Screening and identification of SipC-interacting proteins in Salmonella enteritidis using Gal4 yeast two-hybrid system in duck

Ethical statement

All animal experiments used in this study were approved by the Institutional Animal Care and Use Committee of Yangzhou University (Jiangsu, China) and were strictly implemented according to the regulations for experimental animals. An ordinary housing facility was used and was consistent with the national standard, Laboratory Animal Requirements of Environment and Housing Facilities (GB 14925-2001). Laboratory animal care and the animal experiment protocols and conditions conformed to the Jiangsu Administration Rule for Laboratory Animal Use.

Isolation and culture of dGCs

Healthy, Salmonella-free, 26-week-old Shaoxing ducks were collected from the National Waterfowl Conservation Farm (Taizhou, Jiangsu, China). Separation and cultivation of dGCs was performed as previously described (Gilbert et al., 1977). Briefly, 10 to 15 adult prehierarchical follicles (small yellow follicles), which have thicker granulosa cell layer (Yang et al., 2019) of egg-laying ducks, were collected under aseptic conditions and rinsed with Ca²⁺- and Mg²⁺-free phosphate-buffered saline (PBS) to remove the yolks and vitelline membrane as fully as possible. Thereafter, the tissue was cut into 1–2 mm³-sized blocks, digested with 1 mg/ml collagenase (Type II; Sigma Chemical Company, St. Louis, MO, USA) at 37 °C for 5 min, and filtration through a 200-µm nylon filter. Filtered suspensions were centrifuged twice at 67 × g for 5 min. Pellets were washed with M199 media to remove the remaining collagenase and cell debris, resuspended in three mL 50% Percoll, and centrifuged at 421 × g for 15 min, after which the cell layer was aspirated. Granulosa cell suspensions were prepared by adding a pre-configured M199 media (5% fetal calf serum, two mmol/L L-glutamine, five µg/mL transferrin, 10 µg/mL insulin, 1.75 mM HEPES) and counted following staining with 0.1% trypan blue. Suspensions with a cell survival rate greater than 90% were used for experiments. Cells were used for follow-up experiments in disposable culture flasks after 24 h when the completely adherent. Cell purity was initially determined by hematoxylin and eosin (H&E) staining.

Construction of yeast two-hybrid cDNA library of dGCs

cDNA and yeast two-hybrid libraries were constructed in collaboration with Shanghai OE Biotech. Total RNA was extracted from the dGCs using TRIzol (Invitrogen, Carlsbad, CA, USA), and mRNA was isolated using an Oligotex mRNA Midi Kit (Invitrogen). RNA integrity and size distribution range of mRNA were detected by 10 g/L denaturing gel electrophoresis. Total RNA and mRNA concentrations were measured using a Thermo Nanodrop nucleic acid analyzer.

In order to construct a cDNA library containing three reading frames, the corresponding attB linker sequences at the 5′and 3′ends of the corresponding cDNA to ensure that the cDNA was translated in the correct reading frame (Table 1). The CloneMiner II cDNA Library Construction Kit (Invitrogen) was used according to the manufacturer’s instructions. Column chromatography was performed to fractionate and collect cDNA. Collected cDNA fragments were digested using a pDONR 222 and BP Clonase II enzyme mix for BP recombination. Recombination products were then transformed into E. coli DH10B competent cells by electroporation and cultured at 37 °C and 4 × g for 1 h, and primary library bacteria were obtained. The insert size and recombination rates of the positive clones were determined by PCR.

For construction of a yeast two-hybrid cDNA expression library, the primary library containing 5 × 10⁶ to 1 × 10⁷ positive clones was inoculated into 100 mL broth containing kanamycin (final concentration 50 µg/mL) at 30 °C and 4 × g until OD₆₀₀ reached 1.0. The primary library plasmids were extracted using a PureLink HQ Mini Plasmid DNA Purification Kit (Invitrogen), and subsequently diluted to 300 ng/µL, and then added one µL to compatible vectors pGADT7-DEST, LR Clonase II Mix, and ddH₂O to carry out the LR recombination reaction. The recombination products were transformed into E. coli DH10B competent cells by electroporation and cultured at 37 °C and 4 × g for 1 h to obtain bacteria expressing the library. The size of the inserts and recombination rates of the positive clone were determined by PCR.

