Competent E. coli DH5α and BL21 (Thermo Fisher, Shanghai, China) were used for recombinant plasmid construction and protein expression, respectively. Oligonucleotide “TB” coding for the peptide sequence of Leu-Val-Pro-Arg-Gly-Ser, was synthesized by Suzhou GENEWIZ Bio-Technology, Co., Ltd. Then, the artificially synthesized inaQn-TB sequence was inserted into cloning vector pUC57 to create pUC57-inaQn-TB. After re-digestion of plasmids pET28a, pET28a-P (GII.4) and pUC57-inaQn-TB (Table 1), the inaQn-TB and P (GII.4) fragment were inserted into pET-28a to create recombinant plasmid pET28a-inaQn-TB-P (GII.4). Similarly, pET28a-inaQn-TB was constructed from pET28a-inaQn for use as a negative control. All the recombinant plasmids were used to transform bacteria E. coli BL21.

Escherichia coli BL21 was transformed with recombinant plasmids pET28a-inaQn-TB-P (GII.4) or pET28a-inaQn-TB. The recombinant bacteria were each cultured in Luria-Bertani (LB) (0.5% yeast extract, 1.0% trypton, and 1.0% NaCl) liquid medium containing 100 μg/mL kanamycin, at 37°C with shaking (150 rpm) for overnight. One hundred microliters (100 μL) of each overnight culture were added to aliquots of new LB medium (10 mL, w/100 μg/mL kanamycin), and were cultured at 37°C with shaking (150 rpm) until OD₆₀₀ reached ∼0.6. Isopropyl β-D-1-thiogalactopyranoside (IPTG; Merck, Germany) was added to each OD₆₀₀ ∼0.6 culture to a total concentration of 0.5 mM, and incubated at 26°C with shaking (120 rpm) for 12–16 h. The induced-and-expressed cultures were stored at 4°C for further use.

Recombinant E. coli BL21 expressing inaQn-TB-P in culture (as described above) was adjusted to an OD₆₀₀ of 1.0. The bacteria were isolated and washed twice with phosphate-buffered saline (PBS), then resuspended in digestion buffer (1.0 mL, 20 mM Tris-HCl and 150 mM NaCl, pH 8.0). In accordance with manufacturer guidelines for enzymatic activity, Bovine thrombin (Yeason, Shanghai, China) was added at 1: 2,000 (e.g., 2.0 U enzyme was added to 1.0 mg target protein) to each digestion reaction and incubated at 37°C for 3 h. The reaction was centrifuged at 4°C at 8,000 RCF for 5 min, and the supernatant was quantified for P protein by a commercial BCA assay kit (Beyotime, Shanghai, China). The protein was stored at -20°C for further use.

Recombinant E. coli BL21 strains containing plasmid constructs pET28a-inaQn-TB-P, pET28a-inaQn-P, and pET-28a (negative control) were induced with IPTG, washed twice and then resuspended in PBS. Surface-expressed P proteins from induced BL21-pET28a-inaQn-TB-P (GII.4) were released from bacteria by thrombin digestion as described in the previous section. For SDS-PAGE, the IPTG-induced bacteria were dissolved in 2× SDS-PAGE loading buffer (100 mM, pH 6.8 Tris-HCl, 4% SDS, 20% Glycerol, 0.2% Bromophenol Blue, 2% DTT), while the thrombin-released P proteins were dissolved in 5× SDS-PAGE loading buffer (250 mM, pH 6.8 Tris-HCl, 10% SDS, 50% Glycerol, 0.5% Bromophenol Blue, 5% DTT). Each sample was boiled for 5 min, and 10.0 μL from each was loaded and separated in a 12% SDS-PAGE gel, followed by staining with Coomassie Blue R250 (Beyotime, Shanghai, China). Western Blotting was conducted as described in a previous report (Wang et al., 2008). The antibody against GII.4 HuNoV recombinant viral capsid protein (1: 5,000; Niu et al., 2015), and peroxidase-conjugated goat anti-mouse IgG (H+L, 1: 3,000; Yeasen, Shanghai, China) were used as primary and secondary antibodies in Western Blotting as described in our previous publication (Niu et al., 2015). The 3, 3′-diaminobenzidine (Fedbio, Wuhan, China) was used as chromogenic substrate.

