Comparative roles of clpA and clpB in the survival of S. Typhimurium under stress and virulence in poultry

By assisting in the proteolysis, disaggregation and refolding of the aggregated proteins, Caseinolytic proteases (Clps) enhance the cellular survival under stress conditions. In the current study, comparative roles of two such Clps, ClpA (involved in proteolysis) and ClpB (involved in protein disaggregation and refolding) in the survival of Salmonella Typhimurium (S. Typhimurium) under different stresses and in virulence have been investigated. clpA and clpB gene deletion mutant strains (∆clpA and ∆clpB) of S. Typhimurium have been hypersensitive to 42 °C, HOCl and paraquat. However, the ∆clpB strain was comparatively much more susceptible (p < 0.001) to the above stresses than ∆clpA strain. ∆clpB strain also showed reduced survival (p < 0.001) in poultry macrophages. The hypersusceptibilities of ∆clpB strain to oxidants and macrophages were restored in plasmid based complemented (∆clpB + pclpB) strain. Further, the ∆clpB strain was defective for colonization in the poultry caecum and showed decreased dissemination to the spleen and liver. Our findings suggest that the role of ClpB is more important than the role of ClpA for the survival of S. Typhimurium under stress and colonization in chickens.

Food borne-infections account for about 86% of human cases of non-typhoidal salmonellosis 1 . Salmonella enterica serovar Typhimurium (S. Typhimurium) is one of the most frequently isolated serovars of Salmonella from food-borne infections 2 . S. Typhimurium causes mild to moderate gastroenteritis in healthy individuals 3 , however, it is associated with fatal infections in young, old and immunocompromised people 4,5 . In chickens, S. Typhimurium provokes very mild gastroenteritis. However, infected hens serve as chronic carriers without showing any clinical signs of infection but lay contaminated eggs 6 . Amongst the various food sources entailed in human Salmonella infection, poultry products have been reported to be the foremost complicit in outbreaks across the world (CDC, 2013) 7 . After entering via the oral route, S. Typhimurium reaches into the intestine of the host and penetrates epithelial linings. This is brought about with the aid of T3SS encoded by 40 kb region located in the Salmonella genome called Salmonella Pathogenecity Island (SPI) 1. The T3SS effector molecules stimulate the engulfment of bacteria into the epithelial membrane by mediating rearrangement of actin filaments and causing formation of membrane ruffles around bacteria 8 . Macrophages phagocytose the bacteria once Salmonella reaches sub-mucosa. T3SS2 effector molecules secreted by SPI2 helps in the survival and multiplication of Salmonella inside Salmonella containing vacuole (SCV) present within macrophages. Finally, localization occurs through blood circulation and disseminating macrophages into target organs of predilection especially lymph organs (spleen and caeca) and liver 9 . Different studies reveal differences in the extent of colonization of organs depending on the host infected. In poultry, major colonization is usually seen in caeca while spleen and liver harbour four orders of magnitude of bacteria lower than that of caeca 10 .
Inside the host S. Typhimurium encounters several stresses, including oxidants generated by the phagocytes and higher (42 °C) body temperature of poultry 11 . Superoxide anion (O 2 − ), hydrogen peroxide (H 2 O 2 ) and hypochlorous acid (HOCl) are some of the important reactive oxygen species (ROS) generated by phagocytic cells 12 .
The exposure of S. Typhimurium to high temperature and oxidants results in unfolding and aggregation of proteins. These aggregates contain higher amounts of intermolecular β-sheet structures 13 and are functionally inactive. The accumulation of protein aggregates hampers the cellular survival 14 therefore they need to be taken care of at any cost.
To withstand such assaults, bacteria have evolved several mechanisms. These systems are composed of primary antioxidants, protein repair enzymes and various families of proteases and molecular chaperones. Proteases and molecular chaperones are categorized under heat shock proteins (Hsps) which specifically identify aggregated/misfolded proteins. Subsequently, Hsps cause either proteolysis or disaggregation and refolding 15,16 of protein aggregates. Expression of molecular chaperones are found to be induced under stress conditions [17][18][19] .
