Introduction

The shrimp immune system relies solely on the primitive innate immune response that includes humoral and cellular responses against infectious agents1,2,3,4,5. However, evidences of the invertebrate immune system’s capability of discriminating between pathogen at species and even strain level refute the notion that invertebrate innate immune system is naïve in every infection even with the same pathogen6,7,8,9. Thus, there is a growing interest in invertebrate immune research in learning how the invertebrate immune system can have high levels of specificity in the absence of antibody-mediated immune responses6. Having pathogen-specific immune responses requires a diverse array of pathogen recognition receptors (PRRs) that recognize and initiate the immune response through pathogen-associated molecular patterns (PAMPs)10,11. Upon recognition of PAMPs, PRRs trigger a cascade of immune responses such as agglutination, encapsulation, nodulation, phagocytosis, the release of antimicrobial peptides (AMPs), and activation of the pro-phenoloxidase (ProPO) system leading to melanization, all of which promote the degradation and clearance of pathogens12,13,14,15,16,17,18,19. The shrimp immune system has 11 PRRs20 and among these are the galectins, β-1,3-glycan-binding proteins (BGBPs), β-1,3-glycanase-related proteins (BGRPs), and C-type lectins (CTLs), all of which specifically bind to cell surface carbohydrates of pathogens11,20.

Indeed it is intriguing how carbohydrate complexes from glycans, which holds immensely complex biological information not encoded in the genome, are recognized by PRRs21,22. It is then fitting to look into the lectins, which bind and recognize carbohydrates. Among animal lectins, CTLs are the most abundant and diverse superfamily, they are lectins containing one or more C-type carbohydrate recognition domains (C-type CRDs) that function in a calcium-dependent manner21,22. CRDs have been characterized into two groups according to their motif which can be mannose(Man)-binding EPN motif or galactose(Gal)-binding QPD motif 20,23,24. However, variations of these motifs resulting from a mutation of single amino acid were reported, in shrimp CTLs variants include EPK, EPD, EPQ, EPS, QPN, and QPT20.

As literature on both CTLs increased, ambiguity and confusion over the use of terminologies for their classification arose24. To avoid confusion we conform to the previous review on CTLD superfamily24 (24) we will use CTLD-containing proteins (CTLDcps) to refer to proteins containing the carbohydrate-recognition (CRD) domain, which encompasses both the classical Ca2+-dependent carbohydrate-binding C-type lectins (CTLs) containing C-type CRDs, and the non-lectin C-type lectin-like domains (CTLD).

Recently, another binding motif variant with a QAP (Gln-Ala-Pro) sequence has been reported in several shrimp CTLs. These inlcude CTLDs from Litopenaeus vannamei (LvCTLD)25 and Macrobrachium niponense (MnCTLDcp1,2)26 and LdlrLec1,2 from Marsupenaeus japonicus27. The binding specificity of the QAP motif in LvCTLD, was originally reported to be Gal, suggesting QAP motif has the same specificity as QPD motif 25.

In this study, however, we report a different binding specificity in the QAP motif of a novel CTL from M. japonicus (MjGCTL). The present findings provide insights on how CTLs acts as PRRs working as non-self recognition, initiating immune response such as encapsulation by hemocytes.

Results

Molecular characteristics of MjGCTL

The full cDNA sequence of MjGCTL (DDBJ: LC127418) was obtained. Its ORF was comprised of 935 nucleotide bases in length encoding 311 amino acid residues (Fig. 1), with a predicted molecular weight of 35 kDa. Analysis of protein domains by SMART revealed MjGCTL is comprised of a signal peptide, low-density lipoprotein class-A receptor (LDLa) domain, and a carbohydrate recognition domain (CRD). The predicted amino acid sequence of MjGCTL’s CRD has a QAP binding motif. The MjGCTL amino acid sequence has 5 predicted N-glycosilation and 2 GalNAc O-glycosylation sites (Fig. 1). In the phylogenetic analysis of the amino acid sequences of shrimp CRDs (Fig. 2), MjGCTL clustered with other shrimp CTLs with QAP motifs with a 99.9% bootstrap value.

