Investigation of Gram-negative bacterial surface glycans: characterisation of Moraxella bovis lipooligosaccharide and progress towards developing a Nontypeable Haemophilus influenzae/Moraxella catarrhalis vaccine candidate

Moraxella bovis, Moraxella catarrhalis and Nontypeable Haemophilus influenzae (NTHi) are Gram-negative, oxidase positive, pathogenic microorganisms. M. bovis causes Infectious Bovine Keratoconjunctivitis, known (IBK) as ‘pink-eye’ in cattle worldwide. The disease is of economic importance as it leads to substantial economic loss in the cattle and dairy industries. There are antibiotic treatments available to treat M. bovis associated infections, but they tend to be ineffective at controlling disease outbreaks. Importantly, the current series of antibiotics used to treat IBK have shown occurrence of resistance due to beta-lactamase enzyme produced by the bacteria. There does exist a pilin-based vaccine for IBK which has been approved for use in Australia, although it is problematic because it is not protective against all strains of M. bovis. M. catarrhalis and NTHi are human respiratory tract opportunistic pathogens responsible for otitis media in children and exacerbate chronic obstructive pulmonary disease in adults. Similar to M. bovis, both bacteria have been shown to produce β-lactamase, which has led to the emergence of antibiotic resistance. There is no licenced vaccine for M. catarrhalis or NTHi infections. In the past two-decades studies on M. catarrhalis lipooligosaccharide (LOS) have suggested that this cell surface glycolipid could potentially be incorporated into vaccines. This is based on immunogenicity in a mouse model and role in adherence and invasion of host epithelia and serum resistance. It is also conserved among strains. M. bovis lipooligosaccharide (LOS) is not well studied. Structural analysis of wild-type M. bovis strain Epp63 oligosaccharide (OS from LOS) have identified the core OS as containing eleven sugar residues, including Kdo (the number of Kdo residues is still unknown) with a branched structure. Interestingly, this core OS has an unusual terminal open chain (1S)-GalaNAc residue and lacks heptose residues in its inner core. Recent studies have also elucidated the structure of the cell surface glycans in other strains (M. bovis Mb25 capsule and M. bovis Epp63 LOS), but whether the unusual structural features are present in other strains of M. bovis, was not known. This study elucidated the OS structure and identified the presence of capsular polysaccharide in M. bovis strains Mb25 and L183/2. NMR spectroscopy of the OS from M. bovis Mb25 and L183/2 showed that the structural characteristics of Epp63 and Mb25 strains are shared and that L183/2 OS lacks the terminal (1S)-GalaNAc residue. Strain L183/2 has the same capsular polysaccharide as strain Mb25, namely unsulfated chondroitin, whereas strain Epp63 does not express a capsule. The biological activity of M. bovis Epp63 LOS and the role LOS play in causing disease is not known, therefore a series of biological assays were investigated using OS mutants with varying glycan length. It was identified that LOS truncation affected M. bovis Epp63 susceptibility towards the antibiotics novobiocin and chloramphenicol, but not to nalidixic acid, polymyxin B, rifampin, Tween 20, Triton X-100 or vancomycin. Highly truncated Epp63 OS mutant was found to be susceptible to the bactericidal activity of bovine serum. The growth rate of wildtype L183/2 was significantly slower than that of wild-type Mb25 or Epp63, however, the Epp63 OS mutants showed reduced growth rates compared to the wild-type. Limulus amebocyte lysate (LAL) assay revealed that L183/2 (1.3 x104 EU/mL) and Mb25 (8.9 x 103 EU/mL) had significantly higher toxicity than Epp63 (3.8 x 103 EU/mL; p<0.0001 when compared to L183/2 and Mb25) or Epp63 mutant OS strains (p<0.0001 when compared to L183/2 and Mb25). LOS truncation (from mutant Epp63 strains) affected bacterial interaction with mammalian cells, with increasing truncation inversely correlating with bacterial adherence to Chang conjunctival and HeLa cells. Adherence assays revealed that the Epp63 strain had lower adherence levels than Mb25 or L183/2 strains to Chang or HeLa cells, which decreased with OS truncation. L183/2 was found to exhibit higher levels of adherence than Mb25 or Epp63 to the cell lines. These observations collectively indicate that the OS moiety of the LOS is a significant component in facilitating membrane integrity to preserve normal cell growth and colonisation, maintaining complement resistance of the bacteria to bovine serum and susceptibility towards hydrophobic agents/antibiotics. Interestingly, it also suggests that the OS may possibly play a role in toxicity, at least as measured by LAL assay (lipid A structure of the mutants needs to be elucidated in-order to conclusively indicate OS role). Unlike M. bovis, there is already substantial information known about M. catarrhalis LOS. This information has been exploited here towards vaccine studies utilising M. catarrhalis LOS. Firstly, we investigated the biological role of M. catarrhalis LOS to confirm its suitability as a vaccine antigen. The outer membrane glycan from M. catarrhalis 2951 and 3292 wild-type and 2951lgt1/4Δ mutant strains was isolated, O-deacylated and conjugated to an outer membrane protein (rOMP26VTAL) from NTHi via an adipic dihydrazide linker; to produce the vaccine candidates referred to as 2951dLOS-rOMP26VTAL, 3292dLOS-rOMP26VTAL and 2951lgt1/4ΔdLOS-rOMP26VTAL, respectively. This allowed us to design a single vaccine that would be effective against both M. catarrhalis and H. influenzae. Three subcutaneous immunizations using 2951dLOS-rOMP26VTAL and 3292dLOS-rOMP26VTAL induced an antibody response to their respective serotype antigens and to rOMP26VTAL and NTHi. Confirming the efficacy of the conjugate vaccine, three subcutaneous immunizations using 2951lgt1/4ΔdLOS-rOMP26VTAL elicited an antigen-specific IgG response to all antigens tested (antigens from serotypes A, B, rOMP26VTAL and NTHi). Antisera from serotype A and B conjugates induced bactericidal activity against their respective serotype-specific strains, with no cross-reactivity observed. Antisera from 2951lgt1/4ΔdLOS-rOMP26VTAL immunisation showed protective efficacy against serotype A and B strains. These results indicate that 2951lgt1/4ΔdLOS-rOMP26VTAL is a promising vaccine candidate, but requires further investigation in a challenge model. Overall, the studies within this thesis show that M. bovis OS has a role in bacterial attachment, sensitivity towards bactericidal activity of bovine serum and endotoxin activity of tested strains. This thesis also shows that M. catarrhalis LOS is an important bacterial antigen that can be utilised for vaccine development. Additionally, OMP26, being highly-conserved across NTHi strains and immunogenic makes it the ideal carrier protein for respiratory related vaccine development. Combining the LOS from M. catarrhalis and the rOMP26VTAL from NTHi paves the path for the development of a single vaccine that may prevent respiratory infections caused by these two predominant pathogens.