In S. Typhimurium, three types of energy dependant proteases, including Clps family (viz. ClpA, ClpB, ClpX, ClpP), Lon and HslVU are present 20 . Clps are classified under the Hsp100 family proteins and function as both proteases and chaperones 21 . The ClpA and ClpB/Hsp104 chaperones belong to class 1 AAA + (ATPases associated with various cellular activities) protein superfamily 22,23 . ClpA is a part of the two-component protease (ClpAP) that binds protein substrates and presents them to the protease (ClpP) for degradation [24][25][26][27] . Recognition of substrate proteins is undertaken by the N-terminal domain of ClpA which also simultaneously unfolds the proteins 28,29 . This unfolding aids the proteolytic degradation occurring in the main proteolytic site of ClpP which receives these unfolded proteins via translocation through the hexameric body of ClpA 30,31 .
ClpB is an essential protein of heat shock response. Structurally, ClpB possesses a longer middle region known as ClpB/Hsp 104-linker which is essential for the chaperone activity 32,33 . Although sharing a 42% sequence identity and 64% sequence similarity with ClpA 34 , ClpB functions differently than ClpA. Instead of degradation, ClpB causes disaggregation and refolding of protein complexes in cooperation with DnaK, DnaJ and GrpE (KJE) chaperones 35 . ClpB first acts as a molecular chaperone and catalyses fragmentation of large protein aggregates into small fragments 36 . Subsequently, the activities of KJE resolubilise and refold these smaller fragments rendering their conformation more or less akin to their native forms 37 . Another thought suggests that ClpB/KJE act as a bichaperone in which the aggregated proteins are threaded into the ClpB/KJE complex and then unfolded proteins are extracted by the translocation activity of ClpB 38,39 .
ClpA and ClpB ATPases are implicated in stress tolerances of many organisms and are reported to protect vital cellular proteins during stress conditions [40][41][42] . Further, the ClpA and ClpB ATPases are found to play very important roles in the virulence of several bacterial pathogens [43][44][45] . However, the role of ClpA and the comparative importance of ClpA and ClpB in the stress survival and virulence of S. Typhimurium are not known. Here we evaluated and compared the roles of ClpA and ClpB in the survival of S. Typhimurium under stress and virulence in poultry. To accomplish this, we have generated clpA and clpB gene deletion mutants and complemented strains. Then their sensitivities to high temperature, oxidative stresses and intramacrophage survival were assessed. Further, we have investigated the effect of clpA and clpB gene deletions in the colonization of S. Typhimurium in chickens. It is important to analyze the effect of gene deletion on in vitro growth of bacteria. ΔclpA and ΔclpB mutant strains did not exhibit any defective growth in LB broth (Fig. 1). ΔclpA and ΔclpB mutants and WT strains exhibited sigmoidal growth curves.

Contribution of clpA and clpB in the survival of S. Typhimurium at 42 °C and oxidative stress.
The body temperature of poultry is 42 °C, hence S. Typhimurium experience constant thermal stress inside birds 46 . Therefore, the contributions of clpA and clpB genes in the in vitro survival of S. Typhimurium at 42 °C have been evaluated. As compared to WT, ∆clpA and ∆clpB strains did not show any sensitivity to 37 °C ( Fig. 2A). However, ∆clpA and ∆clpB strains were hypersusceptible (p < 0.001) to 42 °C exposure (Fig. 2B). In comparison to ∆clpA strain, ∆clpB strain showed hypersusceptibility to 42 °C (Fig. 2B). The numbers of bacteria recovered following 120 h of incubations [log 10 colony forming unit(s) (CFUs)/ml as mean ± standard deviation (S.D.)] were 8.43 ± 0.032, 7.48 ± 0.008 and 6.15 ± 0.075 for WT, ∆clpA and ∆clpB strains respectively.
