Split-Operon Control of a Prophage Gene*

Transcription

1 Proceeding8 of the National Academy of Sciences Vol. 65, No. 2, pp , February 1970 Split-Operon Control of a Prophage Gene* L. Elizabeth Bertani DEPARTMENT OF MICROBIAL GENETICS, KAROLINSKA INSTITUTET, STOCKHOLM, SWEDEN Communicated by Renato Dulbecco, November 10, 1969 Abstract. Both prophage integration in bacteriophage P2 and the reverse event, prophage excision, are known to require a specific phage gene product, the so-called int function. We find that P2 can integrate efficiently at a free attachment site also in an immune host (i.e., in the presence of phage specific repressor) provided the superinfecting phage is not deficient in int function. Prophage P2, on the other hand, is not excised from the host chromosome even in a derepressed lysogen unless int function is supplied by a superinfecting phage. Thus, the int function of P2 is expressed constitutively by the superinfecting phage, but is not expressed by the prophage even in the absence of phage repressor. It is proposed that the int function of P2 is not controlled by phage repressor, but belongs instead to a constitutive operon that is physically disrupted by prophage integration. Introduction. The model for prophage integration proposed by Campbell' applies equally well to the noninducible phage P2 as to the inducible phage lambda. Circular forms of P2 DNA are known2 and the prophage map of P2 has been found to be a circular permutation of the map of the vegetatively multiplying phage.3 In addition, it has been shown that the integration of P2 into a sensitive host requires a phage gene product similar to the int function described for lambda.4-6 Integraseless (int) mutants of P2 have been isolated.7 8 One a b pj 1234 P2 int c such mutant, which maps close to the i 4 - phage episite or region of homology Infecting or superinfecting with the bacterial chromosome, has phage chromosome been shown to have decreased site- - - specific phage recombination.8 34 P2 int c a b p1 12 According to Campbell's model, prophage integration takes place following an integrase-mediated reciprocal re- FIG. 1.-Different configurations of the combination event at a specific site on phage chromosome. pi, 1234, P2, int = int the phage chromosome. When the operon in which: P1 = primary site, P2 = secondary promoter site, promoter int prophage detaches again at a later time, structural gene for integrase, and 1234 = the sequence of events is presumably phage episite or region of homology with the reversed. As a result, any operon that bacterial direction chromosome. of transcription. Thea, arrows b, c = indicate other extended to both sides of the integra- phage genes. 331

2 332 GENETICS: L. E. BERTANI PRoc. N. A. S. tion point would be automatically disrupted by integration and then reconstituted again following prophage excision (see Fig. 1). This type of regulation, which would depend on the configuration of the phage chromosome rather than on the presence of phage repressor, has so far not been described. Our results suggest, however, that P2 int function is regulated primarily in this way. This is In contrast to the case for lambda where the activity of the int gene has been shown to be under the control of the phage repressor.6' 9 Materials and Methods. The P2 phages, bacterial strains, media, and general techniques used in these experiments have all been described.7' 8, In some experiments the LB-broth may be supplemented with 0.2 M potassium phosphate (KP) or phage neutralizing antiserum (serum) with a neutralization constant of ca. 1 per min. Phage carrying the c allele were distinguished from c+ phages by the appearance of their plaques. Nonlysogenic strains can be distinguished from lysogenic strains by their sensitivity to P2 vir, and singly lysogenic can be distinguished from doubly lysogenic strains by their sensitivity to P2 vir, vir,3."3 Strains sensitive to both phages have been designated as "sensitive"; those immune to P2 virl, but not to P2 vir, vir13, as "low immunity level"; and those showing increased (but not necessarily complete) resistance to P2 vir, viri3 as "high immunity level." Results. Constitutive expression of P2 int function in immune hosts: Prophage P2 has several chromosomal attachment sites in its host Escherichia coli C of which one (location I) is strongly preferred."i By a series of genetic manipulations it is possible to prepare a single lysogen that carries a prophage at one of the secondary attachment sites (location II)."1 Such a strain is immune, but contains an unoccupied location I. It was previously reported by Six" that superinfection of such a strain results in the formation of double lysogens carrying the superinfecting phage at the preferred site. The proportion of such double lysogens depended upon the multiplicity of superinfection used, with an average probability of attachment per superinfecting phage of 0.04 and identical results were obtained regardless of whether the prophage carried the c+ or the c allele. We have repeated Six's experiment using a strain carrying the clear-plaque mutant P2 c at location II arid wild-type P2 as superinfecting phage (Table 1). The proportion of double lysogens-that is, strains carrying both the c allele of the prophage and the c+ allele of the superinfecting phage-was per cent, TABLE 1. Attachment of P2 at an unoccupied site in an immune host.* No. of Colonies Carrying the Following Superinfecting Multiplicity of No. of colonies Prophage Markers phage superinfection tested c onlyt c + onlyt Both c and c+ P2 int P2 int P2 int P2 int, P2 int, P2 WtM * Strain C-67, carrying P2 c at location II was used as host. The superinfected bacteria were diluted to ca. 104 bacteria per milliliter and grown for several generations in LB + KP+ serum to permit segregation before plating for colonies. Each colony was then streaked out and a progeny colony was tested from each streak. t All low immunity level. $ All high immunity level.

