Characterization of Hagfish (Eptatretus burgeri) Variable Lymphocyte Receptor–Based Antibody and Its Potential Role in the Neutralization of Nervous Necrosis Virus

Key Points

• NNV-specific VLRBs were developed using a library screening system.

• The binding ability of this Ab was enhanced through modular engineering.

• VLRB could be an alternative neutralizing Ab against Betanodavirus infection.

Abstract

The variable lymphocyte receptor (VLR) mediates the humoral immune response in jawless vertebrates, including lamprey (Petromyzon marinus) and hagfish (Eptatretus burgeri). Hagfish VLRBs are composed of leucine-rich repeat (LRR) modules, conjugated with a superhydrophobic C-terminal tail, which contributes to low levels of expression in recombinant protein technology. In this study, we screened Ag-specific VLRBs from hagfish immunized with nervous necrosis virus (NNV). The artificially multimerized form of VLRB was constructed using a mammalian expression system. To enhance the level of expression of the Ag-specific VLRB, mutagenesis of the VLRB was achieved in vitro through domain swapping of the LRR C-terminal cap and variable LRR module. The mutant VLRB obtained, with high expression and secretion levels, was able to specifically recognize purified and progeny NNV, and the Ag binding ability of this mutant was increased by at least 250-fold to that of the nonmutant VLRB. Furthermore, preincubation of the Ag-specific VLRB with NNV reduced the infectivity of NNV in E11 cells in vitro, and in vivo experiment. Our results suggest that the newly developed Ag-specific VLRB has the potential to be used as diagnostic and therapeutic reagents for NNV infections in fish.

Introduction

In jawed vertebrates, the adaptive immune system is generated through the expression of TCRs and BCR belonging to the Ig superfamily (1, 2). In jawless vertebrates, which include lamprey (Petromyzon marinus) and hagfish (Eptatretus burgeri), the adaptive immune response is mediated by variable lymphocyte receptors (VLRs), composed of leucine-rich repeats (LRRs) (3). Mature VLR genes are generated by combinatorial assembly with hundreds of different LRR-encoding cassettes from each germline VLR gene, and thereafter the repertoire of Ag-binding receptors (>10¹⁴) are produced through somatic diversification of LRRs (4, 5). Three distinct types of receptors, VLRA, VLRB, and VLRC, have been identified in both lamprey and hagfish. The VLRA and VLRC presented on the surface of lymphocyte-like cells exhibit gene expression and functional resemblance with the αβ and γδ subunits of TCRs, respectively. By contrast, the VLRB, similar to the BCR in mammals, is expressed on the cell membrane and is secreted into the serum as humoral agglutinins, and is a key component of the humoral immune response in lamprey and hagfish with respect to Ag recognition (3, 5, 6).

VLRB, like VLRA and VLRC, consists of the following domains: a signal peptide, an LRR N-terminal cap (LRRNT), the first LRR, multiple variable LRR modules (LRRVs), an end of variable LRR, a connecting peptide, an LRR C-terminal cap (LRRCT), an invariant threonine-proline rich region (stalk), and a hydrophobic tail (2). Many researchers have indicated that VLRBs, composed of a single polypeptide chain, could be useful for gene modification in vitro, such as redesigning the N-terminal capping motif by swapping of the LRRCT region and modular evolution of the binding interface (7–9). Through crystallographic analysis of the interaction between VLRs and Ags, it has been demonstrated that the LRRVs and LRRCT domain can recognize and bind to specific Ags (10–12). Collectively, we hypothesized that modification of the LRRVs and LRRCT module might improve the Ag-binding ability to a specific Ag.