Figure 1: Sequence analysis of MjGCTL.
figure 1

Complete nucleotide cDNA and deduced amino acid sequences of MjGCTL. The signal peptide is bold-faced, and Low-density lipoprotein receptor class A (LDLa) domain is enclosed in a dotted box, the carbohydrate recognition domain is highlighted with its QAP binding motif designated by a solid box. Predicted N- and O- linked glycosylation sites are underlined and encircled, respectively.

Full size image
Figure 2: Phylogenetic analysis of shrimp carbohydrate recognition domains (CRDs) with 1000 bootstrap replicates.
figure 2

MjGCTL is represented with a (•). Each CTLDcps are labeled with their respective sugar-binding motifs.

Full size image

Tissue distribution and gene expression profile of MjGCTL

MjGCTL transcripts were expressed only in gill and stomach tissues of M. japonicus, with expression being strongest in the gills (Fig. 3A) thus we designated the name M. japonicus gill C-type lectin (MjGCTL). MjGCTL mRNA relative expression after WSSV infection, analyzed by qRT-PCR, showed that compared to 0 dpi, there were no significant differences among the mRNA levels of MjGCTL at 1, 3, and 5 dpi (Fig. 3B).

Figure 3: Tissue expression profile of MjGCTL.
figure 3

(A) Tissue distribution of MjGCTL by RT-PCR analysis. Transcripts of MjGCTL together with elongation factor 1α (EF-1α) as internal control was analyzed in gills (g), hepatopancreas (hp), lymphoid organ (lo), hemocyte (hm), muscle (m), stomach (s), heart (h), intestine (i), nerve (n), and eye (e). (B) qRT-PCR analysis of MjGCTL mRNA expression levels after 0, 1, 3, 5 days post-infection (dpi) with WSSV. There was no significant difference between sampling time at pvalue 0.05.

Full size image

Western Blot analysis of rMjGCTL

The IPTG induced E. coli cells yielded a rMjGCTL with a size of approximately 37 kDa (Fig. 4A). While successfully purified eluted rMjGCTL from Drosophila S2 cells detected by anti-V5 was as approximately 50 kDa (Fig. 4B), which is confirmed by MjGCTL detection in gill tissue that was higher than 37 kDa and slightly lower than 50 kDa (40–45 kDa) (Fig. 4C).

Figure 4
figure 4

Western Blot analysis of MjGCTL (A) in IPTG induced E. coli cells detected using anti-His Lane 1: protein marker (M), Lane 2: rMjGCTL showing the detected protein around 37 kDa. (B) Purified rMjGCTL from Drosophila Schneider 2 cells. Lane 1: protein marker (M), Lane 2: rMjGCTL showing the detected protein around 50 kDa. (C) Detection of MjGCTL using anti-MjGCTL in rMjGCTL (lane 2) and in gill tissues (lane 3) showing the detected protein around 50 kDa and slightly below 50 kDa (40–45 kDa), respectively.

Full size image

MjGCTL promotes bacterial agglutination

rMjGCTL expressed by S2 cells agglutinated both Gram-negative bacteria (EGFP-E. coli and Vibrio parahaemolyticus) and a Gram-positive bacterium (Streptococcus agalactiae) (Fig. 5). The bacterial agglutination was concentration-dependent with a minimum agglutination concentration of 12.5 μg/mL.

Figure 5: Bacterial agglutination activity of rMjGCTL with Gram-negative V. parahaemolyticus and Gram-positive S. agalactiae.
figure 5

Bacterial suspensions were incubated with rMjGCTL, BSA as the protein control, and TBS-Ca2+ (TBSCa) as the negative control.

Full size image

Calcium-dependent binding and carbohydrate specificity of MjGCTL

rMjGCTL’s ability to agglutinate fluorescent (EGFP-expressing) E. coli was inhibited by the removal of CaCl2 or by the addition of EDTA (Fig. 6A), confirming the Ca2+-dependency of rMjGCTL’s agglutinating activity. Agglutination inhibition assays using different carbohydrates revealed that Man and Gal did not inhibit rMjGCTL’s ability to agglutinate V. parahaemolyticus but only Glucose (Glc) (Fig. 6B). Agglutination of was also inhibited by other mono- and disaccharides, and by LPS and PGN in varying concentrations (Table 2).

