Introduction

Noroviruses (NoVs), which belong to the family Caliciviridae, are the foremost cause of acute non-bacterial gastroenteritis globally [1]. NoV virions are approximately 30–38 nm in diameter and comprise 180 monomers of the major capsid protein VP1 and one or several monomers of the minor capsid protein VP2 [2]. These capsid proteins encapsulate one copy of the single-stranded, positive-sense, and polyadenylated RNA genome (length: approximately 7.5–7.7 kb). The human NoV genome contains three open reading frames (ORFs). ORF1 encodes a polyprotein that is essential for viral replication. ORF2 and 3 encode VP1 and VP2, respectively. When expressed in vitro, VP1 can self-assemble into virus-like particles (VLPs), and it has been used extensively to study the structure, immunogenicity, and virus–host interactions of NoV [3,4,5,6].

VP1 can be structurally divided into the N-terminal shell (S) domain and the C-terminal protruding (P) domain. The P domain is further divided into P1 and P2 domains, with P1 separated by P2 into two subdomains (P1-1 and P1-2). The P2 domain sequence is surface-exposed and highly variable, and it is responsible for interactions with histo-blood group antigens (HBGAs), a host susceptibility determinant for infections [7]. Furthermore, the C-terminus of the P1-2 domain is essential for binding HBGAs [8]. The N-terminal sequence of the S domain appears to be crucial for determining the sizes of the particles assembled rather than the assembly itself [9].

Our previous study demonstrated that the N-terminus of a GII.4 NoV VP1 could tolerate foreign sequences without any effect on its assembly [10]. However, extensive cleavage was observed in all expressed mutant VP1 proteins. Because wild-type GII.4 VP1 proteins also exhibited cleavage, the cleavage of mutant VP1 proteins was considered to be associated with the instability of GII.4 VP1 proteins. Here, we further investigated the effects of N-terminal fusions using the GII.6 NoV VP1 backbone, which exhibits minimal cleavage and produces a more homogeneous distribution of sizes of the assembled VLPs. Our data can support the mechanistic study of NoV assembly and cleavage.

Materials and methods

Chimeric protein design

The predicted VP2 amino acid (aa) sequence of a Sydney-2012-like GII.4 strain (GenBank accession number KF306214) was used as the fusion sequence. A portion of the aa sequence predicted to contain a linear B-cell epitope was selected. Short peptide sequences (9-74 aa) of increasing length were added to the N-terminus of GII.6 VP1 (Hu/GII.6/Ehime120246/2012/JP, GenBank accession number AB818400) for fusion protein design. Figure 1 shows a diagram of the constructs and the corresponding short peptide sequences added to the GII.6-AB VP1 N-terminus.

Figure 1.
figure 1

Schematic representation of the parental and VP1 fusion protein constructs.

Full size image

Gene synthesis and generation of recombinant baculoviruses

The genes encoding the fusion proteins were optimized based on the codon usage frequency in Spodoptera frugiperda (Sf9) cells, synthesized, and inserted into the pFastBac-Dual vector (Invitrogen, Carlsbad, CA, USA) under the control of the polyhedrin promoter. To initiate translation, an ATG codon was added to the N-terminus of all fusion-protein-coding sequences. The resultant pFastBac-Dual plasmids containing target sequences were used to produce recombinant bacmids as described elsewhere [10]. The bacmids were purified and then used to transfect Sf9 cells to generate recombinant baculoviruses.

Expression and purification of VLPs

Sf9 cells cultured in serum-free medium (Suzhou World-Medium Biotechnology Co., Ltd., Suzhou, China) in vented conical culture flasks were infected with baculoviruses at a rotation speed of 130 rpm. The culture medium was generally harvested 5-7 days after infection. The VLPs in the culture medium were purified as described previously [11].

SDS-PAGE and Western blot (WB) analysis

For SDS-PAGE analysis, purified VLPs or trypsin-digested VLPs were boiled and loaded onto a discontinuous 8-16% precast gel (SurePAGETM, Genscript, China). After separation, the proteins were stained with Coomassie blue stain. For WB analysis, the separated proteins were transferred to a nitrocellulose membrane and detected by the addition of a DAB color-developing agent after sequential incubation with mouse anti-GII VP1 monoclonal antibody 10G7 at 1 µg/mL and horseradish peroxide (HRP)-conjugated goat anti-mouse IgG antibodies. mAb 10G7, which was produced by the hybridoma technique by immunizing mice with GII.6 VP1 proteins (GenBank accession number AB818400), specifically recognizes a linear epitope located in the P1-2 domain of the GII VP1 protein.

