immX Immunity Region of Rhizobium Phage 16-3: Two Overlapping Cistrons of Repressor Function

Abstract

16-3 is a temperate phage of the symbiotic nitrogen-fixing bacterium Rhizobium meliloti 41. Its prophage state and immunity against superinfection by homoimmune phages are governed by a complex set of controls: the immC and immX repressor systems and the avirT element are all located in well-separated, distinct regions which span 25 kb on the bacteriophage chromosome. The anatomy and function of the immC region are well documented; however, fewer analyses have addressed the immX and avirT regions. We focused in this paper on the immX region and dissected it into two major parts: X_U/L and X_V. The X_U/L part (0.6 kb) contained two overlapping cistrons, X_U and X_L, coding for proteins pXU and pXL, respectively. Inactivation of either gene inactivated the repressor function of the immX region. Loss-of-function mutants of X_U and X_L complemented each other in trans in double lysogens. The X_V part (1 kb) contained a target for X_U/L repressor action. Mutations at three sites in X_V led to various degree of ImmX insensitivity in a hierarchic manner. Two sites (X_V1 and X_V3) exhibited the inverted-repeat structures characteristic of many repressor binding sites. However, X_V1 could also be folded into a transcription terminator. Of the two immunity regions of 16-3, immX seems to be unique both in its complex genetic anatomy and in its sequence. To date, no DNA or peptide sequence homologous to that of ImmX has been found in the data banks. In contrast, immC shares properties of a number of immunity systems commonly found in temperate phages.

A bacterium carrying a prophage is immune to superinfection by a homoimmune phage (20). Superinfection immunity is due to the binding of the prophage-encoded repressor protein(s) to specific operator sites of the phage chromosomes. The repressor prevents vegetative phage development by acting on both the resident prophage and the newly injected DNA of a homoimmune phage (34). Although the repressor systems may differ in specificity, topology, and position of the relevant genes on the genetic map, two major types are emerging among temperate phages investigated by genetic analyses: (i) phages with one immunity control region (canonized by coliphage λ) (20, 34) and (ii) phages with dual control regions (the archetype is Salmonella phage P22) (44). Both types have been described for phages of either gram-positive or gram-negative bacteria. Phages A2, TP901-1, and others of lactic acid bacteria (2, 4, 14, 17, 23, 27), L5 of Mycobacterium spp. (3), PBSX of Bacillus subtilis (28), and a number of coliphage can be classified as the “one-immunity-control-region” type. In contrast, Escherichia coli phage 186 and φ105 of B. subtilis possess two immunity regions, like P22 (9, 10, 22, 41). The highly complex P22 immI region encodes an antagonist (Ant) of the C2 immunity repressor and proteins and RNAs that control ant expression (mnt, arc, and sar) (37). An even more complex immunity system is represented by phage P1, where three immunity regions function in a complex network (19).

The superinfection immunity system of the Rhizobium meliloti temperate phage 16-3 has been mapped to three distinct regions: the immC, immX, and avirT regions. The genetic anatomy and function of the immC region are well documented (GenBank accession no. AJ131679). Cistron c of the immC region codes for a “typical” repressor, with operators at its flanks. Mutations in cistron c lead to clear plaques (29, 30). In its sequence-specific DNA binding, the 16-3 C repressor utilizes a helix-turn-helix operator recognition motif with significant homology (and partial cross-functionality) to the helix-turn-helix motif of the CI repressor of coliphage 434. The cognate operators are also 434 type operators (5, 8, 33).

Little is known of the other two elements, immX and avirT. In the host cell the immC and immX regions together provide very high level immunity to superinfection of the homologous phages. Phage titers drop 10⁷ to 10⁹ magnitudes. By itself immX or immC still inhibits phage growth drastically, 10⁵- to 10⁶-fold. Like immC, immX also expresses a repressor activity, but no mutant phenotype has been identified for it (12). The avirT locus has been localized close to the early genes (6). Its function is not known. One mutant allele, avirT1-9, which has been studied in some detail, gives the superinfecting phage full escape from immX immunity and, to a lesser degree, escape from immC immunity (12).

In this article we report on the unusual structure of the immX gene and on the phenotypes of its mutant alleles. This paper also describes DNA sites other than avirT in which mutations confer insensitivity specifically to the ImmX repressor function upon superinfection.

MATERIALS AND METHODS

Bacterial strains, growth conditions, and phage techniques.

R. meliloti 41 (45) was used for the 16-3 phage experiments and as a recipient host for triparental matings. E. coli strain DH5α (18) was used in cloning experiments and served as the host for donor plasmids used for triparental matings. Growth conditions were those described by Semsey et al. (39). The 16-3 phage techniques used have been described previously (29). Superinfection immunity tests were carried out as described by Dallmann et al. and Dorgai et al. (6, 12). Plasmids and phages used in this study are listed in Tables 1 and 2, respectively.

