(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .



Clostridium perfringens chitinases, key enzymes during early stages of necrotic enteritis in broiler chickens [1]

['Evelien Dierick', 'Livestock Gut Health Team', 'Light', 'Ghent', 'Department Of Pathobiology', 'Pharmacology', 'Zoological Medicine', 'Faculty Of Veterinary Medicine', 'Ghent University', 'Merelbeke']

Date: 2024-10

The interaction between bacteria and the intestinal mucus is crucial during the early pathogenesis of many enteric diseases in mammals. A critical step in this process employed by both commensal and pathogenic bacteria focuses on the breakdown of the protective layer presented by the intestinal mucus by mucolytic enzymes. C. perfringens type G, the causative agent of necrotic enteritis in broilers, produces two glycosyl hydrolase family 18 chitinases, ChiA and ChiB, which display distinct substrate preferences. Whereas ChiB preferentially processes linear substrates such as chitin, ChiA prefers larger and more branched substrates, such as carbohydrates presented by the chicken intestinal mucus. Here, we show via crystal structures of ChiA and ChiB in the apo and ligand-bound forms that the two enzymes display structural features that explain their substrate preferences providing a structural blueprint for further interrogation of their function and inhibition. This research focusses on the roles of ChiA and ChiB in bacterial proliferation and mucosal attachment, two processes leading to colonization and invasion of the gut. ChiA and ChiB, either supplemented or produced by the bacteria, led to a significant increase in C. perfringens growth. In addition to nutrient acquisition, the importance of chitinases in bacterial attachment to the mucus layer was shown using an in vitro binding assay of C. perfringens to chicken intestinal mucus. Both an in vivo colonization trial and a necrotic enteritis trial were conducted, demonstrating that a ChiA chitinase mutant strain was less capable to colonize the intestine and was hampered in its disease-causing ability as compared to the wild-type strain. Our findings reveal that the pathogen-specific chitinases produced by C. perfringens type G strains play a fundamental role during colonization, suggesting their potential as vaccine targets.

The intestinal mucus layer protects the intestinal mucosa from invasion by pathogenic bacteria. However, it is precisely this first line of protection that invading bacteria aim to break down en route to colonizing the gut. Clostridium perfringens type G, the causative agent of necrotic enteritis in broiler chickens, an enteric disease characterised by ulcers in the small intestine, employs such a pathogenic strategy. In addition to the well-studied NetB toxin, this bacterium can produce other pathogen-specific enzymes, some of which could play a central role during disease development. Two of these enzymes were identified as putative chitinases, ChiA and ChiB. We have demonstrated that ChiA and ChiB are indeed functional enzymes that can hydrolyze a wide variety of substrates, with ChiA preferring mucus as a substrate. Mucus breakdown by the chitinases induces both bacterial proliferation and attachment to mucus, two processes that are crucial in the first steps of the pathogenesis. In addition, we have shown that a ChiA mutant strain does not efficiently colonize the intestine and causes less severe lesions as compared to the wild-type strain. Thus, we conclude that chitinases provide an advantage to the pathogen to facilitate colonization of the intestine of broiler chickens.

Funding: The researcher ED was supported by Research Foundation Flanders FWO (Fonds Wetenschappelijk Onderzoek Vlaanderen FWO https://www.fwo.be/ ) under grant number [12X8622N]. YB has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement No 945405 ( https://rea.ec.europa.eu/funding-and-grants/horizon-europe-marie-sklodowska-curie-actions_en ) and was supported by a post-doctoral grant by FWO [12S0519N]. SNS acknowledges research support from the Flanders Institute for Biotechnology (VIB) under grant number C0101 ( https://vib.be/nl#/ ). This study received funding from Evonik Operations GmbH, Nutrition & Care ( https://www.evonik.com/en/company/divisions/nutrition-care.html ). SP and SH are employees from Evonik Nutrition & Care. All other funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Data Availability: Crystallographic coordinates and their structure factors were deposited to the PDB with following accession codes: 8otb (ChiA Apo), 8oye (ChiA Chitin), 8ose (ChiA Bisdionin C), 8owf (ChiA Chitosan), 8ovr (ChiB), 8c6z (ChiB Bisdionin C). Raw sequencing data is available at NCBI SRA under BioProject ID PRJNA1071876. All relevant data are within the paper and the S2 Data file.