Construction of the bait plasmid pGBKT7-sipC and detection of its self-activation activity

According to the sequence of SE sipC (Genbank ID: NC_003197.2), specific primers were synthesized with the upstream primer harboring an Nde I restriction site, and the downstream primer harboring a Pst I restriction site and a stop codon. The recombinant pGBKT7-sipC plasmid (bait plasmid) was confirmed by double restriction enzyme digestion and sequencing (Sangon Biotech, Shanghai, China). The recombinant pGBKT7-sipC plasmid was transformed into yeast strain Y2HGold as per the manufacturer protocol of the Yeastmaker™ Library Construction & Screening Kit (Clontech, Mountain View, CA, USA). Transformants were screened on SD/-Trp/X-α-Gal agar plates at 30 °C for 3–5 days. Growth, size, and color changes of colonies were assessed to determine if the bait plasmid had self-activation activity. Positive (pGBKT7-53/pGADT7-T) and negative (pGBKT7-Lam/pGADT7-T) controls were coated on SD/-Trp/-Leu/X-α-Gal media plates, respectively. Only bait plasmids without auto-activation can be used in yeast two-hybrid screening.

Yeast two-hybrid screen based on co-transformation of bait and prey plasmids

In order to screen host proteins that interact with the SipC bait in the yeast two-hybrid system, a cDNA library of dGCs, pGBKT7-sipC, and the library plasmid were co-transformed into Y2HGold using a Yeastmaker™ Library Construction & Screening Kit according to the manufacturer protocol. Co-transformants were then grown on higher-stringency SD/-Ade/-His/-Leu/-Trp/X-α-Gal/ABA agar plates at 30 °C for 3–5 days. Single blue colonies were selected and serially passaged three times on SD/-Ade-His-Leu-Trp/X-α-Gal/ABA plates. Positive colonies were thereafter re-seeded three times in SD/-Ade-His-Leu-Trp liquid media. Prey plasmids were extracted from putatively positive clones using the Easy Yeast Plasmid Isolation Kit (Clontech).

Confirmation of the positive clones of interactions

To confirm interactions, co-transformations of pGBKT7-sipC bait into Y2HGold with each prey plasmid in putatively positive hits were performed. Briefly, the prey plasmids were extracted from putatively positive clones using an Easy Yeast Plasmid Isolation Kit (Clontech, Mountain View, CA, USA). Subsequently, each prey plasmid was transformed into E. coli DH5α competent cells (Transgen, Beijing, China) and purified from transformants growing on selected LB/ampicillin agar plates using a Plasmid Mini Kit I (Omega Bio-Tek, Norcross, GA, USA). Each putatively positive prey plasmid was co-transformed with pGBKT7-sipC bait and pGBKT7 plasmids into Y2HGold. Cotransformants were grown on SD/-Ade-His-Leu-Trp/X-α-Gal/ABA plates to test for interactions. The cotransformant containing pGADT7-T and pGBKT7-Lam grown on an SD/-Ade-His-Leu-Trp/X-α-Gal/ABA plate was used as a negative control, and the cotransformant containing pGADT7-T and pGBKT7-53 grown on SD/-Ade-His-Leu-Trp/X-α-Gal/ABA was used as a positive control. True positive interactions were indicated by blue colonies under these conditions. Positive prey plasmids were sequenced, and results were blasted against NCBI databases.