TABLE 1.

Plasmids used in this study^a

Plasmid(s)	Relevant genotype and use	Source or reference
pLAFR1	Cloning vector (low copy number)	16
pLAFR3	Cloning vector (low copy number)	43
pBBR1MCS-2	Cloning vector (high copy number)	24
pCU999	Source of the Km cassette	32
pDH1, pDH79, pDH114	Cosmid clones of 16-3cti3; marker rescue of immX, EOP assays	12
pCS271	Derivative of pLAFR3; provides ImmX function in trans	This work
pCS295, pCS318	Deletion derivatives of BamHI K fragment of 16-3 chromosome in pBBRMCS-2	This work
pCS337, pCS338	Derivatives of pBBR1MCS-2; provide ImmX function in trans	This work
pCS341	Derivative of pCS338; carries 4-bp insertion as a result of end filling and ligation at the EcoRI site	This work
pCS379	Derivative of pCS338, carrying opal mutation in ORF127	This work
pCS436	Derivative of pCS338, carrying amber mutation in ORF116	This work
pCS367	Carries Km cassette in the overlapping region of XU and XL, to make Km transducing and immX knockout phages	This work
pCS463	Carries Km cassette in XU, to make Km transducing (XU knockout) phage mutant	This work
pCS469	Carries Km cassette in XL, to make Km transducing (XL knockout) phage mutant	This work
pCS375, pCS376, pCS377	Carry cloned fragments of 16-3v17-1 in pBBR1MCS-2	This work

^aPlasmid maps and their construction schemes will be provided upon request.

TABLE 2.

Bacteriophages used in this study

Bacteriophages	Relevant genotype and use	Source or reference
16-3c+	Standard wild type; immX and immC are intact; prophage provides immunity to homoimmune phage superinfection; forms turbid plaques	29
16-3cti3	Derivative of 16-3c+, but carrying cti3, heat-inducible allele in immC region; forms turbid plaques at 28°C, clear at 36°C, used as wild type in this work	29
16-3cti4Sp4	16-3 carries cti4Sp4 mutant c allele in the immC region; forms clear plaques	29
16-3avc17	Avirulent mutant (avirC); forms clear plaques, does not grow on R. meliloti 41 (16-3cti3) lysogen; carries ORC-1 operator mutation in the immC region; 5.4 kb deleted in the c cistron through attP; XU and XL intact.	This work
16-3v17	vir mutant (insensitive to both ImmX and ImmC repressors); grows on R. meliloti 41 (16-3cti3) lysogens; derivative of 16-3avc17 plus XV1 and XV3 mutations in the immX region; forms clear plaques	This work
16-3vB1N	Like 16-3v17, with different XV1 mutation	This work
16-3v17-1	Derivative of v17 plus XV2 mutation in the immX region	This work
16-3cti3KmR6-1, 16-3cti3KmR-U463, 16-3cti3KmR-L469	Derivatives of 16-3cti3 plus Km cassette inserted into the immX gene at various points; form turbid plaques	This work
16-3cti3x17, 16-3cti3x17-1	Derivatives of 16-3v17 and 16-3v17-1, but immC is intact; both form turbid plaques at 28°C	This work
16-3v27, 16-3v31, 16-3v38	ImmX-insensitive stocks with phenotypes similar to those of 16-3v17 and 16-3v17-1	This work
16-3cti3x375, 16-3cti3x377	ImmX-insensitive stocks with phenotypes similar to those of 16-3cti3x17 and 16-3cti3x17-1	This work
16-3cti4Sp4KmR-U463	Recombinant derivative from cross of 16-3cti3KmR-U463 and 16-3cti4Sp4, carrying cti4Sp4 allele in the c cistron and Km insertion in the XU cistron	This work

Triparental mating.

The triparental mating method was used to transfer the different plasmids resident in E. coli harboring pRK2013 (15). The helper plasmid provides the transfer function via the mob region. One-milliliter portions from each cultured bacterium (optical density at 600 nm, 2.0) were mixed, and cells were collected by a brief centrifugation and resuspended in 3 ml of YTB (10 g of tryptone, 1 g of yeast extract, 5 g of NaCl, 1 mM CaCl₂, and 1 mM MgCl₂/1,000 ml). After a brief centrifugation, the cells were resuspended in 100 μl of YTB and loaded onto a 0.2-μm-pore-size nylon filter (Sartorius, Göttingen, Germany), which was placed on the surface of a YTA (10 g of tryptone, 1 g of yeast extract, 5 g of NaCl, 1 mM CaCl₂, 1 mM MgCl₂, and 1.5% agar/1,000 ml) plate. Cells were washed from the filter after overnight incubation at 28°C and spread onto selective plates (YTA supplemented with either kanamycin or tetracycline) to obtain single colonies. Tetracycline was used at a concentration of 15 μg/ml, while the concentration of kanamycin was 30 μg/ml for E. coli and 400 μg/ml for of R. meliloti.

