Identification of a mutation associated with erythromycin resistance in Bordetella pertussis: implications for surveillance of antimicrobial resistance.

Erythromycin treatment failures and in vitro resistance of Bordetella pertussis have been reported on several occasions in the past few years, but the mechanism of resistance has not been described. One potential mechanism, genetic modification of the erythromycin-binding site on the 23S rRNA of the 50S ribosomal subunit, has been observed in other bacteria. To explore this possibility, we amplified the portion of the 23S rRNA gene encoding the central loop of domain V. DNA sequencing and restriction fragment length polymorphism of the PCR products showed that each of the four erythromycin-resistant B. pertussis strains tested contained an A-to-G transition mutation at position 2058 (Escherichia coli numbering) of the 23S rRNA gene. The mutation was not found in seven erythromycin-susceptible isolates tested. Two of the resistant isolates were heterozygous, containing at least one mutant copy and one wild-type copy of the 23S rRNA gene. These results indicate that erythromycin resistance in these strains is likely due to a mutation of the erythromycin-binding site in the 23S rRNA gene. Identification of the resistance mechanism will facilitate development of molecular susceptibility testing methods that can be used directly on clinical specimens in the absence of an isolate.

Bordetella pertussis isolates with a heterogeneous phenotype for erythromycin resistance.

Erythromycin is currently being used for both prophylaxis and treatment of pertussis infections. Erythromycin resistance was first recognized in Bordetella pertussis in Arizona in 1994, and since then, three additional resistant isolates have been identified in the United States. To better assess the potential public health impact of erythromycin-resistant B. pertussis, we used the disk diffusion assay to evaluate the frequency of erythromycin resistance among 1,030 recently circulating U.S. isolates and found the rate of occurrence to be <1%. We also describe a novel heterogeneous phenotype, with erythromycin-resistant colonies appearing only after a 7-day incubation period. To optimize patient management, we recommend that clinicians be alert to potential treatment failures and that laboratorians use a 7-day incubation period when screening for resistance. Our ongoing national surveillance will continue to monitor for resistant B. pertussis isolates and their potential association with changing pertussis epidemiology.

Molecular epidemiology of Bordetella pertussis by pulsed-field gel electrophoresis profile: Cincinnati, 1989-1996.

Reported cases of pertussis have increased in the United States, with peaks occurring every few years. Bordetella pertussis isolates collected in Cincinnati from 1989 to 1996 were analyzed with pulsed-field gel electrophoresis (PFGE), to evaluate trends. Among 496 isolates, 30 PFGE profiles were identified; 32% were CYXXI-010, the profile that predominated each year. Eighteen profiles (198 strains) were identified in 1989-1992, 20 profiles (197 strains) were identified during the 1993 epidemic, and 11 profiles (101 strains) were identified in 1994-1996. From 1989 to 1996, among 42 patients, isolates from household members in 17 (89%) of 19 households had concordant PFGE profiles. There was no association between PFGE profile and seasonality, age, and hospitalization or pneumonia in infants <1 year old. The 1993 epidemic was associated primarily with an increased prevalence of PFGE profiles that circulated before and after 1993, which suggests that the epidemic was due to factors other than the emergence of a novel B. pertussis strain.

Real-time PCR assay targeting IS481 of Bordetella pertussis and molecular basis for detecting Bordetella holmesii.

Detection of Bordetella holmesii by a real-time PCR assay targeting IS481 of Bordetella pertussis is reported. Sequencing of IS481-specific PCR products from B. pertussis and B. holmesii isolates revealed sequence homology. Restriction fragment length polymorphism demonstrated a low copy number of IS481-like sequences in B. holmesii. These results, and culture of B. holmesii from patients with cough, suggest that the specificity and predictive value of IS481-based PCR assays for pertussis may be compromised.