GST pull-down assay

GST pull-down assays were performed as previously described (Hallstrom et al., 2015). Briefly, GST-SipC fusion plasmid (pGEX-6P-GST containing the GST-tagged C-terminus with SipC) was transformed to BL21 (DE3) followed by cultured induction with IPTG to express the GST-SipC fusion protein. Then, the recombinant bacteria BL21 were centrifuged at 4,315 × g. The pellet was resuspended in lysis buffer (100 mM NaCl, 25 mM Tris, 1% TritonX-100, 0.2 mM lysosome, 1 mM PMSF, 10% glycerin, pH 8.0), sonicated five times (1.5 w/s) and centrifuged at 9,710 × g at 4 °C for 10 min. Pre-cleared supernatants were incubated with glutathione-sepharose affinity matrix beads for 2 h at room temperature. Plasmid constructs pEGFP-PERP and pEGFP-TAB2 were transfected into dGCs using Lipofectamine 2000. The transfected dGCs were lysed and incubated with the SipC-GST-bound beads overnight at 4 °C with end-over-end rotation according to the manufacturer instructions. After washing steps with 1 × PBS, GST-SipC protein complexes were eluted with 10 mM reduced glutathione which was dissolved in lysis buffer. The eluates were then diluted in 20 µL 5 × SDS protein-loading buffer (125 mM Tris-HCl, pH 6.8, 4% w/v SDS, 20% glycerol, 100 mM DTT, 0.02% w/v bromophenolblue) boiling for 5 min. SDS-PAGE detection and Western bolting analysis were then performed. PERP polyclonal antibody (1:2,000; Abcam, Cambridge, MA, USA) and TAB2 polyclonal antibody (1:1,000; Abcam, Cambridge, MA, USA) were used as primary antibodies, and goat anti-rabbit IgG-HRP (1:5,000) was used as secondary antibody. Diaminobenzidine (DAB) reaction was used for color development.

Transient transfection

PERP and TAB2 cDNA were amplified by RT-PCR using gene-specific primers (Table 1). The PCR products were cloned into the expression vector pEGFP-N1. A small interfering RNAs (siRNA) targeting PERP (pGPU6/GFP/Neo-PERP) and TAB2 (pGPU6/GFP/Neo-TAB2) (GenePharma, Shanghai, China) were used to decrease PERP and TAB2 expression, respectively. The interference efficiency of the three targeted interference vectors was detected by qPCR, and the most effective siRNA was used for subsequent experiments. One of the three PERP-siRNAs, TAB2-siRNAs or the negative control siRNA were transfected into dGCs using Lipofectamine 2000 (Invitrogen, Grand Island, NY, USA) according to manufacturer instructions, respectively.

Quantitative real-time PCR analysis

Primers for qRT-PCR specific to duck immune response genes IL-6, IL-1β, TNF-α, and IFN-γ are listed in Table 1. The housekeeping genes glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as reference controls to normalize the expression level. Total RNA was extracted using TRIzol Reagent (Takara, Dalian, China), according to the manufacturer’s instructions. We performed qPCR using an ABI 7500 system (Applied Biosystems, Foster City, CA, USA). Each reaction volume contained 10 µL 2 × SYBR Premix Ex Taq II (Takara), 0.4 µL forward primer (10 µM), 0.4 µL reverse primer (10 µM), 0.4 µL 50 × ROX Reference Dye II (Takara), 2 µL cDNA, and ddH₂O to 20 µL. The amplification program was performed as follows: initial denaturation at 95 °C for 30 s, followed by 40 cycles of denaturation at 95 °C for 5 s and annealing at 60 °C for 34 s. To analyze the specificity of the amplified products, we collected multiple information points for melting curve analysis. The program was as follows: 95 °C for 15 s, 60 °C for 1 min, 95 °C for 15 s, and 60 °C for 15 s. Data were analyzed by the 2-^ΔΔCt relative quantitative method.

Adhesion and invasion assays

The adhesion assay was performed as previously described (Jouve et al., 1997; Lane & Mobley, 2007). Briefly, bacteria were grown to an OD₆₀₀ nm of 2.0 in LB broth at 37 °C. dGCs were seeded in 96-well tissue culture plates at a concentration of 1 × 10⁵ cells per well and grown for 24 h. Cell monolayers were washed three times with sterile PBS (pH 7.2). Then, dGCs were incubated with 100 µL bacterial suspension in DMEM with multiplicity of infection (MOI) of 10 per well of a 96-well plate at 37 °C. Infections were performed in triplicate. After 1 h incubation, the infected cell monolayer was gently washed three times with PBS to remove loosely adherent bacteria. For invasion assays, PBS contained 50 µg/mL gentamicin was used to kill noninvasive bacteria. After 1 h, cells were lysed with 0.5% Triton X-100 (Solarbio, Beijing, China) for 30 min. Lysates were serially diluted and plated on LB-agar plates for bacterial growth. The assay was performed in triplicate.

