Novel M23 peptidases Pgp4, Pgp5, and Pgp6 contribute to helical cell shape in Campylobacter jejuni

Abstract

The helical morphology of Campylobacter jejuni is maintained by its peptidoglycan (PG) layer and influences its success as a pathogen. Periplasmic PG hydrolases that cleave the PG glycan backbone and peptide sidechains (such as carboxypeptidases and endopeptidases) are critical for proper cell function and/or growth and are important in the PG remodeling required for cell shape generation and any morphological alterations. The C. jejuni shape is determined by PG hydrolases Pgp1 (DL-carboxypeptidase), Pgp2 (LD-carboxypeptidase) and Pgp3 (DD-carboxypeptidase/DD-endopeptidase), as well as a group of M23 peptidase domain containing proteins with previously uncharacterized activity: CJJ81176_1105, CJJ81176_1228, and CJJ81176_0166. Using a PG cleavage assay, we showed that 1105 and 1228 have DD-carboxypeptidase/DD-endopeptidase activity, and 0166 is a DD-carboxypeptidase. We renamed 1105, 1228, and 0166 to Pgp4 (peptidoglycan peptidase 4), Pgp5, and Pgp6, respectively. Pgp6 is the first described C. jejuni M23 peptidase with substrate selectivity on monomeric pentapeptides. Sequence comparisons between the DD-carboxypeptidase Pgp6 and the DD-carboxypeptidase/DD-endopeptidase Pgp3 (with an available crystal structure) and their corresponding orthologs revealed that Pgp6 contains insertion sequences in the M23 peptidase domain not present in Pgp3. Modeling of Pgp6 predicted that the insertion sequences would restrict the active site groove, only allowing entrance of a smaller substrate. This provides a possible explanation for the lack of Pgp6 DD-endopeptidase activity. To our knowledge, Pgp6 is the first reported DD-carboxypeptidase in the M23 peptidase superfamily. Deletions in pgp4, pgp5, and pgp6 resulted in mutants with varying curved rod morphologies and changes in PG muropeptide profiles in comparison to wild type and each other. Using these mutants, we examined the effect of deleting these genes on C. jejuni properties affecting pathogenesis and survival: motility, biofilm formation, autoagglutination, the ability to transition to a coccoid form, growth under varying pH, susceptibility to antimicrobial compounds, and adherence, invasion and intracellular survival in human epithelial cells. Each mutant showed distinct phenotypic changes to each other, indicating they are not functionally redundant. This also further supports the correlation between C. jejuni morphology and morphology-related genes with pathogenic potential.

Keywords: Campylobacter jejuni; DD-carboxypeptidase; DD-endopeptidase; M23 peptidase; bacterial morphology; pathogenic attributes; peptidoglycan.

Figure 1

In vitro DD-EPase and DD-CPase activity of Pgp4 and Pgp5, and DD-CPase activity of Pgp6. (A) protein domain organization of C. jejuni M23 peptidases Pgp3, Pgp4, Pgp5 and Pgp6 and the recombinant proteins used for protein expression and purification. The Pgp4, Pgp5, and Pgp6 domains were predicated by jackhammr (Frirdich et al., 2023) and Pgp3 domains were assigned based on the available crystal structure (Min et al., 2020). (B) Purified recombinant Pgp4, Pgp5 and Pgp6 proteins separated on 12% SDS-PAGE and stained with Coomassie Brilliant Blue. The predicted molecular weight of each protein is indicated above the band. (C) HPLC chromatograms of E. coli D456 PG incubated with purified Pgp4, Pgp5 and Pgp6 with and without EDTA. E. coli D456 PG alone was used as a control. (D) Schematic diagram of the muropeptide substrates and the Pgp4, Pgp5 and Pgp6 cleavage sites indicated with arrows. G, N-acetylglucosamine; M(r), reduced N-acetylmuramic acid; L-Ala, L-alanine; D-Glu, D-glutamic acid; mDAP, meso-diaminopimelic acid; D-Ala, D-alanine.