3 VOL. 63, 1970 GENETICS: L. E. BERTANI 333 for multiplicities of 7.7 to 15, in good agreement with Six's results. If the formation of such double lysogens were mediated by some recombination system other than P2 integrase, then it should make no difference if the superinfecting phage is P2 or a P2 int mutant. Instead, when P2 int mutants are used, the probability of attachment of the phage to the free site is reduced at least 100-fold (Table 1). Thus, the attachment of the superinfecting phage to an unoccupied site in an immune host depends upon the ability of the phage to make integrase. In an analogous experiment using phage lambda,16 the superinfecting phage is usually found to attach by integration into the already-present prophage rather than at the unoccupied site. In this case, some recombination system other than int is utilized. When a strain carrying a P2 prophage at location I is superinfected with more genetically marked P2, the most common new types found are single lysogens in which the prophage has been replaced by the superinfecting phage (prophage substitution).15 Prophage substitution seems to be relatively rare in lambda.'7 This is understandable since it appears that substitution in P2 also requires int activity (Table 2). The frequencies of substitution (calculated per superinfecting phage) obtained following superinfection of a location I lysogen were 4-7 X 10-3, whereas those obtained with int mutants were at least 25 times lower. TABLE 2. Prophage substitution in an immune host.* No. of Colonies Carrying the Following Superinfecting Multiplicity of No. of colonies Prophage Markers phage superinfection tested c onlyt c+ onlytt Both c and c+ P2 int P2 int P2 int P2 int P2 int * Strain C-147, carrying P2 c at location I was used as host. The experiments were done as described under Table 1. All low immunity level. t The colonies carrying c+ phage only may result from either recombination between the superinfecting phage and the prophage, or complete replacement of the prophage by the superinfecting phage. According to Six,15 the latter is the more common event. This should hold particularly strongly for the c marker, which is known29 to be very close to the episite. The frequencies of prophage substitution calculated per superinfecting particle observed here are about 4 times higher than those observed by Six. All high immunity level. Both these experiments indicate that integrase is synthesized constitutively by superinfecting P2. They also rule out the possibility that the int gene is active in the prophage state. From additional studies not to be reported here, it would seem most likely that the two double lysogens obtained following superinfection with int mutants are the result of the integration of the superinfecting int mutant into the already present prophage. Absence of int activity in a derepressed host: MNIutants of P2 that make temperature labile immunity repressor have been isolated.'4 Lysogenic strains, prepared from such mutants, are abortively induced when shifted from 30 to 42. The bacteria lose their immunity to superinfection, but do not produce phage. The prophage of an abortively induced lysogen is transferred in bacterial crosses as

