ALBERT

Notes: It is shown that Shigella flexneri maintains genetic control over the modal chain length of the O-antigen polysaccharide chains of its lipopolysaccharide (LPS) molecules because such a distribution is required for virulence. The effect of altering O-antigen chain length on S. flexneri virulence was investigated by inserting a kanamycin (Km)-resistance cassette into the rol gene (controlling the modal O-antigen chain length distribution), and into the rfbD gene, whose product is needed for synthesis of dTDP-rhamnose (the precursor of rhamnose in the O-antigen). The mutations had the expected effect on LPS structure. The rol::Km mutation was impaired in the ability to elicit keratoconjunctivitis, as determined by the Serény test. The rol::Km and rfbD::Km mutations prevented plaque formation on HeLa cells, but neither mutation affected the ability of S. flexneri to invade and replicate in HeLa cells. Microscopy of bacteria-infected HeLa cells stained with fluorescein isothiocyanate (FITC)-phalloidin demonstrated that both the rol::Km and rfbD::Km mutants were defective in F-actin tail formation: the latter mutant showed distorted F-actin tails. Plasma-membrane protrusions were occasionally observed. Investigation of the location of IcsA (required for F-actin tail formation) on the cell surface by immunofluorescence and immunogold electron microscopy showed that while most rol mutant bacteria produced little or no cell-surface IcsA, 10% resembled the parental bacterial cell (which had IcsA at one cell pole; the rfbD mutant had IcsA located over its entire cell surface although it was more concentrated at one end of the cell). That the O-antigen chains of the rol::Km mutant did not mask the IcsA protein was demonstrated by using the endorhamnosidase activity of Sf6c phage to digest the O-antigen chains, and comparing untreated and Sf6c-treated cells by immunofluorescence with anti-IcsA serum.

Notes: We have isolated the lysogenic bacteriophage SfII, which mediates glucosylation of Shigella flexneri O-antigen, resulting in expression of the type II antigen. SfII belongs to group A of the Bradley classification and has a genome size of 42.3 kb. DNA sequencing of a 4 kb BamHI subclone identified four open reading frames (ORFs), of which only two were found to be necessary for serotype conversion. These genes were named bgt, which encodes a putative bactoprenol glucosyl transferase, and gtrII, encoding the putative type II antigen determining glucosyl transferase. These genes are adjacent to the integrase gene (int ) and attachment site (attP ), which are highly homologous to those of Salmonella bacteriophage P22. Another ORF encoded a highly hydrophobic protein of 120 amino acids with homologues in Escherichia coli, Salmonella bacteriophage P22 and S. flexneri. Previous studies identified gtrX, the glucosyl transferase gene, of bacteriophage SfX, which also glucosylates the O-antigen specifically. We determined that gtrX-mediated expression of the group 7,8 antigen also requires bgt. This allowed us to identify gtrII as being the serotype antigen II determining glucosyl transferase. Southern hybridization and polymerase chain reaction (PCR) analyses indicated that bgt homologues exist in the genomes of all S. flexneri serotypes and in E. coli K-12, whereas gtrII was only detected in strains of serotype 2. Transposon TnphoA-derived chromosomal mutations of bgt and gtrII in S. flexneri serotype 2a were isolated and characterized. [35S]-methionine labelling and the use of a T7 RNA polymerase expression system identified a protein of 34 kDa corresponding to Bgt. However, GtrII, which has a predicted molecular weight of 55 kDa, was not detected. We propose that the function of Bgt is to transfer the glucose residues from the UDP-glucose onto bactoprenol and GtrII then transfers the glucose onto the O-antigen repeat unit at the rhamnose III position. The chromosomal organization of these serotype-converting genes, when compared with their homologues in E. coli K-12 chromosome and the P22 bacteriophage genome, were very similar. This suggests that the regions encode similar functions in these organisms and have a similar evolutionary origin.

Notes: Using defined rfb mutants, defective in the biosynthesis of the O-antigen of the lipopolysaccharide (LPS), and monoclonal antibodies (MAbs) to the A, B and C LPS antigens, we have examined the distribution of the antigens and the effects of their loss. By immunogold electron microscopy, it has been possible to determine the relative amounts of the A, B and C antigens on Inaba and Ogawa cells, confirming previous studies based upon bacterial agglutination and hemagglutination inhibitions. These antigens are absent from rfb::Tn mutants selected as resistant to phages which have been shown to use the O-antigen as their receptor. These mutants were severely attenuated as measured by both LD50 and their ability to compete with the wild-type parents when analyzed in the infant mouse cholera model. These mutants were unchanged in the export of cholera toxin or other secreted proteins but revealed an altered outer membrane protein profile. The competition defect suggested an effect on TCP (toxin-coregulated pilus). An analysis of the rfb::Tn mutants revealed that they were unable to assemble TCP on their surface, but the major subunit, TcpA, could be found as an intracellular pool. These mutants could be complemented back to wild-type using the cloned rfb region, implying that functional TCP assembly is dependent upon an intact LPS.

