Proc. Natl. Acad. Sci. USA Vol. 75, No. 11, pp. 5594-5598, November 1978 Genetics Analysis of P1 immunity by using X-P1 recombinants constructed in vitro (P1 repressor genes/PI antirepressor gene/PI dnaB analog gene/DNA cloning) NAT STERNBERG, STUART AUSTIN, DANIEL HAMILTON, AND MICHAEL YARMOLINSKY Cancer Biology Program, National Cancer Institute, Frederick Cancer Research Center, P.O. Box B, Frederick, 21701 Communicated by Gary Felsenfeld, August 10, 1978 ABSTRACT We describe the dissection and reconstruction MATERIALS AND METHODS of a complex control circuit, the P1 immunity system, by a method that involves inserting EcoRI-generated fragments of Bacterial and Phage Strains. The Escherichia coli strains P1 DNA into X vectors that can then be sequentially inserted used were NS985 (HfrH, sup+), RW842 [HfrH XA(int-FII)gajT] into a bacterial cell. Using these techniques we have isolated (14), YMC (supF) (15), K175 (supD, Xr) (10), Q1508 X-P1 hybrid phages that express the products of P1 genes cl, c4, [dnaB70(ts), thyAl (12), NS91 (C600 groPA15) (16), and N3072 ant, and ban and, in appropriately constructed lysogens, con- [HfrH, sup+, A (pro-lac) Xlii] (6). The X vector X firmed the roles played by the first three of these products in Dam15b538sr1X3c1857(ts)nin5 contains a single EcoRI site for phage immunity. In addition we have'localized to particular inserting EcoRI fragments of P1 DNA (11).' Hybrid phages P1 fragments the sites requisite for expression and repression obtained by inserting P1 fragments (11) are designated X-Pl:c4, of these gene products. The analysis leads to the conclusion that :cl gpant acts in trans to antagonize repression mediated by gpcl, X-Pl:c1, etc.; :c4 or refers to pertinent genes located on the in support of' one of two proposed models for gpant action. inserted P1 fragment. Other phages used were PlCm, Moreover, two features of the,immunity system are revealed: PlCmcl.100(ts), PlCmcl.55(am), PlCmc4.32, PlCmelrs, and (f) a hitherto unknown component that effects gpcl repression; Plvdrsban-l. Unless otherwise indicated, all of the P1 phages and (if) an unexpected ability of gpc4 to channel a superin- carried the Tn9 transposon (Cm). PlCmvirsant-10 and fecting cl'+ phage into the lysogenic state, which suggests that PlCmc4.32ant-17 were isolated from P1 lysogens made by gpc4 activity regulates the establishment phase of lysogeny. infecting strain NS985 with either PlCmvdrs or PlCmc4.32. The presence of the ant- mutation in these phages was con- Analysis of complex genetic'controls requires methods that firmed by their inability to make plaques on strain.K175(Plcry) permit parts of the control circuits to be isolated, characterized, (5). Mutational defects are indicated here by a minus sign su- and variously reassembled. We show here how the insertion of perscript unless an allele numbei is specified. defined DNA segments into X vectors that can be successively Media and Growth Conditions. Media and commonly used integrated into the chromosome of a suitable bacterium permits microbiological procedures for P1 and N were as described (13, the analysis of several interacting genes by reconstruction as 17). Because the attX-int region of the X vector has been re- well as dissection experiments. Because of the phenomenon of moved' by the b538 deletion (18), homologous recombination incompatibility, this type of analysis would not be possible if with the cryptic X prophage of 'strain RW842 was used to ly- the genes to be studied were located in a plasmid. sogenize with primary X-P1 hybrids. The subject of our analysis is the immunity to superinfection The experiments described in this report were carried out conferred by the plasmid prophage of coliphage Pl. This im- at an EKi-Pi containment level in accordance with the'stipu- munity is thought to' require the presence of two repressor lations specified in FCRC MUA# 1. proteins, the products of genes ci (1) and c4 (2) (gpcl and