Thrombin-released P proteins were visualized by EM as previously described by Tan and Jiang (2005b) with minor modifications. Thrombin-released P protein was quantitated by BCA assay kit (Beyotime, Shanghai, China). Twenty microliters of thrombin-released P protein (200 μg/mL) was loaded onto the support grid (Mainstream, Shanghai, China) and allowed to bind for 2 min. The remaining solution was wicked away with wedges of Whatman filter paper (General Electric, Beijing, China). Twenty microliters of 1.0% phosphotungstic acid (Ted Pella, Redding, CA, United States) was added to the support grid and allowed to stain for 2 min. The remaining stain was wicked away with wedges of Whatman filter paper. The support grid was washed with 20 μL of ddH₂O and dried immediately by wicking with Whatman filter paper. The support grid was examined under an H-7650 microscope (Hitachi, Japan) at 50,000× magnification and 80 kV.

Thrombin-released P protein was serially-diluted in digestion buffer to a series spanning 2 to 200 μg/mL. One hundred microliters of each dilution were added into immunoassay wells (Nunc Immuno Module, VWR, San Francisco, CA, United States) and incubated at 4°C for overnight. The coated wells were washed three times with PBS, then blocked with 1.0% bovine serum albumin (“BSA”; Yeason, Shanghai, China) at 37°C for 1 h. The blocked wells were washed three times with 120.0 μL of PBS-T (PBS containing 0.1% Tween-20, pH 7.2). One hundred microliters of 1.0 mg/mL Type III porcine gastric mucin (“PGM”; Sigma, St. Louis, MI, United States) was added to each well and incubated at 37°C for 30 min. The optimal dilutions of anti-type A HBGA monoclonal antibody BG2 (Covance, Emeryville, CA, United States) and the secondary antibody, a peroxidase-conjugated goat anti-mouse IgG (H+L chains; Yeasen, Shanghai, China) were determined experimentally to be 1: 1,000 and 1: 3,000 in blocking buffer, respectively. Both antibodies were incubated at 37°C for 1 h respectively. In addition, the antibody against the recombinant viral capsid proteins from our previous report was also utilized as a primary antibody to detect P proteins (data not shown) (Niu et al., 2015). Wells were washed four times with PBS-T following each step. After incubation with 100.0 μL of 3,3′,5,5′-tetramethylbenzidine (TMB; Fedbio, Wuhan, China) in the dark for 10 min, the chromogenic reaction was halted using 50.0 μL of 2 mol/L H₂SO₄. The OD₄₅₀ values were measured using a Sunrise Microplate Reader (Tecan Sunrise, Switzerland). Cells transformed with recombinant plasmid pET28a-inaQn-TB were treated in the same way and used as negative controls. Samples were considered positive when the positive to negative (P/N) ratio was greater than 2.0.

IPTG-induced BL21 with recombinant plasmid pET28a-inaQn-TB-P (GII.4) were collected by centrifugation at 5,000 RCF for 5 min. The pellet was washed twice with PBS (pH 7.2) and adjusted to an OD₆₀₀ of 1.0. The washed cultures were incubated at 37°C for 30 min with PGM at final concentrations of 0.1, 0.2, 0.5, and 1.0 mg/mL. The solution of putative bacteria-PGM complex was washed with PBS containing 0.5‱ Tween-20 (pH 7.2) at least three times to remove any un-complexed PGM. The putative bacteria-PGM complex was digested by bovine thrombin (100 U/mL, Yeason, Shanghai, China) at 37°C for 3 h as described above. The supernatant containing the thrombin-released PGM-to-P-protein complex was collected by centrifugation at 4°C at 8,000 RCF for 5 min. Each immunoassay well (Nunc Immuno Module; VWR, San Francisco, CA, United States) was incubated with 100.0 μL of the supernatant at 4°C for overnight. Each well was washed three times with PBS (pH 7.2), blocked with 120.0 μL of 1.0% BSA in PBS at 37°C for 1 h, and then washed three times with PBS again. One hundred microliters of anti-type A HBGA monoclonal antibody BG2 (1: 1,000; Covance, Emeryville, CA, United States) was added to each well. Peroxidase-conjugated goat anti-mouse IgG (H+L chains, 1: 3,000; Yeasen, Shanghai, China) was used as the secondary antibody. All antibody incubation steps were performed at 37°C for 1 h. The wells were washed three times with 120.0 μL of PBS-T after each incubation step. Then, 100.0 μL of TMB (Fedbio, Wuhan, China) was added to each well. After incubating in the dark for 10 min, the chromogenic reaction was halted using 50.0 μL of 2 mol/L H₂SO₄, and the OD₄₅₀ values were measured. Bacteria with recombinant plasmids pET28a-inaQn-TB and pET-28a were also tested in the same way. Cells incubated without PGM were used as negative controls.

