INTRODUCTION

Top Abstract Introduction Materials and Methods Results AND Discussion References

Salmonella spp., in particular Salmonella typhimurium, play a dominant role in food poisoning. Outbreaks of salmonellosis are observed at an increasing rate. In order to trace infections to the manifold possible sources, strain identification is of great importance. Many systems have been developed that are based on the electromobility of enzymes, on the immunological properties of surface antigens, etc. The method of phage typing has the longest tradition. The first system (designated "Scheme 1") was established in 1943 (8); later it was extended ("Scheme 2") (5). Anderson et al. (1) refined it to a form which still is one of the main identification systems used worldwide and which is known as the Anderson phage typing scheme.

The basis of this typing system is a collection of Salmonella phages which were propagated on particular hosts of the species S. typhimurium. All Anderson typing phages used worldwide today are aliquots of phage stocks prepared decades ago. The Salmonella strains used for phage propagation are also used as reference typing strains. By propagation on these hosts, the phages gained new plating properties. This process has been called "modification," a term not necessarily describing what is understood in the sense of "modification-restriction." The genealogy of various typing phages, i.e., which phage results from the infection of a particular propagating strain with a defined "starting phage," is described elsewhere (5). A graphical presentation of the pedigrees for most phages is shown in Fig. 1. All starting phages of the different lineages trace back to supernatants of Salmonella cultures, in most cases propagating strains (all of the phages with the prefix PBA in Fig. 1, followed by the number of the propagating strain according to Callow [5]), resulting thus from the spontaneous induction of residing prophages and the subsequent release of the temperate phages. Callow has noted that "the great majority of S. typhimurium are lysogenic." This has been confirmed by our own observations (12, 13); moreover, we have shown that most of the phages released by various Salmonella isolates are closely related to the well-known phage P22 and therefore belong to the P22 branch of the lambdoid phages. The genomes of this phage group represent similar arrangements of modules, i.e., DNA segments coding for characteristic functions, which can be exchanged among each other, leading thus to high genetic variability (15).

View larger version (15K): [in this window] [in a new window]   FIG. 1.   Genealogy of Anderson typing phages. *, Phage A9 not available. Since A10 is unlike A5, although BA10 has no P22 sequences, A9 should already have undergone the change and should be identical to A10. ?, Propagating strain or infecting phage not available, (therefore no information about homologous sequences). Underlined terms indicate the phage has an EcoRI restriction pattern that is different from the infecting parental phage; nonunderlined terms indicate no change in the restriction pattern compared to the immediate ancestor; e.g., A8 (not underlined) has the same restriction pattern as A18, but A17 (underlined), which was propagated in BA17 with prophage sequences strongly homologous to P22/ES18, has a pattern different from A18. Thin, medium, and thick arrows indicate no homology, weak homology, or strong homology to P22/ES18, respectively.

Consequently, it was of interest to study the Anderson typing phages and their propagating strains in view of the prophage content of the latter and of the relationships among the Anderson phages and the possible prophages residing in their propagating strains. This study should elucidate the nature of the "modifications" altering the plating properties of the various typing phages.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results AND Discussion References

Bacterial strains. Propagating strains of the typing phages are designated by numbers corresponding to the propagating strain number given previously (Table 4, column 3, of reference 5) with the prefix "BA." These strains were kindly provided by L. R. Ward, Colindale, United Kingdom, via W. Rabsch, Wernigerode, Germany. Other strains were as follows: DB21, a wild-type strain of S. typhimurium LT2 positive for all three known modification-restriction systems (LT, SA, and SB [7]); DB21(sie1), DB21 which is lysogenic for P22 sieA sieB ts12.1 ts2.1, a P22 mutant defective in both superinfection exclusion systems (19); DB21(ES18), DB21 lysogenic for phage ES18; and LB5000 and LB5010, derivatives of LT2, defective in all three restriction functions but active in the corresponding modifications (4). LB5010 is a rough mutant. HisHB22(P22), harboring a deletion in the histidine operon, and Trp-8(P22), both of which are P22 lysogenic auxotrophic mutants of S. typhimurium LT2, were used as recipients in transduction experiments.