gpc4) Construction of Xatt+imm21-P1 and Xatt+imm434-P1 that are located in the separate immC and immI regions of the Hybrid Phages. The Xb538cI857(ts)nin5-P1 hybrids (X- genetic map (Fig. 1). Gpcl represses the viral lytic hybrids) -were restructured first by recombination with functions. Gpc4 is thought to repress a second immI' gene, ant Xfmm21nin5 or Ximm434nin5 to replace the immX' region (or reb). The product of the ant gene has been postulated to act with'imm2l or imm434 and then by recombination with either in trans to antagonize gpcl-mediated repression (9) or Xb515b519Nam7cI857(ts) to replace the b538 deletion with in cis to replicate DNA of the superinfecting phage even in the the two smaller deletions, b515 and b519, which leave attX and presence of gpcl (i.e., it acts as a repressor bypass) (10). Phage the X int gene intact (18). The restructured hybrids (designated P7, which has a ci gene functionally identical to the ci gene Ximm2l-Pl or Ximm434-Pl) can lysogenize standard strains of P1 but differs from P1 in the specificity of its c4 repressor at 32 or 42°C. (9), is able to grow in a P1 lysogen because gpant of the in- Isolation of a X-P1:ban Hybrid. We took advantage of the fecting P7 phage is not repressed by gpc4 of the prophage. ability'of the P1 ban gene product to substitute for the dnaB and inter- protein of the host (12) in order to isolate a X-Pl:ban hybrid. We describe here the isolation, characterization, Strain Q1508(X) was infected with a pool of \-P1 hybrids (de- actions of individual X-P1 hybrids bearing EcoRI-generated rived from 1000 plaques). The infected cells were spread on fragments of P1 (11) that express each of the known components thymine-supplemented'Ll plates and incubated for 36 hr at of the P1 immunity system. An additional hybrid that'expresses 42°C. The frequency of infected cells that survived to form the P1 ban gene (12), known to be directly under cl control (13) colonies (3 X 10-6) was about 30-fold greater than the (Fig. 1), provides'an indicator of cl repression. dnaB70(ts) reversion frequency. Phage that formed turbid plaques at 320C and clear plaques at 42°C [phage with the The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "ad- Abbreviations: Cm, chloramphenicol resistance transposon Tn9; am, vertisement" in accordance with 18 U. S. C. §1734 solely to indicate amber, ts, temperature sensitive; gpc4, gpcl, etc., gene product of gene this fact. c4, cl, etc. 5594 Downloaded by guest on September 27, 2021 Genetics: Sternberg et al. Proc. Natl. Acad. Sci. USA 75 (1978) 5595

2 1122 7 expresses should a riE n gpc4 lysogenization-deficient Plc4- I EoRI fragments r-,n i m permit phage to lysogenize. Of 50 independent RW842 (X-P1) lysogens Cm imml immC v virs screened, 2 showed a 100- to 1000-fold increase in the capacity l /~~~~~~~/\,'fl for being lysogenized by Plc4- (PlCmc4.32). The X-P1 phages gpf2 gpc4 gpant-.- gpban gpc1 (X-Pl:c4a and X-Pl:c4b) obtained from these lysogens were made att + int + and then integrated into another host (NS985). In NS985, as in the original host, they conferred proficiency to be lysogenized by Plc4.32 at high efficiency. Because 90-95% O1la ~ of the P1 phage recovered from the NS985 (X-Pl:c4, FIG. 1. P1 with the positions ofthe two immunity regions PlCmc4.32) lysogens still carried the c4.32 allele, lysogenization and the Cm (Tn9) transposon (3,4) above the bar and pertinent genes, by Plc4.32 was not due to recombination between the c4+ gene whose products are designated in the figure, below the bar. The extent of the X-P1 prophage and the c4- gene of the P1. (The re- of the deletions in the Plcry (5), Pldpro8a (6), and Pldprolla (6) maining 5-10% of the recovered P1 phage were Plc4+, by prophages are indicated by the black areas. The location of EcoRI- rescue of c4.32+ from the X-Pl:c4 prophages). Thus, the generated P1 fragments is based in part on the map of Bachi and Arber (7) and on our own marker rescue experiments