Molecular epidemiology of Bordetella pertussis by DNA fingerprinting with pulsed-filled gel electrophoresis, 1989-1996

Pertussis is an endemic disease in the United States of America, with epidemics occurring every three to four years. In Cincinnati, Bordetella pertussis isolates collected from 1989 to 1996 were analysed by genomic subtyping with pulsed-field gel electrophoresis (PFGE) to evaluate the B pertussis population before, during and after a large epidemic of epidemiologically relevant changes. Among the 496 B pertussis isolates, 31 PFGE profiles were identified; 32 percent of isolates were CYXXI-010 and this profile predominated in each year. Nineteen, 20 and 12 PFGE profiles were identified in the pre-epidemic period (n=198), during the epidemic (n = 197) and in the post-epidemic period (n = 101), resulting in genotypic diversities of 0.82, 0.83 and 0.76 respectively. From 1989 to 1996, among 19 households clusters of 42 patients, 17 (89 percent) households had concordant PFGE profiles among isolates from household members. There was no association between PFGE type and seasonality, age, hospitalisation or pneumonia in infants. The 1993 epidemic was primarily associated with increased prevalence of B pertussis PFGE profiles that circulated before and after the epidemic, suggesting increased susceptibility to pertussis rather than a novel strain as a cause of the outbreak.(Au)

A Bordetella pertussis fepA homologue required for utilization of exogenous ferric enterobactin.

The bfeA (Bordetella ferric enterobactin) receptor gene was cloned from a Bordetella pertussis chromosomal library by using a screen in Escherichia coli to detect iron-repressed genes encoding exported proteins translationally fused to the E. coli phoA gene. The bfeA gene encoded a protein with a molecular mass of approximately 80 kDa and about 50% amino acid sequence identity to both the fepA- and pfeA-encoded enterobactin receptors of E. coli and Pseudomonas aeruginosa, respectively. Enterobactin prepared from iron-starved E. coli cultures supported growth of B. pertussis and Bordetella bronchiseptica in the presence of the iron chelator ethylenediamine-di-(o-hydroxyphenylacetic acid) (EDDA). Expression of the bfeA gene was induced by low iron availability, and iron-regulated expression appeared to be dependent upon the presence of the sequence contained within 370 bp upstream of the bfeA structural gene. An internal fragment of the bfeA structural gene and flanking regions were shown by Southern analysis to be highly conserved among Bordetella species. Insertional inactivation of bfeA in both B. pertussis and B. bronchiseptica greatly impaired their ability to grow in the presence of enterobactin and EDDA. These findings suggest that enterobactin produced by other respiratory flora could aid in the colonization of the respiratory tract by Bordetella species.

Analysis of Bordetella pertussis isolates from an epidemic by pulsed-field gel electrophoresis.

We examined genetic variation among 78 clinical isolates of Bordetella pertussis, including 54 strains recovered during a 1986 pertussis epidemic. A total of 16 pulsed-field gel electrophoresis (PFGE) profiles, generated with each of three different enzymes (XbaI, SpeI, and DraI), were obtained from the epidemic and sporadic isolates included in the study. Indistinguishable profiles were seen among strains unrelated temporally or geographically, as well as among strains isolated sporadically from the same geographic areas. All isolates from the epidemic had indistinguishable PFGE profiles. The PFGE pattern of the epidemic strains was shared with only 1 of 25 strains isolated independently of the outbreak. This isolate was cultured from a specimen from a laboratory scientist who had been working with the epidemic strains, further implicating the usefulness of PFGE for the epidemiologic study of clinical strains of B. pertussis. Differences in PFGE profiles for single epidemic strains occurred occasionally upon repeated passage on agar medium, suggesting that subculturing of initial isolates should be minimized before pulsed-field analysis.

Cloning and initial characterization of the Bordetella pertussis fur gene.

In several Gram-negative pathogens the fur (ferric uptake regulator) gene product controls the expression of many genes involved in iron uptake and virulence. To facilitate the study of iron-regulated gene expression in Bordetella pertussis, we cloned the fur gene from this organism. The B. pertussis fur gene product was 54% identical to the Escherichia coli Fur and complemented two E. coli fur mutants. As with the E. coli fur gene, sequences upstream of the B. pertussis fur were homologous to the consensus Fur-binding site and to the consensus catabolite activator protein binding site.