Figure 6: MjGCTL functions as a classical C-type lectin.
figure 6

(A) Ca2+-dependent agglutination activity of rMjGCTL (50 μg/ml) from Drosophila S2 cells with EGFP-expressing E. coli as influenced by the presence (TBSCa) and absence (TBS(−)) of Ca2+ and the addition of a chelating agent EDTA. (B) Agglutination inhibition by monosaccharaides Man, Gal, and Glc with Vibrio parahaemolyticus

Full size image
Table 1 Primer sequences used in this study.
Full size table
Table 2 Carbohydrate specificity of MjGCTL.
Full size table

rMjGCTL’s opsonic effect enhance hemocyte encapsulation

Hemocytes encapsulated rMjGCTL-Agarose beads immediately after they were added (0 hours post incubation (hpi)) and the binding increased from 2, 4, to 8 hpi (Fig. 7). On the other hand, no binding was observed with BSA-coated beads (protein control) or uncoated beads in TNS buffer (negative control). Melanization (indicated by the brown coloration on beads) also increased with incubation time for rMjGCTL-coated beads but not in the protein and negative controls. Results of hemocyte encapsulation inhibition by carbohydrates (Fig. 8) showed encapsulation by hemocytes were inhibited by Glc and fucose (Fuc), and not by Man and Gal. These results also show that encapsulation by hemocytes results from the binding of rMjGCTL carbohydrate recognition domain to hemocytes as demonstrated by the inhibition through competitive binding of the carbohydrate substrates Glc and Fuc.

Figure 7: MjGCTL facilitates M. japonicus hemocytes encapsulation of Ni-NTA agarose beads.
figure 7

Incubated with rMjGCTL (50 μg/ml) from Drosophila S2 cells, BSA (protein control), and TNS buffer (negative control) demonstrate encapsulation by hemocyte after 0, 2, 4 and 8 hours post-incubation (hpi).

Full size image
Figure 8: Promotion of hemocyte encapsulation is attributed to MjGCTL’s carbohydrate-binding activity.
figure 8

MjGCTL’s binding to hemocytes and the activation of encapsulation are inhibited by MjGCTL’s carbohydrate substrates, which is demonstrated by incubating Man, Gal, Glc, and Fuc at 100 mM with agarose beads bound to rMjGCTL (100 μg/ml) from Drosophila S2 cells after 8 hpi.

Full size image

Discussion

Diversification of shrimp CTLDcps is manifested in their tissue distribution, domain architecture, and sugar specificity20. In contrast to most invertebrate lectins that are mostly expressed in the hepatopancreas, the localization of MjGCTL is in gills and stomach (Fig. 3A) which may suggest MjGCTL reside in the main entry sites and target organs of major shrimp pathogens28,29,30,31. MjGCTL functions as a classical CTL binding to carbohydrates and bacterial components (Table 2), allowing it to cause agglutination (Fig. 5) in a Ca2+-dependent manner (Fig. 6A). MjGCTL may also act as an opsonin, as clearly shown by the encapsulation assay (Fig. 7) and its inhibition by Glc and Fuc (Fig. 8) indicate that rMjGCTL CRD can bind to hemocytes surface carbohydrates facilitating encapsulation. MjGCTL’s being a PRR also give us an insight on the poorly understood mechanism of invertebrate hemocyte encapsulation, where MjGCTL was observed to activate the adhesive state of and recruit hemocytes to encapsulate agarose beads by binding to hemocytes (Fig. 8).