Clostridium perfringens type G is the causative agent of necrotic enteritis (NE) in broilers, an enteric disease with great economic consequences [ 22 ]. C. perfringens is part of the normal gut microbiota of vertebrates, reaching numbers up to 10 5 cfu/g in small intestinal content of healthy chickens [ 23 ]. Dysregulation of the intestinal microbiota can trigger proliferation of a virulent C. perfringens strain, eventually followed by massive toxin production. In contrast to non-pathogenic C. perfringens strains, the type G strains harbour three specific pathogenicity loci associated to their virulence potential: NELoc-1 (42kb; plasmid encoded), NELoc-2 (11.2kb; chromosomal) and NELoc-3 (5.6kb; plasmid- encoded) [ 24 ]. Using a netB-knockout mutant, the NetB toxin, located on NELoc-1, has been identified as a critical virulence factor in NE pathogenesis [ 25 ]. Deletion of the plasmid harbouring the entire NELoc-1 locus including the netB gene, results in complete loss of virulence, which is, however, only partially restored by complementation of this mutant with the netB gene, suggesting the importance of other potential virulence genes located on the plasmid [ 26 ]. In addition to the netB gene, NELoc-1 harbours an additional 36 genes from which two have been identified as putative chitinases, ChiA and ChiB. Interestingly, a study focussing on the C. perfringens gene expression during initial colonization of the intestinal tract, showed that both chitinase genes were upregulated during the early stages of pathogenesis, while expression of none of the 34 other genes located on the plasmid (incl. NetB) was increased, suggesting a possible involvement of these putative chitinases during colonization [ 27 ]. The aim of this research was to elucidate whether these chitinase genes encode functionally active proteins and to study their potential involvement during the early stages of NE pathogenesis. Structural studies were performed using X-ray crystallography, resulting in the description of the binding pockets and potential carbohydrate ligand interactions. In vitro binding studies as well as enzymatic activity studies were performed towards multiple substrates (pseudo-chitin substrates, chitin, mucus), thereby questioning the allocated nomenclature as true chitinases. To determine the biological relevance of chitinases, a combination of in vitro assays, either using C. perfringens mutant strains or recombinantly produced enzymes, were conducted studying the role of chitinases during either nutrient acquisition or bacterial binding to intestinal mucus. Finally, the reduced ability of a ChiA mutant strain to colonize the small intestine and induce NE was assessed using an in vivo colonization assay and NE challenge trial, respectively.

Both commensal and pathogenic bacteria are able to produce a large array of glycosyl hydrolases, enzymes that are able to cleave glycosidic bonds of glycans. Chitinases are glycosyl hydrolases that hydrolyze the β-(1–4)-linkage of N-acetyl-D-glucosamine (GlcNAc) units which are present amongst others in chitin, a linear polymer of GlcNAc. Chitinases are produced by many different bacteria and other life forms like plants, mammals, insects and fungi [ 10 ]. Bacterial chitinases are predominantly members of the glycosyl hydrolase families 18 or 19 [ 11 ]. Chitin is the second most abundant biopolymer in nature and forms the main component of fungal cell walls, crustacean shells and arthropod exoskeletons [ 12 ]. Vertebrates, with the exception of certain fish and amphibian species, do not have the ability to synthesize chitin [ 13 , 14 ]. Despite the lack of chitin in the intestinal tract, various enzymes and proteins which are annotated as chitinases and chitin-binding proteins have been linked to pathogenesis of enteric bacterial diseases [ 11 ]. Potential targets, other than chitin, are the β-1,4-linkage in GlcNAc-containing glycolipids and glycoproteins, which are omnipresent in the gastrointestinal tract [ 11 , 15 ]. Involvement of bacterial chitinases and chitin-binding proteins has been shown in mucus breakdown, bacterial translocation, suppression of host innate immune system and bacterial colonization through attachment to the intestinal epithelium, highlighting the crucial role of these enzymes as key virulence factors in a range of bacterial intestinal diseases [ 11 , 16 – 21 ].

The intestinal microbiota can utilize these glycans in different ways. Mucin glycans can be broken down to metabolizable oligomers that can act as carbon or energy sources [ 6 ]. Also, these carbohydrate structures can provide initial attachment sites for bacteria [ 7 ]. Pathogens have evolved to exploit this niche using the breakdown of or attachment to glycans to their advantage, aiding the initial steps of colonization and proliferation [ 8 , 9 ]. Furthermore, detection of glycan residues or monosaccharides in the environment can act as a chemical cue to help the pathogens sense their surroundings. As a response, pathogens can initiate the expression of virulence factors which may further impair the protective mucus layer, leading to colonization and eventually infection [ 6 , 9 ].

Glycans are ubiquitous throughout nature and play a fundamental role in many biological processes. These carbohydrate residues are diverse in structure and composition and can be part of glycoproteins, glycolipids or other glycoconjugates [ 1 ]. Throughout the gastrointestinal tract of vertebrates, glycans are omnipresent. First, the gut epithelium is covered with a mucus layer composed of highly glycosylated proteins, called mucins. This mucus layer acts as a first line of defence against food particles, gut microbiota, chemicals, enzymes and host or microbial secreted products, and contributes to the symbiosis between the host and the microbiota [ 2 ]. Furthermore, the host produces glycoproteins and glycolipids that are attached to the apical side of epithelial cells, called the glycocalyx, inhibiting bacterial adherence through steric hinderance [ 3 ]. Lastly, almost all crucial molecules involved in the innate and adaptive immune system are highly glycosylated [ 4 ]. Both direct and indirect functions of glycans in the immune system have been described, ranging from their action as ligands, immunogens, and antigens to other more complex processes like cell-cell recognition and antibody glycosylation [ 5 ].

