Bacteriophage-mediated transduction of antibiotic resistance in enterococci.

AIMS: Temperate bacteriophages are bacterial viruses that transfer genetic information between bacteria. This phenomenon is known as transduction, and it is important in acquisition of bacterial virulence genes and antimicrobial resistance determinants. The aim of this study was to demonstrate the role of bacteriophages in gene transfer (antibiotic resistance) in enterococci. METHODS AND RESULTS: Three bacteriophages from environmental samples isolated on pig host strains of Enterococcus gallinarum and Enterococcus faecalis were evaluated in transduction experiments. Antibiotic resistance was transferred from Ent. gallinarum to Ent. faecalis (tetracycline resistance) and from Ent. faecalis to Enterococcus faecium, Enterococcus hirae/durans and Enterococcus casseliflavus (gentamicin resistance). CONCLUSIONS: Bacteriophages play a role in transfer of antibiotic resistance determinants in enterococci. SIGNIFICANCE AND IMPACT OF THE STUDY: This study confirms previous suggestions on transduction in enterococci, in particular on interspecies transduction. Interspecies transduction is significant because it widens the range of recipients involved in antimicrobial resistance transfer.

A study of the transfer of tetracycline resistance genes between Escherichia coli in the intestinal tract of a mouse and a chicken model.

Experiments to demonstrate the transfer of genes within a natural environment are technically difficult because of the unknown numbers and strains of bacteria present, as well as difficulties designing adequate control experiments. The results of such studies should be viewed within the limits of the experimental design. Most experiments to date have been based on artificial models, which only give approximations of the real-life situation. The current study uses more natural models and provides information about tetracycline resistance as it occurs in wild-type bacteria within the environment of the normal intestinal tract of an animal. Tetracycline sensitive, nalidixic acid resistant Escherichia coli isolates of human origin were administered to mice and chicken animal models. They were monitored for acquisition of tetracycline resistance from indigenous or administered donor E. coli. Five sets of in vivo experiments demonstrated unequivocal transfer of tetracycline resistance to tetracycline sensitive recipients. The addition of tetracycline in the drinking water of the animals increased the probability of transfer between E. coli strains originating from the same animal species. The co-transfer of unselected antibiotic resistance in animal models was also demonstrated.

Clinical evaluation of a peptide-ELISA based upon N-terminal B-cell epitope of the VapA protein for diagnosis of Rhodococcus equi pneumonia in foals.

A total of 227 field samples from naturally exposed foals aged between 3 weeks and 6 months were used in an evaluation of a peptide-based enzyme-linked immunosorbent assay (ELISA) for diagnosis of Rhodococcus equi infection. A biotinylated peptide derived from the virulence-associated protein A (VapA) of R. equi, a horse pathogen, was synthesized and designated as PN11-14. The peptide corresponds to the N-terminal B-cell epitope TSLNLQKDEPNGRASDTAGQ of the VapA protein. Based upon a serum immunoglobulin (Ig)G titre of 512 as a positive cut-off value for the R. equi infection, the ELISA provided the overall sensitivity of 47.62%, specificity of 69.67% and an accuracy of 59.47% with a positive predictive value of 57.47% for true R. equi pneumonia. The assay was improved by detecting VapA-specific IgGb antibodies against N-terminal B-cell epitope of the VapA protein rather than IgG antibodies. The VapA-IgGb ELISA showed the overall sensitivity of 70.47%, specificity of 72.13% and accuracy of 71.36% with a positive predictive value of 68.52%. Diagnosis of R. equi disease in 6-week-old foals showed that the VapA-IgGb ELISA provided an increasing trend (P=0.0572) in sensitivity of 82.4% in comparison with the VapA-IgG ELISA which showed the sensitivity of 58.8%. However, differences in specificity of both tests were statistically insignificant (P=0.357) as analysed by the McNemar test. These results indicated that detection of VapA-specific IgGb antibodies may be a better predictor of R. equi disease in foals.