Figure 2

Pgp6 sequence analysis. (A) Phylogenetic tree representing Pgp6/0166 and its orthologs used for conservation analysis. Sequences are designated by the species name and the gene locus tag in brackets. The tree was constructed from a multiple sequence alignment of Pgp6 and its orthologs using the maximum likelihood method in MegaX. (B) Sequence logo constructed using Weblogo3 (Crooks et al., 2004) representing the amino acid distribution at a position created from the alignment used in (A). The numbering on the x-axis corresponds to the residue number in Pgp6. The graph indicates a consensus sequence Fx₃Nx₃Rx₂Nx₃I before the start of the M23 peptidase domain. (C) Multiple sequence alignment of Pgp3, Pgp6, and their orthologs (shown by gene locus tag) from Sulfurimonas autotrophica DSM 16294 (SA) and Sulfuricurvum kujiense DSM 16994 (SK) plotted using ENDscript 3.0 (Robert and Gouet, 2014). Four insertion sequences (SEQ1-4), boxed and labeled, are unique to Pgp6. The blue triangle indicates a conserved residue within or close to the insertion sequences.

Figure 3

The AlphaFold model of Pgp6 and insertion sequences (SEQ1-SEQ4). (A) Ribbon diagram of the Pgp6 AlphaFold model. The N-terminal helix, domain 1, and domain 2 are shown in pink, the linker between domain 2 and M23 peptidase domain in red, the M23 peptidase domain in green, and the C-terminal helix in light orange. Residues of the Zn binding motif (H336, D340, H415, and His417) are shown in stick format. The secondary structure elements of the linker and M23 peptidase domain are labeled. (B) Zoomed in view of insertion sequences SEQ1 and SEQ3. The insertion sequences are shown in black. Side chains of residues in the consensus sequence Fx₃Nx₃Rx₂Nx₃I (F271, N275, R279, N282, and I286) are facing toward the catalytic Zn binding motif. Aromatic residues F271 in SEQ1 and Y324 in SEQ3 cluster with a conserved residue F246. (C) Diagram of the partial Pgp6 model showing that SEQ2 and SEQ4 form contacts that anchor the M23 peptidase domain to domain 2. The I298 residue in SEQ2 is hydrogen bonded to the main chain of P188 and Y189; N441 in SEQ4 is hydrogen bonded to the hydroxyl group of S151; and W437 in SEQ4 forms a hydrophobic core with residues W432 and A159. Hydrogen bonds are indicated as dash lines.

Figure 4

Comparison of M23 peptidase domain proteins with different PG substrates. Ribbon diagrams showing the architecture of M23 peptidase domains from enzymes that cleave the terminal D-Ala of pentapeptide side chains and/or the 4-3 cross-links. Pgp6 is a predicted structure retrieved from the AlphaFold database. C. jejuni Pgp3 (Min et al., 2020), H. pylori Csd3 (An et al., 2015), H. pylori Csd1 (An et al., 2016), and V. cholerae ShyA (Shin et al., 2020) are experimental structures obtained by X-ray crystallography. The M23 peptidase domain is shown in green, the linker that connects the N-terminal region and the M23 peptidase domain in red, and the remainder of the model in pink. The zinc coordinating residues are shown as stick models, and the zinc metal ion, if available, is drawn as a sphere. The functional activity and the identity of the amino acids of the peptide substrate of each enzyme are written above the enzyme models. Note the functional activity of Csd1 was determined from the muropeptide profile of the mutant and not by biochemical activity assay.

