Riemerella anatipestifer strains WJ4 (serotype 1), Yb2 (serotype 2), and HXb2 (serotype 10) were isolated by Qinghai Hu from sick ducks in China in 2000 (Hu et al., 2010), and their full genome sequences have been deposited in GenBank under the accession number CP041029, CP007204, and CP011859. The suicide plasmid pYT354 (Zhu et al., 2017) was generously provided by Professor Mark J. McBride (University of Wisconsin–Milwaukee, Milwaukee, WI, USA). The R. anatipestifer strains and their derivatives were routinely grown in tryptic soybean broth (TSB; Difco, Detroit, MI, USA) at 37^°C with shaking or on tryptic soybean agar (TSA). Escherichia coli cells were routinely grown in Luria–Bertani (LB) broth (Difco) at 37^°C with shaking. The strains, plasmids, and primers used in this study are listed in Table 1. When required, antibiotics were added at the following concentrations: ampicillin (100 μg/mL), chloramphenicol (50 μg/mL), erythromycin (0.5 μg/mL), or kanamycin (50 μg/mL).

The proteins in the culture supernatant of strain Yb2 were precipitated with the trichloroacetic acid (TCA)–acetone method, as described previously (Flaugnatti and Journet, 2017), with some modification. Briefly, R. anatipestifer strain Yb2 was grown to mid-log phase, at an optical density at a wavelength of 600 nm (OD₆₀₀) of 1.0, in TSB at 37^°C with shaking. The cells were pelleted by centrifugation at 7,000 × g for 10 min, and the culture supernatant was filtered through a 0.22 μm pore-size polyvinylidene difluoride (PVDF) filter (Merck Millipore, Darmstadt, Germany) to remove residual cells. TCA was then added to the filtered culture supernatant to a final concentration of 10% (w/v), mixed gently, and incubated at 4^°C overnight. The precipitated proteins were collected by centrifugation at 16,000 × g for 1 h at 4^°C, and the pellet was washed with cold acetone. The pellets were air dried and resuspended in solubilization buffer (7 M urea, 2 M thiourea, 4% CHAPS). The precipitated proteins were quantified with the Bradford method using Bio-Rad Protein Assay Dye Reagent (Bio-Rad, Hercules, CA, USA), and the samples were stored at -80^°C until further analysis.

Samples containing 200 μg of proteins precipitated from culture supernatants were passively rehydrated onto 13 cm immobilized pH gradient IPG strips (GE Healthcare Life Sciences, Piscataway, NJ, USA) in a pH range of 3–10 for 12–16 h. The immunoproteomic analysis of the precipitated proteins was performed as described previously (Hu et al., 2012), and included isoelectric focusing (IEF), sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), western blotting, image processing, matrix-assisted laser desorption/ionization time of flight mass spectrometry (MALDI-TOF MS), database searches, and bioinformatic analyses. Convalescent duck serum against R. anatipestifer strain Yb2 was used in the western blotting analysis.

In this study, proteins AS87_RS06600 (designated “PaR1”), AS87_RS09020, and AS87_RS09965 appeared in more than three spots on the western-blotted membrane in the immunoproteomic analysis. The genes encoding the three proteins were amplified from serotype 2 R. anatipestifer strain Yb2 and cloned into the pET28a vector (Novagen, Darmstadt, Germany). The recombinant proteins were expressed in E. coli BL21(DE3) cells (Novagen, Darmstadt, Germany) and purified with Ni-IDA affinity chromatography (DetaiBio, Nanjing, China) as described previously (Xu et al., 2020). The recombinant proteins rPaR1, rAS87_RS09020, and rAS87_RS09965 were detected on western blots probed with convalescent duck serum directed against R. anatipestifer strain Yb2.

One-day-old Cherry Valley ducklings were purchased from Lijia Duck Farm (Taixin, Jiangsu, China), and no serum antibodies against R. anatipestifer strains WJ4 (serotype 1), Yb2 (serotype 2), or HXb2 (serotype 10) were detected with a whole-cell ELISA (Prieto et al., 2003). The ducklings were housed in cages under a 12 h light/dark cycle with free access to food and water. All applicable international, national, and institutional guidelines for the care and use of the animals were followed. The protocol was approved by the Committee on the Ethics of Animal Experiments of Shanghai Veterinary Research Institute, Chinese Academy of Agricultural Sciences (permit number: SHVRI-ZD-2018-026).