Recognition of a B-cell epitope of the VapA protein of Rhodococcus equi in newborn and experimentally infected foals.

The aim of this study was to evaluate the usefulness of the previously identified B-cell epitope TSLNLQKDEPNGRASDTAGQ of the VapA protein of Rhodococcus equi and its association with R. equi pneumonia. A modified peptide designated PN11-14 corresponding to the epitope was recognized by all sera from experimentally infected foals with virulent R. equi ATCC103+ containing the virulence plasmid but not by its plasmid-cured derivative ATCC103- strain. Marked levels of VapA-specific immunoglobulin (Ig)G were detected in all sera from the ATCC103+ infected foals at 2 weeks after the infection. One control animal had high titres as determined by the peptide enzyme-linked immunosorbent assay (ELISA), indicating the ELISA may not absolutely differentiate between foals with R. equi pneumonia and healthy exposed foals in farms where the prevalence of disease is high. However, numbers of animals used were small. Further evaluation of the peptide ELISA with field samples is necessary to determine whether the assay is diagnostically useful. This study showed that levels of passive transfer of maternal IgG antibodies to the epitope in newborn foals could be measured. Interestingly, the maternally derived antibodies were found to significantly (P<0.05 by Student's t-test) decline 2 weeks after birth. Seroconversion against naturally occurring VapA expressing R. equi could be detected in some foals at 4 weeks of age. Antibodies to the epitope peaked and were significantly (P<0.05) greater in foals aged between 6 and 8 weeks. These results indicated that the peptide ELISA could be used to monitor anti-VapA antibodies in foals, particularly those at the age of 4-6 weeks. It is possible that the ELISA may be of some use as a diagnostic test on farms where R. equi is non-endemic. Further studies using large number of field samples are needed to verify this assumption.

Use of AFLP and PFGE to discriminate between Salmonella enterica serovar Typhimurium DT126 isolates from separate food-related outbreaks in Australia.

In 2001 Salmonella enterica serovar Typhimurium definitive phage-type (DT) 126 was isolated at higher frequency in Australia compared to other S. Typhimurium phage types and in comparison to previous years. Associated with this increase was the implication of this phage type in a number of food-related outbreaks. We compared fluorescent amplified fragment length polymorphism (FAFLP) to pulsed-field gel electrophoresis (PFGE), the current 'gold standard' for molecular typing of Salmonella for the discrimination between outbreak-associated isolates and epidemiologically unrelated DT126 strains. FAFLP showed a greater ability to discriminate between isolates than PFGE, with 16 groups of clusters or individual isolates with < 90% similarity to each other compared to three groups as determined by PFGE. Both methods were able to discriminate between isolates from two separate outbreaks in South Australia and isolates associated with an outbreak at a restaurant in New South Wales. The resolving power of both methods was not sufficient to separate all epidemiologically unrelated DT126 isolates from the outbreak isolates. We conclude that amplified fragment length polymorphism is a useful tool to assist in the discrimination of S. Typhimurium DT126 isolates.

Antimicrobial resistance in Campylobacter spp., Escherichia coli and enterococci associated with pigs in Australia.

The major influences on the amplification and spread of antibiotic-resistant bacteria are the therapeutic use of antibiotics in human medicine and their use in livestock for therapy, prophylaxis and growth promotion. The use of veterinary antibiotics has many benefits to the livestock industries ensuring animal health and welfare, but use at subtherapeutic levels also exerts great selective pressure on emergence of resistant bacteria. The possible effect on human health is a problem of current debate. This study involved sampling pig carcasses, pig meat and assessing the level of resistance in zoonotic enteric bacteria of concern to human health. In South Australian pigs, thermophilic Campylobacter species showed widespread resistance (60-100%) to tylosin, erythromycin, lincomycin, ampicillin and tetracycline. No resistance was seen to ciprofloxacin. The enterococci demonstrated little resistance (0-30%) to vancomycin or virginiamycin, but the overall results from the antibiotic sensitivity testing of the enterococci have demonstrated how widespread their resistance has become. Escherichia coli strains showed widespread resistance to tetracycline and moderately common resistance (30-60%) to ampicillin and sulphadiazine. Resistance to more than one antibiotic was common. Pigs from New South Wales were also sampled and differences in resistance patterns were noted, perhaps reflecting different antibiotic use regimens in that state.

