INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Atrophic rhinitis (AR) is a costly and widely prevalent respiratory disease of swine. Severe, progressive AR results from concurrent infection with both toxigenic Pasteurella multocida and Bordetella bronchiseptica. Toxigenic P. multocida alone may also cause severe disease, while infection with B. bronchiseptica alone generally results in a moderate to mild reversible form. P. multocida and B. bronchiseptica may also cause bronchopneumonia in swine.

Traditionally, a definitive diagnosis of infection with P. multocida or B. bronchiseptica is established from clinical signs and isolation of these agents from nasal swabs or biopsy specimens. Suspect colonies are subcultured and subjected to biochemical testing. P. multocida isolates must subsequently be tested for toxin production, since nontoxigenic strains are not believed to play a role in AR (3). Conventional identification methods suffer from a lack of sensitivity, since both B. bronchiseptica and P. multocida are often found in low numbers compared to the numbers of other bacteria found in clinical specimens. Additionally, since B. bronchiseptica grows more slowly than most other bacteria commonly found in the upper respiratory tract, its presence may be masked by overgrowth of other organisms.

Previously, we reported on a simple, nonradioactive colony hybridization assay for detection of B. bronchiseptica from primary culture plates (19). However, an assay that could detect both toxigenic P. multocida and B. bronchiseptica would be of great benefit to diagnostic laboratories and would facilitate experimental studies. Here we describe a two-color, nonradioactive colony lift-hybridization assay for simultaneous detection of these pathogens and demonstrate its utility with primary isolation plates derived from clinical samples.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strains and growth conditions. The toxigenic strain P. multocida P-4533 (serotype D:3) has been described previously (20). Strain P-1059 (serotype A:3) was determined to be nontoxigenic with a monoclonal antibody specific for the P. multocida dermonecrotic toxin in a colony blot assay, as reported previously (13). This assay was also used to determine the toxin-producing abilities of the 60 field strains from swine (provided by Richard B. Rimler, National Animal Disease Center, Agricultural Research Service [ARS], U.S. Department of Agriculture [USDA], Ames, Iowa) included in Table 1. The capsular serogroup of these strains was determined by indirect hemagglutination (20). Somatic serotypes were determined by gel diffusion precipitin tests (7). B. bronchiseptica KM22 was obtained from a swine herd with AR and has frequently been used as a virulent challenge strain in vaccine studies (12). Except when indicated otherwise, P. multocida was grown on dextrose starch agar and B. bronchiseptica was grown on Bordet-Gengou agar with 10% sheep blood at 37°C for 18 to 36 h.

Probes. A fragment of the P. multocida toxA gene consisting of nucleotides 1983 through 3183 was generated by PCR as reported elsewhere (14) and was cloned into the plasmid vector pCRII (Invitrogen, San Diego, Calif.). This plasmid, designated pPMT, was introduced into E. coli INVF' by electroporation.

Colony lifts and hybridization. Nasal swabs obtained from swine with clinical signs of AR were used to streak sheep blood agar plates. Colony lifts were prepared from plates up to 3 weeks old using positively charged nylon membranes (Boehringer Mannheim, Indianapolis, Ind.), according to the manufacturer's recommendations. Nucleic acid was fixed to the membranes by baking at 80°C for 2 h. The membranes were stored in sealed plastic bags at room temperature for up to 2 months prior to hybridization. Cellular debris was removed by incubation of the membranes in 3× standard saline citrate (1× standard saline citrate contains 150 mM NaCl plus 15 mM sodium citrate [pH 7.0])-0.1% (wt/vol) sodium dodecyl sulfate for 1 to 3 h at 68°C, followed by gentle wiping of the membrane surface with a moistened laboratory tissue. Prehybridization and hybridization were carried out at 42°C by a standard protocol (2). Hybridization solutions contained either 15 to 45 ng each of fluorescein-ToxA and digoxigenin-AlcA (for multicolor detection) per ml or 15 ng of ToxA or AlcA alone (for standard colorimetric or chemiluminescent detection) per ml in 5× standard saline citrate-50% formamide-0.02% sodium dodecyl sulfate-N-lauroylsarcosine-2% Genius blocking reagent-20 mM sodium maleate. Posthybridization washes were performed as described previously (2).