Before immunization and challenge experiments, the median lethal doses (LD_50s) for R. anatipestifer strains WJ4 (serotype 1), Yb2 (serotype 2), and HXb2 (serotype 10) were measured in ducklings as 1.97 × 10⁸ CFU, 1.07 × 10⁶CFU, and 1.89 × 10⁵ CFU, respectively. In this study, the immunized ducklings were challenged with 10 LD50 doses of each strain.

To evaluate the protective immunity conferred by the recombinant proteins rPaR1, rAS87_RS09020, and rAS87_RS09965, 8 days-old Cherry Valley ducklings were divided into four groups (12 ducklings per group) and immunized once intramuscularly with 100 μg of each recombinant protein or phosphate-buffered saline (PBS), each emulsified with Montanide ISA 70 V adjuvant (Seppic, Paris, France). At 2 weeks postimmunization, all the ducklings were challenged with 1 × 10⁷CFU of wild-type Yb2. Ducklings that became moribund were killed humanely and counted as dead, and then subjected to R. anatipestifer identification. The mortality of the ducklings was recorded daily for a period of 10 days after challenge.

To investigate whether rPaR1-immunized ducklings were protected from virulent challenge with R. anatipestifer strains of different serotypes, 8 days-old Cherry Valley ducklings were immunized with ISA-70-V-adjuvant-emulsified rPaR1 (n = 30 ducklings for three groups), ISA-70-V-adjuvant-emulsified PBS (n = 30 ducklings for three groups), or ISA-70-V-adjuvant-emulsified inactivated Yb2 bacterin (n = 10 ducklings). On day 15, the sera of the immunized ducklings from each group were collected and screened for rPaR1 antibodies with an ELISA, using 1 μg/mL rPaR1 as the coating antigen and horseradish peroxidase (HRP)-labeled rabbit anti-duck IgY (IgG) (Biodragon, Suzhou, China) as the secondary detection antibody. The rPaR1- and PBS-immunized ducklings (10 rPaR1-immunized and 10 PBS-immunized ducklings for each strain) were challenged with 2 × 10⁹ CFU of serotype 1 strain WJ4,10⁷ CFU of serotype 2 strain Yb2, or 2 × 10⁶ CFU of serotype 10 strain HXb2. Ten ducklings immunized with inactivated Yb2 bacterin were challenged with 10⁷ CFU of Yb2. The mortality of the ducklings was recorded daily, as above.

The natural ability of mouse complement (Sigma-Aldrich, St. Louis, MO, USA) and inactivated mouse complement (56^°C for 30 min) to kill serotype 1 strain WJ4, serotype 2 strain Yb2, and serotype 10 strain HXb2, was determined. To each well of a 96-well flat-bottom plate (Corning, NY, USA) containing 45 μL of sterile PBS, 25 μL of mouse complement or inactivated mouse complement was added, and then 30 μL of each bacterial suspension (10⁴ CFU/mL in PBS). After incubation at 37^°C for 30 min, 100 μL of the mixture was plated onto TSA plates. Percentage killing was calculated as the geometric mean value of [1-(CFU remaining after mouse complement/CFU remaining after inactivated complement)] × 100 from three trials, and the strain for which the natural bactericidal rate of mouse complement was <15% at a complement concentration of 25% (Martin et al., 2005) was used in the next bactericidal assay. Rabbit antiserum against rPaR1 and naïve rabbit serum were inactivated at 56^°C for 30 min and 2-fold serial diluted in PBS. Then the bactericidal activity against serotype 2 strain Yb2 was performed as described previously (Gu et al., 1996). For each well of a 96-well flat-bottom plate, 50 μL of diluted rabbit antiserum against rPaR1 or naïve rabbit serum, 30 μL of bacterial suspension (10⁴ CFU/mL in PBS) and 15 μL of mouse complement serum were added. The plates were incubated at 37^°C for 30 min, and 50 μL of the mixture in each well was spread onto TSA plates. The bactericidal assay was performed in triplicate. The highest serum dilution causing 50% killing was expressed as the bactericidal titer.

The nucleotide acid sequence of paR1 (AS87_RS06600) from R. anatipestifer strains of different serotypes were retrieved from the GenBank database. The derived amino acid sequences of PaR1 from these strains were aligned and analyzed with Clustal W in the MegAlign program of the Lasergene 7.01 software (DNASTAR Inc., Madison, WI, USA).

Rabbit antiserum against rPaR1 was obtained by immunizing 2 months-old New Zealand rabbits subcutaneously (s.c.) with 100 μg of purified rPaR1 protein emulsified with 50% (vol/vol) Montanide ISA 50 V adjuvant (Seppic) and boosted twice s.c. with 100 μg of rPaR1 at two weekly intervals. The rabbit sera were collected 7 days after the third immunization.

