Figure 5
figure 5

Binding properties of the SjGIPC3-PDZ domain. The four amino acids from the extreme carboxyl terminus of positive clones isolated from Y2H assays with the SjGIPC3-PDZ domain as bait (A), or full-length SjGIPC3 as bait (B), respectively, were aligned and the occurrence for each amino acid at each position was statistically analyzed. The positions were numbered from the carboxyl terminus (position 0). The tables described the percentage of each amino acid type at each given position (blank: percentage = 0%).

Full size image

In the second assay, full-length SjGIPC3 was probed to screen the random peptide library, and a more significant binding specificity emerged. As in the first assay, hydrophobic amino acids, especially Ile (42.5%), Leu (21.3%), or Val (14.2%), were predominantly preferred at P0, with a slight bias toward the aromatic amino acid Phe (9.5%). At P-1, the preference for Glu (45.2%) was dramatically increased when compared with the first assay, along with a rare selection of other residues: His (11.9), Ser (11.9%), Asn (11.9%), or Asp (9.5%). The P-2 position demonstrated a dramatic preference for the polar amino acid Ser (54.8%) and the hydrophobic amino acid Ile (21.4%); other amino acids, such as Thr (11.9%), Val (7.1%), or Leu (4.8%), also occurred at a low rate. At P-3, Glu (45.2%) was predominantly selected, followed by Cys (19.0%) and Asp (14.3%) (Figure 5B). Similar to the first assay, a consensus-binding sequence can be inferred from the second Y2H assay: -[D/E/C]-[S/T/I]-[H/S/N/D/E]-[V/I/L]*. Finally, based on both assays, the consensus-binding sequence of the SjGIPC3-PDZ domain could be reduced to -[SDEC]-[STIL]-[HSNQDE]-[VIL]*, which fits in both class I and class II binding motifs.

Prediction and identification of native ligands of SjGIPC3

We searched for the potential native ligands of SjGIPC3 in the S. japonicum predicted proteomic database (sjr2_protein.fasta) with the Tailfit program using the consensus-binding sequence -[SDEC]-[STIL]-[HSNQDE]-[VIL]* as bait; 92 proteins that physically fit the consensus binding sequence were fished out. We also performed BLAST searching on NCBI within taxa 6182 directly; using each unique carboxy-terminal sequence (the extreme four amino acids) obtained from both Y2H assays as bait. A panel of proteins whose C-terminal tail matched the unique sequences was retrieved. The protein sequences retrieved from both bioinformatics pipelines were used to perform BLASTP searches on NCBI to manually filter those peptides lacking the C-terminus. We also used integrated biological information, such as transmembrane region prediction, potential molecular functions, and the C-terminal sequence consensus between S. japonicum and S. mansoni to determine the most promising candidate ligands. Finally, six proteins were selected as ligand candidates of SjGIPC3 (Table 2). Among them, four were further confirmed to be potential ligands of SjGIPC3 in the Y2H system.

Table 2 Validation of the predicted ligand candidates of SjGIPC3 in the Y2H system
Full size table

Discussion

Domain loss events are extensively widespread in S. japonicum, which serves as a notable consequence of the adoption of parasitism [4]. It is estimated that 1,000 protein domains have been abandoned by S. japonicum during long-term evolution, which suggests that the remaining domains are necessary for parasitic growth, development, and maturation. The proteomic information revealed that at least 25 PDZ domains are distributed among 16 S. japonicum proteins, suggesting that the PDZ domain is one of the important modules reserved in the parasite to mediate multiple biological processes. In the present study, we have characterized one of the PDZ domain-containing proteins from S. japonicum, GIPC3, for the first time. The ligand binding properties of the SjGIPC3-PDZ domain were determined by Y2H assays, and its potential ligands were identified.

Screening a highly diverse peptide library provided an opportunity to fish out the potential binding property of SjGIPC3. Whether the SjGIPC3-PDZ domain or full-length SjGIPC3 were used as bait, analogous binding specificities can be deduced from the unique C-terminal tails obtained from both Y2H assays. However, the GH1 and GH2 domains of SjGIPC3 may still affect the binding specificity of the PDZ domain to some degree, because the ligand motifs were more canonical when full-length SjGIPC3 was used as bait. For instance, at P-1, the acidic amino acid Glu was more preferentially selected in the second assay (45.2%) than in the first assay (13.5%). Similarly, at P-3, 16 types of amino acids emerged in the first assay, while only eight kinds of amino acids were presented in the second assay. Also at this position, the occurrence of Glu was 50.0% when screened against the full-length SjGIPC3, twice the occurrence rate as when the SjGIPC3-PDZ domain was used as bait. Thus, the binding specificity screened using full-length SjGIPC3 as bait may be more authentic to its native status.