Results

Subcellular localization of C. perfringens chitinases ChiA and ChiB The environment of a protein provides, at least in part, the relevant context necessary for its function. Therefore, the subcellular localization of a protein provides information about its biological function. Both putative C. perfringens chitinases, ChiA and ChiB, were predicted to be extracellular proteins by the CELLO subcellular localization predictor (S1 Table). Further analysis indicated that secretion of the chitinases ChiA and ChiB might occur through different pathways. SecretomeP predictions of the putative chitinases showed a SecP score > 0.8 for both putative chitinases, indicating possible secretion via a Sec-independent pathway (non-classical secretion). As SecretomeP can give high SecP scores to proteins containing a signal peptide, this information was combined with SignalP predictions to identify possible conventional signal peptides. For ChiA, a “standard" secretory signal peptide transported by the Sec translocon and cleaved by Signal Peptidase I (Sec/SPI) was identified with 91.1% probability (S1 Table). The cleavage site of the signal peptide was predicted to be located between AA34 and 35 (TKA-KE; 68.33% probability of the cleavage site prediction). No signal peptide was identified for ChiB, indicating that this enzyme is predicted to be secreted by a non-classical secretion mechanism. The chitinase genes are located on the NELoc-1 pathogenicity locus, in close proximity to each other and having the same orientation. Only 85 nucleotides are located in between chiA and chiB, containing a potential ribosome binding site (Shine-Dalgarno sequence; S1 Fig). The expression of the chitinases genes on RNA level was assessed using qPCR. The presence of RNA of both chitinases was demonstrated when C. perfringens CP56 was grown in nutrient rich medium and mucus-supplemented medium, with the expression being the highest in the latter (S2 Fig).

In silico physiochemical properties analysis The alignment of the amino acid sequence of both chitinases using protein-protein BLAST identified 29.06% sequence identity (with 69% coverage) between both enzymes. This clearly indicates that the presence of both chitinases in the C. perfringens genome is not the result of a gene duplication event. The physicochemical properties of the putative chitinases ChiA and ChiB were predicted by the ExPASy ProtParam tool (S2 Table). The molecular weight of both chitinases was around 65–66 kDa. Both chitinases have a similar theoretical isoelectric point (pI) around 5, predicting them to be acidic in nature. Protein stability was assessed by focussing on the instability index, aliphatic index and the grand average of hydrophobicity (GRAVY) index. Both chitinases are classified as stable proteins (instability index below 40), with good thermostability (as indicated by a high aliphatic index, which is a measure for the thermostability of globular proteins). As a similar aliphatic index was obtained for both putative chitinases, similar temperature optima are expected. The obtained GRAVY index around -0.5 indicated that both chitinases are hydrophilic enzymes.

ChiA and ChiB are members of the glycosyl hydrolase family 18 To gain a better understanding of the function of ChiA and ChiB we sought to obtain their structures. We succeeded in obtaining several crystal structures of recombinant ChiA and ChiB in apo (unbound) and ligand-bound forms to resolutions between 1.30 Å and 1.85 Å (S3 Table). Both ChiA and ChiB encode for an N-terminal glycosyl hydrolase family 18 (GH18) Trios-phosphate Isomerase (TIM) barrel followed by a C-terminal putative carbohydrate binding domain (CBD) (Fig 1A and 1C). This C-terminal CBD organization has previously been observed in a chitinase of Chromobacterium violaceum (PDB 4txg) (Figs 1B and S3). When comparing the overall structures of the chitinases, the unexpected location of the ChiB CBD becomes apparent. As compared to ChiA and the Chromobacterium violaceum chitinase, the CBD of ChiB is located on the opposite side of the TIM barrel. This CBD-location of ChiB is similar to the one occupied by the Serratia marcescens ChiA CBD. However, in Serratia marcescens this CBD is located at the N-terminus of the protein (Figs 1D and S3). For both chitinases, the CBD leads to a cleft in the GH18 TIM barrel domain lined with Trp residues which harbours at its base the canonical chitinase active site with a DxDxE motif. The cleft of ChiB is covered by a loop of the TIM barrel leaving only the sides open to the solvent (Fig 1E and 1F). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 1. Crystal structures of chitinases ChiA and ChiB. (A, C) Schematic overview of protein domain organization and overall structures of ChiA and ChiB, both consisting of an N-terminal glycosyl hydrolase 18 (GH18) Trios-phosphate Isomerase (TIM) barrel followed by a putative carbohydrate binding domain (CBD). (B, D) Comparison of overall protein structures of ChiA and ChiB with those of previously studied chitinases (Chromobacterium violaceum and Serratia marcescens). (E,F) Surface representation of the protein crystal structures of ChiA and ChiB identifying the glycosyl hydrolase 18 domain (purple), carbohydrate binding domain (orange) and their open and tunnel shaped binding cleft, respectively, containing the DxDxE binding motif. The transparent purple volume in the cartoon representation of the crystal structures describes the ligand-binding pocket. (G,H, I, J, K, L) Ligand interactions of ChiA or ChiB with substrates: chitin, chitosan or bisdionin C. Ligand OMIT maps are shown as blue mesh representing the difference electron density Fourier coefficient mFo-DFc contoured at + 3 σ carved 2A around the ligands. Maps were generated by performing reciprocal space coordinate refinement after randomizing all atoms on average 0.4 Å in the absence of the ligands. https://doi.org/10.1371/journal.ppat.1012560.g001

