INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The study of temperate phage that infect pathogenic bacteria has provided numerous insights into the virulence of the host bacterium. The role of a temperate phage in virulence can be quite direct, in that some temperate phage carry toxin genes required for the host bacterium (lysogen) to cause disease (5, 26, 38, 39). In other cases, lysogeny affects virulence in more subtle ways that have to do with cell surface alterations (2, 3, 28, 34). Mutations to phage resistance often affect virulence properties by altering the cell surface, rendering the bacterium impaired not only in its ability to interact with phage but also with the host (8, 9, 30, 35). In addition to providing insights into the pathogenic process, temperate phage are often found to be generalized transducing phage (31). Such transducing phage provide valuable tools to characterize pathogenic bacteria genetically, especially in those pathogens where the means for genetic exchange via other mechanisms are limited (17, 18).

Bordetella avium causes bordetellosis, an upper respiratory tract disease in birds. Commercially raised turkeys are particularly susceptible (33). The disease involves colonization of the trachea, resulting in the death of ciliated tracheal cells. Experimentally infected birds normally recover after several weeks (36), but naturally infected birds are subject to a variety of secondary infections which cause severe economic losses in all poultry-producing regions of the world (33). The virulence-associated factors of B. avium are not well defined.

In this communication, we report the discovery of a B. avium temperate transducing phage (Ba1). We determined the host range of the phage and its biochemical and physical properties. Also, we characterized the nature of the Ba1 prophage state. In addition, we present evidence that transduction is generalized, having successfully transduced five randomly chosen insertion mutations. Finally, we identify procedures to increase transduction frequency and reduce lysogenization of transductants.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Strains and growth conditions. Bacteria and bacteriophage used in this study are listed in Table 1. B. avium and Bordetella hinzii broth cultures were grown at 37°C with shaking in brain heart infusion (BHI) medium (Difco) as described previously (37). BHI agar plates consisted of BHI broth with 1.5% agar added. Soft agar for preparing top-agar lysates contained 0.7% agar. Bacteria to be phage infected were grown overnight to stationary phase. The preparation of 5-bromo-4-chloro-3-indolyl--D-galactopyranoside (X-Gal) indicator agar and minimal medium for B. avium has been previously described (37). Bordetella pertussis and Bordetella parapertussis were maintained on Bordet-Gengou agar (BBL Microbiology Systems, Cockeysville, Md.) with 15% sheep's blood. Phage sensitivity spot tests with these two strains employed a minimal medium (Stainer-Scholte) described by Hewlett and Wolff (16).

Preparation of Ba1 lysates for phage purification. A portion of an overnight culture of B. avium 197N (typically 0.2 ml containing approximately 10⁹ CFU) was mixed with a portion of a Ba1 lysate in a 15-ml test tube to produce a ratio of approximately 10⁴ PFU: 1 CFU in a total volume of 0.3 ml. After a 20-min adsorption at 37°C, equal volumes (3 ml each) of soft agar and BHI broth were added, and the contents were mixed and poured onto a freshly made BHI agar plate. After a 10- to 11-h incubation period at 37°C, lysates were harvested by scraping the soft agar layer into a 40-ml centrifuge tube, adding approximately 0.2 ml of CHCl₃, and vigorously mixing the contents on a vortex mixer for 30 s. Following a 20-min incubation at room temperature, agar and cell debris were removed by centrifugation at 7,800 × g for 10 min. The resulting supernatant, referred to as crude lysate, had an average titer of approximately 5 × 10⁹ PFU/ml (titering was performed essentially as for lysate preparation except that 3 ml of soft agar overlay was employed with serial dilutions of the lysate in 0.15 N NaCl). The addition of either MgCl₂ or CaCl₂ during adsorption had no effect upon the resulting titer. For further purification, phage were isolated from the crude lysate by centrifugation for 3 h at 31,200 × g. The resulting phage pellet was resuspended in 10 mM Tris-HCl-5 mM MgCl₂ (-dil [10]), and the phage was isolated on a CsCl step gradient.