Phages. Anderson typing phages were designated with the number according to column 1 of Table 4 in reference 5 with the prefix "A"; S. paratyphi phage 3b, being the direct or indirect ancestor of many typing phages, was also included in the study. All phages were kindly provided by L. R. Ward via W. Rabsch. Original lysates used in the typing assays were plated in suitable dilutions on DB21, and lysates were prepared on the same strain by a single plaque infection. Salmonella phages P22H5, a c2-mutant, and ES18h1, which is related to P22 (14), were also used. ES18 also infects, in contrast to P22, rough strains (11). Phages released from the Anderson propagating strains (i.e., the BA strains) are designated by the name of their host strain with a preceding "P"; e.g., PBA22 is the phage released from strain BA22, the propagating strain for Anderson phage A22.

Media. Luria-Bertani (LB) medium was used for bacterial growth and phage propagation. M9 agar (16) was used for selection of the prototrophic transductants.

Transductions. Transductions were performed as described elsewhere (10).

DNA preparation and hybridization. Phage DNA was isolated from liquid phage lysates as described previously (2). DNA was digested with the restriction enzyme EcoRI (Boehringer Mannheim) as recommended by the supplier. DNA fragments were separated by agarose gel electrophoresis at 10 V/cm¹ in 0.8% (wt/vol) agarose gels (UltraPure; Gibco BRL) prepared with TBE buffer (89 mM Tris base, 89 mM boric acid). The DNA fragments were transferred to Biodyne A transfer membranes (Pall) by Southern blotting (17) and cross-linked to the membrane by exposure to UV light (254 nm, 120 mJ, 30 s) in a Stratalinker 1800 (Stratagene). DNAs from phages P22 and ES18, respectively, were used as probes and labelled with digoxigenin-11-dUTP by using the Random Primed Labelling Kit (Boehringer Mannheim). Prehybridizations, washes, and detection of hybridization signals were performed as recommended by Boehringer Mannheim. The hybridization temperature was 42°C.

	RESULTS AND Discussion

Top Abstract Introduction Materials and Methods Results AND Discussion References

Growth of typing phages on DB21. The original lysates of the Anderson phages used for typing assays and the lysate of phage 3b as obtained from L. Ward were prepared on specific propagating strains as described previously (5). In order to keep host influences as low as possible, they were tested for their ability to plate on our standard phage indicator DB21. All phages formed plaques; however, the efficiencies of plating varied to a great extent, possibly due to the restriction systems active in the LT2 strain DB21. This is supported by the observation that especially phage A8 yielded very few plaques on DB21. However, after isolation of a plaque from DB21, high yields of A8 could be obtained on this strain. Most of the phages formed plaques of similar size to those produced by P22, only phage A8 produced very small plaques. Although the plaques of several phage strains appeared rather clear, they undoubtedly were plaques of temperate phages because all of them could be traced back to prophages. Phage lysate A11 was a mixture of two different phages distinguishable by plaque morphology (clear, turbid). The different types were propagated separately (Table 1). Study of the DNA restriction patterns confirmed that we were dealing with different phages (as described below).

Transduction. All phages were assayed for their capacity to encapsulate and to transfer host genetic material. The auxotrophic mutants HisB22(P22) and Trp-8(P22) were infected with phage particles of the various Anderson typing phages, and wild-type transductants were selected on minimal medium. All phages appeared to be transducers, although with different efficiencies, a result obviously due to different capacities for lysogenization (Table 1). Phages which formed "clear" plaques, produced considerably fewer wild-type transductants, and a poor background lawn of nontransduced recipient cells indicated extensive lytic reactions. Since both markers, his⁺ and trp⁺, could be efficiently transduced, although located at different positions on the Salmonella chromosome, these phages are considered to be generalized transducers (13).

Immunity. As will be shown later, all of the typing phages and phage 3b are relatives of phage P22 with only two exceptions (A12 and A13). Therefore, it was of interest to determine their immunity. Phage P22 has a C-immunity system which is comparable in organization and function to the immunity region of phage lambda. In addition, P22 has a second system, immunity I. This system codes for an antirepressor which is able to inactivate the C repressor (18). Phage ES18, also related to P22, is homoimmune to P22, i.e., it has the same immunity as P22, but it has no antirepressor (14). When a new phage is plated on indicator strain DB21 which is lysogenic for P22 or ES18, three reactions are possible. (i) If the unknown phage lyses both lysogenic strains, it is heteroimmune to P22/ES18. (ii) If it lyses none of them, it is homoimmune to them, and it has no antirepressor (no imm-I region). (iii) If it lyses DB21(ES18) but not DB21(P22), it is homoimmune to P22/ES18 andlike P22expresses an antirepressor. This biological assay is not suitable for testing heteroimmune phages for the presence of an antirepressor. A complication results from the existence of two superinfection exclusion systems (sieA, sieB) (19) in addition to the C immunity expressed by P22 but not by ES18: P22 prophage would exclude other phages even being heteroimmune. Therefore, this test can only be performed with cells lysogenic for a P22 mutant defective in both systems, i.e., P22 sie1. The results of this study are included in Table 1. They show that only A8, A17, and A18 are homoimmune to P22 and that these phages do not have the imm-I system. All other phages are heteroimmune to P22/ES18, and the existence of imm-I cannot be assayed biologically. At present we do not know to how many different immunity systems these P22-heteroimmune phages belong.