Access and ligand interactions within the binding pocket The accessibility of the CBD could impact ligand interaction and therefor the functionality of ChiA and ChiB. In the Apo form of ChiA, the binding pocket is occupied by somewhat disordered molecules originating from the crystallization condition (Fig 1G). The active site motif is trapped in the active conformation with D194 and E196 sidechains engaging in a low-barrier hydrogen bond. The E196Q point mutant of ChiA was generated to capture the ligand within a co-crystal structure. E196 is expected to be protonated in the catalytically competent state and act as a general acid/base to protonate the glycosidic bond during the substrate hydrolysis [28,29]. The chitin oligosaccharide (GlcNAc 6 ) occupies two conformations (Fig 1H). In the first, bound conformation, the oligosaccharide has descended towards the active site occupying positions -2 to +3. The GlcNAc at position -1 is in the boat conformation, priming it for the formation of the oxazolinium reaction intermediate. In the second, encounter conformation, the elongated oligosaccharide is located at the surface of the cleft and only engages the CBM side of the cleft. This second conformation resembles the chitosan oligosaccharide (GlcN) bound conformation (Fig 1I). The missing acetyl group is required for the formation of the oxazolinium reaction intermediate and can consequently not be hydrolyzed. The -2 and -1 positions, where the product disaccharide is bound, are also the location where the inhibitor Bisdionin C binds (Fig 1J). The inhibitor molecule forces D194 into its resting conformation where it engages D192 in a low-barrier hydrogen bond as the other conformation would cause steric clashes. A second low occupancy or highly disordered Bisdionin C molecule might be bound in the vicinity of W257, more towards the CBD. Overall ChiA displays the hallmarks of a chitinase processing chitin from its non-reducing end towards the reducing end. As in ChiA, the sizeable substrate binding pocket of Apo ChiB contained a PIPES buffer molecule originating from the crystallization condition (Fig 1K). The makeup of the binding pocket is conserved, except for the longer loop closing of the cleft and the entry of the cleft near the ChiB CBD. Neither differences should prohibit the binding of a linear chitin oligosaccharide. The active site motif is trapped in the resting conformation with a low-barrier hydrogen bond between D156 and D158. Attempts at obtaining substrate-bound structures failed but the substrate binding mode can be somewhat deduced from the structure of ChiB in complex with inhibitor Bisdionin C (Fig 1L). Two Bisdionin C molecules are observed in the ChiB binding site. The first molecule binds at nearly the same position as in ChiA but shifted approximately 2 Å towards the surface of the cleft. This position enables the Bisdionin C molecule to participate in the coordination of a divalent cation. The cation is part of the crystallization solution and is not present in the Apo structure. The second Bisdionin C molecule occupies the other half of the binding pocket. Each inhibitor molecule engages a side of the Trp120 indole ring with extensive van der Waals contacts which switched changes rotamer from t-105 to m0 causing it to project itself into the binding cleft. Both inhibitor molecules together occupy the ligand binding site from position -2 to +3. Overall, ChiB displays the hallmarks of a chitinase processing chitin from its reducing end towards the non-reducing end.

Glycosyl hydrolase family 18 enzymes are associated with pathogenic C. perfringens type G strains Possible functional redundancy of the chitinases in the C. perfringens genome was assessed by searching both the CAZy database (www.cazy.org) and the NCBI database for the presence of other C. perfringens enzymes belonging to glycosyl hydrolase families 18 (GH18) or 19 (GH19), which both contain chitinases. No GH19 family members were found in C. perfringens, whereas 22.2% (4/18) of the genomes in the CAZy database and 7.2% (56/778) of the C. perfringens genomes in the NCBI dataset contained GH18 containing proteins. The large majority of these GH18 domain containing proteins were found on the netB-plasmid of C. perfringens type G strains (100% (4/4, CAZy database) or 85.7% (48/56, NCBI dataset) of the C. perfringens strains with GH18 also contained NetB). In 8 netB-negative C. perfringens strains from the NCBI dataset GH18 family proteins were identified. These strains were isolated from pigs (n = 2), goats (n = 1), manure treated soil from a research farm (n = 1) or chickens (n = 4) in Australia, Belgium, China or Canada. In three of these chicken isolates, both the ChiA and the ChiB protein were identified, whereas the GH18 containing proteins from the other isolates showed only limited identity with ChiA or ChiB (S4 Table). To confirm the finding that the ChiA and ChiB isolates are solely found in chicken isolates and are linked to the netB-plasmid of type G strains, a diverse collection of C. perfringens strains, obtained from different hosts and geographical locations (Belgium, Denmark,…), was screened using PCR to assess the prevalence of chiA and chiB. All pathogenic netB-positive type G strains isolated from broilers (30/30) tested positive for the presence of both genes. In contrast, the genes chiA and chiB were not present in the commensal netB-negative type A strains isolated from broilers (0/48). In addition, the strains isolated from non-broiler hosts all tested negative: layer (0/7), cattle (enterotoxaemia 0/6; healthy calves 0/8, ruminating cattle 0/3), sheep (0/7), horse (0/5), dog (0/2), pig (0/1), human (gangrene 0/1), deer (0/1) and goat (0/3) (S5 Table).

