Genome-Wide Analysis and Characterization of the Riemerella anatipestifer Putative T9SS Secretory Proteins with a Conserved C-Terminal Domain

ABSTRACT

Riemerella anatipestifer is a major pathogenic agent of duck septicemic and exudative diseases. Recent studies have shown that the R. anatipestifer type IX secretion system (T9SS) acts as a crucial virulence factor. We previously identified two T9SS component proteins, GldK and GldM, and one T9SS effector metallophosphoesterase, which play important roles in bacterial virulence. In this study, 19 T9SS-secreted proteins that contained a conserved T9SS C-terminal domain (CTD) were predicted in R. anatipestifer strain Yb2 by searching for CTD-encoding sequences in the whole genome. The proteins were confirmed with a liquid chromatography–tandem mass spectrometry analysis of the bacterial culture supernatant. Nine of them were reported in our previous study. We generated recombinant proteins and mouse antisera for the 19 predicted proteins to confirm their expression in the bacterial culture supernatant and in bacterial cells. Western blotting indicated that the levels of 14 proteins were significantly reduced in the T9SS mutant Yb2ΔgldM culture medium but were increased in the bacterial cells. RT–qPCR indicated that the expression of these genes did not differ between the wild-type strain Yb2 and the T9SS mutant Yb2ΔgldM. Nineteen mutant strains were successfully constructed to determine their virulence and proteolytic activity, which indicated that seven proteins are associated with bacterial virulence, and two proteins, AS87_RS04190 and AS87_RS07295, are protease-activity-associated virulence factors. In summary, we have identified at least 19 genes encoding T9SS-secreted proteins in the R. anatipestifer strain Yb2 genome, which encode multiple functions associated with the bacterium’s virulence and proteolytic activity.

IMPORTANCE Riemerella anatipestifer T9SS plays an important role in bacterial virulence. We have previously reported nine R. anatipestifer T9SS-secreted proteins and clarified the function of the metallophosphoesterase. In this study, we identified 10 more secreted proteins associated with the R. anatipestifer T9SS, in addition to the nine previously reported. Of these, 14 proteins showed significantly reduced secretion into the bacterial culture medium but increased expression in the bacterial cells of the T9SS mutant Yb2ΔgldM; seven proteins were shown to be associated with bacterial virulence; and two proteins, AS87_RS04190 and AS87_RS07295, were shown to be protease-activity-associated virulence factors. Thus, we have demonstrated that multiple R. anatipestifer T9SS-secreted proteins function in virulence and proteolytic activity.

INTRODUCTION

Secretion systems perform numerous physiological functions essential for cell propagation and fitness, including facilitating nutrient acquisition, communication with the environment, attachment to various surfaces, defense against host antimicrobial systems, and the delivery of virulence factors at precise locations, such as eukaryotic cells (1–6). The recently discovered type IX secretion system (T9SS), a complex transporter only found in many species of the phylum Bacteroidetes, is associated with bacterial gliding motility and protein secretion, and is considered to be a virulence factor in many pathogens (7–12).

The Bacteroidetes T9SS is common in, but apparently confined to, this phylum (13, 14). The core components of the T9SS include GldK, GldL, GldM, GldN, SprA, SprE, and SprT, and the deletion of a single component of T9SS results in defective secretion of proteins, such as the cell-surface components of the gliding motility machinery, extracellular or cell-surface enzymes, adhesins, and virulence factors (7, 9, 15–17). The proteins secreted by the T9SS rely on the Sec system to cross the cytoplasmic membrane, and they typically have a conserved C-terminal domain (CTD), which targets them to the T9SS for secretion across the outer membrane (18). Therefore, the deletion of T9SS genes results in the accumulation of effector proteins in the periplasm, instead of their exportation outside the cell (19).

Riemerella anatipestifer is a member of the phylum Bacteroidetes, and is the type species of the genus Riemerella in the family Flavobacteriaceae (20). It mainly infects domestic ducks, geese, turkeys, and other poultry, causing diseases characterized by serositis and sepsis. Prevalent R. anatipestifer infections can lead to high mortality rates and significant economic losses (20, 21). So far, only a few virulence factors have been reported in R. anatipestifer strains, including lipopolysaccharide, outer membrane protein, CAMP cohemolysin, TonB-dependent receptor Tbdr1, and nicotinamidase PncA (22–25). T9SS has been reported to be a virulence factor and to function in the secretion of various proteins (26, 27). In our previous study, R. anatipestifer T9SS mutant strains Yb2ΔgldM and Yb2ΔgldK showed significantly attenuated virulence, and defective secretion of proteins with a conserved T9SS CTD, compared with the wild-type strain Yb2 (28, 29). A secreted metallophosphatase that was absent from the secreted proteins of Yb2ΔgldK and Yb2ΔgldM was confirmed to be associated with virulence (30).