Most members of the phylum Bacteroidetes secrete proteins across the outer membrane via the type IX secretion system (T9SS) (Sato et al., 2010; McBride and Zhu, 2013). T9SS is a multiprotein system consisting of at least 20 proteins, and GldL is one of the core components of T9SS (Hennell James et al., 2021). To construct a T9SS-defective mutant, a 444 bp (nt 58–501) fragment of the strain Yb2 gldL ORF (687 bp) was deleted in-frame with the suicide plasmid pYT354, as described previously (Barbier et al., 2020). Briefly, a 2,599 bp region corresponding to the first 57 bp of gldL together with the directly upstream region was amplified with primers gldL-L P1 and gldL-L P2. A 2,773 bp region corresponding to the final 186 bp of gldL together with the directly downstream region was amplified with primers gldL-R P1 and gldL-R P2. The two fragments were ligated into suicide vector pYT354, generating the recombinant suicide plasmid p354-gldL-LR. This plasmid was then transferred into wild-type Yb2 by conjugation, and transformants were selected on TSA agar containing erythromycin and kanamycin. One positive colony was grown in TSB without antibiotics at 37^°C overnight, and the cells were then plated on TSA containing 8% sucrose. The positive mutant colonies were screened by PCR using primers gldL ORF F plus gldL ORF R, and 16S rRNA F plus 16S rRNA R, to detect the 16S rRNA, and gldL ORF, respectively. A gldL deletion mutant, which was positive for 16S rRNA and has a smaller PCR product (about 250 bp) for amplification of gldL ORF than that in the wild type Yb2, was designated as ΔgldL.

The subcellular localization of PaR1 predicted by PSORTb v.3.0¹ was “unknown,” suggesting that R. anatipestifer may be not uniquely located.

To investigate whether the specific antibodies to rPaR1 in rabbit antiserum against rPaR1 could bound to the outer membrane of R. anatipestifer strainYb2, a whole-cell ELISA using whole Yb2 cells as the coating antigen was performed as described previously (Prieto et al., 2003). Naïve rabbit serum and convalescent duck serum against R. anatipestifer strain Yb2 were used as the negative and positive controls, and HRP-labeled goat anti-rabbit IgG (Solarbio, Beijing, China) and HRP-labeled rabbit anti-duck IgY (IgG) (Biodragon) as the secondary detection antibodies for detection of rabbit sera and duck serum, respectively. At the same time, PaR1-specific antibodies in the rPaR1 immunized rabbit serum and naïve rabbit serum were detected with an indirect ELISA using 1 μg/mL rPaR1 as the coating antigen. All samples were tested in triplicate. On the other hand, for indirect immunofluorescence assay, rabbit antiserum against rPaR1 was incubated with fresh whole Yb2 cells at room temperature for 2 h. After the cells were washed with PBS, the cells were incubated with Alexa Fluor 488-labeled goat anti-rabbit IgG antibody (ThermoFisher, USA) at room temperature for 1 h. After the washes, the cells were observed on a fluorescence microscope (Nikon Eclipse 80i, Japan). Rabbit serum against Yb2 and naïve rabbit serum were used as positive and negative controls.

The inner (cytoplasmic)-membrane proteins of wild-type Yb2 and the mutant ΔgldL were prepared as described previously (Spratt and Pardee, 1975), with some modification. Briefly, R. anatipestifer cells were grown to mid-log phase (OD₆₀₀ = 1.0) in TSB at 37^°C with shaking. The cells were centrifuged and washed twice with 0.01 M PBS (pH 7.0), and the pellet was resuspended in 0.01 M PBS (pH 7.0) containing 1 mM phenylmethanesulfonyl fluoride. After sonication, any unbroken cells were removed by centrifugation at 7,000 × g for 10 min. The membrane fraction was then collected by centrifugation at 100,000 × g for 30 min. The pellets were resuspended in 0.01 M PBS (pH 7.0) to which 1% (w/v) sarkosyl (Macklin, Shanghai, China) was added, and incubated at room temperature for 20 min. The membrane samples were then separated into the soluble (inner membrane) and insoluble (outer membrane) fractions by centrifugation at 100,000 × g for 30 min.

To determine whether protein PaR1 is secreted to the outside of R. anatipestifer cells through T9SS, the proteins in the culture supernatant of WT Yb2 and ΔgldL were precipitated with the TCA–acetone method, as described above. PaR1 was detected in the whole Yb2 cells, the inner membrane proteins, the normal cell-free culture supernatant, and the precipitated proteins in the culture supernatant with SDS-PAGE and western blotting, using rPaR1 immunized rabbit serum as the primary antibody. The blots were developed using an enhanced chemiluminescence detection kit (ECM Biotech, Suzhou, China).