Fresh garlic: a possible vehicle for Salmonella Virchow.

A sustained increase in Salmonella enterica serovar Virchow notifications in South Eastern Australia between September 1997 and May 1998 instigated a case-control study and environmental investigations. Cases were defined as having locally acquired culture-confirmed S. Virchow phage-type 8 infection and diarrhoeal disease. Matched controls were selected by progressive digit dialling based on cases' telephone numbers. An exposure and food history questionnaire was administered by telephone. Phage typing and pulse field gel electrophoresis were performed on case and environmental isolates. Thirty-two notifications of S. Virchow infection met the case definition, 37% reported bloody diarrhoea and S. Virchow was isolated from blood in 13% of cases. Twelve patients were admitted to hospital and one died. Fresh garlic (OR 4.1, 95% CI 1.3-12.8) and semi-dried tomatoes (OR 12.6, 95% CI 1.5-103.1) were associated with these cases. The associations remained significant after adjusting for sex and age. S. Virchow (PT 8) was cultured from two brands of semi-dried tomatoes associated with cases in two different states. We provide sufficient evidence for semi-dried tomatoes and fresh garlic to be considered as potential risk foods in future Salmonella outbreak investigations.

Characterization of the CsfC and CsfD proteins involved in the biogenesis of CS5 pili from enterotoxigenic Escherichia coli.

The region required for biosynthesis of CS5 pili consists of six csf genes, with csfA encoding the major subunit. In this study, we describe the characterization of two of the genes constituting the region, csfC and csfD, but also identify the true morphology of the CS5 pilus by high resolution electron microscopy. CsfD was shown to be essential in the initiation of CS5 pilus biogenesis, did not possess any chaperone-like activity for the major subunit, and was an integral minor component of the pilus structure. Studies on CsfD translocation across the outer membrane in Escherichia coli K-12 using a csfA mutant also showed that CsfD is likely to be the first pilin subunit assembled. A specific in-frame deletion in the csfC gene resulted in the complete absence of cell surface CS5 pili and prevented the translocation of CsfA and CsfD pilins across the outer membrane. Specific cell localization studies showed an accumulation of CsfC in the outer membranes of E. coli K-12, while complementation experiments with homologous outer membrane assembly genes from CS1 and CFA/I pili systems were unable to restore assembly of CS5 pili. The CS5 pilus was shown to be a 2 nm flexible fibrillar structure, which adopted a predominantly open helical conformation under the electron microscope.

Comparison of virulence of Legionella longbeachae strains in guinea pigs and U937 macrophage-like cells.

A guinea pig model of experimental legionellosis was established for assessment of virulence of isolates of Legionella longbeachae. The results showed that there were distinct virulence groupings of L. longbeachae serogroup 1 strains based on the severity of disease produced in this model. Statistical analysis of the animal model data suggests that Australian isolates of L. longbeachae may be inherently more virulent than non-Australian strains. Infection studies performed with U937 cells were consistent with the animal model studies and showed that isolates of this species were capable of multiplying within these phagocytic cells. Electron microscopy studies of infected lung tissue were also undertaken to determine the intracellular nature of L. longbeachae serogroup 1 infection. The data showed that phagosomes containing virulent L. longbeachae serogroup 1 appeared bloated, contained cellular debris and had an apparent rim of ribosomes while those containing avirulent L. longbeachae serogroup 1 were compact, clear and smooth.