To further clarify the functions and mechanisms of the R. anatipestifer T9SS in bacterial pathogenesis, we predicted T9SS-secreted proteins with a conserved CTD in R. anatipestifer strain Yb2 by searching for CTD-encoding sequences in the whole genome. We confirmed their secretion with a liquid chromatography–tandem mass spectrometry (LC–MS/MS) analysis of the bacterial culture supernatant, and further investigated their roles in bacterial virulence and proteolytic activity.

RESULTS

Bioinformatics analysis.

Proteins with a CTD sequence were identified with a BLAST analysis of the R. anatipestifer Yb2 genome. In total, 19 genes encoding a conserved CTD were predicted to encode T9SS-secreted proteins (Table 1). Nine of them have been reported previously (28).

TABLE 1.

Predicted R. anatipestifer Yb2 proteins secreted by T9SS, identified with sequence BLAST and LC–MS/MS^a

Locus tagb	Molecular mass (KDa)	Uniprot ID	Proteins
AS87_RS00835	161.5	E4T8U9	Peptidase s8 and s53 subtilisin kexinsedolisin
AS87_RS00980	68.225	E4T8X7	Metallophosphoesterase
AS87_RS02020	83.131	E4T9B5	Immunoreactive 84 kDa antigen pg93
AS87_RS02625	39.812	E4T9N2	Endonuclease I
AS87_RS02840	55.6	J9QTI6	Uncharacterized protein
AS87_RS02875	104.3	J9QTH8	Metallophosphoesterase
AS87_RS02950	26.7	E4T9T9	Carbohydrate binding domain
AS87_RS02955	26	E4T9U0	Uncharacterized protein
AS87_RS03090	13.49	E4T9W5	Uncharacterized protein
AS87_RS03200	117.12	E4TBI9	Uncharacterized protein
AS87_RS04190	77.52	H8MA42	Subtilisin-like serine protease
AS87_RS04975	48.4	E4TCP5	Por_Secre_tail domain-containing protein
AS87_RS06600	117.29	E4TA79	Pkd domain containing protein
AS87_RS07160	17.65	E4TC80	Uncharacterized protein
AS87_RS07295	93.945	E4TC54	Fibronectin type iii domain protein
AS87_RS07755	153.791	E4TBW6	Uncharacterized protein
AS87_RS08215	73.9	E4TD92	Por_Secre_tail domain-containing protein
AS87_RS09020	28.6	E4TDQ0	Carbohydrate-binding cenc domain protein
AS87_RS09040	40.9	E4TDQ4	Leucine-rich repeat-containing protein

^aProteins in cell-free culture fluid from Yb2 were analyzed with LC–MS/MS.

^bProteins with a CTD sequence were identified with the conserved domain analysis.

Recombinant T9SS-secreted proteins and their mouse antisera were successfully generated.

Non-full-length gene sequences of the T9SS-secreted proteins were ligated into the pET-30a(+), pCold I, pCold TF, or pGEX-4T vector and used to transform Escherichia coli BL21(DE3). After induction and purification, 19 recombinant proteins were obtained. Escherichia coli BL21(DE3) cells transformed with the empty pET-30a(+), pCo1d I, pCo1d TF, or pGEX-4T vector generated no specific bands, and were used as negative controls (Fig. S1 in the supplemental material).

The ELISA titers of the mouse antisera directed against the 19 recombinant proteins were >1:8000, which were qualified for Western blotting analysis.

Confirming R. anatipestifer secretory proteins with Western blotting.

The results of a BLAST analysis of the R. anatipestifer Yb2 genome and the differential secretory proteins, which were identified by LC–MS/MS analysis in the bacterial culture supernatant of the wild-type strain Yb2 and its T9SS mutant Yb2ΔgldM (Table S1), revealed 19 proteins with a CTD sequence and their Uniprot identification (Table 1). To confirm that the gene products were T9SS-secreted proteins, we analyzed the bacterial culture supernatant of the wild-type strain Yb2 and its T9SS mutant strain Yb2ΔgldM. After ultrafiltration, 20 μg of the secretory proteins of each strain were separated with sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and analyzed with Western blotting to confirm the secretion of the 19 predicted T9SS-secreted proteins. As shown in Fig. 1A, all 19 proteins showed a specific band in samples of the bacterial culture supernatants, and their secretion levels were altered in the T9SS mutant Yb2ΔgldM.

FIG 1.