Conventionally, CTLs are classified by their CRDs into the mannose-type EPN motif or Gal-type QPD motif. In the phylogenetic analysis of the shrimp CRDs (Fig. 2), MjGCTL clustered closely with four other CTLDs containing a QAP binding motif: LvCTLD25, MjLdlrLec 1 and 227, and MnCTLDcp 1 and 226. Among these, MjGCTL shares highest identity (77%) with LvCTLD (Supporting information, Fig. 1). LvCTLD has an affinity for Gal, suggesting that QAP is a variant of QPD motif25. Contrary to the report on LvCTLD, our results suggest that the QAP is a new motif with a different carbohydrate-binding specificity from EPN and QPD motif. The position of MjGCTL’s CRD in the tree (Fig. 2) infers it is not associated with neither EPN and QPD nor their variants (QPT, EPS, etc.). This is corroborated by the binding specificity of MjGCTL, where the finding that MjGCTL does not bind either Man or Gal (Table 2) suggests that its specificity is completely different from that of CTLs with EPN and QPD motifs. Comparing the specificity of MjGCTL to the X-ray crystallographic analyses suggest that the binding specificity of the QAP motif is more similar to that of EPN than to that of QPD, where the QPD motif has an affinity for an equatorial/axial 3-OH/4-OH configuration of carbohydrate residues (Gal and N-acetylgalactosamine)32,33, while the EPN motif has an affinity for an equatorial/equatorial 3-OH/4-OH configuration of carbohydrate residues (Man, Glc, N-Acetylglucosamine, and fucose)33,34. Interestingly, all reported CTLDcps with QAP binding motifs are only found in shrimps, including MjGCTL, and they all contain a low-density lipoprotein receptor class A (LDLa) domain. Shrimp CTLDcps with an LDLa domain interact with viral particles25,26 or respond to viral infection such as WSSV26. However, in the case of MjGCTL, mRNA levels remained unchanged after WSSV infection (Fig. 3B), possibly because the LDLa domain does not interact with WSSV particles. Moreover, shrimp CTLDs not containing LDLa were found to bind with viral particles or change their mRNA transcript levels in response to viral infections15,35,36,37,38,39, thus suggesting that the LDLa domain of MjGCTL does not necessarily imply that the molecule binds or responds to viral particles, which is a characteristic of CTLD molecules. Thus, it is possible that in the case of CTLs LDLs serve as Ca2+-binding site, as they have been known to sequester and regulate Ca2+ levels29,40. However, the unchanged mRNA levels also do not mean MjGCTL does not respond to invading pathogens. This is shown by the finding that knockdown of MjHeCL, a CTL that is constitutively expressed in hemocytes, caused the bacteria to proliferate in hemolymph41. It is possible that just like MjHeCL, MjGCTL is constantly expressed in high levels as readily available defense molecules of the shrimp immune system.

Characterization of the domain architecture of MjGCTL demonstrates its uniqueness and binding-specificity is attributed not only to its binding motif but more importantly to its proper folding and glycosylation6. Here we speculate that the shift in the molecular weight detected of the purified rMjGCTL from its predicted size based on its amino acid sequence (35 kDa), may be due to the addition of carbohydrate moieties that is most likely to the presence of glycosylation. Glycosylation of MjGCTL by Drosophila S2 cells was expected as the sequence contains five potential N-glycosylation sites and two potential O-glycosylation sites (Fig. 1). Interestingly, the attempted production of rMjGCTL by E. coli expression system yielded a rMjGCTL with a protein size of about 37 kDa (Fig. 4A). This may be possible, as the E. coli expression system is known to have the disadvantage of producing misfolded proteins and cannot carry out post-translational modifications such as N- or O- glycosylations42. On the other hand, the eukaryotic expression system of Drosophila S2 cells are have been demonstrated to be a reliable system for proteins requiring N-glycosylations for their proper production as compared to E. coli43. This is confirmed by the detection of MjGCTL in gill tissues that revealed its actual size slightly lower than rMjGCTL from Drosophila S2 cells (Fig. 4C) as the latter is tagged with both V5- and His- epitope which is predicted to have a size of 3.4 kDa.

Despite the high identity to MjGCTL, LvCTLD does not function as a CTL. CTLDs are technically not lectins, and like LvCTLD they are Ca2+-independent. This is supported by the 3D models of MjGCTL and LvCTLD (Supporting information, Fig. 1) that show, unlike LvCTLD, the predicted locations of the Ca2+-binding site are co-located with QAP motif at the 5th-sheet. This is in accord with the structures of classical CTLs, in which the Ca2+-binding site is coupled with the sugar-binding motif in the CRD23,24. This comparison of MjGCTL and LvCTLD therefore show clearly the demarcation of CTLs and CTLDs, and thus provide a good reference for characterization of the growing number of the diverse superfamily of CTLDcps.