DNA procedures.

The basic DNA manipulations and molecular techniques used have been described elsewhere (35). DNA fragments were extracted from agarose gels by using the QIAEX II Gel Extraction kit (Qiagen). Total bacterial DNA was prepared by a method described previously (1). PCR primers are listed in Table 3. PCR-mediated DNA amplifications were carried out by using Pfu polymerase (Promega) to generate DNA fragments for cloning. After 30 cycles of 50 s at 94°C, 50 s at 50°C, and 2 min at 72°C, the PCR mixtures were extracted with phenol and precipitated with ethanol. The DNAs were then resuspended in TE buffer (10 mM Tris-1 mM EDTA) and digested with the appropriate restriction enzyme(s) to generate the required ends of the fragments. DNA fragments were purified by extraction from agarose gels prior to cloning. PCR mutagenesis was performed according to the method of Landt et al. (26). Nucleotide sequence determination was performed by the dideoxy chain termination method (36) using the TaqTrack Sequencing kit and the fmol DNA Cycle Sequencing System kit (both from Promega).

TABLE 3.

List of primers

Primer	Sequencea	Use
1	CAgatatcCGGCTTTAGTCGATGAACT	Cloning the XU/L region, and detecting XU−XL+, XU−XL−, XU+XL−, and XU− XL+ allelic combinations
2	gaggatccTTGACTAAAGCCGCGTCTC
3	CTTTCCCGCtGATTGGAGAG	Combining with primers 1 and 2 for the construction of the opal mutation in XL
4	CGGAGAGGCCATGCTaTCC	Combining with primers 1 and 2 for the construction of the amber mutation in XU
5	AATAATCGGGGGATTGTTTCAGC	Detection of attR
6	GGTCGTTTTTATTGCCGTGG
7	CGCAGTGTAGCGCAGTCTGGT	Detection of attL
8	ATCTAGAAGTAAATCATTGCCGTAAT
9	ctctagaTTCGCGAGGGAGACA	Sequencing of the XV region
10	ACTGGCAGATTTCAACAACTT	Sequencing of the XV region

^aRestriction sites used in cloning procedures (EcoRV, BamHI, and XbaI in primers 1, 2 and 9, respectively) are boldfaced. Lowercase letters represent bases that are not present in the original 16-3 sequence.

Southern hybridization was performed as described previously (35, 42). DNA probes for Southern hybridization were as follows: the 1.7-kb BamHI-HindIII fragment from the immC region and the 1.2-kb kanamycin resistance (Km^r) cassette carried by Km^r transducing 16-3 mutants.

Total-protein analysis.

Total-protein samples were analyzed on a discontinuous sodium dodecyl sulfate-polyacrylamide gel electrophoresis system (25) and blotted to a polyvinylidene difluoride membrane. The N-terminal sequences for the protein of interest were determined by sequencing with an Applied Biosystems protein sequencer (model 471) with an Edman degradation sequenator program (21).

Generation of 16-3v and 16-3cti3x mutants. (i) 16-3v mutants.

As described earlier, the turbid-plaque-forming avirulent mutant strain 16-3c⁺avirT_1-9 mutated spontaneously toward virulence at a frequency of 10⁻⁴ to 10⁻⁵ (6). One of these (second-site) virulent mutants (forming clear plaques) was backcrossed with wild-type 16-3, and avirulent clear recombinants were scored by the method of Dallmann et al. (6). One of these, 16-3avc17, was kept for sequencing of its immC region. It carried an O_R^C-1 mutation (33) and a 5.4-kb deletion from the middle of the c cistron through attP. 16-3v17 (and 16-3vB1N1) was an ImmX-insensitive spontaneous mutant derivative of 16-3avc17, obtained on an R. meliloti 41(pDH1) bacterial lawn. 16-3v17-1 was a spontaneous mutant derivative of 16-3v17. The parental 16-3v17 developed clear lytic patch when spotted onto R. meliloti 41(pCS271) and a blurry one on R. meliloti 41(pCS338). 16-3v17-1 appeared (at a frequency of 10⁻⁸, and recurred several times independently) in this blurry background as a completely clear plaque. The other 16-3v strains were obtained from independently mutagenized (by N-methyl-N′-nitro-N-nitrosoguanidine [NTG] and by UV irradiation) phage populations (described by Dallmann et al. [6]).

(ii) 16-3cti3x mutants.

