Changes in flagellin glycosylation affect Campylobacter autoagglutination and virulence

Abstract

Analysis of the complete flagellin glycosylation locus of Campylobacter jejuni strain 81-176 revealed a less complex genomic organization than the corresponding region in the genome strain, C. jejuni NCTC 11168. Twenty-four of the 45 genes found between Cj1293 and Cj1337 in NCTC 11168 are missing in 81-176. Mutation of six new genes, in addition to three previously reported, resulted in a non-motile phenotype, consistent with a role in synthesis of pseudaminic acid (PseAc) or transfer of PseAc to flagellin. Mutation of Cj1316c or pseA had been shown to result in loss of the acetamidino form of pseudaminic acid (PseAm). Mutation of a second gene also resulted in loss of PseAm, as well as a minor modification that appears to be PseAm extended with N-acetyl-glutamic acid. Previously described mutants in C. jejuni 81-176 and Campylobacter coli VC167 that produced flagella lacking PseAm or PseAc failed to autoagglutinate. This suggests that interactions between modifications on adjacent flagella filaments are required for autoagglutination. Mutants (81-176) defective in autoagglutination showed a modest reduction in adherence and invasion of INT407 cells. However, there was a qualitative difference in binding patterns to INT407 cells using GFP-labelled 81-176 and mutants lacking PseAm. A mutant lacking PseAm was attenuated in the ferret diarrhoeal disease model.

Fig. 1. Pathway of pseudaminic acid synthesis in C. jejuni from UDP-GlcNAc

The enzymatic activity of the products of PseB (Cj1293) and PseC (Cj1294) have been described in Schoenhofen et al. 2005 and that of PseI (Cj1317c) by Chou et al., (2005). Broken arrows indicate biosynthetic steps for which the respective enzymatic activities have still to be assigned to a campylobacter gene. The presumed activity of PseF is based on homology to the corresponding enzyme in the sialic acid biosynthetic pathway, NeuA (Vann et al., 1987).

Fig. 2. Isoelectric focusing pattern of flagellins

Purified flagellins from 81–176 and various mutants were separated on an ampholyte mixture of pH 4–6 and stained with Gel Code Blue. A. Characterization of flagellin from the pseD mutant. Lane 1, 81–176 pseA::cat (Thibault et al., 2001); lane 2, wildtype 81–176; lane 3, 81–176 pseD::cat; lane 4, 81–176 pseD::cat (pRY107/28+pseD). B. Complementation of pseA by insertion into atsA. Lane 1, 81–176 atsA::aphA3; lane 2, wildtype 81–176; lane 3, 81–176 pseA::cat; lane 4, 81–176 pseA::cat, atsA::pseA+aphA3. The position of IEF markers is indicated on the right.

Fig. 3. Structural analysis of flagellin from a mutant in Cj1333

A. Molecular mass determination of intact flagellin from the mutant in pseD. B. Tandem mass spectrometry analysis of the multiply charged ion at m/z 985 from flagellin from the pseD mutant. C. Second generation product ion spectra of oxonium ion m/z 487 obtained from in source dissociation of multiply charged flagellin ion. Conditions: In-source fragmentation was obtained by increasing the RF Lens 1 from 50 to 125 V to produce oxonium fragment ions (m/z 487) from the native flagellin in the orifice/skimmer region. Argon was used as target gas at collision energy of 30 V.

Fig. 4. Structural analysis of flagellin from 81–176 pseD (pRY107/pseD)

A. Molecular mass determination of intact flagellin from the complemented strain. B. Second generation product ion spectra of oxonium ion m/z 486 obtained from in source dissociation of a multiply charged flagellin ion. The presence of m/z 316 product ion confirms the restoration of both groups on this flagellin (see Schirm et al., 2005). Conditions used for in source fragmentation are as described in figure 3.

Fig. 5. Autoagglutination of Campylobacters

Suspensions of bacteria were set to an OD₆₀₀ of 1.0 and incubated without shaking. After 18 h incubation at room temperature, the OD₆₀₀ of the top 1 ml of the tube determined. Mutants used in this study are listed in Table 2. A. Comparison of wildtype 81–176 with isogenic mutants. Column 1, wild-type 81–176; column 2, 81–176 flaA::aph3; column 3, 81–176, pflA::aph3; column 4, 81–176 kpsM::aph3; column 5, 81–176 neuC1::cat; column 6, 81–176 luxS::cat. B. Flagellin glycosylation mutants of C. jejuni 81–176. Column 1, wild-type 81–176; column 2, 81–176 pseD::cat; column 3, 81–176 pseD::cat (pRY107/pseD); column 4, 81–176 pseA::cat; column 5, 81–176 pseA::cat, astA::pseA, aph3; column 6, 81–176 astA::aph3; column 7, 81–176 CjB1301::cat. C. Flagellin glycosylation mutants of C. coli VC167 T2. Column 1, wild-type VC167 T2; column 2, VC167 T2 ptmD::aph3; column 3, VC167 T1 ptmD::aph3 (pRY111/ptmD); column 4, VC167 T2 pseB::cat; column 5, VC167 T2 pseB::cat (pRY107/pseB). The graphs show the mean and standard deviations of 2–8 independent determinations.

Fig. 6. The effect of loss of PseAm on adherence and invasion of INT407 cells

A. Adherence. B. Invasion. Column 1, wild-type 81–176; column 2, 81–176 pseA::cat; column 3, 81–176 pseA::cat, astA::pseA+aphA3; column 4, 81–176 atsA::aphA3.

Fig. 7. Comparison of adherence patterns to INT407 cells of GFP tagged wildtype 81–176 with GFP tagged 1906 mutant

Panel A shows the adherence patterns of wildtype 81–176 and the pseA mutant following 18 h incubation. Panels B and C show adherence patterns following 24 h incubation. Fluorescence images are on the left and phase images are on the right. Final magnification of panels A and B, 1000X; final magnification of panel C, 1500X.

Fig. 8. Virulence in ferrets

Animals were fed C. jejuni at doses ranging from 3.8 x 10¹⁰–4.5 x 10¹⁰ and monitored for diarrheal disease for 4 days. The data are presented as the cumulative % animals that developed diarrhea over time. Grey bars, 81–176; white bars, 81–176 pseA::cat; black bars, 81–176 pseA::cat, astA::pseA + aph3. The data represent the two separate experiments with 8 animals per group or a total number of 16 animals per strain.