Western blotting of T9SS effectors in culture supernatants and whole-bacterial lysates. (A) Samples from bacterial culture supernatant, 20 μg/sample. (B) Samples of whole-bacterial lysates (50 μg/sample). Lane M, PageRule Prestained Protein Ladder (Thermo Scientific, Waltham, MA, USA); lane 1: wild-type strain Yb2; lane 2: T9SS mutant strain Yb2ΔgldM; lane 3: complementation strain cYb2ΔgldM. The target effector protein is indicated in the figure as ASS87_RS02625, etc. TonB was used as the protein loading control.

Deletion of R. anatipestifer T9SS component affects the secretion, but not the transcription of T9SS substrates.

We investigated the expression levels of the 19 T9SS substrates in whole bacterial lysates. The results shown in Fig. 1B indicate that the deletion of the gldM gene caused most substrates to accumulate in the bacteria, except for the effector proteins AS87_RS06600, AS87_RS2840, and AS87_RS04975, which decreased in the whole-cell lysates.

A subsequent quantitative PCR (qPCR) analysis indicated that the deletion of gene gldM did not alter the transcription levels of those substrates (Fig. 2). These results demonstrate that all 19 secretory proteins are T9SS substrates, and that the deletion of the R. anatipestifer T9SS component protein GldM impaired the secretion function.

FIG 2.

qPCR verification of the transcription levels of the 19 genes. The mRNA of all 19 genes in the wild-type strain Yb2, the mutant strain Yb2ΔgldM, and the complementation strain cYb2ΔgldM were quantified with RT–qPCR to determine their transcription levels. The results showed no differences in their transcription among Yb2, Yb2ΔgldM, and cYb2ΔgldM. ns, P > 0.05.

Establishment of R. anatipestifer T9SS-secreted protein mutant strains.

Natural transformation was used to construct mutant R. anatipestifer strains as described previously (30). The erythromycin-resistance gene (erm^r) was amplified from R. anatipestifer strain HXb2. After transformation with the PCR fragments, the mutants were selected based on their erythromycin resistance, and were identified with analytical PCR and DNA sequencing. The primers used for PCR amplification are listed in Table S2. As the results, 19 gene mutants showed a 750-bp fragment of erm^r and a different fragment from the wild-type strain Yb2 were constructed successfully (Fig. 3). A 496-bp fragment of R. anatipestifer 16S rRNA was detected in all strains as a control. The 19 mutant strains were designated Yb2Δ04190, Yb2Δ06600, Yb2Δ02020, Yb2Δ07295, Yb2Δ00835, Yb2Δ00980, Yb2Δ02625, Yb2Δ03200, Yb2Δ03090, Yb2Δ09020, Yb2Δ09040, Yb2Δ02950, Yb2Δ02955, Yb2Δ02840, Yb2Δ02875, Yb2Δ08215, Yb2Δ07755, Yb2Δ04975, and Yb2Δ07160.

FIG 3.

PCR verification of R. anatipestifer strains with T9SS effector mutations. M: DL2000 DNA Marker (Vazyme, Nanjing, China); lane 1: wild-type strain Yb2; lane 2: mutant strain; lane 3: negative control. 16S rRNA and erm^r were amplified as 496-bp and 750-bp fragments, respectively, in all mutant strains. Target genes were amplified as the predicted fragment lengths indicated in the figure (fragment lengths: AS87_RS04190, 2133 bp; AS87_RS06600, 3120 bp; AS87_RS02020, 2235 bp; AS87_RS07295, 2634 bp; AS87_RS00980, 1818 bp; AS87_RS00835, 4428 bp; AS87_RS02625, 1050 bp; AS87_RS03200, 2172 bp; AS87_RS07755, 4047 bp; AS87_RS09020, 720 bp; AS87_RS09040, 1119 bp; AS87_RS02950, 732 bp; AS87_RS02955, 741 bp; AS87_RS02840, 1542 bp; AS87_RS02875, 2793 bp; AS87_RS08215, 1953 bp; AS87_RS04975, 1080 bp; AS87_RS07160, 462 bp; and AS87_RS03090, 357 bp. Gene names are indicated in the figure as AS87_RS04190, etc.

Determination of the bacterial median lethal dose (LD₅₀).

To clarify the effects of the R. anatipestifer T9SS-secreted proteins on bacterial virulence, the LD₅₀ of the 19 mutant and seven complementation strains was assessed in Cherry Valley ducks. Deaths were monitored from day 1 to day 7 postinfection. The LD₅₀ vales were calculated as described previously (25) and are summarized in Table 2. Of the 19 mutants, seven were attenuated more than 100-fold, including three that were attenuated >1,000-fold, >10,000-fold, and >200,000-fold compared with wild-type strain Yb2. These results indicate that the T9SS-secreted proteins are important for the virulence of R. anatipestifer strain Yb2.

TABLE 2.