using am X-Pl:c4 phages express gpc4 constitutively in the prophage state mutations (8). The virs mutation, located in the immI region (2, 9), and this product can act in trans to permit Plc4- to form stable permits the phage to express gpant constitutively even in the presence lysogens. of gpc4. Thus, when a P1 lysogen is infected with Plvirs, gpant is X-Pl:c4 Lysogens Are Immune to P1. Although expressed, repression of lytic functions is lifted, and the infecting Plcl.55(am) formed plaques efficiently and yielded a signif- phage grows. For the same reason, P7, whose ant gene expression is icant number of phage after infection of lysogens of either insensitive to repression by gpc4 of P1, will grow in a P1 lysogen. A-Pl:c4a or X-Pl:c4b, P1 wild type did not (Table 1, line 2). This immunity is probably a consequence of gpc4-mediated c1857(ts) allele] were isolated from these lysogens. Infection repression of the ant gene of the superinfecting phage, because with these phages (X-Pl:ban) increased by about 300-fold the PlIrs and P7, whose ant gene is not sensitive to repression by survival of Q1508(X) at 42°C. A Xatt +imm21-Pl:ban phage the gpc4 of P1, grew well in the X-Pl:c4 lysogens was constructed, and Q1508 lysogens containing this prophage (Table 1, line survived with unit efficiency at 42°C. The X-Pl:ban hybrid 2). phage carried P1 EcoRI fragment 3(6 X 106 daltons). Expression of the ant Gene Product (gpant) by X-P1:c4 Isolation of a X-Pl:cl.100(ts) Hybrid. Strain NS91 Phages. Expression of the ant gene by a superinfecting P1 (Ximm21-Pl:ban) contains a groP mutation, but Ximm434 phage induces a resident P1 prophage (2). Because ant is re- forms plaques with the same efficiency on this strain as it does pressed by gpc4, ant expression by a superinfecting P1 phage on a groP+ strain. Thus, ban gene expression (12, 16) from the requires that the phage be insensitive to gpc4 (e.g., virs) or that Ximm21-Pl:ban prophage can-suppress the inhibitory effect the resident prophage be c4- (Table 2, Exp. A). When a of the groP mutation on phage A growth. Because the P1 cl Plc4-ant- lysogen was superinfected with X-Pl:c4a, the P1 gene product can repress ban gene expression (13) gpcl activity prophage was induced. This hybrid phage was unable to induce was assayed as the restoration of the groP- phenotype to strain a c4+ prophage. We conclude that X-Pl:c4a, like P1, expresses NS91 (Ximm21-Pl:ban). This strain was infected with a pool ant on infection and that this expression is subject to repression of X-P1 hybrid phages and the resulting NS91 (Ximm21-Pl:ban; by gpc4. The X-Pl:c4b hybrid does not express ant. X-P1) lysogens were tested for their ability to allow Ximm434 to form a plaque. On 1 of the 30 lysogens tested, Ximm434 Table 1. Reconstruction of P1 immunity by using A-P1 hybrids formed plaques with '400th to 'A0th the efficiency that was Genes Phage production found on the others. The X-P1 prophage present in that lysogen Prophage present Pici+ Plcl- P7ci+ was recovered and designated X-Pl:cl.100(ts). Analysis of Phage DNA by Agarose Gel Electrophoresis. None None +(40) +(48) +(29) X-P1 DNA was isolated and subjected to EcoRI endonuclease A-PI:c4 c4 -(3) +(60) +(24) digestion and agarose gel electrophoresis as described (11), X-Pi:ci cl -(3.8) +(39) -(3) except that A-P1:2 f2 + + + horizontal slab gel electrophoresis (19) was carried A\-Pl:c4; A-P1:2 c4; f2 -(4.3) +(52) +(33) out in 40 mM Tris acetate/20 mM sodium acetate/2 mM A-Pl:cl; X-P1:2 cl;f2 +(22) +(43) +(19) EDTA, pH 7.8 (20). EcoRI digests containing 1-2 .g of DNA Plcry cl;f2 +(50) +(60) 4-(32) in 25-50 ,ul were applied to the gel and run at 30 V for 18 hr. X-Pl:cl; X-Pl:c4 cl; c4 -(0.4) -(0.6) -(4) The gels were stained with ethidium bromide and photo- Pldprolla cl; c4 -(1.7) -(1.1) -(4.5) graphed under UV light. Plcry; X-Pl:c4 cl; c4; f2 -(0.8) -(2) +(22) Mapping of X:X-P1 Heteroduplexes by Electron Micros- A-Pl:ci, A-Pl:c4 copy. Heteroduplex