Concurrent outbreaks of pertussis and Mycoplasma pneumoniae infection: clinical and epidemiological characteristics of illnesses manifested by cough.

Concurrent outbreaks of illnesses that were manifested by cough and that were suspected to be due to Bordetella pertussis and Mycoplasma pneumoniae infection were investigated in a midwestern town in Illinois. Three studies were conducted: questionnaires on the clinical and epidemiological characteristics of illness were administered to patients; serological tests were performed to confirm the presence of each pathogen and to develop case definitions for each illness; and case definitions were applied to responses to a mail-in questionnaire for estimating the magnitude of both outbreaks. In 135 cases of suspected pertussis and 42 cases of suspected mycoplasmal infection, subjects had a cough for > or = 14 days (the pertussis outbreak case definition). Among 20 laboratory-confirmed cases, a cough for > or = 14 days had a specificity of 20% for pertussis, and a cough for > or = 28 days plus whoop and/or vomiting had a specificity of 90% for pertussis. Six hundred-seventeen pertussis cases per 100,000 population and 1,179 cases of M. pneumoniae infection per 100,000 population occurred. In this setting, the standard outbreak case definition for pertussis lacked adequate specificity to distinguish pertussis from mycoplasmal infection. The magnitude of each outbreak was greater than the number of reported cases suggested.

Viability of Bordetella pertussis in four suspending solutions at three temperatures.

We studied the survival of Bordetella pertussis in four suspending solutions (Casamino Acids broth, deionized water, phosphate-buffered saline, and serum inositol), subjected to three storage temperatures (4, -20, and -70 degrees C) and two freezing methods (direct freezing and fast-freezing in an ethanol-dry-ice bath). Recovery rates were higher for longer periods for suspensions stored at -70 degrees C than those stored at -20 or 4 degrees C. Serum inositol showed the highest recovery rates for all experimental conditions, followed by Casamino Acids, deionized water, and phosphate-buffered saline. Cell viability was significantly reduced in phosphate-buffered saline suspensions fast-frozen before storage. These results identify optimal conditions for storing B. pertussis cells and are applicable to the collection, transport, and storage of aspirated nasopharyngeal samples for use in the laboratory diagnosis of pertussis.

Rapid immunodot technique for identifying Bordetella pertussis.

We developed and evaluated a rapid test with monoclonal antibodies to identify cultures of Bordetella pertussis. Samples of 5 microliters of cells suspended in formalin-saline were dried onto a nitrocellulose disk. The disk was placed in a filtration device, and 5-microliters volumes of murine monoclonal antibody directed against B. pertussis lipooligosaccharide and peroxidase conjugate were added consecutively, with washing after each addition. The disk was removed and immersed in peroxidase substrate solution. All of 66 B. pertussis isolates confirmed by direct fluorescent-antibody assay were correctly identified by using four different monoclonal antibodies. One of the monoclonal antibodies did not react with over 20 bacterial species tested, including other Bordetella, Acinetobacter, Haemophilus, Moraxella, Mycobacterium, Neisseria, and Staphylococcus spp. This technique detected > or = 2 micrograms of lipooligosaccharide per ml or > or = 5 x 10(8) B. pertussis cells per ml. This rapid procedure used small amounts of reagents, needed less equipment, and was less subjective and more specific than the direct fluorescent-antibody assay.

Incubation of water samples containing amoebae improves detection of legionellae by the culture method.