GraphPad Prism 9.0.0 software (GraphPad Software, Inc., San Diego, CA, USA) was used for statistical analysis and preparation of graphs. The statistical significance of the survival curves was determined with a log-rank (Mantel–Cox) test, and the significant differences of ELISA results between groups were identified using the two-tailed Student’s t-test. A probability (p) value of <0.05 was considered statistically significant.

Samples containing 200 μg of proteins precipitated from the cell-free culture supernatant of R. anatipestifer strain Yb2 were separated with two-dimensional polyacrylamide gel electrophoresis (2-DE) and stained with Coomassie Brilliant Blue G-250 (Sigma-Adrich, MO, USA). The proteins were separated according to isoelectric point (PI) in the first dimension over a pH range of pH 3–10, and most proteins on the 2D gels had molecular weights (MWs) of 10–150 kDa (Figure 1A). The western blotting analysis detected 43 immunogenic protein spots when probed with convalescent duck serum against R. anatipestifer strain Yb2 (Figure 1B), but no appreciable spots appeared on the membrane probed with naïve duck serum (data not shown). Eleven immunogenic proteins were identified with MALDI-TOF-MS and a peptide mass fingerprinting (PMF) analysis (Table 2). Five proteins appeared in more than one spot, and of these, proteins AS87_RS06600 (PaR1), AS87_RS09020, and AS87_RS09965 appeared on 23, eight, and four spots, respectively. Moreover, both AS87_RS06600 and AS87_RS09020 were predicted to contain a type A carboxy-terminal domain (CTD) of the type IX secretion system (T9SS) (Kulkarni et al., 2017), suggesting they were secreted by T9SS. In addition, it is noteworthy that 11 immunogenic proteins identified in the culture supernatant of R. anatipestifer strain Yb2 in this study were not identified in the whole cells (Hu et al., 2012; Zhai et al., 2012) or culture supernatant (Zhai et al., 2013) of R. anatipestifer in immunoproteomic analyses in previous studies. The subcellular locations of these proteins predicted with the PSORTb v.3.0 software showed that six proteins (54.55%, 6/11) were extracellular proteins and the locations of five (45.45%, 5/11) were “unknown” (Table 2).

In this study, three proteins (PaR1, AS87_RS09020, and AS87_RS09965), which appeared in more than three spots on PVDF membranes in a western blotting analysis, were expressed in E. coli and purified. The immunogenicity of the recombinant proteins was investigated with western blotting using duck antiserum against Yb2 as the primary antibody. The three recombinant proteins reacted with duck antiserum against Yb2 (Supplementary Figure 1), suggesting that they are immunogenic proteins of R. anatipestifer.

As shown in Figure 2A, five (41.67%) of 12 ducklings immunized once with the soluble rPaR1 protein were protected from wild-type Yb2 challenge, whereas all the ducklings immunized with rAS87_RS09020, rAS87_RS09965, or PBS died within 4 days of challenge. There was a significant difference in the protective immunity conferred by rPaR1 and the PBS control (p < 0.01). To investigate whether rPaR1 conferred protective immunity against challenge with R. anatipestifer strains of different serotypes, ducklings were challenged with serotype 1 strain WJ4, serotype 2 strain Yb2, or serotype 10 strain HXb2 at 2 weeks after rPaR1 immunization. All the ducklings inoculated with PBS emulsified with ISA 70 V adjuvant died within 6 days of challenge with strains WJ4, Yb2, or HXb2, whereas 2/10 (20%), 4/10 (40%), and 0% of ducklings immunized with rPaR1 emulsified with ISA 70 V adjuvant survived challenge with serotype 1 strain WJ4 (p > 0.05, versus PBS group challenge with WJ4), serotype 2 strain Yb2 (p < 0.05, versus group challenge with Yb2), and serotype 10 strain HXb2 challenge, respectively (Figure 2B), suggesting that rPaR1 did not confer protection against serotype 10 strain HXb2 or serotype 1 strain WJ4. Before the challenge, the sera from the immunized ducklings were collected and the PaR1-specific antibodies measured with an ELISA using rPaR1 as the coating antigen. As shown in Figure 2C, the average OD₄₉₀ values for antibodies to PaR1 in the serum samples obtained from ducks immunized with ISA-70 V-emulsified rPaR1 was 1.39 ± 0.17, whereas that in sera from ducklings immunized with ISA-70 V-emulsified inactivated Yb2 bacterin was 0.24 ± 0.06.