formation and formamide spreading for A-Pl:f2 cl, c4; f2 -(2.2) -(3) +(32) electron microscopic analysis were carried out as described by Pldprolla;X-P1:2 cl;c4;f2 -(1.8) -(3) +(29) Broker and Chow (21); a Hitachi HU 12A electron microscope Pldpro8a cl;c4;f2 - - + was used at 50 kV. Heteroduplexes were analyzed by enlarging P1 cl;c4;f2 -(1.2) -(1.4) +(24) photographic negatives X20 with a Seamco model 210 micro- P7 cl;c4;f2 + + - fiche reader and the images were measured with a Numonic The host used in this experiment was either N3072 (for Pldpro electronic graphics calculator. lysogens) or NS985. The X-Pl:cl hybrid was Ximm434-P1:c1+, the A-Pl:c4 hybrid was Aimm21-Pl:c4b (identical results were obtained RESULTS with Ximm2l-Pl:c4a), and the X-P1:2 hybrid was XcI857(ts)-P1:2. Isolation of X-P1 Hybrid Phages that Express gpc4. Hybrid The Plcl- phage used was Plcl.55(am). (+), Plaque-forming effi- phages were isolated by inserting EcoRI-generated fragments ciency of 0.5-1.0; -, plaque-forming efficiency of < 10-3 relative to the plaque-forming efficiency with strain NS985. Plvirs formed of Plcl.100(ts) DNA into a A vector (11). A lysate obtained from plaques with near-unit efficiency on all of the lysogens. The values a pool of 500 hybrid phage plaques was usedlo lysogenize strain in parentheses are the phage yields per infected cell in Li broth at 120 RW842. We expect that a X-P1 prophage that constitutively min after infection at 32°C (13). Downloaded by guest on September 27, 2021 5596 Genetics: Sternberg et al. Proc. Natl. Acad. Sci. USA 75 (1978) Table 2. Gpant production by X-Pl:c4a A B C D E F G Infecting phage Yield of phage/infected cell Experiment A NS985(Plc4-ant-) NS985(Plc4+ant-) 22, Plvirs 24 (9ant-;15ant+) 19 (10ant-;9ant+) 3 P1 16 (6ant-;10ant+) 0.3 -

X-Pl:c4a 10 0.08 5.6 - X-Pl:c4b 0.16 0.04 8 r Experiment B 944* Q1508 Q1508(Plc4-ant-) Q1508(Plc4+ant-) 10 - Plvirsban- 0.2 22 (9antC;13ant+) 16 (7ant-;9ant+) 11r-- Plban- 0.4 24 (12ant-;12ant+) 0.2 12.13

X-Pl:c4a 0.1 12 0.2 - 14.4 r X-Pl:c4b 0.05 0.3 0.2 15 - 16 - Bacterial strains for phase infection were grown at 320C in Li broth with maltose (0.2%) and, if carrying P1 prophage, chloramphenicol 19 - (25,gg/ml). P1 was assayed on strain K175(Xr) and X was assayed on strain YMC by using TB plates (P1 will not form plaque on TB plates). Plant+ and ant- phage in the yield were distinguished on the basis of their differential plaque-forming ability on strain K175 (Plcry) (5). Experiment A: The multiplicity of phage infection was 2.5 and the yield of P1 at 38°C was assayed 150 min after infection. Procedures for P1 (13) and X (17) infection have been described. The infecting X-P1 phage have the immunity of phage 21, and the NS985 strains all carry a Ximm2l prophage. The yield of the infecting FIG. 2. Electrophoresis was carried out in a 0.7% agarose slab gel. Ximm2l-P1 hybrid phages under these homoimmune conditions was Lanes contained EcoRI-digested DNA as follows: A, from XD-srIX3 0.25-0.5 phage per cell. Experiment B: The multiplicity of phage in- vector; B, PlCmcl.100; C, X-Pl:ban; D, X-P1:2; E, X-Pl:c4a; F, fection was 2.5 phage per cell, and the yield of phage at 40°C was as- X-Pl:c4b; G, X-Pl:cl.100(ts). The two common high molecular weight sayed 120 min after phage infection. bands present in all of the lanes with X DNA are the two arms of the X vector (11). The minor band above these two bands corresponds to Expression of ant from X-Pl:c4a was confirmed in studies the complex of left and right vector arms joined by the X cohesive on the of a P1 which is termini. The PlCmcl.100 DNA (lane B) that we used to clone frag- expression gene (ban), repressed by gpcl ments contains a tandem duplication of the Tn9 transposon. This in lysogens (13). Expression of ban can be detected by the ability transposon contains an EcoRI site and is located in P1 EcoRI frag- of gpban to suppress host dnaB(ts) mutations and thus allow ment 4 (7). Consequently, fragment 