Some protozoans isolated from aquatic habitats, including domestic water supplies, can support the intracellular replication of autochthonous legionellae in vitro. We studied the effect of incubating water samples containing amoebae on the sensitivity of culture for legionellae. Samples collected during investigations of legionellosis epidemics and shown by conventional culture procedures to contain amoebae, but not legionellae, were incubated at 35 degrees C and replated. Legionellae were recovered from 59 of 144 such samples. Species isolated included L. pneumophila, L. anisa, L. bozemanii, L. gormanii, L. micdadei, L. rubrilucens, L. sainthelensi, L. steigerwaltii, and an unnamed species. Acanthamoeba polyphaga, Acanthamoeba hatchetti, a Rosculus sp., Hartmannella vermiformis, and Vahlkampfia spp. were among the autochthonous amoebae identified. Legionellae were recovered by this procedure from only 3 of 63 samples that were negative for amoebae by conventional culture procedures. These results show that water samples negative for legionellae, but positive for amoebae, by standard culture techniques should be incubated and replated to maximize the sensitivity of culture for legionellae.

Characterization of an axenic strain of Hartmannella vermiformis obtained from an investigation of nosocomial legionellosis.

A free-living amoeba identified as Hartmannella vermiformis was isolated from a water sample obtained during an investigation of nosocomial legionellosis. Hartmannella vermiformis is known to support the intracellular multiplication of Legionella pneumophila. This strain of H. vermiformis, designated CDC-19, was cloned and established in axenic culture to develop a model for the study of the pathogenicity of legionellae. Isoenzyme patterns of axenically-cultivated strain CDC-19 were compared with two strains of H. vermiformis derived from the type strain, one axenic (ATCC 50236) and the other grown in the presence of bacteria (ATCC 30966). Enzyme patterns suggested that all three strains are assignable to the species H. vermiformis. Axenic H. vermiformis strain CDC-19 has been deposited with the American Type Culture Collection (ATCC 50237) and should prove useful in the study of protozoan-bacterial interaction.

Virulence of a Legionella anisa strain associated with Pontiac fever: an evaluation using protozoan, cell culture, and guinea pig models.

Legionella anisa and the amoeba Hartmannella vermiformis were isolated from an indoor fountain implicated as the infectious reservoir in an outbreak of Pontiac fever. We evaluated the ability of this strain of L. anisa to multiply in cultures of an amoeba (H. vermiformis), a ciliated protozoan (Tetrahymena pyriformis), and human mononuclear cells and to infect guinea pigs. These bacteria multiplied in the culture of H. vermiformis but failed to infect guinea pigs or the cultures of T. pyriformis and human mononuclear cells. These findings suggest that some Legionella spp. may multiply only in specific protozoan hosts. The inability of this strain of L. anisa to multiply in human phagocytic cells may be related to the development of Pontiac fever rather than pneumonic legionellosis in exposed individuals. Further studies are necessary to determine whether the ability of legionellae to infect certain host cells can be correlated to differences in human disease.

Increased recovery of Legionella micdadei and Legionella bozemanii on buffered charcoal yeast extract agar supplemented with albumin.

The recovery of Legionella micdadei and L. bozemanii serogroups 1 and 2 from infected guinea pig spleens was evaluated by using two culture media: buffered charcoal yeast extract agar with 0.1% alpha-ketoglutarate (BCYE alpha) and the same medium supplemented with 1% bovine serum albumin (ABCYE alpha). At the lowest dilution of spleen tissue (10(-1)), recovery of all strains of L. micdadei and L. bozemanii was more efficient on ABCYE alpha than on BCYE alpha. L. micdadei strains had higher recovery rates on ABCYE alpha after another 10-fold dilution, but recoveries of L. bozemanii were similar on both media. Recovery rates for most test strains were comparable on BCYE alpha and ABCYE alpha at the highest dilution (10(-3)) of tissue tested. The presence of albumin in BCYE alpha increased the recovery rate of L. micdadei more than that of L. bozemanii. The use of ABCYE alpha medium in place of BCYE alpha may improve the recovery of L. micdadei and L. bozemanii from clinical specimens. Preliminary studies indicate that this medium also enhances recovery of certain Legionella spp. from environmental samples.

Effects of transport temperature and medium on recovery of Bordetella pertussis from nasopharyngeal swabs.