Bacterial LD₅₀s determination^a

Strain(s)	Disrupted geneb	Predictive protein function	LD50 (CFU)	Attenuation (fold)
Yb2Δ06600	AS87_RS06600	Pkd domain containing protein	4.62 × 1010	227,459
cYb2Δ06600			1.13 × 108	557
Yb2Δ04190	AS87_RS04190	Serine-type endopeptidase activity	2.25 × 109	11,066
cYb2Δ04190			2.05 × 107	101
Yb2Δ00980	AS87_RS00980	Metallophosphoesterase	2.71 × 108	1,334
cYb2Δ00980			5.28 × 106	26
Yb2Δ03090	AS87_RS03090	Unknown	1.23 × 108	604
cYb2Δ03090			5.28 × 106	26
Yb2Δ02625	AS87_RS02625	Endonuclease activity	7.7 × 107	255
cYb2Δ02625			2.44 × 107	120
Yb2Δ02955	AS87_RS02955	Unknown	4.06 × 107	200
cYb2Δ02955			1.14 × 107	56
Yb2Δ02950	AS87_RS02950	Carbohydrate binding domain	4.04 × 107	190
cYb2Δ02950			9.14 × 106	45
Yb2Δ00835	AS87_RS00835	Serine-type endopeptidase activity	1.64 × 107	81
Yb2Δ02875	AS87_RS02875	Metallophosphoesterase	1.30 × 107	64
Yb2Δ07160	AS87_RS07160	Unknown	5.3 × 106	26
Yb2Δ02840	AS87_RS02840	Unknown	3.86 × 106	19
Yb2Δ02020	AS87_RS02020	Immunoreactive 84 kDa antigen pg93	3.45 × 106	17
Yb2Δ07295	AS87_RS07295	Serine-type endopeptidase activity	3.3 × 106	16
Yb2Δ08215	AS87_RS08215	Unknown	3.01 × 106	15
Yb2Δ03200	AS87_RS03200	Unknown	2.43 × 106	12
Yb2Δ09020	AS87_RS09020	Carbohydrate binding domain	2.03 × 106	10
Yb2Δ07755	AS87_RS07755	Unknown	2.03 × 106	10
Yb2Δ04975	AS87_RS04975	Unknown	1.62 × 106	8
Yb2Δ09040	AS87_RS09040	Class 1 internalin InlJ	8.12 × 105	4

^aThe LD₅₀ of the wild-type strain Yb2 was determined as 2.03 × 10⁵ CFU. Virulence attenuation was expressed as “fold” and calculated as Fold = LD₅₀ CFU of the mutants or complementation strains/LD₅₀ CFU of the wild-type strain Yb2 (2.03 × 10⁵).

^bBased on the R. anatipestifer Yb2 genome (accession no. CP007204).

The seven mutant strains whose virulence was attenuated >100-fold were conjugated with a shuttle plasmid containing the deleted gene sequence to construct complementation strains. The virulence of the complementation strains was largely restored (Table 2).

Protease activity measurements.

The cell-free culture fluids of the wild-type strain Yb2 and the 19 mutant strains were collected, and their protease activities were determined with azocasein (Sigma) as the substrate. The results indicated that mutant strains Yb2Δ04190 and Yb2Δ07295 were thoroughly defective in protein digestion (Fig. 4), consistent with the fact that the enzymes involved in the digestion of such substrate are secreted by T9SS (12, 31).

FIG 4.

Determination of proteolytic activity. The protease activities of the cell-free culture fluids of the wild-type strain Yb2 and 19 mutant strains were determined with azocasein (Sigma) as the substrate. The deletion of genes AS87_RS04190 and AS87_RS07295 significantly reduced the bacterial protease activities. Error bars represent standard deviations, which were calculated from three independent experiments performed in triplicate (****, P < 0.001; ns, P > 0.05).

DISCUSSION

In many members of the phylum Bacteroidetes, the T9SS has been shown to be associated with various activities, such as gliding motility, protein secretion, and the virulence of many pathogens (8, 14, 16, 28). T9SS is important in the secretion of at least 30 proteins by the periodontal pathogen Porphyromonas gingivalis, including its major virulence factors, called gingipains (Kgp, RgpA, and RgpB) (13, 32, 33). It also allows the fish pathogen Flavobacterium psychrophilum to secrete toxins that play a role in the pathogenesis of cold-water disease (12). T9SS-secreted proteins are known virulence factors in other bacteria (12, 14). In our previous study, the deletion of gldK and gldM, which encode core components of the R. anatipestifer T9SS, attenuated bacterial virulence and reduced the secretion of proteins (28, 29). These results are similar with those observed for F. johnsoniae and P. gingivalis, and support the roles of GldK and GldM in protein secretion (16, 34–36).