IBM SPSS statistics software (version 19) was used for statistical analysis. Each experiment was performed in triplicate (N = 3), which in turn was independently repeated three times (n = 3). One-way ANOVA was utilized for data analysis. Differences in means were considered significant when p < 0.05.

SDS-PAGE and Western Blot were used to characterize the expression of the fusion proteins and the thrombin-released P protein. The thrombin-released P protein, InaQn-P and InaQn-TB-P fusion proteins were expected to be segregated to positions roughly corresponding to 35, 58, and 70 kDa, respectively. All proteins were visualized by Coomassie Blue staining in SDS-PAGE (Figure 2A), and confirmed by Western Blot (Figure 2B). We found that the thrombin-released P proteins did not exhibit a significant background of bacterial proteins (Figure 2A).

The morphology of the thrombin-released P proteins was observed by negative-staining EM. The P protein was revealed to have a spherical structure of ∼15 nm (Figure 3).

A modified ELISA was developed to quantitate the type-A HBGA-binding capacity of thrombin-released P proteins. The binding capacity was presented by OD readings (Supplementary Table S1) and P/N ratios. A positive correlation was observed between OD₄₅₀ and the concentration of thrombin-released P proteins, until a plateau was reached at 128 μg/mL of the latter (Figure 4). The OD₄₅₀ P/N ratios of P-protein-bound-HBGAs vs. pET28a-inaQn-TB/BL21 background proteins were: 2.29, 2.58, 3.04, 3.43, 4.50, 5.02, 4.82, and 3.88 for thrombin-released P proteins at concentration of 2, 4, 8, 16, 32, 64, 128, and 200 μg/mL, respectively. The optimal concentration of P protein to bind HBGAs was 64 μg/mL. There was no significant binding of HBGA to the thrombin-digested expression product from pET28a-inaQn-TB/BL21, which does not express any HuNoV P protein domains (Figure 4). As shown in Figure 4, the P proteins could be directly collected by low speed centrifugation after thrombin treatment, and maintain its ability to bind to its receptors.

Co-incubation of viral receptor (HBGA) and induced recombinant bacteria resulted in the binding of HBGA to surface-displayed HuNoV P proteins. The bacteria-HuNoV P proteins-HBGAs complex could be isolated by a low-speed centrifugation step (5,000 RCF for 5 min). The HuNoV P protein-HBGA complex can be released and isolated from bacteria by thrombin-digestion, followed by low speed centrifugation (8000 RCF, 5 min), as indicated in Figure 1. A modified ELISA was developed to detect type A HBGA isolated from the complex (Figure 5). The amount of the viral receptor in the complex was reflected by OD₄₅₀ readings with MAb against type A HBGA (BG2) (Supplementary Table S2). The P/N ratio of recombinant cells incubated with or without PGM was calculated. Various concentrations of PGM were used to test for an optimal concentration to produce the lowest receptor background for the HBGAs-HuNoV-P proteins–bacteria complex. Although there was a positive correlation between OD₄₅₀ and PGM concentration, OD₄₅₀ of the background increased significantly when concentration of the latter exceeded 0.5 mg/mL. Incubation of PGM concentrations of 0.1, 0.2, 0.5, and 1.0 mg/mL with pET28a-inaQn-TB-P (GII.4)/BL21 produced P/N ratios of 1.88, 4.49, 7.66, and 9.61, respectively; similarly, pET28a-inaQn-TB/BL21 produced P/N ratios of 1.20, 1.38, 1.65, and 2.48, respectively; and that of pET-28a/BL21 produced P/N ratios of 1.13, 1.39, 1.61, and 2.44, respectively. These results indicate that the optimum concentration of PGM for the binding test was about 0.5 mg/mL. The absorbance of type A HBGA recovered from the complex was significantly higher than that of pET28a-inaQn-TB/BL21 and pET-28a/BL21 (p < 0.05). There was no significant difference between signals of E. coli BL21 cells transformed with plasmid pET28a-inaQn-TB and that of plasmid pET-28a with or without PGM treatment (p > 0.05). These results indicated that InaQn-TB-P protein could be displayed on the surface of bacterial cells, retained its receptor-binding ability, and that the receptor-to-P protein complex could be easily isolated from the system.