ChiA and ChiB show preferential substrate binding Substrate binding is a prerequisite for an enzyme to exert its enzymatic activity. Characterisation of the binding properties of enzymes aids the identification of their potential substrates. To examine whether the C. perfringens chitinases interact with various polysaccharides, a binding assay between the recombinant chitinases and either crystalline chitin, colloidal chitin or GlcNAc-coated beads was performed. The amount of bound and unbound protein to the substrate was visualised after which the fraction of bound protein was calculated (Fig 2A). For this assay, ChiB was able to bind to crystalline chitin (73%), colloidal chitin (82%) and GlcNAc-coated beads (74%), whereas ChiA showed only limited binding to the substrates (9%, 5% or 1% to respectively crystalline chitin, colloidal chitin or GlcNAc). As C. perfringens type G strains are intestinal pathogens, and chitin is not present in the intestinal mucosa, chitin is probably not the primary substrate of the putative chitinases in vivo. As both putative chitinases were classified as O-glycosyl hydrolases, they might exert their action on mucin-type O-glycans in the intestinal mucus. The ability of the recombinant enzymes to interact with crude chicken intestinal mucus or porcine mucus was assessed using a dot blot assay (Fig 2B). None of the enzymes were able to bind to porcine type II or III mucus. No binding of ChiB could be detected to the substrates. This contrasts with ChiA, where no binding was observed towards either of the porcine mucins, but a clear interaction of ChiA with the crude chicken mucus was observed (5 out of 6 biological replicates of chicken mucus). These findings indicate that both chitinases have different substrate preferences: ChiB has a higher affinity towards chitin whereas ChiA binds to small intestinal chicken mucus. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. Binding of the inactive chitinases ChiA or ChiB towards (A) crystalline chitin, colloidal chitin or GlcNAc-covered beads or (B) crude chicken intestinal mucus. (A) In solution binding assay: The amount of bound and unbound protein to the substrate was visualised using SDS-PAGE and subsequent Coomassie staining. B = bound fraction, S = supernatants representing the unbound protein fraction. Intensity of the bands was quantified as OD/mm2. The bound protein fraction was calculated by dividing the intensity of bound protein by total protein for each substrate. Assay was performed twice, one representative figure is shown. (B) Dotblot: Either porcine mucus type II (1), porcine mucus type III (2) or crude chicken mucus (3–8; 6 biological replicates) were spotted onto the membrane. Membranes were incubated with either recombinant ChiA or ChiB. The amount of bound enzyme was visualized using anti-HIS-antibody. Assay was performed twice, one representative figure is shown. https://doi.org/10.1371/journal.ppat.1012560.g002

Activity of chitinases towards colloidal chitin and 4-MU-(GlcNAc) 1-3 substrates The putative chitinases belong to the glycosyl hydrolases family 18, a family that harbours both active chitinases that hydrolyse glycosidic bonds as well as chitinase-like proteins which only bind to but do not cleave the substrate. To assess whether the chiA and chiB genes encode enzymatically active proteins, both putative chitinases were recombinantly expressed and their enzymatic activity towards colloidal chitin was characterized (Table 1). Both chitinases demonstrated activity towards colloidal chitin. However, a higher hydrolysis rate and higher catalytic efficiency (k cat /K M ) was observed for chitinase ChiB as compared to ChiA. To elucidate whether the observed enzymatic activity towards chitin was caused by exo- or endo-glycosidase activity (cleaving either terminal or internal glycosidic linkages in the polymer, respectively) of the chitinases, further assays using the fluorescently labelled chitin pseudo-substrates 4-MU-(GlcNAc) 1-3 were performed. Both chitinases were inactive towards 4-MU-GlcNAc, whereas both 4-MU-GlcNAc 2 and 4-MU-GlcNAc 3 were suitable substrates (Table 1). This indicates that both chitinases function as endo-chitinases. Overall, a higher hydrolysis rate and higher catalytic efficiency (k cat /K M ) was observed for chitinase ChiB as compared to ChiA. Due to unambiguous model fitting, no kinetic parameters could be determined for ChiB using 4-MU-(GlcNAc) 3 as a substrate. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Table 1. Kinetic parameters of the fitted models to the enzymatic activity data of ChiA or ChiB towards the pseudo-chitin substrates 4-MU-GlcNAc 2 , 4-MU-GlcNAc 3 or colloidal chitin. K cat = catalytic constant, K m = substrate concentration given that the reaction rate reaches ½ V max , K cat /K m = Specificity constant, ND: kinetic parameters not determined due to unambiguous model fitting for ChiB using 4-MU-(GlcNAc) 3 . BDL: “Below detection limit”. https://doi.org/10.1371/journal.ppat.1012560.t001