Phage isolation and nucleic acid extraction and manipulation. An eight-step 17-ml CsCl gradient (41) subjected to centrifugation for 20 h at 116,000 × g in a Beckman SW28 rotor was used to estimate the buoyant density of Ba1 (ca. 1.55 g/ml), and a working gradient was established that consisted of three steps (1.36 g/ml, 1.50 g/ml, and 1.60 g/ml). The three-step gradient produced a well-isolated, readily visible phage band when approximately 10¹¹ PFU were applied to the gradient. Needle aspiration of the band resulted in the recovery of an average of 75% ± 15% of the PFU applied to the gradient. Phage DNA was extracted as indicated for bacteriophage  (24). Chromosomal and plasmid DNA was isolated as described for Escherichia coli (32). Restriction endonuclease digestion and agarose gel electrophoresis were performed as described by Davis et al. (10). DNA probes were isolated from agarose gels and then labeled with digoxigenin (DIG) as described by the manufacturer (Boehringer Mannheim). Transfer of DNA to nitrocellulose was accomplished as directed by Maniatis et al. (23), and bands were detected as directed in the DIG high-prime DNA labeling and detection starter kit II (Boehringer Mannheim).

Electron microscopy. Phage morphology measurements were conducted using CsCl purified phage. For contour length phage genome measurements, dilutions of phage DNA (extracted as directed by Dykstra [12]) were mixed with dilutions of purified pBR322 DNA (used as a size standard). These mixtures were placed on Formvar-coated grids and treated as described by Dykstra (12). Grids were examined with a Philips 410 transmission electron microscope. Contour lengths were determined for 10 full-length phage genomes using a map measure and were compared to the lengths of relaxed forms of pBR322.

Immunological techniques. Antibody to Ba1 was raised by injection of a New Zealand White rabbit with approximately 10¹¹ PFU of CsCl-purified phage in complete Freund's adjuvant. Six monthly booster doses of approximately 10¹⁰ PFU in incomplete Freund's adjuvant were given, and blood was withdrawn biweekly. The neutralization constant (K) value (42) of the antiserum after 2 months was stable at approximately 470 min¹.

Genetic techniques. Coreversion analysis of a motility-negative (Mot) mini-Tn5lacZ insertion mutant (strain G146 [Table 1 and reference 37]) that was phenotypically Lac⁺ on X-Gal plates was accomplished by first patching 20 colonies into a 0.35% soft agar layer on BHI agar plates. This was followed by examination of the plates after 48 h at 31°C for flares of motile revertants surrounding the original site of inoculation. Twenty independently isolated, motile revertants were examined for their Kan and Lac phenotype. All were Kan^s, Lac and Mot⁺.

Statistical analysis. Means, the standard deviation of the means, and the statistical significance of mean differences (using Student's t test) were determined with the aid of the statistical package included with Microsoft Excel, version 4.0. P values of <0.05 were taken as significantly different.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Discovery. Ba1 was discovered along with another B. avium temperate phage (Ba2) in a screen that involved cross-spotting supernatants from four B. avium strains (197N, GOBL271, ATCC 35086, and Wampler [Table 1]) onto lawns of each of the strains. Supernatants from broth cultures of the ATCC 35086 strain and the Wampler strain produced plaques on our laboratory B. avium strain, 197N. The Wampler phage was designated Ba1, and the ATCC 35086 phage, Ba2. Both phage appeared to be temperate, judging from their turbid plaques. Also, both phage were morphologically indistinguishable when viewed in the electron microscope (data not shown). However, lysogens constructed from each phage in strain 197N were heteroimmune and were distinguishable by other criteria (e.g., plaque morphology [our unpublished observations]). Ba1 was the only phage found to transduce via allelic replacement (described in more detail below) and was consequently chosen for further study.

Host range. A spot test, in which at least 10⁸ PFU of Ba1 (in approximately 20 µl) was dispensed onto lawns of Bordetella bronchiseptica, B. pertussis, B. hinzii, and B. parapertussis, was used to determine if they were sensitive to Ba1-mediated lysis. None of the species tested (Table 1) showed any evidence of sensitivity to Ba1. However, six of nine B. avium strains, obtained from clinical cases of bordetellosis (Table 1), were sensitive to Ba1 in the spot test. (Ba1 was subsequently shown to form plaques on these strains with a plaquing efficiency similar to that seen with strain 197N.) Of the three resistant strains, one (Ba177) appeared to be lysogenic for a bacteriophage morphologically similar to Ba1, raising the possibility that resistance in this strain was due to immunity or to superinfection exclusion (2). However, no experiments were performed to confirm this possibility.