In lamprey, VLRBs, like Igs, are expressed on the lymphocyte-like cells through a GPI–anchor domain and secreted to the serum (5). Secreted VLRBs are present as pentamers or tetramers conjugated to a C-terminal cysteine rich tail, which are structurally similar to IgM (13, 14). Several studies have demonstrated that Ag-specific VLRB proteins are produced in lamprey in response to particular Ags, such as Brucella abortus, sheep RBCs, Bacillus anthracis, and erythrocytes, and secreted VLRB proteins have a similar function to Abs in jawed vertebrates (4, 15–20). It has been shown that Ag-specific agglutinins and immunological memory can be generated by the adaptive immune system of lamprey (15, 21, 22). However, only a few studies have been conducted on hagfish VLRB with respect to its structure and functional similarities with lamprey. Recently, hagfish monoclonal VLRBs have been developed against viruses such as avian influenza (AI) virus and viral hemorrhagic septicemia virus (VHSV) (13, 23).

Nervous necrosis virus (NNV), also known as Betanodavirus, causes viral encephalopathy and retinopathy, which is an infectious disease responsible for high levels of mortality in more than 40 species of marine and freshwater fish. Infection by this virus leads to abnormal behavior and visual dysfunction, such as spiral swimming along with extensive vacuolation and neuronal degeneration (24–26). Studies cited in the previous sentence have mAbs from seabream (Sparus aurata) and giant grouper (Epinephelus lanceolatus), but our mAb against NNV was produced from hagfish (Eptatretus burgeri). In the current study, we produced a recombinant monoclonal VLRB from hagfish immunized with NNV, in which we describe the characteristics of this VLRB (e.g., the specificity and binding ability to NNV). In this study, an Ag-specific VLRB Ab was produced from screening system using C4-binding protein (C4bp) domain derived from human C4bp that plays a crucial role in the oligomerization of seven α-chains to form a heptameric complex (27). We tried to change the binding ability of this Ab through modular engineering of the LRRCT and LRRV and by doing so produced a VLRB that could potentially be used as an alternative neutralizing Ab against Betanodavirus infection.

Materials and Methods

Cells and viruses

Human embryonic kidney (HEK) 293F cells (Life Technologies) were maintained in high glucose DMEM (Life Technologies) containing 10% FBS at 37°C with 5% CO₂. E11 cells (catalog no. 01110916; European Collection of Authenticated Cell Cultures, U.K.) were cultured at 25°C in Leibovitz (L-15) medium supplemented with 1% penicillin–streptomycin and 10% FBS. For the preparation of the virus stock, confluent epithelioma papulosum cyprinid cells and E11 cells were infected with the VHSV and NNV, respectively, at a multiplicity of infection of 1. After 1 h, the inoculum was removed and L-15 medium containing 2% FBS was added. The viral supernatant was harvested once an extensive cytopathic effect (CPE) was observed and the cellular debris was removed by low speed centrifugation. Virus titers were determined based on 50% of the tissue culture infective dose (TCID₅₀). The AI viruses were prepared as previously described (23).

Hagfish immunization

Inshore hagfish (Eptatretus burgeri), ∼20–30 cm in length, were purchased from commercial fishermen (Bogyeong Hagfish Serves) and maintained in tanks at 14–15°C. The fish anesthetized by immersion in 3-aminobenzoate methanesulfonic acid (0.1 g/L; Sigma-Aldrich) were i.p. injected with 30 μg of NNV in 0.67× PBS, four times every 2 wk. Peripheral blood was collected and diluted with 0.67× PBS containing 10 mM EDTA. These blood samples were layered onto a 28% Percoll (GE Healthcare) gradient and centrifuged at 400 × g for 20 min at 4°C. Hagfish leukocytes obtained from the centrifugation were stored at 4°C for RNA extraction.