The R. anatipestifer T9SS mutant Yb2ΔgldM showed abnormal secretion levels of nine proteins containing the T9SS CTD (28, 29). Further study revealed that the deletion of the T9SS-secreted protein metallophosphatase attenuated the bacterium’s virulence and caused a defect in phosphatase activity, indicating that the T9SS effector protein acts as an enzyme and plays an important role in the pathogenic processes of R. anatipestifer (30). To detect more T9SS effectors encoded in the genome, we used BLAST to search the whole R. anatipestifer Yb2 genome for genes encoding proteins with a T9SS CTD sequence. We then investigated the secretion of these proteins with LC–MS/MS and confirmed the differential secretion of these proteins by Yb2 and its T9SS mutant (Yb2ΔgldM) (Table S1). We found 19 proteins with a T9SS CTD sequence, which were confirmed to be differentially secreted proteins in the T9SS mutants in this study. In addition to nine genes that we have reported previously (28), 10 novel genes encoding a CTD sequence were detected in the Yb2 genome, which were predicted to encode effector proteins of the T9SS: AS87_RS02840, AS87_RS08215, AS87_RS02875, AS87_RS09020, AS87_RS09040, AS87_RS02950, AS87_RS02955, AS87_RS04975, AS87_RS07160, and AS87_RS07755.

To confirm the secretion of these proteins, recombinant proteins and mouse antisera were generated. A Western blotting analysis indicated that the secretion levels of 16 proteins were significantly reduced in the T9SS mutant strain Yb2ΔgldM, whereas those of three proteins were increased. These results were almost identical to the LC–MS/MS results reported for the gldM mutant, further confirming that all the effector proteins are secreted by T9SS, and that this process is related to the GldM protein.

According to other reports, the deletion of a T9SS gene results in the accumulation of effector proteins in bacteria rather than their exportation outside the cell (11, 19). To confirm that this occurs in the R. anatipestifer T9SS mutant, we detected the expression of the 19 proteins in whole bacterial cells with a Western blotting analysis. As expected, 14 proteins accumulated in the bacteria, evident as increased protein levels in the bacterial cells of the T9SS mutant strain Yb2ΔgldM, and their reduced secretion into the culture supernatant. AS87_RS07160 and AS87_RS03090 also increased in the bacteria, whereas both the secretion and the expression in whole bacteria of AS87_RS06600 and AS87_RS02840 were reduced. The secretion of AS87_RS04975 was increased, but its presence in the bacteria was reduced. Quantitative PCR showed no difference in the transcription of the 19 genes in the wild-type strain Yb2 and T9SS mutant Yb2ΔgldM, suggesting that in addition to the known function of T9SS in exporting effector proteins outside the bacterial cell, other mechanism may affect the secretion of T9SS effectors. This warrants further investigation.

Of the 19 R. anatipestifer T9SS effectors, AS87_RS00980 has been identified as a virulence factor (30). To examine whether the other 18 T9SS effectors are associated with R. anatipestifer virulence, natural transformation was used to construct 18 gene mutant strains in this study. Their LD₅₀ values indicated that the virulence of the AS87_RS02950, AS87_RS02955, AS87_RS02625, AS87_RS03090, AS87_RS00980, AS87_RS04190, and AS87_RS06600 gene deletion mutants was attenuated more than 100-fold compared with that of the wild-type strain Yb2. Of these secreted proteins, AS87_RS00980 has been shown to have metallophosphoesterase activity (30). AS87_RS04190 is a predicted subtilisin-like serine protease, which belongs to the peptidase S8 family domain of the Kp43 proteases, the members of the peptidase S8, or subtilase clan of proteases. These proteases have an Asp/His/Ser catalytic triad, a similar domain to that found in trypsin-like proteases, but do not share their three-dimensional structure. Kp43 has three Ca²⁺-binding sites, which differ from the corresponding sites in the other known subtilisin-like proteases, and is known to be an oxidation-resistant protease compared with the other subtilisin-like proteases (37, 38). AS87_RS02625 is a predicted endonuclease I. In Bacillus subtilis, endonuclease I is thought to normally generate double-stranded breaks in DNA, which includes a nonspecific, Mg²⁺-activated RNase precursor and is inhibited by different RNA species (39). AS87_RS06600 is a PKD-domain-containing protein. Identification of its conserved domains showed that polycystic kidney disease 1 (PKD1) protein contains 14 repeats, which also occur elsewhere, such as in microbial collagenases (40). AS87_RS02950 is a carbohydrate-binding domain (CBD)-containing protein, and the tertiary structure of the CBD is strikingly similar to that of the bacterial 1,3-1,4-beta-glucanases and other sugar-binding proteins with jelly-roll folds (41). The function of AS87_RS02955 is unknown. The secretion of AS87_RS03090 was significantly increased in the T9SS mutant Yb2ΔgldM, and its LD₅₀ was attenuated >500-fold, suggesting that this protein plays a regulatory or synergistic role with other effectors in pathogenic processes. These results strongly indicate that the proteins secreted by the R. anatipestifer T9SS are required for bacterial virulence.