Detection of probes. The Genius Multicolor Detection kit (Boehringer Mannheim) was used for development of colony lift membranes hybridized simultaneously with a mixture of fluorescein-ToxA and digoxigenin-AlcA. The protocol supplied by the manufacturer was followed. Briefly, after posthybridization washes, probe-target DNA hybrids were fixed to the membranes by exposure to UV light (254 nm) for 3 min. The membranes were incubated in 1% Genius blocking reagent for 30 min, followed by an additional 30-min incubation in a 1:5,000 dilution of anti-fluorescein-alkaline phosphatase (anti-fluorescein-AP) for detection of fluorescein-ToxA. After washing, the membranes were incubated in freshly prepared "red" AP substrate solution (containing naphthol-AS-phosphate and fast red TR). They were allowed to develop, without agitation and with protection from light, for 15 to 45 min. Residual AP was inactivated by incubation of the membranes in 50 mM EDTA (pH 8.0) for 10 min at 85°C. This temperature is expected to melt DNA hybrids; however, the previous UV cross-linking step prevents bound but undeveloped digoxigenin-AlcA from being washed away. A second round of detection identical to the first one was carried out, except that anti-digoxigenin-AP and "blue" substrate solution (containing naphthol-AS-phosphate and fast blue B) were used. Following development of digoxigenin-AlcA, membranes were washed briefly in 10 mM Tris-1 mM EDTA (pH 8.0), allowed to air dry, and stored at room temperature. In one series of experiments "green" substrate reagent (containing naphthol-AS-Gr-phosphate and fast blue B), also included in the Genius Multicolor Detection Kit, was substituted for the red or blue substrate.

Conventional identification of B. bronchiseptica and P. multocida. Modified MacConkey/Bordetella agar plates (MicroBiologics, St. Cloud, Minn.) were streaked with nasal swabs obtained from swine with clinical signs of AR and were incubated at 37°C for 48 h. Nonfermentative colonies were subcultured on sheep blood agar and were subsequently used for biochemical testing. Isolates positive for oxidase, citrate, urea, and growth in 6.5% NaCl, negative for indole, and causing no change or an alkaline reaction on triple sugar iron slants were identified as B. bronchiseptica (17).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Optimal assay parameters. Preliminary experiments were conducted with colony lifts from sheep blood agar plates streaked with a mixture of toxigenic P. multocida (strain 4533), nontoxigenic P. multocida (strain 1059), and B. bronchiseptica. Membranes were subjected to simultaneous hybridization with fluorescein-labeled ToxA and digoxigenin-labeled AlcA and were developed by the multicolor detection procedure. As expected, development of fluorescein-ToxA with the red substrate reagent produced a pink hybridization signal only from the toxigenic isolate of P. multocida, and development of digoxigenin-AlcA with the blue substrate reagent produced a purple hybridization signal only from B. bronchiseptica (Fig. 1). Hybridization signals were usually visible within 20 min. However, the colors continued to intensify for at least another 15 min. Therefore, 45 min was chosen as the standard development time. Use of both probes, each at a concentration of 30 ng/ml, resulted in strong color development in the absence of background. Increasing the probe concentration further did not increase the background, but neither did it increase the intensities of the signals or reduce the time required for development. Probes were used at 30 ng/ml each for the remainder of the study. Identical results were obtained when the order of detection was reversed. The color of the hybridization signals from toxigenic P. multocida and B. bronchiseptica was reversed when the AlcA probe was developed with the red reagent and the ToxA probe was developed with the blue reagent.

View larger version (115K): [in this window] [in a new window]   FIG. 1.   Colony lift from a plate streaked with toxigenic P. multocida (A), nontoxigenic P. multocida (B), and B. bronchiseptica (C), hybridized with 30 ng each of fluorescein-ToxA and digoxigenin-AlcA per ml, and developed by the multicolor detection procedure.

Specificity of ToxA. The ToxA probe used in the present study is derived from the central portion of the toxA gene. One previous investigation in which a PCR assay for identification of toxigenic P. multocida was described suggested that E. coli and S. typhimurium each contain sequence homologous to this region (8). Other reports failed to confirm this observation (9, 11, 14). To further establish the specificity of the ToxA probe used here, colony lifts of six swine isolates of E. coli and single isolates of S. typhimurium, S. choleraesuis, S. anatum, S. heidelberg, S. infantis, S. montevideo, and S. derby were hybridized with digoxigenin-ToxA. Negative reactions were obtained with all strains, both from membranes developed by the standard colorimetric procedure and from duplicate membranes developed with the more sensitive chemiluminescent substrate CSPD (Fig. 2). Identical results were observed when ToxA labeled with fluorescein was substituted for digoxigenin-ToxA.