4, which normally overlaps growth of phages such as X or Plban -, which require the dnaB fragments 5 and 6, is cut by the endonuclease to give three new frag- product to replicate. In Table 2, Exp. B, we show that X-Pl:c4a ments, 4*, 4** (overlaps P1 fragment 9), and 4*** (overlaps P1 frag- but not X-Pl:c4b superinfection of a suitable P1 lysogen induces ment 14). The P1 fragments in X-P1:c4a are derived from Plc 1.100, the prophage ban gene. Again, this ant-mediated induction a phage that lacks Tn9, so the P1 fragments seen in lane E must be 9 and 14 and not 4** and 4***. Although X-Pl:c4b carries a Pi frag- requires that the resident prophage be c4-. ment from PlCmcl.100, this fragment must be 9 and not 4** because P1 EcoRI Fragments Obtained from X-P1:c4a and it contains the c4 gene and the wild-type allele of am3.21. The size X-P1:c4b. DNA isolated from X-Pl:c4a and X-Pl:c4b was of the P1 fragments was determined in a separate gel with EcoRI cleaved by restriction enzyme EcoRI and analyzed by agarose fragments of X DNA and Hae III fragments of simian virus 40 DNA gel electrophoresis (Fig. 2). Both hybrid phages carried P1 as standards (11). fragment 9 (2.4 X 106 daltons) which contains the wild-type allele of P1 am3.21. In addition, X-Pl:c4a carried fragments pected X map coordinate (0.656) of the unique srIX3 site into 14 (1.0 X 106 daltons) and 23 (0.18 X 106 daltons), whereas which the P1 fragments are inserted (11). With X-Pl:c4b, the X-Pl:c4b carried fragment 22 (0.21 X 106 daltons) (data not addition formed a single-stranded loop corresponding to 2.64 shown for fragments 22 and 23). By EcoRI cleavage and liga- X 106 daltons of double-stranded DNA, as expected for frag- tion of X-Pl:c4b DNA, we isolated a X-P1 hybrid that contained ments 9 plus 22 (2.6 X 106 daltons). With X-Pl:c4a, the addition only P1 fragment 9. The properties of this hybrid were indis- assumed a stem-loop structure. The double-stranded stem tinguishable from those of X-Pl:c4b. Thus, fragment 9 contains corresponded to 2.45 X 106 daltons of DNA, the approximate the c4 gene and a promoter sufficient for its expression. The size of fragment 9, and the single-stranded loop corresponded wild-type alleles of some ant - mutations can be rescued from to 1.16 X 106 daltons of double-stranded DNA, the sum of fragment 9 and other ant - mutations from fragment 14 (un- fragments 14 and 23 (1.2 X 106 daltons). Thus, X-Pl:c4a con- published data). Because an allele for constitutive ant synthesis tains two copies of fragment 9 in inverted order flanking (virs) is present on fragment 9 and the ant structural gene spans fragments 14 and 23. This arrangement of fragment 9 DNA is the 9-14 junction, transcription of ant probably starts in frag- probably due to the chance association of EcoRI fragments ment 9 and proceeds into fragment 14 (Fig. 1). If oirs defines during the in vitro ligation reaction because no stem-loop the site of gpc4-mediated repression of ant, this site must also structure can be generated in P1 DNA and only one copy of be on fragment 9. It is unlikely that ant expression requires EcoRI fragment 9 is seen per P1 DNA molecule by gel analy- fragment 23, which is present on X-Pl:c4a but is derived from SiS. a relatively distant region to the right of ant. Characterization of a X-Pl:cl Hybrid. A X-Pl:cl.100 hybrid Heteroduplex Analysis of X-Pl:c4 Phages. Heteroduplex phage was isolated, and the Xatt + imm434-Pl:cl.100(ts) hy- structures obtained when DNA from either of the two X-Pl:c4 brid was then constructed and the Plcl+ allele was crossed into hybrids was denatured and reannealed to DNA from the X this phage. The following properties indicate that this vector are shown in Fig. 3. An addition was found at the ex- Ximm434-P1:c1+ hybrid phage expresses the cl gene. (i) Downloaded by guest on September 27, 2021 Genetics: Sternberg et al. Proc. Natl. Acad. Sci. USA 75 (1978) 5597 We therefore lysogenized the