∆clpA and ∆clpB strains accumulate more aggregated proteins. As ClpA and ClpB inhibit protein aggregations, we hypothesized higher levels of protein aggregates in ∆clpA and ∆clpB strains. Following incubations with PBS or 1.5 mM HOCl, we observed more amounts of protein aggregates in ∆clpA and ∆clpB strains as compared to WT counterpart. The amount of loading was normalized in terms of CFUs and was equivalent to 45 × 10 7 CFUs/lane. Interestingly, the amount of aggregates was higher in ∆clpA strain as compared to ∆clpB strain (Fig. 6). Following 3 mM of HOCl exposure, we observed higher levels of aggregates in WT, ∆clpA and ∆clpB strains of S. Typhimurium. The amount of loading per lane was equivalent to aggregates isolated from 3 × 10 5 CFUs. Some of these aggregates failed to enter in the gel (marked by arrow in the figure). WT strain did not show much increase in protein aggregates following HOCl treatment. We observed HOCl-dose dependent increase in amount of aggregates in ∆clpB strain. ∆clpA strain showed more amounts of aggregates in 0 and 1.5 mM HOCl treated samples than WT and ∆clpB strains. Mid-log grown cultures of WT, ∆clpA, ∆clpB mutants and ∆clpA + pclpA (complemented) strains of S. Typhimurium were exposed to 0 or 1% paraquat. Cultures were then ten fold serially diluted and plated on HE agar plates. CFUs/ml were calculated after overnight incubation of plates. Data are presented as mean ± S.D. of three replicates (***denotes p < 0.001).  (Table 1). In birds infected with ∆clpA strain, 4/4 (100%) caeca were positive for Salmonella on 7 and 14 days post infection. After Figure 5. Sensitivities of WT, ∆clpA, ∆clpB, ∆clpA + pclpA and ∆clpB + pclpB strains to poultry macrophages stimulated by LPS. The mid log grown cultures were suspended in RPMI-1640. Macrophages were stimulated with LPS as described in materials and methods. The poultry macrophages were infected at MOI of 1:50 (macrophage: bacteria) for 2 h at 37 °C, 5% CO 2 . Macrophages were lysed by 0.1% TritonX-100. Bacteria were serially diluted and plated on HE agar plates and the number of bacteria recovered (log 10 CFUs/ml as mean S.D.) were counted the next day. Data is representative of two individual experiments (n = 3) (***denotes p < 0.001). Typhimurium. Midlog grown cultures of various strains were exposed to different concentrations of HOCl for 30 min. Aggregated proteins from such cells were prepared and analysed on 10% SDS-gel. Relative quantification was done using ImageJ software(NIH) bundled with 32 bit Java 1.8.0 and expressed as fold difference from protein aggregates isolated from WT strain incubated with 0 mM HOCl. Quantification is indicated below each lane. Experiment was performed two times.   Table). At different times post infection, the birds were dissected and presence of S. Typhimurium and mutant strains were confirmed by PCR (as described in materials and methods).  (Table 1). After concluding the role of clpB in the caecal colonization of S. Typhimurium, next, we analyzed the contribution of clpB in the dissemination of S. Typhimurium to poultry spleen and liver. We determined the bacterial loads in the spleen and liver on 7, 14 and 21 days post infection. In the spleen of WT strain infected birds, we obtained bacteria at all times post infection (Fig. 7A). The numbers of Salmonella recovered on 7, 14 and 21 days (log 10 CFUs/spleen as mean ± S.D.) were 2.57 ± 0.56, 1.46 ± 1.69 and 0.400 ± 0.80 respectively. Similarly, bacteria were recovered on all times post infection from the spleen of ∆clpA strain infected chickens. The counts were (log 10 CFUs/spleen as mean ± S.D.) 1.35 ± 1.57, 1.31 ± 1.52 and 0.40 ± 0.80 on 7, 14 and 21 days post infection (Fig. 7A). Interestingly as compared to that in WT strain infected birds, the bacterial loads in the spleen of ∆clpB strain inoculated chicks reduced significantly (p < 0.01) on 7 and 14 days post infection. On 21 days post infection, we did not recover any bacteria in the spleen of ∆clpB strain infected birds (Fig. 7A). The recovered numbers of bacteria (log 10 CFUs/spleen as mean ± S.D.) on 7 and 14 days post infection were 0.96 ± 1.22 and 0.40 ± 0.80 respectively.