Cloning of VLRB cDNA library

A partial VLRB fragment library from LRRNT to the stalk region was amplified from total RNA of the NNV-immunized hagfish using primers LRRNT Sfi I forward and Stalk Sfi I reverse (LRRNT Sfi I forward: 5′-AGGCCACCGGGGCCTGTCCTTCACGGTGTTCCTG-3′, Stalk Sfi I reverse: 5′-TGGCCCCAGAGGCCCGCGTTCATGACACGGCCGA-3′). The PCR product obtained was digested with Sfi I and ligated into the Sfi I sites of plasmid pKepta/ccdB containing C4bp oligomerization domain derived from human C4bp (hC4bp) (GenBank accession no. NM_000715, https://www.ncbi.nlm.nih.gov/nuccore/NM_000715) (14, 27). To construct the plasmid encoding the monomeric VLRBs, VLRB genes were amplified with respective primers, LRRNT Sfi I forward and Stalk stop Sfi I reverse, including a stop codon inserted before the C4bp domain (LRRNT Sfi I forward, Stalk stop Sfi I Reverse: 5′-TGGCCCCAGAGGCCCTCAGCGTTCATGACACGGCCGA-3′). Amplified genes were cloned into the Sfi I sites of pKINGeo/ccdB. All underlined sequences represent the restriction enzyme sites mentioned in the primer names.

Transfection

The constructed plasmids were purified using DNA spin mini-prep kits (iNtRON Biotechnology) and quantified using a NanoDrop spectrophotometer. For transfection, HEK 293F cells were seeded into 96-well or 24-well plates, grown to 90% confluence, and transfected with the plasmids using Lipofectamine 2000 (Invitrogen Life Technologies) according to the manufacturer’s instruction. After 4 h, the DNA–lipofectamine complexes were replaced with DMEM containing 2% FBS. After 72 h, each supernatant was harvested and centrifuged for removal of cells and debris.

Screening of the VLRB library

To screen for NNV-specific recombinant VLRBs, 200 ng/well of NNV as well as VHSV (used as a negative control) were coated onto 96-well plates (Corning) and incubated overnight at 4°C. The Ags were washed with 1× TBST (10 mM Tris-HCL, 150 mM NaCl, 0.5% Tween 20 [pH 8]) and blocked with 5% skimmed milk in 1× TBST for 1 h at room temperature (25°C). Supernatants harvested at 72 h posttransfection were incubated for 1 h at room temperature (25°C). Binding of recombinant VLRBs to the Ag was detected with a mouse anti-VLRB IgG1 (11G5) diluted in 5% skimmed milk, followed by HRP-conjugated goat anti-moues IgG, after which 100 μl/well of developing buffer containing 42 mM 3,3′,5,5′-tetramethylbenzidine and 1% H₂O₂ was added for 20 min at room temperature, and 50 μl/well 1 M H₂SO₄ was then added to stop the reaction. The OD of the reaction was read at 450 nm using a microtiter plate reader.

Western blot analysis

The secreted recombinant VLRBs, harvested at 72 h posttransfection, were separated on an 8% SDS-PAGE gel under nonreducing and reducing conditions, and separated proteins were transferred to methanol-activated PVDF membranes. The membranes were blocked with 5% skimmed milk in 1× PBS containing 0.1% Tween 20 and then incubated with mAb 11G5 followed by HRP-conjugated goat anti-mouse IgG. Expression of the secreted recombinant VLRBs was visualized using a SuperSignal West Pico Chemiluminescent Substrate kit (Thermo Fisher Scientific).

Assessing recombinant VLRBs ability to bind to NNV by ELISA and immunoblotting

For the ELISA, 200 ng/well of NNV, viral hemorrhagic septicemia, and AI viruses were used to coat the 96-well plates (Corning). For immunoblotting, the viruses were treated with 5× SDS loading buffer and dotted on methanol-activated PVDF membranes. The Ags were blocked with 5% skimmed milk in 1× TBST for ELISA and 0.1% Tween 20 in PBS for immunoblotting. Culture supernatants containing VLRB were incubated for 1 h at room temperature, and the rest of the procedure was conducted as described above for ELISA screening and Western blot analysis. When a competitive ELISA was used, 100 times dilution of recombinant VLRB supernatants were preincubated with a variety of concentrations of NNV or VHSV and subjected to competitive reactions on NNV-coated plates. The remainder of the ELISA procedure was performed as described above for the ELISA in screening of the cDNA library section.