Morphology and genome length of CsCl-purified phage. Measurements of 10 virions negatively stained and examined via transmission electron microscopy produced the following morphological measurements: (i) head size, measured parallel and perpendicular to the tail, was 55 ± 4 nm and 53 ± 2 nm, respectively; (ii) extended tail length was 85 ± 5 nm; and (iii) extended tail width was 14 ± 2 nm. Electron microscopy further revealed that the phage had a contractile, sheathed tail and faintly visible short tail fibers (Fig. 1A). The length of the double-stranded, linear DNA phage genome was determined by contour length measurements using pBR322 (in its relaxed circular form) as a size standard (Fig. 1B). Measurements of 10 phage genomes produced a length corresponding to a size of 46.5 ± 0 kb. This size is similar to size estimates based on electrophoretic migration of phage DNA digested with three different restriction endonucleases (XhoI, HindIII and EcoRI), which gave an average size of 48.6 ± 2.0 kb (data not shown).

View larger version (138K): [in this window] [in a new window]   FIG. 1.   (A) Transmission electron micrograph of negatively stained Ba1 phage particles. One phage (arrowhead) appears to have an empty head. (Note its electron-darkened head and contracted tail sheath exposing the inner tail core.) (B) Transmission electron micrograph of Ba1 DNA prepared from purified phage as described in Materials and Methods. Arrowheads denote a single linear genome. The relaxed form of pBR322 DNA (4.36 kb) was used as a size standard (box).

Ba1 prophage state. Ba1 lysogeny was stable in the absence of external reinfection, as demonstrated by the growth of a 197N lysogen for over 25 generations in the presence of Ba1 antiserum sufficient to eliminate all detectable external phage (K = 47 min¹). Cultures were maintained in logarithmic growth by periodic dilution. Examination of 100 isolated colonies at the beginning and at the end of the experiment for Ba1 production and Ba1 immunity revealed that all 200 colonies examined were lysogenic. From the foregoing, we concluded that Ba1 formed a true lysogen rather than existing as a pseudolysogen (produced as a result of a persistent infection and exemplified by certain filamentous phages [2]).

View larger version (56K): [in this window] [in a new window]   FIG. 2.   (A) An autoradiograph of NotI-digested chromosome and phage DNA that was hybridized with DIG-labeled vegetative Ba1 DNA. Lanes: 1, nonlysogenic strain (197N); 2, lysogenic strain (AP21); 3, vegetative Ba1. The numbers to the left denote the fragment sizes in kilobases of the bands lettered on the right. The letters (a to e) identify specific vegetative phage fragments. The c' in the lysogen lane (2) denotes a new fragment in the lysogen not present in the vegetative phage. (B) An autoradiograph of NotI-digested chromosomal and phage DNA hybridized with a DIG-labeled EcoRI restriction endonuclease fragment internal to the vegetative Ba1 NotI c fragment. Lanes: 1, the nonlysogenic strain (197N); 2, lysogenic strain (AP21); 3, vegetative Ba1. The numbers to the left denote the molecular weights of the putative attachment site-containing bands (c in the vegetative phage, and c' and c" in the lysogen) in kilobases.

Ba1-mediated transduction of markers at five randomly chosen loci. In order to test whether Ba1 was a transducing phage, we first established the association of a motility mutation with a transposon insertion using a coreversion test (Materials and Methods). We then used the Mot Kan^r Lac⁺ mutant as a donor in transduction experiments. The recipient was a spontaneous streptomycin-resistant (Str^r) mutant of strain 197N. Selecting for Kan^r, we readily found Kan^r transductants, all having the phenotype Str^r Kan^r Mot Lac⁺, as opposed to the donor phenotype, Str^s Kan^r Mot Lac⁺, or the recipient phenotype, Str^r Kan^s Mot⁺ Lac. Interestingly, the Ba2 phage, when tested in this manner, transduced the antibiotic resistance marker, but approximately half of the transductants were still Mot⁺. As stated earlier, we have not characterized Ba2 further. Subsequent transduction experiments using Ba1, a hemagglutination-defective (Hag) Lac⁺ mutant, and three additional mutants with insertions at undefined locations supported the generalized nature of the transduction in that all five markers tested were transduced, albeit some at different (i.e., statistically distinguishable) frequencies than others (Table 2).

View larger version (18K): [in this window] [in a new window]   FIG. 3.   Autoradiograph showing EcoRI restriction endonuclease fragments of donor and recipient strains in transduction experiments using a neoR gene probe as described in Materials and Methods. Lanes: 1, mot::mini-Tn5lacZ donor strain (G146); 2, recipient strain (197N2); 3, mot::mini-Tn5lacZ transductant; 4, hag::mini-Tn5lacZ donor strain (G145); 5, recipient strain (197N2); 6, hag::mini-Tn5lacZ transductant. Arrowheads denote the size of the reacting bands.