Ximm2l-Pl:c4b, Ximm434-PI:cl+ lysogen with various X-P1 hybrid phages containing P1 fragments from the left 20% of the P1 map (Fig. 1). A X-P1 prophage containing P1 fragment 2 (6.7 X 106 dal- tons) (Fig. 2) was found to restore normal immunity to the ly- sogen (Table 1, line 11). We suggest that this fragment contains a gene (designated f2) whose product is required for normal immunity. The significance of this product as a component of the immunity system was confirmed by demonstrating that the . Y" . "extended" immunity conferred by Pldprol la was abolished by the presence of X-PL:f2. The resulting lysogen exhibited an immunity indistinguishable from that of P1 wild type (Table 1, lines 12 and 14). Note that Pldpro8a, which retains fragment 2, showed normal immunity (Table 1, line 13). The synthesis of gpf2 may also determine the immunity properties of lysogens FIG. 3. (A) Heteroduplex between XD-srIX3 DNA and X-P1:c4b containing the c4-deleted prophage Plcry (Fig. 1). Although DNA. The mean (±SEM) length of the single-stranded (ss) addition A-Pl:cl+ alone confers immunity to a cl+ infecting phage, loop (P1 fragments 9 and 22) corresponds to 4.07 ± 0.05 kilobases of Plcry cl+ does not (Table 1, lines 3 and 7). This difference can DNA. The short arm is 37.19 ± 0.14% of the entire X homoduplex. This be attributed to the synthesis of gpf2 from Plcry because a places the ss loop at X map position 0.657, in good agreement with the position of the srIX3 site (22). (B) Heteroduplex between XD-srIX3 A-Pl:cl+, X-P1:2 lysogen (Table 1, line 6) resembles a Plcry DNA and X-Pl:c4a DNA. The double-stranded (ds) stem (fragment lysogen in having no immunity to a cl+ infecting phage. Re- 9) is 10.47 ± 0.19% (3.8 kilobase pairs) of the length of the X homo- cently, two mutations, c6.101 (23) and bof (24) that affect duplex and is located 37.2 ± 0.22% in from one end of the DNA mol- prophage immunity have been isolated. Both are located in the ecule (X map position 0.656). The length of the ss loop (fragments 14 fragment 2 region of the genetic map (Fig. 1). Although these and 23) corresponds to 1.79 ± 0.05 kilobases of DNA. The ss standard mutations remain to be fully characterized, we note that P1 used on this grid was OX174 DNA (5.27 kilobases). (X23,000.) bof- lysogens exhibit "extended" immunity (24). Plc4-ant- will not form a plaque on a Ximm434-Pl:c1+ ly- DISCUSSION sogen whereas Plc4-ant + will. This is because gpant antag- onizes gpcl-mediated repression. (ii) A dnaBts (Ximm2l-Pl: The use of X-P1 hybrids to dissect and reconstruct the P1 im- ban, Ximm434-Pl:c1+) lysogen dies at 420C and a groP- munity system affords two major advantages over standard (Ximm21-Pl:ban, Ximm434-P1:cI+) lysogen exhibits a groP- genetic approaches to the study of P1 immunity: (i) the ex- phenotype [AcI857(ts) fails to plaque on this host]. These results pression of individual components of immunity can be studied indicate that ban gene expression from the Ximm2l-Pl:ban without interference from extraneous viral genes and can be hybrid is being repressed by gpcl from the Ximm434-P1:c1+ shown to be independent of distant regions of phage DNA; and phage. (iii) Plcl.55(am), although unable to lysogenize normal (ii) the interactions between individual components of im- sup+ hosts, lysogenizes NS985sup + (Ximm434-Pl:c1+) at the munity located on separate DNA segments can be studied same frequency as P1 wild type (0.05-0.15). This is due to cl without encountering the incompatibility barrier to their stable complementation rather than recombination because 96% of coexistence. X vectors are particularly suited for such studies, the P1 phage recovered from such lysogens carried the cl.55 because X hybrids bearing any of several immunities (e.g., allele. The remaining P1 phage carried the cl+ allele, showing immX, imm434, imm2l) can be constructed and corresponding that this allele is present on Ximm434-Pl:cl+. singly or multiply immune lysogens can be