Mutation of the LRRCT and LRRV domain

To swap the LRRCT domain, the Ag-binding domain was amplified from VLR76 with primers (LRRNT Sfi I forward and LRRCT library reverse: 5′-GCTGGTCAGGCGATCAAA-3′). The LRRCT domains library was amplified from VLRB cDNA with primers (LRRCT library forward: 5′-TTTGATCGCCTGACCAGC-3′, and Stalk Sfi I reverse). The resulting products were constructed by overlapping PCR using primers LRRNT Sfi I forward and Stalk Sfi I reverse, and the assembled fragment was cloned into plasmid pKepta/ccdB. For modifying the first LRRV domain, a combined library was constructed using an overlapping PCR with the following mutagenic primer (LRRV module 1 mutagenic primer reverse: 5′–ATGAGGAATTGACTGGAACTTGTTMNNMNNCAGMNNCAGAAATGTGAGACTGGTAAGTTTATCAAACACTCCACTAGGAAGAGACTGCAGCTTATTTACATG-3′) in the reverse direction. The resulting amplicons were cloned into plasmid pKepta/ccdB. The subsequent transfection and screening processes were described above.

Immunocytochemistry

The E11 cells were grown to 80% confluence on an eight-chamber slide incubated for 24 h and infected with 1× 10² TCID₅₀ of NNV. After 1 h absorption, the infected cells were washed with 1× PBS and cultured in Leibovitz (L-15) medium supplemented with 2% FBS until the appearance of a CPE. When an extensive CPE was observed, the infected cells were fixed with 4% paraformaldehyde in 1× PBS, blocked with 0.1% BSA in 1× PBS for 30 min, and incubated with the recombinant VLRB supernatant for 1 h, followed by mAb 11G5 for 1 h. Immunofluorescent analysis was performed by incubating the cells with FITC-conjugated AffiniPure goat anti-mouse IgG for 30 min. Cells were then stained with DAPI for 15 min. The slides were mounted with Vectashield H-1000 (Vector Labs), and the images were examined under the Olympus FV 1000 fluorescence microscope (Olympus).

Flow cytometry

The E11 cells were grown to 90% confluence in a 24-well plate and infected with 1 × 10² TCID₅₀ of NNV. When an extensive CPE was observed, the infected cells were harvested and blocked with 0.1% BSA in 1× PBS for 30 min. Cells were then incubated with the recombinant VLRB supernatant for 1 h and mAb 11G5 for 1 h, followed by FITC-conjugated AffiniPure goat anti-mouse IgG for 30 min. An inhibition test was performed in which NNV was preincubated with the recombinant VLRB supernatant for 1 h, and then used to infect cells. After 48 h, the infected cells were incubated with mouse mAb 3A11, previously produced in our laboratory against NNV, for 1 h, followed by FITC for 30 min. The cells were washed twice with 1× PBS between each step, resuspended with 1× PBS, and analyzed with a FACSCalibur (BD Biosciences).

Infection study

The recombinant VLRB supernatant was purified using the Strep-Tactin XT Superflow high capacity (IBA Lifesciences), followed by construction of the plasmid encoding the VLRB fragment into pStrepTwin vector, which was prepared in our laboratory. In assessing the ability of the VLRBs to reduce virus infectivity, 30 olive flounders (Paralichthys olivaceus) were artificially infected by injecting them with 1 × 10⁷ TCID₅₀/ml of NNV i.m., and 30 more fish were injected with 1 × 10⁷ TCID₅₀/ml of NNV preincubated with 50 μg of recombinant VLRB Ab. Mortality was recorded for 28 d and the dead fish were collected for necropsy. Total RNA was obtained from kidneys, and the presence of the virus was confirmed by RT-PCR. The PCR products were obtained with respective primers, NNV coat protein (GenBank accession no. DQ116037, https://www.ncbi.nlm.nih.gov/nuccore/DQ116037) T2 region (forward: 5′-GGATTTGGACGTGGGACCAA-3′, reverse: 5′-CGAGTCAACCCTGGTGCAGA-3′) and T4 region (forward: 5′-GTGTCGGTGCTGTGTCGCTG-3′, reverse: 5′-CGAGTCAACCCTGGTGCAGA-3′).