We compared relative recoveries of Bordetella pertussis from simulated nasopharyngeal (NP) specimens incubated in three separate transport media at different temperatures. Transport media included one-half-strength Regan-Lowe (RL.5), Regan-Lowe with one-half-strength agar (RL.5A), and buffered charcoal-yeast extract agar supplemented with alpha-ketoglutarate, lincomycin, and anisomycin (BCYE alpha LA). For each transport medium, recovery of B. pertussis was least efficient after storage at 25 degrees C. The highest recovery of B. pertussis from a mixed culture was achieved with RL.5 at 4 degrees C. Overall, RL.5 and RL.5A were comparable as transport media whether held at 4 or 25 degrees C, but fewer organisms were recovered from BCYE alpha LA. In addition, Regan-Lowe (RL), Bordet-Gengou, and cyclodextrin media were compared as primary isolation media for recovering B. pertussis from simulated NP swabs held at 4 and 35 degrees C in RL.5 medium. The highest recovery of B. pertussis was obtained on RL primary isolation medium. Bordet-Gengou medium recovered only 80% and cyclodextrin medium recovered less than 60% of the numbers recovered on RL medium. Based on these results, refrigeration (4 degrees C) of NP swabs shipped in RL.5 transport medium and using RL as the primary isolation medium are recommended for recovering B. pertussis from swab specimens.

Immunoblot analysis of humoral immune responses following infection with Bordetella pertussis or immunization with diphtheria-tetanus-pertussis vaccine.

To help develop better diagnostic tests for pertussis, we examined the serologic response to whole-cell proteins of Bordetella pertussis after natural infection or vaccination with diphtheria-tetanus-pertussis vaccine. Serum specimens collected during a pertussis outbreak investigation and from uninfected persons were used in Western blot (immunoblot) analyses to determine the presence of immunoglobulin G (IgG) and IgA antibodies to specific B. pertussis proteins. IgG antibodies to proteins of molecular masses 220 and 210 kilodaltons (kDa) were detected in 14 of 18 serum samples obtained from patients with culture-confirmed pertussis greater than or equal to 40 days after the onset of coughing. IgA antibodies were detected in 15 of the 18 samples. Of 19 serum samples obtained from patients who had not been ill with pertussis, 6 contained IgG antibodies to these proteins and 1 contained IgA antibodies. The two proteins bound antiserum specific for filamentous hemagglutinin and comigrated with purified filamentous hemagglutinin. IgG antibodies to two additional protein bands of molecular masses 84 and 75 kDa were associated with previous vaccination. Antibody to the 84-kDa protein was detected in 15 of 17 vaccinated, never-infected persons, and antibody to the 75-kDa protein was detected in 16 of the 17. None of 11 nonvaccinated, never-infected persons tested had antibodies to either protein. All seven fully vaccinated persons with culture-documented infection had antibodies to both proteins. Antibodies to the 84-kDa protein were detected in 6 of 22 nonvaccinated and infected persons, and antibodies to the 75-kDa protein were detected in 8 of the 22. Use of Western blot analysis in this study allowed us to distinguish antibody responses to infection and immunization.

Comparison of guinea pig and protozoan models for determining virulence of Legionella species.

Legionella pneumophila organisms are able to infect and multiply within the ciliated protozoan Tetrahymena pyriformis. This ability may be associated with virulence, because an attenuated strain of L. pneumophila fails to multiply within this protozoan, whereas a virulent strain increases 10,000-fold in number when coincubated with T. pyriformis. Seventeen strains (11 species) of legionellae were evaluated for virulence by intraperitoneal injection of guinea pigs and inoculation of protozoan cultures. Analysis of the data indicates that there are four categories of legionellae with respect to virulence as follows: organisms that infect and kill guinea pigs and multiply in T. pyriformis; organisms that infect but do not kill guinea pigs and multiply in T. pyriformis; organisms that do not infect guinea pigs but are lethal at high concentrations and multiply in T. pyriformis; and organisms that neither infect nor kill guinea pigs and fail to multiply in T. pyriformis. Evidence suggests that these distinctions are based on two virulence factors: intracellular multiplication in a host and toxic activity.