In summary, the characterization of MjGCTL provides further evidence of the complexity and diversity of invertebrate CTLDs. MjGCTL is a classical C-type lectin that binds to carbohydrates in a Ca2+-dependent manner. It functions as a PRR, recognizing non-self, through bacterial agglutination and it acts as an opsonin by binding directly to hemocytes to activate their adhesive state for their encapsulation. The sugar specificity of the QAP binding motif of MjGCTL is different from that of EPN and QPD motifs, suggesting QAP as a new binding motif. The finding that MjGCTL is almost identical to LvCTLD, yet possesses a different function, clearly demonstrates CTLDcps superfamily can be more accurately classified by their function rather than solely by their binding motif they possess. The divergent binding-motif and specificity of MjGCTL provides an insight on how invertebrate immune system possesses unique and highly specific pathogen recognition receptors through the CTLDcps.

Methods

Experimental Shrimp

Marsupenaeus japonicus shrimp weighing 10 grams were obtained from a farm in Miyazaki prefecture Japan and were acclimatized for at least 3 days before the experiment. Shrimp were kept in tanks with recirculating water maintained at 25 °C μg/mL and 35 ppt salinity.

RNA isolation and cDNA synthesis

Total RNA was isolated from various tissues (gills, hepatopancreas, lymphoid organ, hemocytes, muscle, stomach, heart, intestine, nerve, and eye) of M. japonicus using RNAiso (TAKARA, Japan) that was used as template (1 μl) for cDNA synthesis using High capacity cDNA reverse transcription kit (Applied Biosystems, USA).

Detection and sequencing of full cDNA sequence of MjGCTL

MjGCTL primers (Table 1) were designed using EST data from M. japonicus cDNA library. Full-length MjGCTL was amplified, cloned to pGEM T-easy vector (Promega, USA) and cloned into competent Escherichia coli JM109 cells and sequenced in both directions using M13 primers (Table 1). The sequences were aligned and assembled using GENETYX Ver.11.04 (GENETYX, Japan).

Analysis of MjGCTL expression profile

MjGCTL transcripts were detected through semi-quantitative RT-PCR using 1 μl of cDNA from various tissues of M. japonicus using MjGCTL primers and elongation factor 1α (EF-1α) as an internal control (Table 1). Thermal cycling conditions were pre-denaturation at 95 °C for 5 min, 28 cycles of 95 °C for 30 sec, 55 °C for 30 sec and 72 °C for 1 min, and followed by final extension at 72 °C for 5 min. The PCR products were then viewed by 1% agarose gel electrophoresis.

Cell culture of Drosophila Schneider 2 (S2) cells

Drosophila Schneider 2 (S2) cells were cultured in Schnieder Drosophila medium (SDM) (Invitrogen) supplemented with 10% fetal bovine serum (FBS) and antibiotics in a culture flask maintained without CO2 at 28 °C. S2 cells were passaged several times until reaching the optimum cell growth rate following the manufacturer’s protocol.

Production of recombinant MjGCTL (rMjGCTL)

The cDNA sequence of MjGCTL was amplified using rMjGCTL primers (Table 1), where EcoRI and NotI sites were added to the forward and reverse primers (5′ and 3′ end), respectively. The rMjGCTL PCR was cloned then cloned into the restriction sites of expression vector pMT:BiP:V5-His C (Invitrogen, Carlsbad, CA, USA). Recombinant plasmid was then transfected using Effectene Transfection Reagent (Qiagen, Germany) following manufacturer’s instructions. Stable cell lines were selected by passaging the cells several times on SDM with 125 mg/ml of blasticidin. Following the large-scale production of the stable cell-lines, protein expression was then induced by CuSO4 (600 μM). One day after induction, the cells were centrifuged and rMjGCTL was purified from the supernatant using Ni-NTA agarose (Qiagen) purification column with 500 mM imidazole. The purified protein was quantified using DC protein assay (BIO-RAD) following the manufacturer’s protocol.

Prior to using S2 cells, production of rMjGCTL through bacterial cell system was also attempted through cloning MjGCTL to pCold II DNA (Takara bio, Japan) and transforming it to Escherichia coli (Origami B(DE3)), however the purification of the recombinant protein from large-scale production was unsuccessful.

