Heads or tails? An insight into the nature of antibacterial structures of an entomopathogenic bacterium Brevibacillus laterosporus : A thesis submitted in partial fulfilment of the requirements for the Degree of Doctor of Philosophy at Lincoln University

Abstract

Bacterium eating viruses” (phages) are one of the biotic factors that can disrupt the mass production of bacteria through the lysis of the cells during growth. Brevibacillus laterosporus (Laubach) is an entomopathogenic, gram-positive, and spore-forming bacterium found around the world, but individual strains can differ greatly in their virulence and hosts. Since 2012, three B. laterosporus strains (Bl 1821L, Bl 1951, Bl Rsp) have been discovered and characterised in New Zealand. All the strains exhibited pathogenicity against diamondback moth larvae and mosquitoes. However, during culturing, the cultures often lost virulence, which was perceived to be due to growth issues and the presence of potential bacteriophages. Though isolates Bl 1821L and Bl 1951 are under development as a biopesticide, the lack of consistent production has hindered their development. Based on these issues this research project commenced to isolate and characterise the suspected bacteriophages, cure the bacteria of the invading phage particles, compare the virulence of phage cured and uncured strains against diamondback moth, and finally develop a production protocol that would allow a biopesticide to be developed from this insect pathogenic bacterium free of phages. However, initial work involving classical phage isolation and enumeration assays (plaque and serial dilutions) and an electron microscopic examination could not substantiate the presence of putative phage particles, despite bioinformatic predictions indicating the presence of intact phages in Bl 1821L, Bl 1951, and Bl Rsp genomes. Assessment of electron micrographs of mitomycin C induced cultures (Bl 1821L & Bl 1951) and bioactivity tests of polyethylene glycol precipitated cultures validated the presence of incomplete phage particles with hexagonal or phage head (encapsulating) and contractile tail-sheath like structures. Subsequent N-terminal sequencing of a prominent ~48 kD protein of SDS-PAGE after mitomycin C induction led to the discovery of putative antibacterial phage tail-like bacteriocins (PTLBs) in the Bl 1821L genome. The putative Bl 1821L PTLB displayed a broader spectrum of activity than the Bl 1951 PTLB. N-terminal sequencing of purified ~48 kD protein of Bl 1821L identified a phage-like element PBSX protein XkdK in the Bl 1821L genome. BLASTp analysis of Bl 1821L 48 kD identified protein accessions with >90% amino acid similarity to the phage tail-sheath protein of Bl LMG 15441. Using the same methodology an XkdK homolog was identified in the Bl 1951 genome. Although the translated product of the Bl 1821L xkdK gene encoding region exhibited amino acid identity to the analogous region of Bl 1951, the bioinformatic analysis revealed some differences in the operon surrounding the gene, a region which corresponded to the PBSX region in Bacillus subtilis. Bioinformatically, the PBSX-like region in Bl 1951 encodes imperfect repeats of glycine rich proteins (1700 bp long) while a putative phage region resides in the analogous Bl 1821L region. A second putative antibacterial protein (bacteriocin) of ~30 kD was also found in SDS PAGE analysis of purified Bl 1821L and Bl 1951 putative antibacterial protein. N-terminal sequencing of purified ~30 kD protein identified matches to both a 25 kD hypothetical and a 30 kD putative encapsulating protein homologs residing in each of the genomes. Various purification methods employed in this study enabled the purification of one putative antibacterial protein (~30 kD) of Bl 1951 and two putative antibacterial proteins (~30 kD & ~48 kD) of Bl 1821L. Subsequent TEM examination of purified antibacterial Bl 1821L proteins revealed the presence of phage head-like encapsulin (~30 kD) and polysheath-like (~48 kD) structures. Although only the ~30 kD protein was purified from Bl 1951, both the phage head-like encapsulin and polysheath-like structures were observed under an electron microscope. SDS-PAGE analysis of spontaneously produced