In liver, we recovered bacteria from WT strain infected birds at all times post infection (Fig. 7B). The numbers of WT bacteria recovered (log 10 CFUs/gm of liver as mean ± S.D.) were 3.02 ± 0.68, 1.88 ± 1.36 and 0.5 ± 1.0 on 7, 14 and 21 days post infection respectively. In ∆clpA strain infected birds, we recovered 2.20 ± 1.49 and 0.75 ± 1.5 bacteria (log 10 CFUs/gm of liver as mean ± S.D.) on 7 and 14 days post infection (Fig. 7B). However, the bacterial loads in the liver of ∆clpB strain infected birds were significantly (p < 0.01) less on 7 days post infection (0.75 ± 1.5). Following 14 and 21 days post infection; we did not recover any bacteria from the liver of ∆clpB strain infected birds (Fig. 7B).

Discussion
S. Typhimurium encounters numerous stresses inside the host which primarily affect the integrity of cellular proteins. Molecular chaperones and proteases can refold or remove these abnormal proteins, thus play very important roles in maintaining the cellular homeostasis 48 . Clp proteases belong to AAA + class of proteins which are classified in to Class I and Class II types. Class I protease which include ClpA, ClpB and ClpC have two AAA + domains and degrade/disaggregate larger substrates while class II AAA + proteases like ClpX have only one AAA + domain and can only deal with smaller substrates. These proteins require various accessory proteins to execute their functions. ClpB complexes with DnaK, DnaJ and GrpE. ClpA, ClpX and ClpC associate with ClpP but each require a different set of adapter proteins. Like ClpS assists ClpA; MecA and YpbH coordinate with ClpC while ClpX complexes with SspB, RssB and UmuD 49 . Out of the known Clp proteases, we sought to analyse the relative importance of degradation versus disaggregation of protein aggregates in the survival of S. Typhimurium. In the current study, we have evaluated the comparative roles of ClpA (protein degradation chaperone) and ClpB (protein disaggregation chaperone) in the survival of S. Typhimurium under in vitro stress and in virulence.
Our SDS-gel analysis revealed greater amounts of protein aggregates in ∆clpA and ∆clpB strains than that in WT strain of S. Typhimurium. Interestingly, we observed higher amounts of protein aggregates in ∆clpA strain than that in ∆clpB strain (Fig. 6). Following incubation of clpB gene deletion strain of E. coli at 42 °C, increased aggregation of pre S2-β-galactosidase was observed 70 . In a separate study, degradation of green fluorescent protein aggregates at 42 °C was observed in ∆clpB and WT strains but not in ∆clpA strain of Brucella suis 52 . ClpA might be actively involved in degrading the protein aggregates in ∆clpB strain that could be the reason we did not observe higher levels of aggregates in ∆clpB strain as compared to ∆clpA strain.
Following exposure of 3 mM HOCl we have observed more protein aggregates in WT, ∆clpA and ∆clpB strains of S. Typhimurium. Some of these aggregates were resistant to SDS and beta-mercaptoethanol (β-ME) and failed to enter in stacking gel (Fig. 6, marked by arrow). Cell has a limited capacity to refold/remove aggregated proteins. The protein aggregates formed under severe stress (3 mM HOCl in our experiment) might be beyond the repair/ removal capabilities of cellular chaperone/protease systems 71 . Further, we have observed significant amount of killing of WT as well as of mutant strains following exposure to 3 mM HOCl (Fig. 4).
S. Typhimurium primarily colonizes in the caecum of young chicks 72 and disseminates to spleen and liver. Eventually these birds serve as a carrier and lay contaminated eggs. As compared to WT and ∆clpA strains, ∆clpB strain was highly defective in caecal colonization and dissemination to spleen and liver (Table 1 and Fig. 7). Similarly, clpA deletion did not affect the colonization of B. suis and H. pylori in mice 51,73 . However, following inoculation of a pool of transposon mutants in chickens, defective recovery of a clpB gene insertion mutant was  75 . clpB gene deletion strains of Listeria monocytogenes 21 , Francisella tularensis 66 and Leptospira interrogans 61 exhibited attenuated virulence and defective survival in animal models. In brief, our data and previous reports suggest that clpB is more important than clpA for the survival of bacterial pathogens in the host.
