Attachment and invasion of Neisseria meningitidis to host cells is related to surface hydrophobicity, bacterial cell size and capsule

Abstract

We compared exemplar strains from two hypervirulent clonal complexes, strain NMB-CDC from ST-8/11 cc and strain MC58 from ST-32/269 cc, in host cell attachment and invasion. Strain NMB-CDC attached to and invaded host cells at a significantly greater frequency than strain MC58. Type IV pili retained the primary role for initial attachment to host cells for both isolates regardless of pilin class and glycosylation pattern. In strain MC58, the serogroup B capsule was the major inhibitory determinant affecting both bacterial attachment to and invasion of host cells. Removal of terminal sialylation of lipooligosaccharide (LOS) in the presence of capsule did not influence rates of attachment or invasion for strain MC58. However, removal of either serogroup B capsule or LOS sialylation in strain NMB-CDC increased bacterial attachment to host cells to the same extent. Although the level of inhibition of attachment by capsule was different between these strains, the regulation of the capsule synthesis locus by the two-component response regulator MisR, and the level of surface capsule determined by flow cytometry were not significantly different. However, the diplococci of strain NMB-CDC were shown to have a 1.89-fold greater surface area than strain MC58 by flow cytometry. It was proposed that the increase in surface area without changing the amount of anchored glycolipid capsule in the outer membrane would result in a sparser capsule and increase surface hydrophobicity. Strain NMB-CDC was shown to be more hydrophobic than strain MC58 using hydrophobicity interaction chromatography and microbial adhesion-to-solvents assays. In conclusion, improved levels of adherence of strain NMB-CDC to cell lines was associated with increased bacterial cell surface and surface hydrophobicity. This study shows that there is diversity in bacterial cell surface area and surface hydrophobicity within N. meningitidis which influence steps in meningococcal pathogenesis.

Figure 1. The surface area of strain NMB-CDC is larger than that of strain MC58.

A representative flow cytometry histogram of the forward scatter (FSC) from strains MC58 (white area under the curve) and NMB-CDC (grey shaded area) analysed with the BD Influx cytometer. The mean of the FSC of NMB-CDC was 19,000 and the mean of the FSC of MC58 was 8,000 in this experiment.

Figure 2. Strain NMB-CDC associates with and invades epithelial cell lines at greater rates than strain MC58.

The rates of association and invasion of strains NMB-CDC and MC58 into Detroit 562 epithelial cells (Panel A) and human bronchial epithelial cells 16HBE14σ- (Panel B) were assessed. Attachment as percent of the inoculum (black bars read off the left y-axis) and invasion as the percentage of the associated bacteria (white bars read off the right y-axis) is shown. The average rate (+/–SEM) from three biological repeats in triplicate following 6 hrs co-incubation is shown. *: p<0.001 determined by Mann-Whitney t-test compared to strain NMB-CDC, ND: Not Detected, viable counts were below the limit of detection.

Figure 3. The role of Type IV pili is conserved in bacterial association with host cells.

Panel A. The ability of strains lacking pili (ΔpilE) to attach to Detroit 562 cells was compared to the parental wild-type. Strains NMB-CDC and MC58 and their respective pilE mutant strains are shown in black and white, respectively. All strains were Opa+, Cap+ and LNT+. The average rate (+/−SEM) of association as a percentage of the inoculum of three biological repeats in triplicate as determined by viable counts following 6 hr co-incubation is shown. Panel B. The role of the class of pilin in association with host cells. Strain NMB-CDC was modified to express class I pilin from the iga locus in the absence of an intact pilEII locus (CKNM397). All strains were Opa+, Cap+ and LNT+. The ability of CKNM397 to associate with Detroit 562 cells was compared to parental wild-type strain NMB-CDC and CKNM394 containing an inactivated iga locus. The rate of association as a percentage of the inoculum of three biological repeats in triplicate was determined by viable counts following 1 hr co-incubation. The relative rate of association of each strain was normalised to strain NMB-CDC (value of 1) and is plotted as the average fold change (+/−SEM). *: p<0.005 determined by Mann-Whitney t-test. Panel C. The expression of a pilin glycan containing either GATDH (strain NMB-CDC) or DATDH (strain MC58) retains the same role in bacterial invasion for both strains of meningococci. The biosynthesis of the pilin glycan was interrupted by insertional inactivation of pglF. All strains were Pil+, Opa+, Cap+ and LNT+. The rate of association was determined by viable counts following 6 hr co-incubation. The relative rate of association of each strain was normalised to parental wild-type (value of 1) and is plotted as the average fold change (+/−SEM). Rates of attachment are shown in black bars (left y-axis) and invasion rates are shown in white bars (right y-axis). *: p<0.02 determined by Mann-Whitney t-test.

Figure 4. The effect of capsule and total sialic acid on attachment and invasion of strains NMB-CDC and MC58.

Panel A and B show the rates of association and invasion, respectively, of capsule and sialic acid mutants relative to strain MC58. Parental wild-type (white bars), capsule mutants (ΔsynD, light grey bars), LOS sialylation mutants (Δlst, dark grey bars) and sialic acid mutants (ΔsynB, black bars) in strain NMB-CDC (solid) and MC58 (diagonally stripes bars) are shown. The rate of association and invasion was determined by viable counts following 6 hr co-incubation. The relative rate of association of each strain was normalised to strain MC58 (value of 1) and values were plotted as the average fold change (+/−SEM). *: p<0.05, **:p<0.005, ***:p<0.0001 as determined by Mann-Whitney t-test).

Figure 5. MisR directly binds to the synA-ctrA promoter region.

Panel A. MisR interacts with the cps intergenic probe containing both the synA and ctrA promoters. Phoshorylated MisR (MisR∼P, Lanes 2–4) and MisR (lanes 5–7) binds the probe in a dose-dependent manner. Lane 1 contains probe alone, Lanes 2 and 5 contain 68 pmol of protein; lanes 3 and 6 contain 136 pmol of protein; lanes 4 and 7 contain 204 pmol of protein. Panel B. Competition EMSA. The mobility shift of the labeled probe by MisR∼P (136 pmol, lane 2) relative to labeled probe without protein (lane 1) was competed away by unlabelled probe (1 µg in lane 3 and 2 µg in lane 4). This interaction between MisR∼P and the labeled probe was specific as the complex remained intact and could not be competed away when cold, unlabeled non-specific DNA was added (1 µg in lane 5 and 2 µg in lane 6).

Figure 6. Analysis of the capsule polymer and surface hydrophobicity profiles of strains NMB-CDC and MC58.

Panel A. Physical modality and length of the capsule polymers of both strains are the same. Purified capsule preparations (5 µg [lanes 1 and 3] and 10 µg [lanes 2 and 4]) from strains MC58 (lanes 1 and 2) and NMB (lanes 3 and 4) were separated by DOC-PAGE and silver stained. Panel B. Comparison of surface hydrophobicity of strains NMB-CDC and MC58 using HIC. Surface hydrophobicity of strains NMB-CDC (A) and MC58 (B) with their respective isogenic mutants (non-encapsulated, M7 and MC58ΔsynB and hyper-encapsulated variants, NMBΔmisR and MC58ΔmisR) were assessed by interaction chromatography. The proportion of bacteria eluted from the column was determined by optical density and is presented relative to the isogenic non-encapsulated control (M7 for NMB and MC58ΔsynB for MC58). The average of at least three biological repeats (+/−SEM) for each strain is shown. *: p<0.01 determined by unpaired t-test compared to the non-encapsulated controls.