EcoRI restriction pattern and hybridization with P22 and ES18 DNA. DNA from all typing phages was prepared, and EcoRI restriction patterns were monitored, followed by blotting for hybridization. Several typing phages show identical restriction patterns (Table 1). For example, phages 3b, A1, and A7 exhibit the same pattern. A closer look at Fig. 1 shows that these individual typing phages derive from each other and differ only in the bacterial hosts used for propagation. Therefore, phage 3b is the ancestor of A1, which was obtained by propagating phage 3b in strain BA1; phage A7 was obtained by propagating phage A1 in bacterial strain BA7. This demonstrates that these phages form a direct genealogic line and did not undergo visible changes in their DNA compositions by passage through different bacterial strains.

View larger version (52K): [in this window] [in a new window]   FIG. 2.   EcoRI restriction patterns of phage DNAs as indicated.

Propagation strains for Anderson typing phages: prophages and their DNA. It is a characteristic feature of the typing phages that any that have been propagated on a particular host strain undergo specific "modifications." At least in some cases these modifications included also changes in the DNA restriction patterns (Fig. 1). These changes are possibly due to recombination with residing prophages. We have shown that most, if not all, natural Salmonella isolates are lysogenic at least for one prophage (12) and that most of them are related to phage P22 (15). Therefore, it was interesting to examine whether this holds true also for the Anderson propagating strains. If the situation here were similar, at least some of the "modifications" of the typing phages could result from recombination with the resident prophages as indicated by the EcoRI patterns, thus producing new module combinations and consequently different biological properties. Through acquisition of new base plate modules a different host range might result (e.g., ES18 which infects rough strains), and exchange of immunity regions would also influence plating properties, depending on the prophage residing in the strain to be tested.

Interactions between infecting typing phages and prophages. To test this hypothesis, the EcoRI restriction patterns of infecting typing phages, of the phages released by the respective propagation strains, and of the new typing phages resulting from these infections were compared. Figure 3 shows a very clear example of an altered DNA pattern due to recombination. Phage A14 results from propagation of phage 3b in strain BA14. This propagation strain harbors a P22-related prophage: PBA14. "Progeny" phage A14 shows three bands from "parent" 3b (arrows pointing to 3b) and four bands from the other "parent" PBA14 (arrows pointing to PBA14). One of the bands and the underrepresented band resulting from DNA packaging (9) have all three phages in common. Not all lineages are as clear as this one. Figure 4 shows the next step leading from typing phage A14 to the new typing phage A26 by propagation in strain BA26. Once more the DNA of newly arising phage shows changes, although most bands of "parent" A14 are conserved. However, three bands are missing, but no bands from the other "parent" PBA26 appear. In many cases the recombination processes may be even more complicated, possibly due to repeated recombination events during a propagation cycle. Also, the changes of the restriction patterns from phage 3b to A4 and from A1 to A5 cannot be interpreted in a simple way because propagating strains BA4 and BA5 did not release a detectable phage and because hybridization with P22 and ES18 probes did not indicate homology with these genomes. It cannot be excluded that the Salmonella typing strains which, according to a previous study (5), are also the propagation strains and which came to us recently have been replaced since the time of propagation of the typing phage stocks.

View larger version (69K): [in this window] [in a new window]   FIG. 3.   EcoRI restriction patterns of phage DNAs as indicated. DNA of phage A14 in the center is flanked by the DNA of its ancestor phage (phage 3b) and the DNA of the phage (PBA14) residing in the propagation strain (BA14). The arrows indicate bands retained from the respective "parent."

View larger version (54K): [in this window] [in a new window]   FIG. 4.   EcoRI restriction patterns of phage DNAs as indicated. DNA of phage A26 in the center is flanked by the DNA of its ancestor phage (phage A14) and the DNA of the phage (PBA26) residing in the propagation strain (BA26).