Both ClpA and ClpB play important roles in preventing protein aggregations in the cell. As ClpA degrades protein aggregates, protein pool in the cell needs to be replenished via translational synthesis. While ClpB is involved in disaggregation and refolding of existing protein aggregates which would be a rapid and energy efficient way for restoration of protein function(s) in the cell (Supplementary Fig. S4). Therefore, ClpB would play more important role than ClpA in the cellular survival under stress conditions. Consistence to this hypothesis, in our experiments we observed that clpB gene deletion strain of S. Typhimurium was much more susceptible to different stresses in vitro (than ∆clpA strain) and showed defective virulence in chickens. Culturing of S. Typhimurium. S. Typhimurium, its isogenic mutants and complemented strains were grown in Luria Bertani (LB) broth or Hektoen Enteric (HE) agar as described earlier 76 . Ampicillin at the concentration of 100 µg/ml was included while culturing the complemented strain.

Methods
Construction of clpA and clpB gene deletion mutants and complemented strains. Primers utilized in this study are listed in Table 2. The clpA and clpB gene deletion mutants and complemented strains in S.  Typhimurium were constructed as described earlier 76 . Briefly, the FRT flanked kanamycin cassettes were amplified using pKD4 plasmid as template. The kanamycin cassettes were electroporated in lambda red recombinase expressing S. Typhimurium. Positive recombinants were selected on kanamycin plates and confirmed by PCR. The kanamycin cassettes were then removed by flp recombinase. The clpA and clpB deletions mutants were confirmed by c and d primers located about 200 bp upstream and downstream from clpA and clpB genes. The gene deletion mutant strains were designated as ∆clpA and ∆clpB. For complementations, the clpA and clpB genes were amplified by primers e and f and cloned into plasmid pQE60 (clpA at HindIII and BamHI restriction sites and clpB at XhoI and BamHI restriction sites). The positive recombinant plasmids were introduced into the ΔclpA and ΔclpB strains by electroporation. The positive colonies were confirmed by PCR using g and h primers ( Table 2). The complemented strains were designated as ∆clpA + pclpA and ∆clpB + pclpB.

Confirmation of transcription in complemented strains by Reverse Transcriptase (RT) -
Polymerase chain reaction (PCR). Overnight grown cultures of WT, ∆clpA, ∆clpB, ∆clpA + pclpA and ∆clpB + pclpB strains of S. Typhimurium were harvested by centrifugation. RNA from such pellets were isolated using Trizol reagent. RNA samples were treated with RNase-free DNase I and then dissolved in nuclease free water. All RNA samples were tested for contamination of DNA using S. Typhimurium clpA and clpB gene specific primers ( Table 2 j and k).
RT-PCR was performed according to the protocol as described in Superscript VILO cDNA synthesis kit (Invitrogen). In brief, in 20 µl reactions, 0.5 µg of RNA samples were mixed with 4 µl of 5 × VILO reaction mix and 2 µl of 10 × Superscript enzyme mix. cDNA was synthesized by incubation of the above mix at 25 °C for 10 minutes (min) followed by 42 °C for 60 min and termination at 85 °C for 5 min. Part of clpA and clpB genes were PCR amplified using these cDNA samples as templates and j and k primers ( Table 2).
Growth curve study. The growth of WT, ΔclpA and ΔclpB strains was analyzed as described earlier 77

Evaluation of in vitro susceptibilities of ΔclpA and ΔclpB mutant strains to different oxidants.
Overnight cultures of different strains were sub-cultured in fresh LB broth (at the ratio of 1:100) and incubated in a shaking incubator at 37 °C. The mid log phase grown cultures were then exposed to different concentrations of paraquat, H 2 O 2 and HOCl (sodium hypochlorite, NaOCl, Sigma) for 2 h. Aliquots were withdrawn, serially diluted and plated on HE agar plates. CFUs/ml were calculated following incubation of plates at 37 °C for overnight.