16-3cti3x17 and 16-3cti3x17-1 were obtained from crosses between 16-3cti3Km^R6-1 and 16-3v17 or 16-3v17-1, respectively (see Fig. 3A). 16-3cti3x375 and 16-3cti3x377 were isolated as insensitive (ins) recombinants in marker rescue crosses of 16-3cti3 with pCS375 or pCS377, respectively (see Fig. 3C).

FIG. 3.

Mapping of ImmX targets (X_V region). (A) Interval mapping by three-point crosses using ins/sens as a nonselective marker pair. ins, insensitive to ImmX repressor; sens, sensitive. Filled triangle, site of inserted Km^r cassette; open bar, deletion in the immC region. (B) Marker rescue for ImmX-sensitive (sens) phenotype. (C) Marker rescue for ImmX-insensitive (ins) phenotype. +, ins recombinants appeared at a frequency of >10⁻⁶; −, no ins recombinants found among 10⁸ progeny. (D) Mutant sites for ImmX insensitivity in the X_V region. Arrowhead indicates a putative promoter.

Production of plasmids carrying 16-3 X_U/L cistrons with Km^r cassettes inserted.

Plasmids carrying 16-3 X_U/L cistrons with Km^r cassettes inserted were constructed for targeted translocations of the Km^r cassette into the 16-3 chromosome by homologous recombination (see the next section). Plasmid pCU999 (32) was the source of the Km^r cassette (with multiple symmetric restriction sites at either end). Briefly, the cassette was first removed by appropriate restriction cleavages and then inserted into the 16-3 target sequence [bordered by the XbaI(40)-BamHI(42) sites]. The 16-3 target had previously been cloned into a modified pBluescript II KS(+) vector (Stratagene) (EcoRI site destroyed). The Km^r cassette and the flanking 16-3 seqences were then removed from pBluescript II KS(+) and translocated into the conjugative plasmid pBBR1MCS-2 (the precondition for transferring the Km^r insertion X_U/L allele from E. coli to R. meliloti). For pCS367, EcoRI cleavage was used to retrieve the Km^r cassette, which was then inserted into the [pBluescript II KS(+)-derived] 16-3 XbaI(40)-BamHI(42) region at the EcoRI(41) site (mapped in the overlapping region of X_U and X_L cistrons). For pCS463, the Km^r cassette was retrieved by HincII digestion. In the pBluescript II KS(+)-derived 16-3 XbaI(40)-BamHI(42) fragment, an NcoI site was introduced by PCR mutagenesis (26) in the nonoverlapping part of the X_U cistron. The cassette was then inserted at this NcoI site (previously filled up). For pCS469, a similar procedure was followed except that the NcoI site was introduced into the nonoverlapping part of the X_L cistron. A full description of the construction of the plasmids will be provided upon request.

Constructing Km^r transducing phages.

Plasmids pCS367, pCS463, and pCS469 carrying Km^r cassettes were introduced in the lysogenic strain R. meliloti 41(16-3cti3) by triparental crosses. The Km^r transconjugants were propagated (optical density at 600 nm, 0.3) and heat treated for 30 min at 36°C in order to induce the prophage. The culture was vigorously shaken for 3 h at 28°C to achieve lysis before chloroform was added to kill the remaining bacteria. Progeny phages were plated with R. meliloti 41 on YTA with kanamycin (200 μg/ml). Km^r colonies emerged after 3 days at 30°C, and colonies were checked for phage production (after heat induction). The phage DNA yields were tested by restriction enzymes (EcoRI, EcoRV) and by DNA sequencing. The isolates, which carried the Km^r cassette and were flanked by the immX sequences, have been kept for further work. All of the 27 independent isolates carried the Km^r cassette with the expected flanking sequences, demonstrating that the Km^r cassette was translocated from the donor plasmids to the phage chromosomes by double homologous crossovers. Phage isolates 16-3cti3Km^R-U463, 16-3cti3Km^R-L469, and 16-3cti3Km^R6-1 were kept for further experiments. It is noteworthy that the phage chromosomes which incorporated the Km^r cassette together with the carrier plasmid (by single crossover) were too large for packaging (11). The clear-plaque-forming Km^r transducing phage 16-3cti4Sp4Km^R-U463 was constructed by a standard cross between 16-3cti4Sp4 and 16-3cti3Km^R-U463. Clear-plaque-forming progeny (i.e., carrying the cti4Sp4 allele) were checked for the presence of the Km^r cassette by restriction analyses and DNA sequencing. Furthermore, the progeny was checked by the transduction test. The lysogenic strain R. meliloti 41(16-3cti3) conferred kanamycin resistance after being transduced with the clear plaque mutants containing Km^r.

Marker rescue analysis. (i) Rescue of the ImmX sensitivity phenotype.