Measurement of the protease activities of the 19 mutant strains showed that Yb2Δ04190 and Yb2Δ07295 were significantly defective in the digestion of their substrate proteins. More interestingly, the LD₅₀ of the AS87_RS04190 mutant strain was attenuated 11,066-fold, whereas the AS87_RS07295 mutation had little effect on bacterial virulence. These two gene products differ in that AS87_RS07295 contains an additional fibronectin type 3 domain (FN3) and an extra bacterial prepeptidase C-terminal domain (PPC), although whether the additional FN3 and PPC structures limit the protein’s function requires further clarification.

Subtilisin-like serine proteases have a great many predicted functions, such as the degradation of gelatin, the hydrolysis of the Aα chain of fibrinogen, and the modulation of cytokine secretion, and one has recently been shown to be an important virulence factor in Streptococcus suis (42–44). In Treponema denticola, a subtilisin-like serine protease has been implicated in complement evasion by inducing C3 cleavage (45). In the present study, R. anatipestifer AS87_RS04190 was identified as a T9SS effector, an important virulence factor, and a putative peptidase S8 subfamily member in the Kp43 protease family. The protease activity of the AS87_RS04190 mutant strain was significantly defective in the digestion of substrate proteins, which is consistent with the fact that the enzymes involved in the digestion of these substrates are secreted by T9SS (12, 31, 46, 47). However, how AS87_RS04190 functions during R. anatipestifer infection is unclear, and the mechanism of this protease in R. anatipestifer pathogenesis warrants further investigation.

In summary, the results presented here demonstrate that R. anatipestifer Yb2 encodes at least 19 T9SS-secreted proteins in its genome, which have multiple functions in bacterial virulence. The potential mechanisms of these effectors may contribute to the pathogenicity of R. anatipestifer, which need to be further verified experimentally.

MATERIALS AND METHODS

Bacterial strains and growth media.

The bacterial strains and plasmids used in this study are listed in Table S2. R. anatipestifer serotype 2 strain Yb2 was the wild-type strain and R. anatipestifer serotype 10 strain HXb2 provided the erm^r element. T9SS mutant Yb2ΔgldM and complementation strain cYb2ΔgldM were constructed in our previous study (25, 28). The R. anatipestifer strains were grown at 37°C in tryptic soy broth (TSB; Difco, Franklin Lakes, NJ, USA) or on solid tryptic soy agar (TSA) plates. The E. coli strain was grown at 37°C on Luria–Bertani (LB) plates or in LB broth. Antibiotics were used at the following concentrations when required, unless otherwise indicated: ampicillin (100 μg/mL), erythromycin (1.0 μg/mL), kanamycin (50 μg/mL), streptomycin (50 μg/mL), and cefoxitin (5 μg/mL).

Bioinformatics analysis of R. anatipestifer genes with a conserved T9SS CTD.

The complete genomic gene annotation (https://www.ncbi.nlm.nih.gov/genome/2536?genome_assembly_id=235107) and the conserved domain analysis (https://www.ncbi.nlm.nih.gov/Structure/cdd) were used to identify genes encoding proteins with a CTD in R. anatipestifer strain Yb2. The predicted functions of the proteins were analyzed at https://www.uniprot.org/uniprot.

LC–MS/MS analyses.

The secreted proteins of R. anatipestifer wild-type strain Yb2, mutant strain Yb2ΔgldM, and complementation strain cYb2ΔgldM were analyzed with LC–MS/MS in our previous study (28). In brief, the strains were grown in 200 mL of ADCF medium at 37°C with shaking until an optical density at a wavelength of 600 nm (OD₆₀₀) of 0.8. The cultures were centrifuged at 10,000 × g for 10 min at 4°C, and the resulting supernatants were purified by passage through 0.22 μm polyvinylidene difluoride (PVDF) filters. The proteins in the supernatants were collected with 3-kDa Amicon Ultra Centrifugal Filter Units (Sigma) and stored at −80°C until analysis. The protein quality was determined with SDS-PAGE followed by Coomassie brilliant blue staining. Enzymatic in-gel digestion was performed at the Madison Mass Spectrometry Facility of Shanghai Applied Protein Technology Co. Ltd (Shanghai, China).

Expression of the 19 recombinant proteins with a CTD and mouse antiserum production.