SDS-PAGE and Western blot analysis

Eluted proteins from Drosophila S2 cells were subjected to 15% SDS-polyacrylamide gel electrophoresis and transferred to a nitrocellulose membrane. The proteins were then detected with mouse anti-V5 monoclonal antibody (Thermo Fisher Scientific, USA) as primary antibody diluted to 1:5000 in 5% BSA and 0.05% Tween 20 in TBS (TBST) (blocking solution), and with anti-mouse IgG (H + L), AP conjugate (Promega, USA) diluted to 1:10000 in blocking solution. Bands were then visualized with substrate solution containing 5-bromo-4-chloro-3-indolyphosphate and nitroblue tetrazolium (BCIP/NBT, Sigma) for 5 min and washed with distilled water. On the other hand, rMjGCTL from IPTG induced E. coli cells were detected through 12% SDS-PAGE and Western blot analysis using monoclonal anti-His mouse IgG, however rMjGCTL were not detected upon purification.

To determine the actual size of MjGCTL in shrimp, total protein were extracted from gill tissues and was subjected to 12% SDS-PAGE and Western Blot analysis, as described above, using rabbit polyclonal anti-MjGCTL diluted to 1:5000 with blocking solution (2.5% skimmed milk in TBST) as the primary antibody and anti-Rabbit IgG(Fc), AP Conjugate (1:5000) (Promega, USA) as the second antibody.

Bacterial agglutination assay

Agglutination was assayed by a modification of the method of Luo et al.44. The bacteria used in the assay were two Gram-negative Vibrio parahaemolyticus and EGFP-expressing Escherichia coli and Gram-positive Streptococcus agalactiae. Bacteria were resuspended in TBS-Ca2+ (Tris-HCl, pH 7, 100 mM NaCl, and 10 mM CaCl2) at 1 × 109 cells/ml. Ten microliters of bacteria suspension was then incubated with same amount rMjGCTL (50 μg/ml) at room temperature (25 °C) for 1 h. Agglutination were then viewed under light microscope (Nikon) and fluorescence microscopy for GFP-expressing E. coli. The minimum concentration of rMjGCTL that would agglutinate was determined with serial dilutions of rMjGCTL in TBS-Ca2+ (10 mM) and was used in agglutination assay as mentioned above. To test the calcium-dependency of rMjGCTL, agglutination was assayed using only TBS (without CaCl2) and also by adding EDTA with final concentration of 10 mM into TBS.

Agglutination inhibition assay of carbohydrates

The sugar specificity of rMjGCTL was expressed as the minimum concentration of a sugar needed to inhibit agglutination. rMjGCTL was incubated with different concentrations of several carbohydrates for 1 h at room temperature prior to the agglutination assay described above. All assays were done in triplicates.

in vitro encapsulation assay

The assay was done as described by Ma et al.45. BSA-coated beads and uncoated beads in TNS buffer (10 mM Tris-HCl, 140 mM NaCl, 20 mM CaCl2, pH 7.9) were used as positive and negative controls, respectively. Prior to adding the hemocytes, the beads were washed three times with 1 ml TNS buffer, incubated with 1 ml of His-tagged rMjGCTL or BSA (300 μM) in TNS overnight at 4 °C with gentle rotation, washed five times with 1 ml TNS and resuspended in 1 ml TNS as a 50% slurry. Hemolymph was collected from M. japonicus with a 23-gauge needle and syringe with 1.5 ml PBS containing 10 mM EDTA. The hemocytes were then collected, washed three times, and resuspended in 1 ml 2x Leibovitz’s L-15 medium (Invitrogen, USA). Encapsulation was assayed by adding 10 μl of the hemocyte suspension to the wells of a 24-well cell culture plate coated with 500 μl 1% agarose. One (1) μl of the beads slurry (incubated with rMjGCTL or BSA) was added to each well. Beads suspended in TNS buffer were used as a negative control. Encapsulation by hemocytes was then observed at 0, 2, 4 and 8 h post incubation (hpi) under a light microscope (Nikon).

Hemocyte encapsulation inhibition by carbohydrates

Prior to the encapsulation assay, 100 mM of Man, Gal, Glc, and Fuc were incubated with the bound rMjGCTL and was viewed at 8 hpi. Experimentally added carbohydrates with affinity to MjGCTL will compete with that of the carbohydrate substrates of hemocytes therefore disrupting the encapsulation of the agarose beads. All experiments were done in triplicates.

Additional Information

How to cite this article: Alenton, R. R. R. et al. Pathogen recognition of a novel C-type lectin from Marsupenaeus japonicus reveals the divergent sugar-binding specificity of QAP motif. Sci. Rep. 7, 45818; doi: 10.1038/srep45818 (2017).

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.