B-Cell epitope mapping of the VapA protein of Rhodococcus equi: implications for early detection of R. equi disease in foals.

Linear B-cell epitopes of the Rhodococcus equi virulence-associated protein (VapA) were mapped using a synthetic peptide bank in this study. The peptides were screened in an enzyme-linked immunosorbent assay (ELISA) with a total of 70 sera from foals with current R. equi disease (51 sera), as well as from foals that had either recovered from R. equi infection 10 months previously (3 sera) or that had no known history of R. equi disease (16 sera). An epitope with the sequence NLQKDEPNGRA was identified and was universally recognized by all 51 sera from foals with R. equi disease and was not recognized by any of the other sera. There was poor reactivity between all sera and peptides relating to other areas of the VapA protein. It is proposed that an ELISA based upon a defined peptide epitope may be used in an improved serological diagnostic test for R. equi infection in foals.

CS5 pilus biosynthesis genes from enterotoxigenic Escherichia coli O115:H40.

We have sequenced the entire region of DNA required for the biosynthesis of CS5 pili from enterotoxigenic Escherichia coli O115:H40 downstream of the major subunit gene, designated csfA (for coli surface factor five A). Five more open reading frames (ORFs) (csfB, csfC, csfE, csfF, and csfD) which are transcribed in the same direction as the major subunit and are flanked by a number of insertion sequence regions have been identified. T7 polymerase-mediated overexpression of the cloned csf ORFs confirmed protein sizes based on the DNA sequences that encode them. The expression of only the csf region in E. coli K-12 resulted in the hemagglutination of human erythrocytes and the cell surface expression of CS5 pili, suggesting that the cluster contains all necessary information for CS5 pilus biogenesis and function.

Distribution of two hemolytic toxin genes in clinical and environmental isolates of Aeromonas spp.: correlation with virulence in a suckling mouse model.

Previous studies have shown that two hemolytic toxins, HlyA and AerA, contribute to the virulence of Aeromonas hydrophila. A survey was performed to gauge the distribution of hlyA and aerA genes in clinical and environmental Aeromonas isolates. For A. hydrophila, A. veronii biotype sobria and A caviae, 96%, 12% and 35% of strains, respectively, were hlyA positive, whereas, 78%, 97%, 41%, respectively, were aerA positive. All virulent A. hydrophila isolates were hlyA+ aerA+. This genotype was most common in A. hydrophila (75.4%) followed by A. caviae (29.4%) and A. veronii biotype sobria (9.6%). For A. hydrophila, a two-hemolytic toxin model of virulence provides the best prediction of virulence in an animal model.

Sequence-based classification scheme for the genus Legionella targeting the mip gene.

The identification and speciation of strains of Legionella is often difficult, and even the more successful chromatographic classification techniques have struggled to discriminate newly described species. A sequence-based genotypic classification scheme is reported, targeting approximately 700 nucleotide bases of the mip gene and utilizing gene amplification and direct amplicon sequencing. With the exception of Legionella geestiana, for which an amplicon was not produced, the scheme clearly and unambiguously discriminated among the remaining 39 Legionella species and correctly grouped 26 additional serogroup and reference strains within those species. Additionally, the genotypic classification of approximately 150 wild strains from several continents was consistent with their phenotypic classification, with the exception of a few strains where serological cross-reactivity was complex, potentially confusing the latter classification. Strains thought to represent currently uncharacterized species were also found to be genotypically unique. The scheme is technically simple for a laboratory with even basic molecular capabilities and equipment, if access to a sequencing laboratory is available.

Sequence analysis of the mip gene of the soilborne pathogen Legionella longbeachae.