To express the predicted CTD-containing proteins, recombinant plasmids were constructed using the appropriate primers and gene sequences (Table S2). The six genes AS87_RS00980, AS87_RS06600, AS87_RS03200, AS87_RS02020, AS87_RS07295, and AS87_RS00835 were amplified with the respective primers and inserted into the pET-30a(+) vector. The eight genes AS87_RS08215, AS87_RS02625, AS87_RS03090, AS87_RS02950, AS87_RS02955, AS87_RS09020, AS87_RS07160, and AS87_RS02875 were ligated into the pCold I vector. The four genes AS87_RS07755, AS87_RS09040, AS87_RS02840, and AS87_RS04975 were integrated into the pCold TF vector. AS87_RS04190 was inserted into the pGEX-4T vector. The cloning sites in all the vectors were unitive choose to the BamHI and SalI sites. Escherichia coli host cells were transformed with each individual plasmid and cultured, and recombinant protein expression was induced with 1 mM isopropyl β-d-1-thiogalactopyranoside. After culture, the cells were harvested by centrifugation at 10,000 × g for 5 min at 4°C, resuspended in lysis buffer (20 mM Na₃PO₄, 0.5 M NaCl, pH 7.4), and purified with HisTra affinity columns (GE Healthcare, Uppsala, Sweden) or BeaverBeads GSH (Beaver, Boston, MA, USA), according to the manufacturer’s protocol. Aliquots of the fractions obtained were analyzed with SDS-PAGE. The protein concentrations were measured with a BCA Protein assay kit (Beyotime, Shanghai, China), with bovine serum albumin (BSA) as the standard.

Six-week-old BALB/c mice were immunized three times with the purified recombinant proteins at 2-week intervals at a dose of 0.2 mg of purified protein in the same volume of Montanide ISA 50 V adjuvant (Seppic, Paris, France). The antiserum titers were measured with ELISAs to be >1:8000.

Western blotting.

The R. anatipestifer wild-type strain Yb2, mutant strain Yb2ΔgldM, and complementation strain cYb2ΔgldM were grown in 400 mL of ADCF-MAb medium (HyClone, USA) at 37°C with shaking until an OD₆₀₀ of 0.8. The cultures were centrifuged at 8,000 × g for 10 min at 4°C. The supernatants were collected and passed through 0.22 μm PVDF filters to remove the bacterial residue before the detection of the secreted proteins. The bacterial pellets were collected to detect the proteins in the whole bacterial cells. The secretory proteins in the supernatants were collected with 3-kDa Amicon Ultra Centrifugal Filter Units (Sigma), and the final protein concentrations were measured with a BCA Protein assay kit (Beyotime, Shanghai, China). In total, 20 μg of secreted proteins or 50 μg of whole bacterial cells from each sample were separated with SDS-PAGE and then electrophoretically transferred onto nitrocellulose membranes (Millipore, Billerica, MA, USA) for Western blotting analysis. Mouse antisera against each recombinant protein were used as the primary antibodies, and a horseradish peroxidase-conjugated goat anti-mouse IgG polyclonal antibody (Bio-Rad Laboratories, Hercules, CA, USA) was used as the secondary antibody. A mouse anti-TonB-dependent receptor antibody was used as the control for protein loading. The specific bands were developed with the Basic Luminol Chemiluminescent Kit (SB-Wb001) (Share-bio, Shanghai, China), visualized with the Tanon 5200 automatic chemiluminescence image analysis system (Tanon, Shanghai, China), and quantified with the ImageJ software (National Institutes of Health, Rockville, USA).

Quantitative PCR.

The total RNA of R. anatipestifer wild-type strain Yb2, mutant strain Yb2ΔgldM, and complementation strain cYb2ΔgldM was extracted with TRIzol Reagent (Invitrogen, Carlsbad, CA, USA), as described previously (48). The purified RNA was then treated with amplification-grade DNase I (Invitrogen) to remove contaminating DNA and reverse transcribed into cDNA with an iScript cDNA Synthesis Kit (Invitrogen). Quantitative PCR was performed with gene-specific primers to confirm the transcription levels of the T9SS effector proteins in R. anatipestifer (Table S2). The data were analyzed with the cycle threshold (2^−ΔΔCT) method after normalization to the expression of the reference gene encoding l-lactate dehydrogenase (49).

Construction of mutant and complementation strains.