R. meliloti 41 carrying pDH1 or pDH114 was infected with ImmX-insensitive phages (16-3v17 or 16-3v17-1). Phage progeny were first plated on R. meliloti 41 before individual plaques were tested for growth on plates seeded with R. meliloti 41(pLAFR3) and R. meliloti 41(pCS271). Those phages which grew on R. meliloti 41(pLAFR3) but not on R. meliloti 41(pCS271) were scored.

(ii) Rescue of the ImmX-insensitive phenotype.

Derivatives of R. meliloti 41 carrying plasmids with cloned DNA fragments from 16-3v17-1 were infected with 16-3cti3 (ImmX sensitive). Phage progeny were plated on R. meliloti 41(pCS271) in order to select for plaques of ImmX-insensitive recombinants (which emerged at a frequency of ∼10⁻⁶). Some of the recombinants were isolated, purified, and carefully retested for ImmX sensitivity or insensitivity on R. meliloti 41(pLAFR3) and R. meliloti 41(pCS271). Recombinants, which inherited ImmX insensitivity, showed lysis on R. meliloti 41(pCS271) and R. meliloti 41(pLAFR3).

Nucleotide sequence accession number.

The nucleotide sequence of the EcoRI(37)-EcoRI(43) region of phage 16-3 has been deposited in GenBank under accession no. AJ519534.

RESULTS

Mapping and cloning of the repressor function of the immX gene.

Earlier studies localized the immX gene in the EcoRI L and H fragments of the 16-3 chromosome (12). By setting out from this premise, the position of the gene was narrowed down to a 442-bp region, which included the EcoRI(41) site (Fig. 1A, pCS337). The 442-bp region provided complete ImmX function from trans in the carrier cells, as shown by a 5- to 7-order-of-magnitude reduction in the efficiency of plating (EOP) for the superinfecting 16-3 strains (Table 4, 16-3cti3 and 16-3cti4Sp4). Sequencing of this locus revealed two overlapping open reading frames (ORFs) of 116 and 127 putative codons (Fig. 1B). The two ORFs were in the same frame with opposite directions.

FIG. 1.

The immX region. (A) Localization of ImmX repressor function. +, ImmX repressor function expressed; −, no expression; open bars, deletions. Numbers at restriction sites refer to positions in the 16-3 chromosome map (13). Tentative physical positions for the following reference ts mutants are marked (based on unpublished marker rescues): ts2 [BamHI(29-34)], ts5 and h [BamHI(34-36)], ts216 and ts5124 [EcoRI(65-87)], and ts518 [BamHI(94-98)]. avirT was closely linked to ts5124 (6). Early and late regions were defined by ts mutants (31), and silent regions were defined by their absence (6, 7). Note that on the recombination map the order was defined as silent region-c gene-early region-late region, since recombination frequencies for the c gene and late ts mutants were much higher than those for the c gene and early ts mutants, or for early and late ts mutants (31). ORFs mentioned in the section on sequence analysis are indicated. (B) Site-directed mutations in plasmids and phages. Filled triangles show the locations of Km^r cassette insertions in phages 16-3cti3Km^R-L469 (a), 16-3cti3Km^R6-1 (b), and 16-3cti3Km^R-U463 (c). Open triangle, 4-bp insertion in plasmid pCS341. Filled and open circles represent the transversion G/C to T/A in plasmid pCS436: the filled circle indicates the stop codon in X_U (Glu GAG→amber), and the open circle indicates the synonymous codon in X_L (Leu CTC→Leu CTA). Filled and open diamonds represent the transition C/G to T/A in pCS379: the filled diamond indicates the stop codon in X_L (Arg CGA→opal), and the open diamond indicates the synonymous codon in X_U (Ser TCG→Ser TCA). ORF116 and ORF127 correspond to X_U and X_L, respectively. Arrows indicate the directions of the ORFs. The positions of the mutations [distances in base pairs from the EcoRI(41) site] are as follows: 24 bp leftward, 0 bp, and 304 bp rightward for Km^R-L469, Km^R6-1,and Km^R-U463, respectively; 19 bp rightward for the G/C-to-T/A transversion in pCS436; and 159 bp rightward for the C/G-to-T/A transition in pCS379. In pCS341 the filled-up EcoRI(41) site itself was the site for the 4-bp frameshift.

TABLE 4.