Results

Selection of Ag-specific recombinant VLRBs

To produce NNV-specific VLRBs, the cDNA corresponding to the mature VLRBs from NNV-immunized hagfish were screened by ELISA. From a total of 600 clones, 2 clones (VLR67 and VLR76) were selected, which demonstrated specific binding ability against NNV and exhibited no significant response to VHSV and AIV. VLR2, used as a negative control, displayed no binding ability against the three viruses, NNV, VHSV, and AIV. The binding ability values of the NNV-specific VLRBs did not reach that of the anti-NNV–specific mAb, 3A11, however (Fig. 1A). We evaluated the multimeric structure and secretion levels of VLR67 and VLR76 by separating them by SDS-PAGE under nonreducing and reducing conditions and subjected them to Western blot analysis with the mouse anti-VLRB mAb, 11G5, detecting the invariant stalk region. The secreted VLRBs were observed between 35 and 40 kDa under reducing conditions, corresponding with their estimated molecular weights. Because conjugating the VLRB with C4bp leads to its oligomerization, multimerized VLRBs were detected at a size >170 kDa under nonreducing conditions. The negative VLRB, VLR2, with no reactivity to the viruses, expressed a similar pattern of reactivity to the NNV-specific VLRBs in Western blotting (Fig. 1B). Consistent results were obtained for immunoblotting analysis, which revealed that VLR67 and VLR76 recognized NNV but did not exhibit reactivity to other viruses, VHSV and AIV, whereas VLR2 showed no specific response to NNV, VHSV, and AIV (Fig. 1C). To compare the binding ability and expression levels between monomeric VLRBs and multimerized VLRBs, we deleted the C4bp oligomerization domain from VLR67 and VLR76 and transfected this to HEK 293F cells. The monomeric VLR67 and VLR76 migrated as monomeric units at approximately ∼ 40 kDa under both nonreducing and reducing conditions, and these monovalent VLRBs showed no reactivity to the Ags, NNV, VHSV, and AIV, as expected (Fig. 1D) (27, 28). These results indicated that the Ag binding ability of the multimerized form of VLRBs was much higher than that of monomeric VLRBs.

FIGURE 1.

Characterization of produced rAg-specific VLRBs. (A) Direct response of the recombinant VLRBs to NNV, VHSV, and AIV when screened by ELISA. (B) Expression levels of secreted VLRBs analyzed by Western blotting under nonreducing and reducing conditions. (C) Response of VLRBs in the viruses in immunoblotting, VLR2 served as a negative control. (D) Expression levels and Ag binding affinities of monomeric VLRBs (VLR67^mono, VLR76^mono) deleted C4bp oligomerization domain compared with multimeric VLRBs. VLR2^mono was used as a negative control and multivalent VLRBs served as positive control. Error bars indicate SEM for five analysis.

Binding ability maturation by LRRCT domain mutation

Although recombinant VLRBs specifically detected NNV, the binding ability of these VLRBs did not perform any better than that of the mouse mAb. A previous study had established that the LRRCT module of VLRs containing a highly variable insert contributed to the ability of VLRs to recognize Ags, and swapping of this module might lead to the enhancement of the Ag-binding ability. To improve the binding ability of recombinant VLRBs, we attempted some mutations of the LRRCT domain by amplifying the LRRCT repertoire from VLRB cDNA library using LRRCT library primers and assembling these fragments with the corresponding region of VLR76 (Fig. 2A). Of the 200 LRRCT mutant clones that were screened, VLR18 showed the high binding ability to NNV, so it was selected (Fig. 2B). Furthermore, VLR18 diluted 500 times had higher binding ability than VLR76 diluted two times (Fig. 2C), indicating that the mutation in the LRRCT was involved in the enhancing the binding ability of the VLR. The sequence of the construct generated was confirmed as correct through DNA sequencing (Supplemental Fig. 1).