4 334 GENETICS: L. E. BERTANI PROC. N. A. S. if it were still attached to the bacterial chromosome. These results suggest that integrase activity is absent from P2 lysogens, even when they are derepressed. Detachment of a prophage in abortively induced bacteria can be brought about by superinfecting phage and can be detected either as prophage loss in bacterial crosses or as an increase in the production of phage carrying prophage markers." The effect is particularly strong if the superinfecting phage is first irradiated with ultraviolet light and, at its maximum, amounts to the release of one to two particles with prophage markers by every bacterium. The role of the superinfecting phage could be to supply int function to the system. This interpretation is supported by the observation that int mutants, either irradiated or not, when used to superinfect an abortively induced lysogen (Table 3), are not able to pro- TABLE 3. Prophage detachment in a derepressed host. * Burst size of phage with the Superinfecting Multiplicity of following marker Expt. phaget superinfectiont c + c 1 P2int Irradiated P2 int No phage P2int Irradiated P2 int No phage P2 int, Irradiated P2 int No phage P2 int, Irradiated P2 int, No phage P2intH Irradiated P2 int No phage P2 intao Irradiated P2 ints No phage * Strain C-1097 (grown at 300 in LB + KP + serum), carrying P2 C5 as prophage, was derepressed by placing at 420 for 10 min before superinfection. After an additional 15 min at 420, the superinfected bacteria were transferred back to 300. The amount of phage produced was measured after 1-hr incubation at 300. In expts. 2 and 6, was used as indicator, whereas C-85 was used in the others. t The survivals of the phages to the dose of ultraviolet light used were: 23, 26, 34, 40, 35, and 26% in expts. 1-6, respectively. I The multiplicity based on the total number of particles adsorbed, assuming that the adsorption of ultraviolet-inactivated particles is normal, is given. mote an increase in the production of phage bearing prophage markers. Thus, int function must be supplied to a derepressed lysogen to obtain prophage detachment. Discussion. These experiments show that the expression of P2 int function is largely independent of the presence or absence of the phage immunity repressor. Instead, the int gene appears to be constitutive in the superinfecting phage and inactive in the prophage. It is therefore proposed that it belongs to a constitutive operon that is physically disrupted by prophage integration. Alternatively, one could imagine that it is the int gene itself that is split. It is clear that the int gene is not completely inactive in the prophage state. All P2 lysogens produce some phage spontaneously and since lysogens for P2 int

5 Voi. 65, 1970 GENETICS: L. E. BERTANI 335 mutants have a greatly reduced spontaneous phage production,7 it follows that integrase must be produced occasionally in ordinary lysogens. It has been shown that some transcription of genes in an operon can be initiated with low probability at secondary sites within the operon rather than at the main promoter site."8 Occasional transcription beginning at a secondary site might also explain the presence of low levels of integrase necessary for phage production in P2 lysogens (see Fig. 1). Such a model is also compatible with the idea, already proposed,'9 that spontaneous phage production in P2 is dependent not only on the immunity of the lysogen but also on some other prophage initiated event called "prophage activation." In addition, mutation of the secondary site in such a way that it could no longer function as a promoter site would result in phage with an "excision-less" phenotype; that is, phage with normal attachment, but defective detachment. In the alternative possibility that the int gene itself is split, rare integrase production would be more difficult to explain. Following superinfection of a lambda lysogen, the superinfecting lambda phage is usually found to integrate into the already-present prophage to produce socalled tandem double lysogens with two prophages occupying the same bacterial attachment site.20 In a similar experiment with P2, the relatively rare double lysoxgens that are obtained carry the two prophages at distinctly different sites on the bacterial chromosome." P2 tandem double lysogens, composed of two int+ phages, have never been reported, although recently tandems formed by two P2 int mutants7 or by two P2 phages having different site specificities21 have been described. The absence of stable int+ tandems in P2 would be expected if the int gene is under split operon control, because in tandem lysogens one int operon would remain intact or be reconstituted. Such strains would produce integrase constitutively, most likely resulting in great instability. Restoration of a gene function in double lysogens has been reported for lambda, although the case may be somewhat more complex than that described here. Some of the lambda biotin-transducing phages that contain deletions of the nonessential phage functions occupying the middle part of the lambda chromosome are unable to multiply on a bacterial host that carries a rec A mutation.22 One could assume that wild-type lambda, which multiplies normally on the rec A host, has a gene that can substitute for the rec A function. This gene is not expressed in lambda single lysogens, but Erskine23 has reported that lambda double lysogens, presumably tandems, are able to restore the rec+ phenotype to the rec A host. In some cases, however, an effect of the repressor seems also to be present.24 The possibility that some genes in temperate phages are regulated independently of repressor suggests an explanation for some of the observations made on cancer-producing viruses such as polyoma or SV40. The cells that are transformed by these viruses to cancer cells apparently carry virus in some integrated, largely inactive form, similar to the prophage state, but according to certain tests, do not seem to contain virus specific repressor.2528 One could postulate that the genes for independent virus multiplication lie all together in a constitutive operon that is split when the virus becomes integrated. Alternatively, it might be some critical gene that is split. In any case, the integrated virus would be inactive as a virus even if no repressor were present.