Mass spectrometry analysis

Samples were analyzed by reversed-phase liquid chromatography-mass spectrometry (LC-MS) using an LC-30A HPLC System (Shimadzu Corporation, Kyoto, Japan) coupled to a Triple TOF™ 4600 mass spectrometer (AB SCIEX, USA). The flow rate for separation using an ACQUITY UPLC BEH300 C4 column (2.1 mm × 50 mm, C4, 1.7 μm) (Waters, MA, USA) was 0.4 mL/min with mobile phase A (aqueous solution of 0.1% formic acid) and mobile phase B (0.1% formic acid in acetonitrile). The following solvent gradient system was used: 0–2.5 min, 5–15% B; 2.5–5 min, 15–95% B; 5–7.5 min, 95% B; 7.5-7.6 min, 95–5% B; 7.6–8 min, 5-95% B; 8–8.1 min, 95–5% B; 8.1–10 min, 5% B. The peak chromatograms were acquired in the positive mode, and the precursor ion m/z ratio within the range of 800-3,800 was analyzed.

Transmission electron microscopy (TEM) analysis

The assembly of VLPs was confirmed by TEM (JEM-1400, JEOL, Japan) at 100 kV after negative staining using phosphotungstic acid.

Results

N-terminally modified GII.6 VP1 fusion proteins exhibit different degrees of cleavage

To further investigate the effects of adding foreign sequences to the NoV VP1 N-terminus, a more stable GII.6 VP1 was selected (designated as GII.6-AB). Using the recombinant baculovirus expression system, GII.6-AB VP1 proteins were expressed. These proteins exhibited a single protein band, whereas in most cases, two bands were observed for GII.4 VP1 proteins [11]. The constructed fusion proteins were expressed using the recombinant baculovirus expression system, purified using CsCl gradient density centrifugation, and analyzed using SDS-PAGE (Fig. 2A) and WB (Fig. 2B) analysis. SDS-PAGE analysis revealed that, except for the GII.6-AB/N9 fusion protein, all of the fusion proteins (GII.6-AB/N19, GII.6-AB/N29, GII.6-AB/N39, GII.6-AB/N49, and GII.6-AB/N74) exhibited two or more visible bands. The GII.6-AB/N49 fusion proteins displayed the largest protein bands. The observed molecular weights (MWs) of GII.6-AB/N29 and GII.6-AB/N39 appeared to be smaller than or similar to that of GII.6-AB/N19. In the WB analysis, two or more reactive bands were present for all fusion proteins, which is consistent with the SDS-PAGE results.

Figure 2.
figure 2

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and western blot analysis. Similar amounts of CsCl-density gradient-purified wild-type or fusion VP1 proteins were loaded on precast gels for SDS-PAGE (A) and WB analysis (B). For A, the samples were laoded in the following order: lane 1, GII.6-AB VP1; lane 2, trypsin digested GII.6-AB VP1; lane 3, GII.6-AB/N74 VP1; lane 4, trypsin digested GII.6-AB/N74 VP1; lane 5, GII.6-AB/N49 VP1; lane 6, trypsin digested GII.6-AB/N49 VP1; lane 7, GII.6-AB/N39 VP1; lane 8, trypsin digested GII.6-AB/N39 VP1; lane 9, GII.6-AB/N29 VP1; lane 10, trypsin digested GII.6-AB/N29 VP1; lane 11, GII.6-AB/N19 VP1; lane 12, trypsin digested GII.6-AB/N19 VP1; lane 13, GII.6-AB/N9 VP1; lane 14, trypsin digested GII.6-AB/N9 VP1. For B (10G7), samples in lane 1-14 were added in the same order as in A, except that in lane 15, supernatants from wild-type baculovirus infected sf9 cells was added. Purified mAb 10G7 was used at 1 μg/mL.

Full size image

The cleavage site is most probably located in the N-terminus.

To characterize the cleavage patterns of fusion proteins, WB analysis was performed using a monoclonal antibody (mAb) that recognizes the P1-2 domain of the GII VP1 protein. A primary reactive band corresponding to the P domain was observed for all trypsin-treated wild-type VP1 and fusion proteins (Fig. 2B). The reactive bands for all wild-type VP1 and fusion proteins had similar MWs, which indicated that the observed cleavage was most probably located in the N-terminus.

Mass spectrometry analysis shows the presence of similarly cleaved products

LC-MS was performed to better characterize the observed cleaved products and determine how they were generated. The predicted MWs of the GII.6-AB, GII.6-AB/N9, GII.6-AB/N19, GII.6-AB/N29, GII.6-AB/N39, GII.6-AB/N49, and GII.6-AB/N74 proteins were 60006.17, 61158.47, 62226.71, 63201.69, 64225.84, 65443.26, and 67955.89, respectively. Based on these results (Fig. 3), except for the wild-type GII.6-AB VP1 protein, which displayed highly consistent values (mass determined through LC-MS was nearly identical to the predicted MW), the masses of all other fusion proteins were significantly different from their predicted MWs, indicating extensive cleavage, most probably at the N-terminus. However, all fusion and wild-type proteins had cleaved products with similar or nearly identical masses, such as fragments with masses of 19887.2–19916.1 and 59658.6–59751.4, respectively. The presence of fragments with similar masses, regardless of the length of the added N-terminal sequences, suggests a possible common processing mode. Of note, fragments with masses corresponding to those of the nearly full-length wild-type VP1 protein predominated among all fusion and wild-type VP1 proteins, except for GII.6-AB/N49 proteins. The inconsistency between SDS-PAGE and LC-MS results for GII.6-AB/N49 proteins might be a result of their instability during LC-MS analysis.