EOP for superinfecting phages in R. meliloti 41 strains expressing ImmX function^a

Superinfecting phage strains	EOP in R. meliloti 41 with plasmid:			Plaquec phenotype	Operatord
	pCS271b (immX)	pCS338b (immX)	pDH114 (immC)		OR	OL
16-3cti3	<10−6	<10−6	10−5-10−6	T	wt	wt
16-3cti4Sp4	<10−6	<10−6	10−5-10−6	C	wt	wt
16-3cti3KmR 6-1	<10−6	<10−6	10−5-10−6	T	wt	wt
16-3avc17	<10−6	<10−6	0.3-1	C	ORC-1	Δ
16-3v17-1	0.5-1	0.3-1	0.1-0.5	C	ORC-1	Δ
16-3v17	0.5-1	0.3-1	0.1-0.5	C	ORC-1	Δ
16-3vB1N1	0.5-1	0.3-1	0.1-0.5	C	ORC-1	Δ
16-3cti3x17-1	0.5-1	0.3-1	≈10−6	T	wt	wt
16-3cti3x17	0.5-1	0.3-1	≈10−6	T	wt	wt
16-3cti3x377	10−3-10−4	10−4	≈10−6	T	wt	wt
16-3cti3x375	≈10−4	≈10−5	≈10−6	T	wt	wt
16-3v27	≈10−2	≈10−2	10−1-10−2	C	NT	NT
16-3v31	0.5-1	0.3-1	10−1-10−2	C	NT	NT
16-3v38	10−4-10−5	≈10−6	10−1-10−2	C	NT	NT

^aThe EOPs for R. meliloti 41 and its derivatives carrying the vector plasmid pLAFR1, pLAFR3, or pBBR1MCS-2 were identical and were taken as a reference (i.e., EOP = 1). For these assays, phage titers were set to 10⁹ particles per ml on an R. meliloti 41 lawn.

^bR. meliloti 41 generates similar results with either pCS271 or pDH1, as well as with either pCS337 or pCS338.

^cT, turbid plaques; C, clear plaques.

^dwt, wild-type operator; O_R^C-1, mutant operator allele; Δ, operator deleted; NT, not tested.

Analysis of the sequence of the EcoRI LH region.

In the 5,245-bp EcoRI(37)-EcoRI(43) segment of 16-3 DNA (Fig. 1A), 61 ORFs with 50 or more codons were detected by the EMBOSS program (http://www.bioinformatics.abc.hu). Amongthese only two significant homologies were found: (i) the ORF from bp 1 to 1476 gave a score of 176 (bits) with the Sinorhizobium rkpY putative protein kinase gene (this value corresponded to ∼60% homology at the local stretch), and (ii) the ORF from bp 2599 to 3138 gave a score of 96 (bits) with the Xanthomonas carboxypeptidase gene. Homology was found at very low scores (bits) for the rest of the 59 ORFs (37, 54, and 61 for 3 ORFs and below 35 for 56 ORFs). Thirteen putative promoter sites and 15 putative transcription terminator sites were predicted by using a program available at http://www.fruitfly.org/seq-tools/promoter.html and by the Genetics Computer Group (GCG) Wisconsin package (available at http://www.embnet.abc.hu). Further details are available upon request.

The gene for ImmX function is assigned to two overlapping cistrons.

Products encoded by pCS337 and pCS338 (in trans) provided ImmX function, while a 4-bp insertion in ORF116 or ORF127 eliminated the above function. In order to determine the active product(s), R. meliloti 41 strains carrying pCS338 and pCS341 were tested for protein expression. Two proteins expressed from pCS338 were identified by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and confirmed by N-terminal sequencing (data not shown). The proteins corresponded to ORF116 and ORF127. ORF127 and the corresponding protein followed the general rule (i.e., ATG for the start), while the 116-amino-acid protein and the cognate ORF (ORF116) started at a CTG (Leu) codon, frequently utilized in rhizobia. No proteins were detected from the pCS341 allele.

In order to provide further evidence of the action of the two proteins, various mutations were introduced at specific sites in the two ORFs (Fig. 1B). In pCS379, ORF127 was destroyed by a nonsense (opal) mutation at codon 45 (arginine, CGA→TGA), while this change led to a synonymous codon in ORF116 (serine, TCG→TCA). In pCS436, ORF116 was destroyed by an amber mutation at codon 10 (glutamic acid, GAG→TAG), while this mutation led to a synonymous codon in ORF127 (leucine, CTC→CTA). In pCS341, both ORF127 and ORF116 were destroyed by a 4-bp insertion leading to a frameshift mutation. Since the ImmX function was inactivated by these changes, we concluded that two cistrons (and two proteins) were involved in ImmX activity: X_U and X_L (corresponding to ORF116 and ORF127, respectively [Fig. 1B]). The results from the protein expressions are fully in agreement with these mutation analyses, confirming the existence of two products (data not shown). It is noteworthy that no homologous sequences for X_U and X_L have appeared so far in the EMBL, SwissProt, and PDB databanks.

Km^r transducing phage mutants in which the immX gene is knocked out.