FIGURE 2.

Maturation of the recombinant VLRBs by mutation of LRRCT domain in vitro. (A) Scheme for the LRRCT mutagenesis by amplifying the LRRCT repertoire from VLRB cDNA library using LRRCT library primers. (B) Specificity of the mutant VLRBs confirmed by ELISA in response to NNV, VHSV, and AIV. (C) ELISA response using various dilutions of mutant VLRBs compared with the original VLRB template.

Enhancement of the binding ability by LRRV module evolution

The LRR β-strands of the LRRVs comprising the inner concave surface were involved in binding to specific Ags. With the hypothesis that modification of the LRRV domain could increase the binding ability of VLR, we tried to amplify the VLRB cDNA library using LRRV mutagenic primer in the reverse direction using an overlapping PCR with VLR18 (Fig. 3A). In the modular evolution, out of a total of 200 clones, various clones showed higher binding ability than VLR18 (data not shown). Among these, VLR6, having only different LRRV1 sequence, displayed a significant increase in the binding ability and performed even better than VLR18 and was therefore selected (Fig. 3B). Furthermore, the sequence of VLR6 was identified by DNA sequencing (Supplemental Fig. 1). When all the recombinant VLRBs, including VLR76, VLR18, VLR6, and VLR2 (negative control), were subjected to SDS-PAGE, the secreted VLRBs migrated between 35 and 40 kDa under reducing conditions, and multimerized VLRBs were observed at a size >170 kDa under nonreducing conditions (Fig. 3C). In addition, VLR6 diluted 16 times had higher binding ability to NNV than VLR18 diluted two times (Fig. 3D), suggesting that the improved binding ability was attributed to the modification of the LRRV module. We separated secreted VLR76, VLR18, and VLR6 by SDS-PAGE and Western blotting with mAb 11G5 to examine their levels of expression, As expected, the secreted VLRBs were observed around 40 kDa under reducing conditions and higher than 170 kDa under nonreducing conditions (Fig. 3D). When immunoblotted against NNV, VLR76, VLR18, and VLR6 exclusively detected the virus and did not recognize VHSV or AIV (Fig. 3E). Moreover, to compare the binding ability of VLRB Abs and mouse mAb (3A11) produced in our laboratory, ELISA was performed with various concentration of NNV starting with virus concentrations (5 μg) in which all Abs were saturated and various dilution factor of VLRB Abs. As a result, VLR18 and VLR6 Abs showed higher binding ability than mouse Abs at the same concentration (Supplemental Fig. 2).

FIGURE 3.

Modular evolution of LRRV domain through overlapping PCR method using LRRV mutagenic reverse primer. (A) Schematic design of LRRV domain mutation established for modulating binding ability. (B) Specificity of LRRV modified form of VLRB in response to NNV, VHSV, and AIV. (C) Expression levels of VLRBs in SDS-PAGE and Western blotting under nonreducing and reducing conditions. (D) Binding ability with various dilutions of Ag based on ELISA. (E) Immunoblotting of the recombinant VLRBs was analyzed to various viruses, NNV, VHSV, and AIV.

Detection of the progeny NNV from generated VLRBs

Because it was demonstrated that the recombinant VLRBs generated are able to detected NNV by ELISA, Western blotting, and immunoblotting, we examined whether the recombinant VLRBs could recognize NNV in infected cells in vitro. Progeny virus grown in E11 cells infected with NNV and screened by immunocytochemistry and flow cytometry using the recombinant VLRBs as primary Abs. In the NNV-specific VLRBs groups, cells were positively stained with FITC, shown by green fluorescence (Fig. 4A), whereas no fluorescence staining was seen in negative control, VLR2. Moreover, NNV-infected cells treated with NNV-specific VLRBs showed a higher FITC population in flow cytometry, compared with very few FITC stained cells with the negative control (Fig. 4B).