putative antibacterial proteins of Bl 1821L and Bl 1951 upon purification also yielded two prominent bands of ~30 kD and ~48 kD. Assaying size exclusion chromatographic fractions of Bl 1951 harbouring the 30 kD against Bl 1821L resulted in the presence of small Bl 1821L cells that may be indicative of persister cell formation in the population of Bl 1821L. BLASTp analysis of the homologs Bl 1821L and Bl 1951 30 kD amino acid sequences identified >97% amino acid identity to the Linocin M18 bacteriocin family protein of Bl LMG 15441 and Bl GI-9 which are known as encapsulating proteins. Using the bioinformatic tools AMPA and CellPPD motifs relating to the bactericidal activity and cell penetrating peptides were identified. Antibacterial activity of the identified 25 kD hypothetical and 30 kD putative encapsulating proteins of Bl 1821L was further validated through gene expression in a gram-positive bacterium Bacillus subtilis WB800N and the subsequent assay tests and SDS-PAGE analysis of purified proteins from the recombinants. A preliminary assessment of the expressed 25 kD hypothetical gene (pHT01-hypo) exhibited an increased effect against Bl 1821L and Bl 1951 compared to the 30 kD putative encapsulating gene (pHT01-encap). The putative antibacterial proteins (bacteriocins) of Bl 1821L and Bl 1951 sustained the antagonistic activity against B. laterosporus over a wide range of pHs and temperatures. A loss in the antagonistic activity of putative bacteriocins of Bl 1821L and Bl 1951 after treatment with proteolytic enzymes authenticated their proteinaceous nature. However, catalase treatment could not abrogate the inhibitory action of the putative antibacterial proteins of Bl 1821L and Bl 1951, which showed that growth inhibition was not due to hydrogen peroxide production. This study describes the first examples of the spontaneous induction of high molecular weight (HMW) bacteriocins of Bl 1821L and Bl 1951 from the genus Brevibacillus. Spontaneously induced HMW bacteriocins of Bl 1951 affected the growth of bacterium by causing a significant decline in the number of viable cells after 18 hours of culture inoculation that corresponded to the highest antagonistic activity against Bl 1951 and Bl 1821L. Likewise, though not at a significant level spontaneously induced HMW bacteriocins of Bl 1821L decreased the number of viable cells of Bl 1821L after 18 hours of growth. Antagonistic activity of putative antibacterial proteins in growth assays not only varied between the different proteins but also with the state of the proteins (crude or purified). Crude lysate of Bl 1821L and Bl 1951 harbouring both the putative encapsulating (30 kD) and phage tail-like (48 kD) proteins prominently caused autocidal activity with a decrease of 30.1% and 48.4% in the number of viable cells. The purified ~30 kD putative encapsulating protein of Bl 1821L and Bl 1951 exhibited bactericidal activity against both strains. Numerous plausible killing mechanisms of 30 kD putative encapsulating protein of Bl 1821L and Bl 1951 are proposed. These include the activation of stress relevant transcriptional regulator family proteins (PadR & MarR) or YtxJ protein under some unknown stresses, iron malnutrition, failure of ferritin protein to detoxify iron, and cell penetrating peptides activity. The purified ~48 kD putative phage tail-sheath protein of the PBSX-like region of Bl 1821L was active against Bl 1951. Bactericidal activity of ~48 kD purified phage tail-sheath protein of Bl 1821L was likely due to the contractile injection system by forming pores in susceptible cells. Overall, this research identified and characterised two bactericidal proteins of ~30 kD (encapsulating) and ~48 kD (phage tail-sheath) in Bl 1821L and one ~30 kD encapsulating protein in Bl 1951 that were found to be implicated in the growth issues of the insect pathogenic strains Bl 1821L and Bl 1951. A hypothetical protein (25 kD) with more potent putative antibacterial activity in Bl 1821L and Bl 1951 was also identified and expressed late in the project but requires further investigation. The findings provided a wealth of knowledge that will be useful in the future development of a biopesticide from this beneficial bacterium.