Km^r cassettes were inserted at the nonoverlapping regions of X_U and X_L in order to disrupt the action of X_U (pCS463) or X_L (pCS469). These allelic variants were then transferred to the 16-3 chromosome (for details, see Materials and Methods), resulting in Km^r transducing phage mutants (16-3cti3Km^R-U463 and 16-3cti3Km^R-L469). Similarly, a third mutant allele of immX was also constructed, but the Km^r cassette was inserted into the overlapping region of X_U and X_L (it was first inserted into plasmid pCS367 and then transferred to the phage chromosome by recombination), destroying both cistrons (16-3cti3Km^R6-1).

These mutations were not lethal, since the phages formed normal turbid plaques on R. meliloti 41 with an appearance somewhat different from that of 16-3cti3 (a strong, point-like turbid center with a pale turbid belt around it). Upon lysogenization, the three phages were able to integrate into the bacterial chromosome via the 16-3 int/att integrative recombination pathway. Phage integration was confirmed by PCR amplification of attL and attR by use of the primers listed in Table 3 (40); i.e., both attL and attR appeared in the lysogenic cells. Although the immC region was intact in the immX mutants, the lysogenic R. meliloti 41 derivatives which carried them as prophage have shown sensitivity to superinfection of homoimmune phages. Phages such as 16-3cti3 or 16-3cti4Sp4 plated with high EOPs (>0.1) in R. meliloti 41 lysogenic for immX mutant phages. Furthermore, these immX mutants were not immunity insensitive, i.e., they did not grow on either the lysogenic strain R. meliloti 41(16-3cti3) or R. meliloti 41(pCS271).

Isolation of double lysogens for complementation tests.

Two kinds of double lysogenic R. meliloti 41 strains were constructed with respect to the X_U and X_L alleles: one for cis arrangement (X_U+L+/X_U−L−) and the other for trans arrangement (X_U−L+/X_U+L−).

For the “cis” double lysogen, a lawn of the lysogenic strain R. meliloti 41(16-3cti3Km^R6-1) (i.e., X_U−L− genotype) was spotted with the clear-plaque mutant 16-3cti4Sp4 (i.e., X_U+L+ genotype). After a 36-h incubation, massive bacterial growth was detected at the site of the spot. Five independent colonies were then isolated, grown (with the remaining phages removed by serial washings), and then tested (i) for the two prophage, (ii) for the alleles of the X_U and X_L cistrons, and (iii) for immunity to superinfection (see the detailed description in the next section for genetic complementations). All five isolates behaved alike. As expected, the double lysogens liberated massive bursts after heat induction, consisting of both clear- and turbid-plaque-forming progenies in comparable amounts. DNA isolated from them supported the amplification of those fragments from the X_U/L region, which could be obtained only from the two parental prophages (i.e., no traces of recombinations were detected in the immX region).

For the “trans” double lysogen, a lawn of the lysogenic strain R. meliloti 41(16-3cti3Km^R-L469) (i.e., X_U+L− genotype) was spotted with clear-plaque mutant phage 16-3cti4Sp4Km^R-U463 (i.e., X_U−L+ genotype). Here, too, as with the cis double lysogen, massive bacterial growth was observed at the site of the spot. From this point on, the procedure described above for the cis double lysogen was used. All five isolates upon bursting liberated equal amounts of clear- and turbid-plaque formers; PCRs supported exclusively the amplification of those fragments of the X_U/L regions which were inherited from the parental prophages, and again no traces of recombination were recorded.

These observations have proven that (i) the double lysogens were stabilized by the (trans) dominance of the cti3 allele over the nonfunctioning allele cti4Sp4 and that (ii) only the “parental” allele combinations of X_U and X_L were in the cells; hence, the trans double lysogens, which were to provide the critical answer for the complementation (see the next section), were not biased by X_U+L+ wild-type recombinants, i.e., by functional gene copies.

Genetic complementation between X_U and X_L mutants.

Our conclusion that ImmX function was due to two cistrons, X_U and X_L, was probed by genetic complementation tests with the Km^r transducing (“immX knockout”) phage mutants carrying Km^R-U463, Km^R-L469, and Km^R6-1 insertions and X_U/L wild type alleles. As described in the preceding section, two different double-lysogenic R. meliloti 41 strains were constructed: R. meliloti 41(16-3cti4Sp4 16-3cti3Km^R6-1) for cis heterozygotic arrangement of the X_U and X_L alleles and R. meliloti 41(16-3cti4Sp4Km^R-U463 16-3cti3Km^R-L469) for trans heterozygotic arrangement. The two double lysogens were tested for superinfection immunity, i.e., for the ImmX function. The cis and trans heterozygotes were immune to superinfection of 16-3cti3 and 16-3cti4Sp4 but sensitive to the tester ImmX-insensitive 16-3v17-1 strain. Hence, these tests confirmed that the two cistrons (X_U and X_L) were in the genetic background of the ImmX function, since X_U and X_L mutants complemented each other in trans, i.e., the ImmX function was completely restored in the trans double lysogen (Fig. 2)

FIG. 2.