Figure 3
figure 3

LC-MS analysis. CsCl-density gradient-purified VLPs were analyzed by LC-MS.

Full size image

Trypsin treatment leads to more-extensive cleavage

Trypsin digestion of GII.6-AB VP1 proteins resulted in two dominant protein bands that could be recognized by the mAb (Fig. 3). The smaller and larger fragments had masses of 26286.3 and 31835.8, respectively, which is consistent with the results of our previous study [10]. Similar fragments were observed for all trypsin-digested fusion proteins (Fig. 2A and B).

All fusion proteins assembled into VLPs

TEM was performed to determine the assembly status of the expressed fusion proteins. According to the TEM results (Fig. 4), all fusion proteins successfully assembled into VLPs with sizes similar to those of the wild-type GII.6-AB VP1-assembled VLPs, except for the fusion protein GII.6-AB/N49, which exhibited particles or capsomeres of various sizes.

Figure 4
figure 4

Tranmission electron microscopy (TEM) analysis. CsCl-density gradient-purified VLPs were negatively stained and observed by TEM. Scale bar, 200 nm.

Full size image

Discussion

In a previous study, we investigated the effects of N- or C-terminally modified fusion proteins based on the GII.4 NoV VP1 protein backbone [11]. Cleavage is commonly observed for expressed GII.4 VP1 proteins. The instability of the N-terminal sequences of expressed NoV VP1 has been observed in multiple studies [12, 13]. The expressed fusion proteins might be extensively cleaved because of the instability of the backbone proteins. To confirm the aforementioned association and to determine its effect on cleavage and assembly, N-terminal fusion proteins based on the GII.6-AB VP1 protein backbone were designed. In our unpublished expression experiments, unlike expressed GII.4 VP1 proteins, which always produced a doublet band, expressed GII.6-AB VP1 proteins in different batches mostly exhibited a single band, indicating a higher level of resistance to enzymatic cleavage.

In total, six fusion proteins were designed, successfully expressed, and purified. Surprisingly, all fusion proteins exhibited extensive cleavage, indicating that the cleavage was independent of the backbone proteins. By using an mAb targeting the C-terminus, we were able to conclude that the cleavage probably occurred in the N-terminus. Further evidence supporting our hypothesis is that C-terminally modified VP1 fusion proteins tended to be stable [14]. Extensive cleavage did not affect the assembly of these fusion proteins, and this is consistent with another report showing that mutants with N- or C-terminal deletions do not affect particle assembly [15]. GII.6-AB/N9, GII.6-AB/N19, GII.6-AB/N29, GII.6-AB/N39, and GII.6-AB/N74 protein-assembled VLPs had similar sizes and shapes, while GII.6-AB/N49 protein-assembled VLPs were inhomogeneous in both size and shape. A comparison of the gel electrophoresis patterns of all of the fusion proteins revealed that the GII.6-AB/N49 protein-assembled VLPs contained proteins with larger MWs. If the sizes and shapes of the assembled VLPs differed because of the presence of larger proteins, then cleavage must have occurred before VLP assembly, or GII.6-AB/N74 and GII.6-AB/N49 protein-assembled VLPs should have similar sizes and shapes. The significant differences in LC-MS results observed between the GII.6-AB/N49 proteins and other fusion proteins indicate that the GII.6-AB/N49 proteins were highly unstable.

Based on the aforementioned results, we propose that the N-terminus of GII.6 NoV VP1 can tolerate foreign sequences of a certain length without affecting its assembly into regular VLPs. The observed cleavage might be related to instability of the fusion proteins under the conditions employed. Additional stable short peptides can be used to determine the maximal length of sequences that can be appended without affecting VLP assembly and cleavage. Furthermore, all of the wild-type and fusion proteins had fragments with masses (59658.6–59750.7) close to the predicted MW of the full-length wild-type VP1 protein. These masses suggest that approximately two or three amino acids had been lost. N-terminal protein sequencing and peptide mass fingerprinting (PMF) were not performed, so the exact cleavage sites were not determined. The mechanisms underlying the formation of these fragments and whether they were associated with the cleavage of VP1 proteins need to be investigated further. Here, we provide detailed data about cleavage patterns and characteristics of assembled VLPs for N-terminally modified fusion proteins designed based on the GII.6 NoV VP1 backbone. The results might be useful for further study of the assembly and cleavage mechanisms of NoV.