Previously, GIPCs have been shown to interact specifically with Class I (−X-S/T-X Φ*), II (−Φ X Φ*), and III (−X X-C*) binding motifs [43]. Moreover, Kermit, the GIPC orthologue in Drosophila, and GLUT1CBP each interact with Myosin VI through an internal sequence within the cargo binding domain [43, 44]. In this study, we focused on the C-terminal binding specificity of SjGIPC3 and found that the PDZ domain of SjGIPC3 preferred to bind to Class I and II motifs, consistent with the extensive ligand-binding ability of GIPC orthologues in other species [43]. However, regarding the binding of Class II motifs, the only example presented thus far is that the PDZ domain of synectin binds to sydecan-4 through the “-EFYA*” motif, in which the aromatic amino acid Phe was selected at P-2[18]. In our study, it is worth noting that this site can be occupied by the aliphatic amino acid residues Ile and Leu, in addition to Ser and Thr (Table 1). Moreover, in a previous report, the binding motifs of GIPC orthologues trend toward selecting a Val, Ala, or Cys residue at p0[43]; here, the more hydrophobic amino acids, such as Ile and Leu, were predominantly preferred. These diverse binding characteristics could be utilized for drug development against SjGIPC3.

On average, one PDZ protein is capable of interacting with 17 partners [45]. Although a panel of S. japonicum proteins deposited in the public database bears a C-terminal tail fittting in the consensus binding sequence, the bioinformatic hints at molecular function and localization have prevented them from being ligand candidates of SjGIPC3. Using the Y2H assay, four proteins were determined to be potential ligand candidates for SjGIPC3, which suggests that some other potential ligands may not have been discovered yet, likely restricted by the quality of the proteomic database, as truncated fragments have been extensively deposited in the S. japonicum proteomic database.

To adapt to the complexity of the life cycle, schistosomes are equipped with a complex nervous system to sense environmental signals [46]. Transcriptional analysis has revealed that the expression of SjGIPC3 was dramatically up-regulated after invasion of the host, suggesting that it may play an important role (e.g., signal transduction) in the context of the parasite-host interaction. More importantly, the preferential expression of SjGIPC3 in male rather than female adult worms further hinted that the potential function of the protein was related to the processing of environmental information, as male adult worms interface with the host and female adult worms resided in the gynecophoral canal, more stimulating signal would be received by male adult worms [47]. GIPC3 mutations have recently been reported to cause autosomal recessive nonsyndromic sensorineural hearing loss [48–50]. Previously, Yi and his colleagues suggested that the GIPC orthologue in mouse might associate with extrasynaptic NMDA receptors through binding to the C-terminal tail “ESDV”, and may be involved in the organization and trafficking of this population of receptors [51]. Intriguingly, among the four potential ligands of SjGIPC3 identified in this study, one is the glutamate receptor NMDA, which also bears a class I motif “TTTL”; thus, we cannot rule out the possibility that SjGIPC3 may also have a physiological function in the regulation of NMDA receptor trafficking, similar to its orthologue in mouse.

Conclusions

For the first time, we have presented the molecular characterization of the PDZ domain-containing protein GIPC3 from S. japonicum. We defined the ligand binding specificity of the SjGIPC3-PDZ domain by screening a random peptide library constructed with human genomic DNA. The ligand binding ability of the SjGIPC3-PDZ domain is relatively extensive; it can bind both type I and type II ligand motifs, which hints at its potential to play multiple roles during the development of S. japonicum, especially in the adult stage. Our study also provided a practical method for predicting potential ligands of SjGIPC3. Ongoing studies to identify more ligands involved in signal transduction followed by improvement of the quality of the S. japonicum proteomic dataset will assist in the discovery of novel drug targets against the parasite.

Ethics statement

All procedures performed on animals within this study were conducted following animal husbandry guidelines of the Chinese Academy of Medical Sciences and with permission from the Experimental Animal Committee.