Effect of temperature and pH on enzymatic activity towards pseudo-chitin substrates Since it is hypothesized that the putative chitinases play a role during C. perfringens pathogenesis, their ability to be enzymatically active at temperatures and acidity levels representative of the gastrointestinal tract was assessed. The effect of temperature and pH on the enzymatic activity of both enzymes towards 4-MU-GlcNAc 2 and 4-MU-GlcNAc 3 was analysed (Fig 3A). Temperature optima (at a constant pH5) were found at 37°C, except for ChiA cleaving 4-MU-GlcNAc 2 where maximal hydrolysis rate was reached at 30°C. Lower temperatures reduced the activity of the recombinant chitinases, although most often not completely when reaching 4°C. In addition, temperatures exceeding the optimal temperature, reduced the enzymatic activity. However, at 42°C (chicken body temperature), still high activity was measured. Chitinase activity (for both ChiA and ChiB) was highest at a pH between 5 and 6 (at a constant 42°C). An excessively acidic or alkaline environment inhibited ChiA enzymatic activity. The enzymatic activity of ChiB was only hampered when the pH was too low. Both enzymes were highly active at 42°C and pH 5–6, the biologically relevant conditions inside the gastrointestinal tract of broiler chickens. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. Enzymatic activity of chitinases ChiA and ChiB towards (A) pseudo-chitin substrates (effect temperature and pH) or (B) crude chicken mucus. (A) Effect of temperature and pH on chitinase activity towards pseudo-chitin substrates (4-MU-GlcNAc 2 and 4-MU-GlcNAc 3 ). The enzymatic activity of ChiA (left) or ChiB (right) at different temperatures was assessed at a constant pH of 5 (upper panels), whereas the effect of the pH was monitored at a constant temperature of 42°C (lower panels) (no replicates). (B) Chitinase activity towards chicken mucus using turbidity assay. Biological replicates of crude intestinal chicken mucus (150 μg) were incubated with either recombinant ChiA, ChiB (15 μg) or PBS. After a one hour incubation period at 37°C, the turbidity of the mixture was measured at an OD-value of 450 nm. The relative index was calculated by dividing the OD of the chitinase-treated mucus by the OD of the untreated mucus sample for each mucus sample. Lines indicate the means with their respective standard deviations. Significant differences are indicated with ‘*’ (p≤0.05). ‘**’ (p≤0.01) and ‘***’ (p≤0.001). https://doi.org/10.1371/journal.ppat.1012560.g003

Intestinal mucus, a chitinase substrate of biological significance Since chitinase ChiA was able to bind to chicken intestinal mucus, the enzymatic activity of the chitinases towards crude chicken mucus was assessed using a turbidity assay (Fig 3B). With the addition of either chitinase, the turbidity of the mucus-containing mixture increased, indicating mucosal breakdown. The factor to which extent the OD-value increased was calculated relative to the blank conditions. The turbidity index of mucus treated with ChiA or ChiB was significantly different as compared to untreated mucus (p≤0.0001 for ChiA; p = 0.0429 for ChiB). The activity of ChiA on crude chicken mucus was higher as compared to ChiB (p = 0.0429), indicating a higher preference of the ChiA towards chicken mucus. The hypothesis that chitinases aid the breakdown of intestinal mucus to facilitate bacterial growth was validated through the treatment of chicken intestinal mucus with either ChiA, ChiB or PBS (as a negative control). The growth rate of wild type CP56 was assessed in media supplemented with either chitinase- or PBS-treated mucus. The growth rate of the bacteria grown in media supplemented with 5% ChiA-treated mucus increased significantly as compared to the untreated mucus (p = 0.0429) (Fig 4A), indicating that chitinases indeed aid nutrient acquisition of the pathogen to some extent. This effect was lower and not significant when supplementing the medium with ChiB-treated mucus (p = 0.6620). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 4. Effect of mucus on the growth of C. perfringens. (A) Growth analysis of C. perfringens CP56 in media supplemented with 5% chicken intestinal mucus pre-treated with either PBS (= untreated), 15μg of chitinase ChiA or ChiB for one hour. Growth rate is defined as the slope during exponential rate. Bars indicate the means with their respective standard deviations. (B) Growth analysis of wild-type and chitinase mutant strains in different media. Wild type CP56 (Black), CP56ΔchiA1 (light blue), CP56ΔchiA2 (dark blue), CP56ΔchiB1 (light orange) or CP56ΔchiB2 (dark orange) were grown in either nutrient rich medium, nutrient poor medium or nutrient poor medium supplemented with 5% crude chicken mucus. Growth rate is defined as the slope during the exponential phase. Bars indicate the means with their respective standard deviations. Significant differences are indicated with ‘*’ (p≤0.05). ‘**’ (p≤0.01). ‘***’ (p≤0.001) and ‘****’ (p≤0.0001). https://doi.org/10.1371/journal.ppat.1012560.g004