Natural transformation was performed as described previously (30, 50). Briefly, mutagenized PCR fragments were created by joining the approximately 800-bp left (upstream) flanking sequence and the 800-bp right (downstream) flanking sequence of the sequence to be deleted (when the gene size exceeded 3,000 bp, the flanking sequence length was increased to 1,000–1,200 bp) and the gene encoding erythromycin resistance (erm^r). The three PCR fragments (upstream, downstream, and the erm^r cassette) were amplified with overlap PCR using the primer pair gene-up-F and gene-down-R. The corresponding primers are shown in Table S2. The wild-type strain Yb2 was grown in liquid TSB to logarithmic growth phase and centrifuged at 8,000 × g for 10 min at 4°C, and the pellet suspended in fresh TSB buffer to an OD₆₀₀ of 1.0. The bacterial suspensions (0.3 mL) were transferred to sterilized tubes; 1 μg of DNA fragment carrying the erm^r was added and the cells incubated for 1 h at 37°C with shaking at 220 rpm. The bacterial cultures were then plated onto TSA plates containing 1.0 μg/mL erythromycin antibiotic. The bacterial colonies were identified with analytical PCR and DNA sequencing to confirm that they carried the appropriate insertions.

The modified shuttle plasmid pCP29-ompA was used to construct the complementation strains of selected mutants, whose virulence was attenuated >100-fold (according to LD₅₀ determination). Briefly, the open reading frames (ORFs) of the genes were amplified from R. anatipestifer Yb2 genomic DNA with the appropriate primers (Table S2). After digestion with SphI and XhoI, each fragment was inserted into pCP29-ompA at the sites digested with the same enzymes, generating the complementation plasmids. The complementation strains were generated by transferring the appropriate plasmid into the mutant strain by conjugation and were identified with analytical PCR.

LD₅₀ determination.

LD₅₀ was determined as described previously (25). Briefly, 60 14-day-old Cherry Valley ducks were randomly divided into six groups of 10 ducks. The ducks in groups 1–6 were infected intramuscularly with R. anatipestifer at 10⁵, 10⁶, 10⁷, 10⁸, 10⁹, or 10¹⁰ CFU/duck, respectively. The infected ducks were housed in separate cages under a 12-h light–dark cycle with free access to food and water during the study. The clinical symptoms and deaths of the ducks were recorded daily for 7 days, and the bacterial LD₅₀ was calculated with an improved version of Karber’s method (51).

Measurement of protease activity.

The proteolytic activity of each strain was quantified as described previously (28). In brief, the wild-type strain Yb2 and the 19 mutant strains were incubated in ADCF-MAb medium (HyClone) for 8 h at 37°C with shaking at 220 rpm, and 5 mL samples were collected. The cell-free culture medium of the strains was then collected by centrifugation at 19,950 × g for 10 min at 4°C, and the supernatants were purified by passage through 0.22 μM HT Tuffryn syringe filters (Pall Life Sciences, Ann Arbor, MI, USA). The bacterial pellets were dried in an 80°C heat block for 3 h, and the dry weights of the cell pellets were measured to calculate the proteolytic activity. A 2% azocasein (Sigma) solution was prepared in 0.05 M Tris–HCl (pH 7.4). The cell-free supernatant (50 μL) was mixed with 50 μL of substrate (azocasein) and incubated at 37°C for 6 h. Triplicate assays were performed for each supernatant sample and negative control (50 μL of ADCF-MAb medium). After incubation, 130 μL of 10% trichloroacetic acid was added to each sample, then let stand for 10 min at room temperature followed by centrifugation at 19,950 × g for 20 min at 4°C to remove the precipitated azocasein. An aliquot (100 μL) of the soluble supernatant was added to a flat-bottomed 96-well plate, and 200 μL of 1 M NaOH was added and mixed. The OD₄₅₀ was determined with an iMark Microplate Absorbance Reader (Bio-Rad Laboratories). The raw OD₄₅₀ values obtained from triplicate assays were averaged for each supernatant sample, and the mean negative control OD₄₅₀ was subtracted from these values. The proteolytic activity per milligram of dry cells was calculated as (mean sample OD₄₅₀ − mean negative control OD₄₅₀) × 1000 × 100/dry weight of the bacterial pellet.

Statistical analysis.

Statistical analyses were conducted using GraphPad Software version 6.0 (La Jolla, CA, USA). One-way analysis of variance (ANOVA) was used for analyses of protease activity, and two-tailed independent Student's t test was used for analyses of the qPCR assay. P values of <0.05 were considered significant.

Ethics statements.

The study protocol was approved by the Institutional Animal Care and Use Committee of Shanghai Veterinary Research Institute, the Chinese Academy of Agricultural Sciences (approval no. SHVRI-SZ-20200719-01), and was conducted in strict accordance with the recommendations outlined in the Guide for the Care and Use of Laboratory Animals. One-day-old Cherry Valley ducks were obtained from Zhuang Hang Duck Farm (Shanghai, China) and housed in separate cages with a 12-h light–dark cycle and free access to food and water. Six-week BALB/c mice were obtained from Shanghai SLAC Laboratory Animal Corporation Limited (Shanghai, China) and housed in cages at a controlled temperature of 28–30°C under biosafety conditions, with water and food provided ad libitum.