6 336 GENETICS: L. E. BERTANI PROC. N. A. S. The excellent technical assistance of Ewa Forsberg is gratefully acknowledged. * This work was supported by a joint grant from the Swedish Medical and Natural Sciences Research Councils and the Swedish Cancer Society. 1 Campbell, A., Advan. Genet., 11, 101 (1962). 2 Inman, R. B., and G. Bertani, J. Mol. Biol., 44, 533 (1969). 3Calendar, R., and G. Lindahl, Virology, in press. 4Zissler, J., Virology, 31, 189 (1967). 5 Gingery, R., and H. Echols, these PROCEEDINGS, 58, 1507 (1967). 6 Gottesman, M. F., and M. B. Yarmolinsky, J. Mol. Biol., 31, 487 (1968). 7Choe, B.-K., Mol. Gen. Genet., 105, 275 (1969). 8 Lindahl, G., Virology, 39, 861 (1969). 9 Signer, E., Ann. Rev. Microbiol., 22, 305 (1968). 10 Bertani, G., J. Bacteriol., 62, 293 (1951). '1 Bertani, G., and E. Six, Virology, 6, 357 (1958). 12 Bertani, L. E., Virology, 13, 378 (1961). 13 Bertani, G., Virology, 18, 131 (1962). 14 Bertani, L. E., Virology, 36, 87 (1968). 15 Six, E., Virology, 14, 220 (1961). 16 Campbell, A., and J. Zissler, Virology, 28, 659 (1966). 17 Lieb, M., Cold Spring Harbor Symposium on Quantitative Biology, 18, 71 (1953). 18 Imamoto, F., and J. Ito, Nature, 220, 27 (1968). 19 Six, E., Virology, 7, 328 (1959). 20 Calef, E., Genetics, 55, 547 (1967). 21 Six, E., personal communication. 22 Manly, K. F., E. R. Singer, and C. M. Radding, Virology, 37, 177 (1969). 23 Erskine, J. M., J. Gen. Virol., 5, 161 (1969). 24Ibid., 171 (1969). 26 Dulbecco, R., Proc. Royal. Soc., B, 160, 423 (1964). "' Watkins, J. F., and R. Dulbecco, these PROCEEDINGS, 58, 1396 (1967). 27 Sambrook, J., H. Westphal, P. R. Srinivasan, and R. Dulbecco, these PROCEEDINGS, 60, 1288 (1968). 28Jensen, F. C., and H. Koprowski, Virology, 37, 687 (1969). 29 Lindahl, G., Virology, 39, 839 (1969).