View larger version (119K): [in this window] [in a new window]   FIG. 2.   Colony lift from a plate streaked with swine isolates of E. coli (locations 1 to 6) or S. typhimurium (location 7), S. choleraesuis (location 8), S. anatum (location 9), S. heidelberg (location 10), S. infantis (location 11), S. montevideo (location 12), and S. derby (location 13); hybridized with digoxigenin-ToxA; and developed with a chemiluminescent substrate. The location of the toxigenic P. multocida strain used as a positive control is indicated with an asterisk.

Comparison of multicolor detection with standard identification methods. Nonselective primary isolation plates obtained from nasal swabs of swine with clinical signs of AR were used to compare the sensitivity and specificity of the multicolor detection procedure with those of standard identification methods based on a combination of colony morphology, biochemical testing, mouse lethality, and ELISA. The results for a total of 84 specimens derived from 15 herds are summarized in Table 2. Concordant results were obtained for 78 specimens, 59 of which were negative for both toxigenic P. multocida and B. bronchiseptica. Of the remaining samples with concordant results, 10 tested positive for toxigenic P. multocida and 9 tested positive for B. bronchiseptica.

View larger version (36K): [in this window] [in a new window]   FIG. 3.   Colony lifts from a primary isolation plate reported to be negative for toxigenic P. multocida by conventional methods. Membranes were hybridized with 30 ng each of fluorescein-ToxA and digoxigenin-AlcA per ml and developed by the multicolor detection procedure (A) or subjected to a colony blot assay with a monoclonal antibody specific for the P. multocida dermonecrotic toxin (B).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

A previous report from this laboratory described the use of a Bordetella-specific probe in a highly sensitive colony lift-hybridization assay for identification of B. bronchiseptica infection in swine (19). This procedure is faster, more objective, and better suited to the analysis of large numbers of samples than standard identification methods based on colony morphology and biochemical testing. Another advantage is the ability to use primary isolation plates, without the need for pure cultures of suspect colonies. However, information regarding the presence of both toxigenic P. multocida and B. bronchiseptica is required during clinical evaluation of swine herds for AR. Using a commercially available multicolor development kit, we have modified the previously reported colony lift assay such that both toxigenic P. multocida and B. bronchiseptica can be identified from the same colony lift membrane.

The sensitivity and specificity of the AlcA probe for identification of B. bronchiseptica have previously been demonstrated with colony lifts derived from diagnostic samples (19). However, the membranes used in that study were developed with a chemiluminescent substrate, which is approximately 10 times more sensitive than multicolor detection (2). Nevertheless, in the present study all samples reported to contain B. bronchiseptica by conventional methods were also positive by the colony lift assay. Additionally, we identified B. bronchiseptica from five samples that were reported to be negative on the basis of standard methods. Visual examination did not reveal isolated colonies with a morphology typical of that of B. bronchiseptica on the primary isolation plates derived from these samples. This omission is explained by the fact that only a few AlcA-positive colonies were present in an area of nearly confluent growth and illustrates the limitations of conventional methods for identification of B. bronchiseptica. We conclude that the colony lift-hybridization assay described here is more sensitive than standard methods since it permits the detection of even a few colonies whose presence and morphology are masked by overgrowth of other bacteria in the sample. Similar findings were reported in a previous study in which 6 of 77 clinical specimens reported to be negative by conventional identification methods were subsequently shown to contain colonies of B. bronchiseptica in an area of confluent growth (19).

All samples from which toxigenic P. multocida was identified by conventional methods were also positive by the colony lift assay with the ToxA probe, indicating that this technique is at least as sensitive as conventional identification methods. One sample reported as containing nontoxigenic P. multocida by standard methods was found to contain a few colonies positive by the colony lift assay with ToxA. As described above, the positive colonies were located in an area of the plate with nearly confluent growth. Identical results were obtained from a repeat colony lift developed with a chemiluminescent substrate. Furthermore, production of the dermonecrotic toxin by the ToxA-positive colonies was demonstrated with an antitoxin monoclonal antibody. The age of the plate and the excessive contamination that occurred during storage prevented us from isolating toxigenic P. multocida. However, as was previously shown for B. bronchiseptica with the AlcA probe (19), some samples testing negative by conventional methods may contain low numbers of toxigenic P. multocida that are identified only in the colony lift assay. Since the tonsil, rather than the turbinate, is the anatomic site colonized most effectively by P. multocida (1, 3), low numbers of this organism in nasal swabs are not unexpected. Additionally, it is known that swine may be colonized simultaneously with a mixture of toxigenic and nontoxigenic P. multocida (1, 3). Since a single colony was selected for the subculture used for determination of toxin production in this study, false-negative results could occur with samples from swine having mixed infection.