FIGURE 4.

Detection of progeny virus in E11 cells infected with NNV. (A) Immunofluorescence staining of E11 cells infected with NNV captured by VLRBs. Virus-infected E11 cells were stained with recombinant VLRBs followed by FITC-conjugated goat anti-mouse IgG (green), and the overlay images (merge) were shown with DAPI (blue). Scale bars, 400 μm. (B) E11 cells infected with NNV were captured by recombinant VLRBs and then stained with 11G5 followed by FITC-conjugated goat anti-mouse IgG and analyzed by flow cytometry analysis. The cells stained with the recombinant VLRBs were shown by shaded area, and VLR2 served as negative control. Mock cells were incubated without any recombinant VLRB shown by black line.

Inhibition of NNV infection in E11 cells using recombinant VLRB

To assess if the VLRBs generated against NNV were able to inhibit the proliferation of NNV infection in E11 cells, VLR6, diluted 100 times, was preincubated with various concentration of NNV and VHSV, and then these treated viral suspensions were added to the NNV-coated 96-well plates for competitive ELISA. Our results revealed that the binding ability to NNV was gradually reduced from a virus concentration of 0.1 μg/ml then to a basal level of 10 μg/ml, whereas no significant change in the binding ability to NNV was evident in the VLR6–VHSV mixture (Fig. 5A). Furthermore, E11 cells infected with NNV preincubated with VLR6 or VLR2 were then stained with mouse anti-NNV mAb, 3A11, and analyzed by flow cytometry. The cells infected with NNV preincubated with VLR2 showed an infection rate of ∼95.1%, whereas cells infected with virus preincubated with VLR6 exhibited an infection rate of only ∼57.5% (Fig. 5B), indicating that VLR6 has the capacity to reduce the infectivity of NNV in E11 cells by ∼37.6%, compared with VLR2.

FIGURE 5.

Functional assessment of an Ag specific VLRB toward NNV. (A) Competitive ELISA analysis for identifying the ability of VLRB to reduce virus infection to E11 cells. The binding ability of VLR6 to NNV and VHSV. VLR2 served as negative control. (B) Flow cytometry analysis of E11 cells infected with NNV when stained with VLRBs. The ability of NNV to infect E11 cells after treatment with VLR6 or VLR2 used as a negative control. Mock represents E11 cells infected with NNV and incubated with no VLRB.

The capacity of the recombinant VLRB to reduce the infectivity of NNV

To demonstrate whether the injected flounders were infected with the virus, electrophoretic profiles of NNV coat protein amplicons were obtained from mRNA extracted from kidneys randomly selected in each group (Fig. 6A). The survival rate showed a significant difference between the group infected with just the virus and the group preincubated with VLR Ab. In the NNV-infected group, fish started to die 4 d postinfection (dpi), and only 10 of the fish survived at 22 dpi. In contrast, in NNV preincubated with the recombinant VLRB (VLR6) group, fish started to die 9 dpi, and the survival rate was 67% (Fig. 6B).

FIGURE 6.

In vivo assessment of the recombinant VLRB to reduce the infectivity of NNV. (A) Three flounders were randomly selected from both groups, including the NNV-infected group and NNV preincubated with the purified VLR6. The mRNAs encoding NNV coat protein were amplified by NNV coat protein T2 (870 bp) and T4 (420 bp) region specific primers. (B) Thirty olive flounders were infected by i.m. injection with NNV, and 30 more fish were injected with NNV preincubated with the purified VLR6. Mortality rates reached 67% in the NNV infected group (dashed line) and 33% in the NNV preincubated with the purified VLR6 group (solid line).