readily selected. The cl gene product expressed from the original This method has allowed us not only to confirm the roles played Ximm434-Pl:cl.100 isolate is temperature sensitive-i.e., P1 by elements of the P1 immunity system (c4,cl, ant) but also to ant- will form plaques on an NS985 (Ximm434-Pl:cl.100) localize sites needed for expression and control of those elements lysogen at 42°C but not at 320C. The X-P1:cl.100 hybrid and and to reveal a new element and several new features of the its cl+ derivative (data not shown) carried P1 fragments 7 (4 immunity system. X 106 daltons) and 11 (1.9 X 106 daltons) (Fig. 2). The PIcl gene is expressed from a X-P1 hybrid prophage X-Pl:cl+ Lysogens Are Immune to cl+ Superinfecting containing EcoRI fragments 7 and 11. In contrast, a X-P1 hybrid Phage. Plcl+ and P7cl+ did not form plaques on a X-Pl:cl+ containing only fragment 7 carries at least part of the ci gene, lysogen although Plcl- did (Table 1, line 3). We believe that not this effect is due to the of from the by marker rescue tests, but does not express gpcl (data inability gpant produced it is EcoRI the cI into superinfecting phage to prevent repression when gpcl is being shown). unlikely that cleavage splits gene produced from both incoming and resident phages. portions contained on fragments 7 and 11 because: (i) the dis- Reconstruction of the P1 Immunity System. Table 1 (line tribution of am mutations on fragments in the immC region 8) shows that a lysogen containing Ximm2l-Pl:c4b and places at least one fragment (fragment 22) between fragments Ximm434-Pl:cl+ is immune to superinfection by either a 7 and 11 (Figs. 1 and 2). The orientation of fragments 7 and 11 Plcl+ or a Plcl- phage. To our surprise, however, this double relative to each other in the X-Pl:cl hybrid is inverted relative lysogen was also immune to the normally heteroimmune phage to their orientation in P1 DNA (data not shown). Thus, frag- P7. We call this "extended" immunity. An analogous lysogen ment 7 contains the entire cl gene but probably lacks a site (e.g., containing the Ximm2l-Pl:c4b prophage and the deleted P1 a transcription promoter) needed for cl expression. Fragment prophage Plcry cl+ (Fig. 1) exhibited normal immunity (Table 11 apparently can provide a substitute site for this expression. 1, line 10). In contrast, another deleted prophage, P1 dprolla, An alternative possibility, that fragment 11 contains a gene which contains both the c4 region and the cl+ gene (Fig. 1), whose product is needed for cI expression, seems unlikely be- expressed the same "extended" immunity as the lysogen har- cause none of the clear mutations that affect P1 lysogeny is boring X-Pl:c4 and X-P1:cl+ (Table 1, line 9). These results located in the fragment 11 region of the P1 map (1, 23, 25). suggested that an element present on Plcry and absent on The c4 gene and a promoter sufficient for its expression are Pldprolla is required for normal P1 immunity. present on EcoRI fragment 9 as is the wild-type allele of the virs Downloaded by guest on September 27, 2021 5598 Genetics: Sternberg et al. Proc. Natl. Acad. Sci. USA 75 (1978) mutation, a mutation that allows constitutive ant production. 1. Scott, J. L. (1970) Virology 41, 66-71. The ant gene spans the junction between fragments 9 and 14. 2. Chesney, R. H. & Scott, J. R. (1975) Virology 67,375-384. in 9 3. Rosner, J. L.& Gottesman, M. M. (1977) DNA Insertion Ele- The transcription of ant probably originates fragment ments, Plasmids, and Episomes, eds. Bukhari, A., Shapiro, J. & because the only other fragment (fragment 23) present on the Adhya, S. (Cold Spring Harbor Laboratory, Cold Spring Harbor, X-PI:c4a phage, which expresses ant, is derived from a region NY), pp. 213-218. downstream from the origin of ant transcription (Fig. 1). 