Our results also indicate that the ToxA probe is highly specific. Although the plates used for colony lifts contained large numbers of bacteria other than P. multocida, no false-positive results were obtained for the 74 plates reported to be negative for toxigenic P. multocida by conventional methods. Previous studies with probes or PCR primers that overlap the sequence of ToxA likewise demonstrated no cross hybridization with approximately 20 other bacterial species, including E. coli and S. typhimurium (9, 11, 14). In contrast, one additional study did suggest that E. coli and S. typhimurium may each have sequence homologous to a region that is included in the ToxA probe, since PCR products were obtained from both organisms with primers specific for a portion of this region (8). However, in the study presented here, we found no evidence that ToxA cross hybridizes to E. coli or Salmonella spp. isolated from swine.

The specificity of ToxA for toxigenic P. multocida further depends upon the assumption that nontoxigenic strains lack the toxA gene. Most available evidence is consistent with this assumption (4, 5, 9, 10, 11, 14, 15). However, a few reports present conflicting data suggesting that silent or incomplete copies of toxA may occasionally be found in nontoxigenic isolates (8, 9, 11). Our investigation did not detect sequence homologous to ToxA in the 38 diagnostic samples demonstrated to contain nontoxigenic P. multocida or in 43 nontoxigenic strains isolated from locations around the world. Thus, we conclude that the ToxA probe is highly specific for only toxigenic strains of P. multocida.

The assay described in this report includes probes labeled with digoxigenin or fluorescein. Biotinylated probes are also commonly used by many laboratories for nonradioactive detection of nucleic acids and are suggested as another alternative for use with the Genius Multicolor Detection kit. However, we previously found that biotinylated probes are not suitable for identification of toxigenic P. multocida since their use results in false-positive signals from nontoxigenic strains (18). It was further demonstrated that the false-positive results are due to residual avidin-binding proteins in the P. multocida isolates present on the colony lift membrane. These proteins, which occur in both P. multocida and Pasteurella haemolytica, bind the streptavidin-AP conjugate used for development of biotinylated probes. Thus, biotinylated probes should not be substituted for digoxigenin- or fluorescein-labeled probes in this assay or in any colony lift procedure performed with samples suspected of containing P. multocida.

The multicolor detection kit utilized in the present investigation permits a single colony lift to be used for rapid, sequential detection of ToxA and AlcA. However, the multicolor reagents are considerably more expensive than those required for either conventional colorimetric detection or chemiluminescent detection. In addition, we found the green substrate reagent included in the kit to be unsuitable for the assay, as reported here. Since the multicolor reagents are not available individually, users must discard one-third of the purchased kit. Therefore, some laboratories may wish to consider other options for the development of membranes hybridized with ToxA and AlcA. Either standard colorimetric detection or chemiluminescent detection can be performed with duplicate colony lifts, each hybridized to a separate probe. If only a single colony lift is available, chemiluminescent detection is the only viable alternative. In that case, membranes must be sequentially hybridized with each of the probes and stripped between the two detection procedures. This option has a more extended turnaround time than either multicolor detection or standard colorimetric detection with duplicate membranes, since overnight hybridization with ToxA and AlcA must be performed sequentially rather than simultaneously. The best alternative for development of membranes hybridized to ToxA and AlcA will depend on the priorities of the laboratory performing the assay.

In summary, we report here a colony lift-hybridization assay that includes highly sensitive and specific probes useful for identification of B. bronchiseptica and toxigenic P. multocida from primary isolation plates. Multicolor detection provides rapid and definitive results from a single colony lift, although other options for development may be preferred depending on the needs of the laboratory conducting the assay. The colony lift assay provides faster, more objective results than conventional identification of these pathogens and should prove useful in both diagnostic and research settings.