To understand the basis of pathogenesis by Legionella longbeachae serogroup 1, the importance of the Mip protein in this species was examined. Amino-terminal analysis of the purified, cloned L. longbeachae serogroup 1 ATCC 33462 Mip protein confirmed that the cloned gene protein was expressed and processed in an Escherichia coli background. DNA sequence analysis of plasmid pIMVS27, containing the entire L. longbeachae serogroup 1 mip gene, revealed a high degree of homology to the mip gene of Legionella pneumophila serogroup 1, 76% homology at the DNA level and 87% identity at the amino acid level. Primer extension analysis determined that the start site of transcription was the same for both species, with some differences observed for the -10 and -35 promoter regions. Primers designed from the mip gene sequence obtained for L. longbeachae serogroup 1 ATCC 33462 were used to amplify the mip genes from L. longbeachae serogroup 2 ATCC 33484 and an Australian clinical isolate of L. longbeachae serogroup 1 A5H5. The mip gene from A5H5 was 100% identical to the type strain sequence. The serogroup 2 strain of L. longbeachae differed by 2 base pairs in third-codon positions. Allelic exchange mutagenesis was used to generate an isogenic mip mutant in ATCC 33462 and strain A5H5. The ATCC mip mutant was unable to infect a strain of Acanthamoebae sp. both in liquid and in a potting mix coculture system, while the A5H5 mip mutant behaved in a manner siilar to that of L. pneumophila serogroup 1, i.e., it displayed a reduced capacity to infect and multiply within Acanthamoebae. To determine if this mutation resulted in reduced virulence in the guinea pig animal model, the A5H5 mip mutant and its parent strain were assessed for their abilities to establish an infection after aerosol exposure. Unlike the virulent parent strain, the mutant strain did not kill any animals under two different dose regimes. The data indicate that the Mip protein plays an important role in the intracellular life cycle of L. longbeachae serogroup 1 species and is required for full virulence.

Interspecies sequence differences in the Mip protein from the genus Legionella: implications for function and evolutionary relatedness.

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).

Cloning and characterization of two immunophilin-like genes, ilpA and fkpA, on a single 3.9-kilobase fragment of Aeromonas hydrophila genomic DNA.

Antiserum to Aeromonas hydrophila A6 cell envelopes was shown in a previous study (C. Y. F. Wong, G. Mayrhofer, M. W. Heuzenroeder, H. M. Atkinson, D. M. Quinn, and R. L. P. Flower, FEMS Immunol. Med. Microbiol. 15:233-241, 1996) to protect mice against lethal infection by this organism. In this study, colony blot analysis of an A. hydrophila genomic library using antiserum to A. hydrophila A6 cell envelopes revealed a cosmid clone expressing a 30-kDa protein which has not been described previously in aeromonads. The nucleotide sequence of a 3.9-kb fragment derived from this cosmid which expressed the 30-kDa protein revealed two potential open reading frames (ORFs) with homology to known immunophilin proteins. ORF1 encoded a 212-amino-acid protein (molecular mass, 22.4 kDa) with 56% identity to the immunophilin SlyD protein of Escherichia coli. ORF1 was subsequently designated ilpA (immunophilin-like protein). ORF3 encoded a potential gene product of 268 amino acids with a typical signal sequence and a predicted molecular size of 28.7 kDa. The inferred amino acid sequence showed 46% identity with the sequence of the FkpA protein of E. coli and 40% identity with the sequence of the macrophage infectivity potentiator (Mip) protein of Legionella pneumophila. ORF3 was designated fkpA (FK506 binding protein) by analogy with the E. coli FkpA protein. Expression of the FkpA protein was confirmed by Western blot (immunoblot) analysis, which detected a 30-kDa protein, with antiserum to the Mip protein of Legionella longbeachae and a specific antiserum to anA. hydrophila 30-kDa membrane protein. PCR and Southern analysis showed that a DNA sequence encoding FkpA was found in all 178 aeromonads of diverse origins tested. A nonpolar insertion mutation in the fkpA gene did not attenuate virulence in a suckling mouse model nor did it affect the expression of hemolysins or DNase. This suggests that either the fkpA gene is not essential in the virulence of A. hydrophila under these conditions or there are other genes in A. hydrophila coding for proteins with similar functions.