The bactericidal activity of rPaR1 immunized rabbit serum was also determined. A natural bactericidal activity assay showed that the natural bactericidal rates of mouse complement against serotype 1 strain WJ4, serotype 2 strain Yb2, and serotype 10 strain HXb2 were below zero, 8.5% (<30%), and 87.4% (>30%), respectively. Therefore, the mouse complement used in this study was only suitable for the analysis of the bactericidal activity against serotype 2 strain Yb2, but not against serotype 1 strain WJ4 and serotype 10 strain HXb2. The bactericidal assay showed rPaR1 immunized rabbit serum showed bactericidal activity against strain Yb2 at a titer of 1:8 (Figure 2D).

The paR1 gene sequences of 33 R. anatipestifer strains of different serotypes were retrieved from GenBank. Among these, 26 strains carried one copy of the paR1 gene, and the other seven strains (20190305E2-2, 20190109E1-1, 20190609E1-1, RCAD0416, 20190121E1-3, 20190507E1-1, 20190509E1-1) carried 2–3 copies of paR1 homologous. A homology analysis showed that PaR1 of serotype 2 R. anatipestifer strain Yb2 shared 100% identity with those of strains RA-YM (serotype 1), RA-LZ01 (serotype 1), RA-GD (serotype 1), 17 (serotype 17), and RCAD0392, indicating that PaR1 is not a serotype-specific protein of R. anatipestifer (Supplementary Figure 2). In addition, PaR1 of strain Yb2 shared 99.9, 84.5, 83.8, and 85.5% identities with that of strains DSM15868 (= ATCC 11845, type strain), WJ4 (serotype 1), CH-2 (serotype 2), and HXb2 (serotype 10), respectively.

To determine whether PaR1 is located on the outer membrane of R. anatipestifer, whole Yb2 cells were used as the coating antigen in a whole-cell ELISA. The average OD₄₉₀ values for rabbit antiserum against rPaR1 was 0.088 ± 0.001, which was close to that for naïve rabbit serum, whereas that for rabbit antiserum directed against rPaR1 was 1.28 ± 0.02 when rPaR1 protein was used as the coating antigen in ELISA (Figure 3A), indicating that the polyclonal antibodies against rPaR1 could not bind to the whole R. anatipestifer cells. Furthermore, no fluorescence was observed on the R. anatipestifer cells under fluorescence microscopy when rabbit antiserum against rPaR1 was used as the primary antibody and fluorescein-isothiocyanate-labeled goat anti-rabbit IgG antibody as the secondary antibody (data not shown). These results indicate PaR1 does not localize to the R. anatipestifer outer membrane.

The proteins in the culture supernatant of Yb2 and T9SS-defective mutant ΔgldL were detected by SDS-PAGE and western blotting. Compared to the wild type, the amount of proteins in the cell-free culture supernatant of ΔgldL mutant was obviously reduced (Figure 3B), indicating that T9SS-defective mutant ΔgldL was defective in extracellular protein secretion. PaR1 of wild-type Yb2 (proPaR1) was predicted to be 117 kDa in size, according to its amino acid sequence. As shown in Figure 3C, a band close in size to proPaR1 and rPaR1 was detected with rabbit antiserum against rPaR1 from the whole cells of Yb2 and ΔgldL mutant with western blotting. In contrast, as shown in Figures 3B, C, and PaR1 was the major protein component in the culture supernatant of wild-type Yb2, and mature PaR1, which is about 85 kDa, was detected in the normal or 1,000-fold-concentrated cell-free culture supernatant of WT Yb2. At least six other smaller bands, including two obvious bands of about 40 kDa, were also detected. However, PaR1 was absent from both the normal and 1,000-fold-concentrated culture supernatant of the T9SS-defective mutant ΔgldL, indicating that PaR1 protein is mainly secreted to the outside of wild-type strain Yb2 cells via the T9SS. In addition, PaR1 was also detected among the inner-membrane proteins of Yb2 and mutant ΔgldL, and the most obvious bands had a molecular mass close to those of proPaR1 and rPaR1 (Figure 3C), suggesting the CTD is not cleaved yet for the inner membrane PaR1. The N-terminal signal peptide guides the PaR1 protein across the inner membrane in the first step in its secretion from R. anatipestifer cells, therefore, most probably, the PaR1 on the inner membrane is a normal intermediate state of the secretion process. Overall, the localization of PaR1 is extracellular after secretion through T9SS.