Figure 5

The effect of pgp4, pgp5 and pgp6 on motility, biofilm formation, autoagglutination and CFW reactivity. The phenotypic properties of the pgp4, pgp5 and pgp6 mutant, complemented (∆pgp4c, ∆pgp5c, and ∆pgp6c) and overexpression (81–176 + pgp4, 81–176 + pgp5 and 81–176 + pgp6) strains were assessed. (A) Motility assayed by measuring halo diameters in soft agar plates. S.E. (error bars) was calculated from 10 technical replicates. Statistical significance was determined using a one-way ANOVA with a Dunnett’s test for multiple comparisons. Data was representative of three independent experiments (B) biofilm formation assessed by crystal violet staining of standing cultures in borosilicate tubes and quantification of dissolved crystal violet at 570 nm. S.E. values were calculated from triplicate cultures and are representative of three independent experiments. Statistical significance was determined using a two-way ANOVA with a Dunnett’s test for multiple comparisons. (C) Autoagglutination was measured at 0, 3, 6, and 24 h in PBS at 25°C. The data was normalized with the OD₆₀₀ at t = 0 representing 100% and the OD600 at each timepoint calculated as a percentage of that at t = 0. A decrease in percentage represents an increase in autoagglutination. S. E. values were calculated from triplicate cultures and are representative of three independent experiments. Statistical significance was determined using a two-way ANOVA with a Dunnett’s test for multiple comparisons. (D) The fluorescence relative to wild type after 48 h of growth on plates containing 0.002% CFW. The controls included ∆spoT and ∆pgp1 representing a hyperfluorescent and hypofluorescent strain, respectively. Note that in the bottom row, all strains are compared to the wild type shown in the first panel of that row as both panels are cropped from the same image. The asterisk (*) indicates a statistically significant difference in comparison to wild type, with *, **, ***, or *** indicating p < 0.05, p < 0.01, p < 0.001 and p < 0.0001, respectively.

Figure 6

The effect of pgp4, pgp5 and pgp6 on acid survival. The ability of the pgp4, pgp5 and pgp6 mutant, complemented (∆pgp4c, ∆pgp5c, and ∆pgp6c) and overexpression (81–176 + pgp4, 81–176 + pgp5 and 81–176 + pgp6) strains to grow on MH-HCl-TV agar at pH 4.5, 5.0, 5.5, 6.0, 6.5, 7.0 and MH-TV unadjusted for pH was assessed. Only growth on pH 5.0 showed differences in comparison to wild type and is shown here. Overnight cultures were standardized to an OD₆₀₀ of 0.50 OD/mL and serially diluted 10-fold in MH-TV broth in a microtiter plate. 5 μL of each dilution was spot plated with the most concentrated at the top of the plate. The MH-HCl-TV plates were incubated for 2 days at 37°C under microaerophilic conditions to assess growth at each dilution. The data presented is representative of three independent experiments.

Figure 7

The effect of pgp4, pgp5 and pgp6 on the transition to the coccoid form. (A) DIC microscope images of C. jejuni wild type 81–176, ∆pgp4, ∆pgp5 and ∆pgp6 mutant strains grown on solid media at 37°C to follow the transition to the coccoid form over time. Representative cells considered to be helical, coccoid or transitioning to the coccoid form are indicated by a, b or c, respectively. (B) The percentage of helical, coccoid and cells transitioning to the coccoid form as determined from DIC images such as those shown in (A) of the pgp4, pgp5 and pgp6 mutant, complemented (∆pgp4c, ∆pgp5c, and ∆pgp6c) and overexpression (81-176+pgp4, 81-176+pgp5, and 81-176+pgp6) strains. At least three separate fields of view of approximately 100–200 bacteria were counted for each strain at each timepoint and this was carried out in triplicate.

Figure 8

The effect of pgp4, pgp5 and pgp6 on adherence, invasion and intracellular survival in epithelial cells. The adherence, invasion and intracellular survival ability of the pgp4, pgp5 and pgp6 mutant, complemented (∆pgp4c, ∆pgp5c, and ∆pgp6c) and overexpression (81–176 + pgp4, 81–176 + pgp5 and 81–176 + pgp6) strains in the INT407 epithelial cell line was assessed by a gentamicin (Gm) protection assay. Adherence and invasion was assessed at 3 h, invasion at 5 h (post-gentamicin treatment for 2 h), and intracellular survival at 7 h. The data was normalized with the cfu of the inoculum representing 100% and the cfu at each timepoint calculated as a percentage of the inoculum. In all experiments S.E. (error bars) were calculated from triplicate readings and are representative of three independent experiments. Statistical significance was calculated using a two-way ANOVA with the Dunnett’s test for multiple comparisons in comparison to wild type. The asterisk (*) indicates a statistically significant difference in comparison to wild type, with *, **, ***, or *** indicating p < 0.05, p < 0.01, p < 0.001 and p < 0.0001, respectively.