Generation of C. perfringens ChiA and ChiB mutants To further address the role of C. perfringens chitinases ChiA and ChiB, mutant strains from each of the chitinase genes (chiA or chiB) were constructed from the pathogenic C. perfringens type G strain CP56, using the ClosTron mutagenesis system. PCR using a forward primer targeting the ClosTron insert and a reverse primer downstream of the insertion site showed correct insertion of the ClosTron insert in either the chiA or chiB gene. No additional ClosTron insertions were detected using dPCR [30]. Despite multiple attempts, chitinase mutant strains could not be complemented. Large plasmid uptake severely hampered the phenotype of the strains, thereby significantly reducing growth properties. Alternatively, whole genome sequencing was performed to assess potential secondary mutations of the mutant strains (two for each chitinase: CP56ΔchiA1, CP56ΔchiA2, CP56ΔchiB1 and CP56ΔchiB2) that could affect the virulence phenotype and consequently hamper the outcome of subsequent experiments. As compared to the CP56 wild-type genome, no INDELs were found in all mutant strains. A total of seven, five, two and eight SNPs were identified in the genomes from respectively CP56ΔchiA1, CP56ΔchiA2, CP56ΔchiB1 and CP56ΔchiB2. In addition, these limited number of SNPs were not present in more than one strain. An overview of the identified gene products and their predicted function is given in S6 and S7 Tables. The SNPs with a potential impact on mucin degradation, both its carbohydrate and peptide moiety, will be highlighted here [31]. Mutant CP56ΔchiA1 harbours a missense-variant in a hypothetical protein containing a DUF1667 domain. The function of this domain remains unknown, however in a small amount of cases it has been found in archaeal and bacterial hypothetical proteins, some of which have been annotated as potential metal-binding proteins, often oxidoreductases and dehydrogenases. Two cases have been described in literature in which a DUF1667 region was located in glycerol-3-phosphate dehydrogenases, enzymes with a role in lipid metabolism [32,33]. Mutant strain CP56ΔchiA2 harbours six missense variants and one SNP that resulted in a stop codon. One of these SNPs is located in a hypothetical protein that has 97.78% sequence identity with the Zinc-dependant exopeptidase M28, a bacterial leucyl aminopeptidase that can potentially lead to protein degradation of the protein backbone of mucin. The two SNPs in mutant CP56ΔchiB1 are not located in genes with a potential mucin-degrading function. CP56ΔchiB2 harbours a SNP in a DUF4091 domain-containing protein. Although the function of this domain is uncharacterised, it is often conserved in N-acetylgalactosaminidases. These types of enzymes are known to catalyse the hydrolysis of the O-glycosidic bond between GalNAc at the reducing end of a mucin sugar chain and serine/threonine of the proteins [31,34,35]. Except from the ClosTron insertion in the chiA or chiB gene (for respectively CP56ΔchiA1/2 or CP56ΔchiB1/2), no further variants were detected in the well-known toxin genes cpa or netB, either of the pathogenicity loci (NELoc-1, NELoc-2 or NELoc-3) nor the genes encoding the VirSR regulatory system which is known to control the expression of virulence genes. We sought further support for these findings by assessing the NetB producing capacity of the mutant strains. A haemolytic assay was performed in which no difference in NetB activity of the different culture supernatant was quantified (S4 Fig), indicating no impact of the SNPs on the main toxin associated to virulence. In addition, the growth properties of the mutant strains were evaluated to determine a potential overall effect of the detected SNPs. Wild-type and mutant strains had an equal growth rate at mid-exponential phase (CP56: 0.0439 ± 0.0014 min-1; CP56ΔchiA1: 0.0444 ± 0.0026 min-1; CP56ΔchiA2: 0.0437 ± 0.0023 min-1; CP56ΔchiB1: 0.0441 ± 0.0023 min-1; CP56ΔchiB2: 0.0425 ± 0.0024 min-1) (Figs 4B and S5). Taken together, the genetic analysis of the mutant strains (only limited effect of SNPs that are not located in virulence-related regions), their equal NetB activity and their equal growth compared to the wild-type strain gave us confidence that possible differences in phenotype between the strains are unlikely to be the result from the secondary mutations.