Statistical analysis

Each experiment was repeated at least three times and the data were analyzed using an independent-sample t test (SPSS version 13).

Construction of yeast two-hybrid cDNA library of dGCs

The dGCs were in round or oval shapes at first, after 24 h in culture, the monolayers adhered to the wall and spread out like pebbles. Granular material was observed in the center of cells. Small particles were observed by H&E staining (Fig. S1). Purified dGCs (2 × 10⁷) were used for cDNA library construction. A diluted primary library solution (50 µL at 1:1000 dilutions) was spread on kanamycin-resistant LB plates for 14 h. According to the formula, the titer of the primary library was calculated to be 2. 97 × 10⁶CFU/mL, and the total library capacity was 1. 19 × 10⁷CFU. Twenty-four clones were randomly picked for PCR analysis. All clones were positive, with different sizes of fragments inserted. The recombination rate of the primary library was 100%, and the average size of insert was >one kb (Fig. S2A). The diluted expression library (50 µL at 1:1,000) was plated on ampicillin-containing LB plates for 14 h. According to the formula, the expression library titer was 1.36 × 10⁶ CFU/mL, and the total library capacity was 5.44 × 10⁶ CFU. Twenty-four clones were randomly selected for PCR analysis. All selected clones were positive, with different sizes of fragments inserted, indicating that the recombination rate of the expression library was 100%. The average size of insert was >one kb (Fig. S2B). These results indicated that the cDNA library could be used for yeast two-hybrid screening.

Screening of Salmonella enteritidis SipC-interacting proteins using a Gal4 yeast two-hybrid system

pGBKT7-sipC demonstrated successful insertion of the gene into the yeast expression vector pGBKT7 by sequencing. To test the auto-activation activity of the bait protein in yeast cells, the pGBKT7 and pGBKT7-sipC bait plasmids were transformed into Y2HGold and subsequently the transformants were grown on SD/-Trp-Leu/X-α-Gal media plates. As a negative control, the empty vector failed to activate the reporter gene (Fig. 1A). Transformants containing the pGBKT7-53 and pGADT7-T control vectors were grown on SD/-Leu-Trp/X-α-Gal media and were blue in color as a positive control experiment (Fig. 1B). Although transformants containing pGBKT7-sipC grew on SD/-Trp/X- α-Gal media, not blue colonies were observed. The bait gene did not activate the reporter genes Ade2 and His, indicating an absence of self-activation in pGBKT7-sipC (Fig. 1C).

After cotransformation of pGBKT7-sipC and prey plasmids and growth on SD/-Ade-His-Leu-Trp/X-α-Gal/ABA plates, 20 blue clones were obtained. Subsequently, these 20 prey plasmids were isolated from their corresponding colonies and rescued through transformation of E. coli DH5 α cells. To eliminate false positive hits, each of the 20 prey plasmids was co-transformed with pGBKT7-sipC into Y2HGold cells. Cotransformants were cultivated on SD/-Ade-His-Leu-Trp/X-α-Gal/ABA plates. Each of the 20 cotransformants grew as blue colonies on SD/-Ade-His-Leu-Trp/X-α-Gal/ABA plates. Meanwhile, each of the 20 prey plasmids was co-transformed with pGBKT7 plasmid into Y2HGold cells, no colony growth was observed on SD/-Ade-His-Leu-Trp/X-α-Gal/ABA plates. Based on the above results, 12 host proteins were identified to interact with SipC (Figs. 1D–1I). To confirm the nucleotide sequence information of the identified host proteins interacting with SipC, the 20 prey plasmids were sequenced and analyzed using the BLAST tool in NCBI (DOI 10.6084/m9.figshare.7844426.v1). Twelve of the interaction plasmids were identical to SipC (Table 2).