Notes: Abstract Rheumatic fever continues to be a significant problem in Australian Aboriginal communities and developing countries worldwide. Early diagnosis could facilitate the institution of penicillin prophylaxis resulting in the prevention of recurrences of rheumatic fever. An overlapping biotinylated peptide bank of 82 peptides, based on the known sequence of Streptococcus pyogenes M24 protein, was used in a standard enzyme immunoassay. A total of 82 sera were tested from both aboriginal and non-aboriginal subjects with clinically proven rheumatic fever, rheumatic heart disease and matched controls. Two peptides with significant sequence homology at the C-terminal end were found to be discriminatory between aboriginal cases and controls. It is proposed that these peptides could be the basis of a serological test for rheumatic fever.

Notes: Cloning of the rfb genes of Shigella flexneri 2a into Escherichia coli K-12 strain DH1 results in the synthesis of lipopolysaccharides (LPS) with an O-antigen chain having type antigen IV and group antigens 3,4. During genetic studies of these rfb genes in E. coli K-12, we observed that strains harbouring plasmids with certain mutations (inversion and transposon insertions) which should have blocked O-antigen synthesis nevertheless still produced LPS with O-antigen chains. These LPS migrated differently on silver-stained SDS—polyacrylamide gels, compared with the LPS produced by wild-type rfb genes, and the group 3,4 antigens were barely detectable, suggesting that the O-antigen was altered. Investigation of the genetic determinants for production of the altered O-antigen/LPS indicated that: (i) these LPS are produced as a result of mutations which are either polar on rfbF or inactivate rfbF; (ii) the rfbX gene product (or a similar protein in the E. coli K-12 rfb region) is needed for production of the altered O-antigen in the form of LPS; (iii) the rfbG gene product is required for the production of both the parental and altered LPS; (iv) the dTDP-rhamnose biosynthesis genes are required. Additionally, an E. coli K-12 gene product(s) encoded outside the rfb region also contributes to production of the O-antigen of the altered LPS. An antiserum raised to the altered LPS from strain DH1(pPM2217 (rfbX::Tn1725)) was found to cross-react with nearly all S. flexneri serotypes, and with the altered LPS produced by other DH1 strains harbouring plasmids with different rfb mutations, as described above. The reactivity of the altered LPS with a panel of monoclonal antibodies specific for various S. flexneri O-antigen type and group antigens demonstrated that their O-antigen components were closely related to that of S. flexneri serotype 4. The RfbF and RfbG proteins were shown to have similarity to rhamnose transferases, and we identified a motif common to the N-termini of 6-deoxy-hexose nucleotide sugar transferases. We propose that the E. coli K-12 strains harbouring the mutated S. flexneri rfb genes produce LPS with a hybrid O-antigen as a consequence of inactivation of RfbF and complementation by an E. coli K-12 gene product. Analysis of the genetic and immunochemical data suggested a possible structure for the O-antigen component of the altered LPS.

Notes: The nucleotide sequence of the mip genes and their inferred amino acid sequences were determined from 35 Legionella species and compared with the published sequences for L. pneumophila, L. micdadei and L. longbeachae. The sequences were 69–97% conserved at the nucleotide level and 82–99% at the amino acid level, with total conservation of amino acids determined to be associated with sites known to be involved in peptidyl prolyl cis–trans isomerase activity. No apparent difference could be determined in the arrangement of amino acids that would predict a functional difference in Mip from species associated with disease and Mip from species isolated only from the environment. Additionally, a phylogenetic comparison of the sequences with published 16S RNA sequences, using both genetic distance and maximum parsimony methods, was performed. Few relationships were apparent that were well supported by both data sets, the most robust being a clade comprising {[(cincinnatiensis, longbeachae, sainthelensi, santicrucis) gratiana] (moravica, quateirensis, shakespearei, worsleiensis) anisa, bozemanii, cherrii, dumoffii, gormanii, jordanis, parisiensis, pneumophila, steigerwaltii, tucsonensis, and wadsworthii }.