UV irradiation of transducing lysates. One complication of Ba1 transduction was that transductants were frequently lysogenized. Ostensibly, this was because of the relatively high number of phage needed to get a successful transduction. Ba1 antiserum, employed during the transduction, was not effective at improving transduction frequency and eliminating lysogenization (our unpublished results). However, UV irradiation of Ba1, so as to drastically reduce the number of viable phage, not only reduced lysogen formation but also greatly increased transduction frequency (Fig. 4). The best transduction frequency with the fewest lysogens was found after exposure of Ba1 phage to approximately 4.2 × 10⁴ ergs/mm², resulting in approximately 99.9% killing of vegetative phage. A similar level of UV irradiation of a hydroxylamine-derived clear plaque Ba1 mutant (Ba1c1) was extremely effective at improving transduction frequency, and transductant lysogeny was eliminated (data not shown). However, unirradiated Ba1c1 lysates did not transduce at measurable frequencies.

View larger version (22K): [in this window] [in a new window]   FIG. 4.   Effect of UV irradiation on transduction parameters: viable phage, transduction frequency, and transductant lysogeny. Solid squares () denote the percentage of viable Ba1 after UV irradiation of a crude lysate grown on strain G146. Also plotted are the percentage of mot::mini-Tn5lacZ nonlysogenized transductants () and the number of transductants emerging from the irradiated lysate at each UV dose (). Points represent the averages of two experiments. Vertical bars denote standard errors of the means.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In this report we describe the discovery and characterization of Ba1, a temperate, transducing phage for B. avium. Ba1 was isolated along with one other temperate but distinctly different B. avium phage, Ba2. Ba1 was chosen for further study because it was capable of transduction via an allelic replacement mechanism. To our knowledge, Ba1 and Ba2 are the first reported phage discovered for this species of Bordetella.

Other members of the Bordetella genus are known to be infected by temperate phage (20, 21, 29). Both B. pertussis and B. bronchiseptica have temperate phage associated with them, and one phage for B. bronchiseptica has been reported to be a temperate transducing phage (M. Liu and J. F. Miller, Abstr. 97th Gen. Meet. Am. Soc. Microbiol., abstr. B-321, 1997). However, none of the Bordetella phage have been used extensively as genetic tools, nor has an examination of the effect of lysogeny or phage resistance upon virulence been investigated.

Ba1 had an icosahedral head, a sheathed, contractile tail, double-stranded DNA, and a linear genome of a remarkably constant length. Ba1 most closely resembled members of the Myoviridae family morphologically (1). The stability of the prophage state was indicated by using Ba1-specific antiserum to show that lysogeny was maintained in the absence of external reinfection. Results of DNA hybridization experiments with four independently isolated 197N lysogens were consistent with phage circularization and integration at a preferred site in both phage and host replicons (7).

The most useful feature of Ba1 was its ability to transduce genetic markers. This property should allow the transfer of mutations between any Ba1-susceptible B. avium strains. Using nine random clinical B. avium isolates, we found that 67% were susceptible, but none of the other Bordetella species were lysed by Ba1. With regard to transduction frequency, some genetic markers were transduced at frequencies significantly different from those of others. A contributing factor to heterogeneity in marker-based transduction frequencies has been traced to features of the DNA surrounding the marker being transduced (e.g., chi sites [27]). Such features may be a contributing factor here. Also, we do not know the distance between the five transduced markers. We have assumed, since all five markers were chosen at random and all five were transduced, that the transduction is generalized. However, should the markers all turn out to be (coincidentally) closely linked, this conclusion would need to be reevaluated.

As with other transducing phage, transduction frequency was dramatically improved by UV irradiation of the transducing lysate (13). UV treatment reduces the number of viable phage, which are responsible for killing or lysogenizing transductants. With regard to the problem of transductant lysogenization, our Ba1 clear-plaque mutant (Ba1c1), if UV irradiated, eliminated the problem of transductant lysogeny. In addition to the ability of Ba1 to transduce, other properties associated with its interaction with B. avium (e.g., lysogeny or mutations to Ba1 resistance) may prove to be instrumental in better understanding B. avium pathogenesis.