Measurement of virulence of aeromonads using a suckling mouse model of infection.

Virulent and avirulent strains of Aeromonas spp. were identified and virulence quantified using an animal model. Virulence was measured by determining a 50% lethal dose (LD50) 43 h after oral administration of live bacteria. The LD50 of virulent Aeromonas isolates ranged from log10 7.53 (mean) organisms to log10 8.88 (mean). Some isolates were avirulent in this model. Detection of cytotoxic activity in culture supernatants correlated with virulence (Fisher exact test, P = 0.0029). There was no correlation between LD50 and the source of the isolate, beta-haemolysis or lipopolysaccharide (LPS) banding profile on SDS-PAGE. In this animal model, virulence was multifactorial in that: (i) bacterial multiplication in the gut was associated with fatal infection; (ii) the increase in bacterial numbers in the gut of mice administered a lethal dose of bacteria was accompanied by accumulation of fluid; and (iii) there was evidence of extraintestinal spread of infection. Protection of suckling mice by rabbit antiserum to Aeromonas cell envelopes was observed.

Characterization of the small cryptic plasmid, pIMVS1, of Salmonella enterica ser. Typhimurium.

The nucleotide sequence has been determined of a small cryptic plasmid of Salmonella Typhimurium, designated pIMVS1, which correlated with an outbreak of gastroenteritis. This plasmid has a size of 3357 bp with no known functions. It does not encode any protein products but shows homology to several well-characterized plasmids. The replication system with RNAI and RNAII and the replication origin (oriV) is highly similar to that of plasmid p15A. It also has a putative mobilisation origin similar to that of ColE1.

Sequence of a variant Shiga-like toxin type-I operon of Escherichia coli O111:H-.

PCR amplification was used to screen faecal isolates of Escherichia coli from a 12-month-old boy with haemolytic uraemic syndrome for the presence of Shiga-like toxin (SLT)-encoding genes. One isolate, belonging to serotype O111:H-, was positive for SLT-I by this method. UV induction indicated that the strain was lysogenic for a lambdoid bacteriophage, but this did not encode the toxin. Southern hybridization analysis of chromosomal DNA revealed that the SLT-I gene was located on an 8.5-kb EcoRI fragment. SLT-I was further localized to within a 3.0-kb SphI-EcoRI fragment. A separate subclone contained a 3.75-kb HindIII fragment, 1.18 kb of which was common to both. Nucleotide sequence analysis of derivatives of these clones revealed that the SLT-I A subunit gene from E. coli O111:H- differed from the previously published sequences for SLT-I by 5 bp [resulting in two amino acid (aa) changes]. It was more closely related to the gene encoding the A subunit of the Shiga toxin from Shigella dysenteriae type 1, from which it differed by 3 bp (resulting in one aa change). The DNA sequence of the B subunit-encoding gene was identical to that of the other two toxins. The region of DNA upstream from the SLT-I of E. coli O111:H- contained an IS element, as well as a region with strong homology to a portion of the genome of bacteriophage lambda.

Lysogenic phage in Salmonella enterica serovar heidelberg (Salmonella Heidelberg): implications for organism tracing.

A phage typing system using a group of 11 closely related phage (as judged by Southern analysis and restriction fragment length polymorphism analysis) was able to distinguish at least six phage types in Salmonella heidelberg of human and animal origin. Restriction fragment length polymorphism analysis using cosmid probes from S. heidelberg confirmed that most S. heidelberg isolates belong to a single 'clonal' group. Southern analysis using DNA isolated from each of the testing phage group showed that phage types 4, 5 and 6 carry closely related endogenous or lysogenic phage. Induction of a lysogenic phage Hlp-4 (Heidelberg lysogenic phage) from type 4 could become lysogenic and convert phage types 1 and 3 to phage type 4 and phage type 5 to a non-typable phenotype.