4. MacHattie, L. A. & Jackowski, J. B. (1977) DNA Insertion Ele- What is the function of gpant ? The experiments of Table 2 ments, Plasmids,and Episomes, eds.: Bukhari, A., Shapiro, J. & show that gpant can act in trans to induce a P1 prophage, as Adhya, S. (Cold Spring Harbor Laboratory, Cold Spring Harbor, evidenced both by lytic growth of the prophage and by in- NY), pp. 219-228. duction of a specific gene (ban) known to be subject to cI re- 5. Scott, J. R. (1973) Virology 53, 327-336. 6. Rosner, J. L. (1975) Virology 66, 42-55. pression. We believe this result rules out the suggestion by West 7. Bachi, B. & Arber, W. (1977) Mol. Gen. Genet. 153,311-324. and Scott (10) and Scott et al. (26) that the ant gene (renamed 8. Walker, D. H. & Walker, J. T. (1975) J. Virol. 16,525-534. the reb gene by these authors) makes a cis-acting protein that 9. Wandersman, C. & Yarmolinsky, M. (1977) Virology 77,386- allows P1 to replicate lytically in the presence of active gpcl. 400. We favor the hypothesis that gpant antagonizes gpcl-mediated 10. West, B. W. & Scott, J. R. (1977) Virology 78, 267-276. repression. 11. Sternberg, N., Tiemeier, D. & Enquist, L. (1977) Gene 1, We have shown that X-P1:c4 lysogens, although not immune 255-280. to Plcl-, are immune to P1. We a 12. D'Ari, R., Jaffe-Brachet, A., Touati-Schwartz, D. & Yarmolinsky, believe this result reflects M. (1975) J. Mol. Biol. 94, 341-366. previously unsuspected role for gpant in regulating the level 13. Austin, S., Sternberg, N. & Yarmolinsky, M. (1978) J. Mol. Biol. of gpcl activity during the establishment of the P1 prophage. 120,297-309. Gpc4 repression of the ant gene during the early stages of P1 14. Enquist, L. W. & Weisberg, R. A. (1976) Virology 72, 147- infection may channel the infection into a lysogenic response. 153. Note that, if ant gene expression by the infecting phage is not 15. Dennert, G. & Henning, V. (1968) J. Mol. Biol. 33,322-329. repressed by gpc4, as when the X-Pl:c4 lysogen is superinfected 16. Georgopoulos, C. P. & Herskowitz, I. (1971) in The Bacteriophage by P7 or by Plvirs, then phage growth is normal (Table 1, line Lambda, ed. Hershey, A. D. (Cold Spring Harbor Laboratory, 2). If a cell contains gpc4 and gpcl but not the product of a gene Cold Spring Harbor, NY), pp. 553-564. 17. Sternberg, N. (1976) Virology 71, 568-582. contained on P1 EcoRI fragment 2 (gene f2), it is immune to 18. Davis, R. W. & Parkinson, J. S. (1971) J. Mol. Biol. 56, 403- superinfection by Plcl+, Plcl-, and P7 ("extended" immu- 424. nity). Because expression of the P7 ant gene is insensitive to 19. McDonnell, M. W., Simon, M. N. & Studier, F. W. (1977) J. Mol. gpc4 of P1 (2, 9), the extended immunity must reflect an in- Biol. 110, 119-146. ability of gpant to overcome gpcl-mediated repression. The 20. Sklar, J., Weisman, S., Mussp, R., DiLauro, R. & de Crombrugghe, presence of the f2 gene product restores sensitivity to P7 su- B. (1977) J. Biol. Chem. 252, 3538-3547. perinfection. We believe that gpf2 interacts with sites that 21. Broker, T. & Chow, L. (1976) Electron Microscopy of Nucleic regulate cl gene activity or with the cl gene product rather Acids (Cold Spring Harbor Laboratory, Cold Spring Harbor, NY). than with immI sites or products (Fig. 1) because we can detect 22. Haggerty, D. M. & Schlief, R. F. (1976) J. Virol. 18,659-663. a fragment 2-dependent reduction of gpcl-mediated re- 23. Scott, J. R. & Kropf, M. (1977) Virology 82,362-368. pression of the PI ban gene in a cell lacking the imml region 24. Touati-Schwartz, D. (1978) in DNA Synthesis: Present and (data not shown). Future, eds. Molineux, I. & Kohiyama, M. (Plenum, ), We thank R. Musso and R. Yuan for many enlightening discussions pp. 683-692. during the course of this work and S. Shafer and C. Billman for diligent 25. Scott, J. R., Kropf, M. & Mendelson, L. (1977) Virology 76, secretarial assistance. This research was sponsored by the National 39-46. Cancer Institute under Contract NOI-CO-75380 with Litton Bionetics, 26. Scott, J. R., West, B. W. & Laping, J. L. (1978) Virology 85, Inc. 587-600. Downloaded by guest on September 27, 2021