Discussion

The diversity of Ag recognition sites is the result of gene conversion-like mechanism in VLRs, similar to the gene rearrangement observed in Igs. V(D)J recombination is a unique mechanism that occurs in higher vertebrates, whereas VLR gene maturation is generated by the distinct mechanism based on LRR-encoding genes in jawless vertebrates, lamprey, and hagfish (10). As with Ig-based Abs capable of recognizing foreign material, LRR-based VLRs also play a significant role in Ab–Ag interaction and are deemed to be a good alternative for Ig-based Abs in the fields of biotechnology and biomedicine (4, 5, 29). In the current study, we successfully produced Ag-specific monoclonal VLRBs from hagfish, screened from numerous transcripts encoding mature VLRBs after immunostimulation of the fish with NNV, and demonstrated the specificity and binding ability of these recombinant monoclonal VLRBs to the virus by the addition of C4bp oligomerization domain, which might have caused the increased interactions with stability compared with dimeric mouse Abs (30).

The Ag binding properties of the NNV-specific recombinant monoclonal VLRBs obtained from hagfish were compared with that of an anti-mouse NNV-specific mAb (3A11), previously developed in our laboratory using ELISA and immunoblotting. Although the mouse mAb (3A11) showed higher binding ability in ELISA than that of the recombinant VLRBs, it was not possible to know which viral protein reacted in the Western blot analysis. In hagfish, two Ag-binding regions, LRRCT and LRRV, are capable of detecting specific Ag; thus, we attempted to execute modular swapping of the hypervariable loop of LRRCT domain to acquire VLRBs possessing improved binding ability. From a relatively small library of 200 clones from modular engineering of the LRRCT domain, a mutant possessing 16 different sequences and one more amino acid, was found to have high Ag binding ability for NNV, ∼250-fold higher than the original template. This result is consistent with previous observation that the residue substitution in the hypervariable loop of LRRCT, which plays an important role in Ag recognition, contributes to the increased ability of Ag binding (7, 13). Moreover, because the mutagenesis of the other Ag-binding sites, LRRVs, can be used for binding ability maturation (8), the hypervariable region in the LRRV domain was also modified by modular evolution. In the amino acid sequence of VLR18, tyrosine, known to play an important role in Ag binding in terms of protein–protein interaction because of exposure to concave region, is present in LRRV2 (10). Thus, the modular mutations were made to the sequences of LRRV1 and LRRV3, and the mutant VLR (VLR6) was obtained from the mutation of LRRV1. Unfortunately, the modular mutation on LRRV3 did not enhance any Ag binding ability. Among the resulting clones, the best-performing clone, with two different sequences, had a binding ability ∼8-fold higher than the original template selected. One of the advantages of VLRs, determined from previous studies, was that the single chain of VLR can be altered to improve binding ability through somatic gene modification in the LRRNT and LRRCT domains (7, 9). Our results confirmed this and we found that a few number of residue substitutions in the specific VLR region can alter the binding ability from low to high binding (7).

The ability of Abs to neutralize virus is usually confirmed by the lack of a CPE in infected cells, and treatment of the neutralizing Abs can decrease the numbers of cytopathic cells, compared with the untreated conditions (25, 31–33). In general, the capacity of Abs to neutralize a virus in vitro is 1 × 10² TCID₅₀ in a neutralization test; however, the VLRBs produced in this study were able to inhibit the viral infection at <1× 10² TCID₅₀ (data not shown). We demonstrated that when the VLRBs were preincubated with the virus and then added to the cells, the infectivity of the virus was significantly reduced, compared with the negative control in both ELISA and flow cytometry. This could be because the VLRBs prevents attachment of the virus to the cells or they block receptor recognition on the target cells.

In conclusion, we developed NNV-specific VLRBs using a library screening system based on the adaptive immune system of hagfish. We enhanced the specificity and binding ability of these to their target Ags based on modifications to the VLRBs’ DNA sequence. More importantly, these VLRBs offer great potential to be used as alternative Abs to mammalian Igs and the ability of the VLRBs to reduce virus infectivity can be a useful development new treatments for biomedical purposes.

Disclosures

The authors have no financial conflicts of interest.