C. perfringens chitinase mutant strains show attenuated growth in chicken mucus To further investigate the biological significance of the C. perfringens chitinases, their importance during growth of C. perfringens in mucus-containing media was assessed using either a wild-type strain or its isogenic mutants which lack either of the chitinase genes. Growth analysis was performed for the wild-type or chitinase mutant strains (CP56ΔchiA1, CP56ΔchiA2, CP56ΔchiB1 or CP56ΔchiB2) in different media (Figs 4B and S5). When grown in nutrient rich or nutrient poor medium, no difference in growth between the wild-type and either of the chitinase mutant strains could be observed (p = 0.3964 and p = 0.2619, respectively). The wild-type strain grew faster in nutrient poor medium supplemented with 5% crude chicken mucus as compared to non-supplemented nutrient poor medium (p = 0.0006), indicating that crude mucus is a good nutrient source. In mucus-supplemented medium all mutant strains grew slower as compared to the wild-type strain (p = 0.0046 for CP56 vs. CP56ΔchiA1; p = 0.0013 for CP56 vs. CP56ΔchiA2; p = 0.0297 for CP56 vs. CP56ΔchiB1; p = 0.0172 for CP56 vs. CP56ΔchiB2), indicating that the chitinases are advantageous in the utilisation of mucus as an additional growth substrate. To further strengthen these findings, bacterial competition assays between the wild-type and one of either of the chitinase mutant strains in mucus-containing media were performed. Therefore, different media were inoculated with an equal amount of wild-type and mutant strain, either CP56ΔchiA1 or CP56ΔchiB1. For both mixtures (CP56 with CP56ΔchiA1 or CP56 with CP56ΔchiB1), the competition index (ratio mutant:wild type) in nutrient rich medium did not differ from 1 (both during the exponential phase and at saturation), indicating that both strains were growing at an equal rate in nutrient rich medium (Exponential phase: ChiA p = 0.6353 and ChiB p = 0.6917; Overnight: ChiA p = 0.4741 and ChiB p = 4724) (Fig 5A). When mucus is the main nutrient source in the medium (i.e. mucus-supplemented nutrient poor medium), no significant drop in competition index could be observed when growing CP56 together with CP56ΔchiA1 (Exponential phase: p = 0.0704; Overnight p = 0.1056) or CP56ΔchiB1 (Exponential phase: p = 0.2627; Overnight p = 0.9723). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 5. Effect of mucus on colonisation C. perfringens. (A) Competition assay of C. perfringens wild-type strain CP56 and mutant strains CP56ΔchiA1 or CP56ΔchiB1 in either in vitro or in vivo growth conditions. In vitro: An equal mix of wild-type and mutant strain was grown in either nutrient rich medium or nutrient poor medium supplemented with 5% chicken intestinal mucus. Samples were taken at the exponential growth phase and at saturation after overnight incubation. In vivo: 18-days old broiler chickens were inoculated with an equal mix of wild-type and CP56ΔchiA1 mutant strain. After 24 hours, samples were taken from the intestinal content (jejunum or ileum). The amount of wild-type or mutant strain in the samples was determined using dPCR. The competition index is defined as the ratio of the mutant on wild-type strain, divided by the respective ratio in the inoculum. Lines indicate the means. (B) Mucus binding assay of wild-type CP56 C. perfringens strain in media supplemented with recombinant chitinases, either ChiA or ChiB. Washed C. perfringens overnight culture was added to the wells containing a mucus agar layer, supplemented with either 50 μg of recombinant enzyme (ChiA or ChiB) or an equal volume of PBS as a negative control. Wells were anaerobically incubated for 90 min at 37°C after which the bound bacteria were quantified through titration. The binding ratio is defined as the percentage of bacteria bound to the mucus in supplemented media as compared to non-supplemented conditions. Bars indicate the means with their respective standard deviations. https://doi.org/10.1371/journal.ppat.1012560.g005

Role of chitinases during C. perfringens colonization The digestion of mucus by chitinases may alter its composition, which could facilitate bacterial binding and may be a critical step during the initial colonization of the small intestine. To assess whether the chitinases could affect C. perfringens adherence to intestinal mucus, a mucus binding assay was performed. Addition of ChiA to CP56 increased the bacterial binding capacity to intestinal mucus as compared to non-supplemented conditions (Fig 5B, p = 0.0398). The effect of ChiB on the binding of CP56 to chicken mucus could not be statistically proven (p = 0.0763). Since ChiA has more affinity towards mucus, the following in vivo studies were preformed using only the CP56ΔchiA1 mutant strain. To study the importance of this chitinase during colonization, an in vivo colonization assay was performed in which 18-day old broilers were orally inoculated with a 50/50 mixture of wild-type and CP56ΔchiA1 mutant strain. As a control, the inoculum was grown in an in vitro setting in different media, verifying that both strains were growing at an equal rate (Fig 5A). After oral administration to broilers, the mean competition index of the retrieved intestinal content samples significantly dropped both in the jejunum and ileum (Fig 5A; p = 0.0020 for both), indicating that more wild-type strain was present in the intestinal content as compared to mutant.

[END]
---
[1] Url: https://journals.plos.org/plospathogens/article?id=10.1371/journal.ppat.1012560

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.