Susceptibility of WT, ΔclpA and ΔclpB strains to macrophages. The susceptibilities of WT, ΔclpA, ΔclpB and complemented (∆clpA + pclpA and ∆clpB + pclpB) strains of S. Typhimurium to monocyte derived macrophages (MDM) were determined as described earlier 78 with minor modifications. Briefly, heparinised poultry blood was layered over equal amount of Histopaque-1077 (Sigma) and mononuclear cells (MNCs) were recovered by centrifugation at 1300 × g for 30 min at 25 °C. MNCs were washed two times with RPMI-1640 media (HiMedia) supplemented with 2% chicken serum, 8% fetal bovine serum and 1 × antibiotic/antimycotic solution (Gibco). The cells were counted by trypan blue dye exclusion method. MNCs were adjusted to 2 × 10 6 cells/ ml in similar media. Then, the cells were seeded in 24 well cell culture plates at the number of 1 × 10 6 cells/ well and incubated for 6 h at 37 °C/5% CO 2 . Non-adhered cells were removed by washing. Cells were stimulated with Salmonella enterica serovar Typhimurium LPS (Sigma) at 0.5 µg/ml for 48 h. Following incubation, the cells were washed with antibiotic free media and infected with WT, ΔclpA, ΔclpB or complemented strains of S. Typhimurium at multiplicity of infection (MOI) of 1:50 (macrophages:bacteria). The infection was simulated by centrifugation at 120 × g for 10 min at 25 °C. The cell-bacterial mix was incubated for 2 h. To kill non invaded bacteria the mix was incubated in gentamicin (50 μg/ml) containing media for 90 min. The cells were then incubated in gentamicin (10 μg/ml) containing media. Following 24 or 48 h of incubation, the cells were lysed with 0.1% Triton X-100 and lysates were 10 folds serially diluted and plated on HE agar plates. Plates were incubated at 37 °C and colony forming units (CFUs)/ml were determined.
Analysis of protein aggregations. Overnight grown cultures of WT, ΔclpA and ΔclpB strains were diluted in fresh LB broth and incubated at 37 °C and 180 rpm for 3 h. The cultures were then exposed to 0, 1.5 and 3 mM concentrations (final) of HOCl for 30 min. Bacteria were harvested by centrifugation at 7000 rpm for 10 min. Protein aggregates from such exposed cultures were isolated as described earlier 79 . In brief, the bacterial pellets were suspended in 500 µl of 50 mM Tris(hydroxymethyl) aminomethane (Tris) (pH 7.4), 5 mM ethylenediaminetetraacetate (EDTA), 20% sucrose and 1 mg/ml lysozyme. The mixtures were incubated at 25 °C for 10 min and diluted by addition of 5 volumes of 30 mM Tris buffer (pH 7.4). Following 20 seconds of brief sonication, mixtures were supplemented with 10 mM magnesium chloride (MgCl 2 ) and DNaseI. Unbroken cells were removed by centrifugation at 2,000 × g for 2 min. Supernatants were incubated with 0.5% Triton X-100 for 15 min ScIentIfIc RepoRTS | (2018) 8:4481 | DOI:10.1038/s41598-018-22670-6 at 25 °C. The insoluble cell fractions (protein aggregates) were recovered by centrifugation at 8,000 × g for 15 min at 25 °C. The protein aggregates were analyzed by sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) under reducing conditions. Loading was normalized in terms of CFUs/ml. Relative amounts of protein aggregates in different treatments/groups were analysed by ImageJ software (NIH) bundled with 32 bit Java 1.8.0. Sum of peak area in different lanes were calculated using the software. Total sum of peak area in 0 mM HOCl treated WT sample (lane) is considered as one.

Analysis of virulence in poultry.
One day old chicks were procured from ICAR-Central Avian Research Institute (CARI), Izatnagar, India and provided with adlibitum feed and water. The birds were screened for the presence of Salmonella spp. as described earlier 76 . The Salmonella free birds were divided into three groups. At the age of six days (~1 week) they were orally infected with WT or ΔclpA or ΔclpB strains at a dose of 1 × 10 9 CFUs/ bird. Following 7, 14, and 21 days post-infection, 4 birds were sacrificed from each group. Caecal colonization and bacterial burdens in liver and spleen were assessed as described elsewhere 76 . Statistical analysis. Data were analyzed by SPSS. Comparisons between multiple groups were done by using one way analysis of variance (ANOVA) followed by Post -hoc Tukey alpha test. p < 0.05 was considered significant among different test groups.
Data availability. The data analysed and generated during the current study are available from the corresponding author.