Identification of Salmonella enteritidis SipC-interacting proteins PERP/TAB2 in using GST pull-down assays

Detection of protein–protein interaction using the GST fusion protein pull-down technique. Full length SE SipC cDNA was introduced into a GST expression vector, downstream of the glutathione-S-transferase coding part. Following expression in bacteria, the corresponding fusion protein and the transfected cell lysates were prepared, then their purity and interaction relationship were detected by SDS-polyacrylamide gel and Western bolting analysis. GST pull-down assays revealed that GST did not bind to host protein PERP/TAB2. However, GST-SipC bound specifically to PERP/TAB2 with normal expression of each protein. These results support the direct and specific interaction between the SE effector protein SipC and the host protein PERP/TAB2 (Fig. 2).

PERP interacts with SipC to promote SE adhesion and invasion to dGCs

To verify the function of PERP interacted with SipC in SE adhesion and invasion, plasmid pEGFP-PERP transfection was performed with dGCs at 90% confluence using Lipofectamine 2000. After 4 to 6 h transfection, the media was changed to serum-free DMEM. Then cells were treated with MY₁WT and MY₁ΔsipC to detect the amount of bacteria. We found a significant increase after 1 h post-infection (MOI = 10) in the number of SE MY₁WT infecting PERP-overexpressing dGCs compared with that of control dGCs (P < 0.01) (Figs. 3A–3B, Table S1). Then, we applied siRNA-RNAi systems to specifically suppress PERP in dGCs. The ratio of siRNA-treated cells with green fluorescent protein expression exceeded 80%, indicating successful infection. PERP knockdown followed decreased SE adhesion and invasion by treatment with MY₁WT in dGCs, respectively (P <0.01) (Figs. 3C–3D, Table S1). These results suggest that PERP may play a role in promoting the adhesion and invasion of SE MY₁WT.

The expression of duck immune response genes IL-6, IL-1β, TNF-α, and IFN-γ was also be detected when PERP-overexpressing or knockdown in dGCs during treatment with MY₁WT or MY₁ΔsipC, respectively. There was a significant increase level in immune response of IL-1β and TNF-α in infecting MY₁WT in PERP-overexpressing dGCs compared with that of control dGCs (P <0.05) (Fig. 3E, Table S2). When PERP knockdown followed by treatment with MY₁WT in dGCs, there is an opposite trend emerged of the mRNA expression level of TNF-α (P <0.05) (Fig. 3G, Table S2), however, overexpression of PERP or knockdown PERP showed no significant mRNA level change of IL-6, IL-1β, TNF-α, and IFN-γ during MY₁ΔsipC infection (Figs. 3F and 3H, Table S2).

TAB2 interacts with SipC to promote SE activating inflammatory to dGCs

Similarly, we overexpressed and knockdown TAB2 in dGCs treated with MY₁WT or MY₁ΔsipC to determine the effects of SE adhesion and invasion and simultaneously detected the expression of duck immune response genes IL-6, IL-1β, TNF-α, and IFN-γ in MY₁WT- or MY₁ΔsipC-infected dGCs, respectively. We found that there was no significant difference after 1 h post-infection (MOI = 10) in the amount of SE MY₁WT infecting TAB2-overexpressing dGCs compared with that of control dGCs in both adhesion and invasion process (Figs. 4A–4B, Table S1). Analogously, the number of SE in the knockdown group showed no significant difference compared with that of control group (Figs. 4C–4D, Table S1).

For the mRNA expression of some representative inflammatory gene, IL-6, IL-1β, TNF-α, and IFN-γ within TAB2-overexpressing dGCs infected by MY₁WT were significantly higher than that of the control group (P <0.05) (Fig. 4E, Table S2). In contrast, when TAB2 knockdown followed by treatment with MY₁WT in dGCs significantly decreased the expression levels of IL-6, IL-1β, and IFN-γ compared to the control group (P < 0.05) (Fig. 4G, Table S2). Meanwhile, MY₁ΔsipC infection had no influence in the mRNA expression of inflammatory gene regardless of TAB2 knockdown or overexpression (Figs. 4F and 4H, Table S2). These results suggest that TAB2 may play a role in activating the production of inflammatory response during the SE MY₁WT infection.