Schematic structures of double lysogens. Inverted filled triangles, sites of Km^r cassette insertion; cti3 and cti4Sp4, alleles of the c cistrons forming turbid and clear plaques, respectively. (A) cis double lysogen (X_U+L+/X_U−L−); (B) trans double lysogen (X_U−L+/X_U+L−).

Mapping of mutations leading to ImmX insensitivity.

After proving the involvement of the two cistrons in ImmX function, we searched for target elements for the ImmX repressors. We were especially interested in sites which might be near the immX genes. These should be sites other than avirT_1-9, which mapped on the opposite half of the 16-3 chromosome (6, 12). Since the immX genes were tagged physically with the Km^r cassette, and consequently the Km^r cassette was to be useful as a genetic marker for these genes in phage crosses, we hoped that mutations linked to immX could be identified at the sequence level. We tested the vir mutants (see Materials and Methods) for growth on strains carrying the cloned immX genes or the immC gene. Tester bacterial strains were R. meliloti 41(pCS271), R. meliloti 41(pDH1), R. meliloti 41(pCS337), and R. meliloti 41(pCS338) for immX and R. meliloti 41(pDH114) for immC. It was observed that 16-3 vir mutant stocks lysed these bacterial lawns, while wild-type 16-3 strains and the clear-plaque-forming mutant 16-3cti4Sp4 carrying a mutation in the c cistron did not (in agreement with earlier studies [6, 12]). Mutant 16-3avc17, which carried the operator mutation O_R^C-1 along with a deletion (which eliminated half of the c cistron and the O_L operator), lysed only the pDH114-carrying host. The transparency and size of plaques and the EOPs varied widely for the vir mutants in the presence of the ImmX function (Table 4).

In order to assign this ImmX sensitivity-insensitivity phenotype to a major region of the 16-3 chromosome, three-point mappings were done by crosses between vir stocks 16-3v17, 16-3v17-1 (ImmX insensitive), and 16-3cti3Km^R6-1 (ImmX sensitive). First, recombinants were scored by Southern hybridization assays for intact immC regions (DNA marker for the 16-3cti3Km^R6-1 parent) and for the intact immX gene (i.e., absence of the Km^r cassette; DNA marker for the vir parents). Out of 108 randomly chosen progeny in the 16-3v17-1 cross, 14 had both immC and immX intact. The recombination frequency for the 16-3v17 cross was 4 of 72. The recombinants were then tested for the ImmX sensitivity-insensitivity phenotype (i.e., this being the nonselective marker). All the recombinants (14 of 14 and 4 of 4) grew on the ImmX-expressing strain R. meliloti 41(pCS271) and formed turbid (cti3) plaques on R. meliloti 41 (i.e., they had immC intact and were ImmX insensitive, the phenotype of 16-3cti3x17 and 16-3cti3x17-1). This outcome was compatible with a linkage of ImmX sensitivity-insensitivity and the left flanking region of the Km^r marker in 16-3cti3Km^R6-1. Furthermore, the manifestation of the ImmX insensitivity phenotype of these recombinants indicates independence from the function of the immC region (Fig. 3A). Mutant pairs, one with an intact immC region, the other with the immC region deleted, plated with equal efficiency in ImmX-expressing R. meliloti 41 strains (Table 4).

The results of the phage crosses were confirmed by marker rescue analyses on the 16-3v17 and 16-3v17-1 phage chromosomes: double crossovers between pDH1 and chromosomes of the above vir stocks resulted in few progeny (4 of 2,500 for 16-3v17 and 5 of 2,500 for 16-3v17-1), and these were unable to grow on R. meliloti 41(pCS271) (i.e., they were ImmX sensitive) (Fig. 3B).

Combinations of mutations for ImmX insensitivity in the vicinity of the immX genes.

Refined mapping of the ImmX insensitivity mutations of 16-3v17 and 16-3v17-1 was carried out by a series of marker rescue experiments where cloned fragments from 16-3v17-1 were probed against the 16-3cti3 phage chromosome. Segments from 16-3v17-1 providing ImmX insensitivity were readily exchanged in the 16-3cti3 chromosome (by double crossover [for details, see Materials and Methods]). The mutations carried by these recombinants were ascribed to a 1,302-bp section of the phage chromosome (Fig. 3C). Sequencing of this section revealed three mutant sites (X_V1, X_V2, and X_V3) outside but in the close vicinity of the X_U/L cistrons: X_V1 X_V2 X_V3 in 16-3v17-1 and 16-3cti3x17-1; X_V1 X_V3 in 16-3v17, 16-3cti3x17, and 16-3vB1N1; X_V1 X_V2 in 16-3cti3x377; and X_V1 in 16-3cti3x375 (Fig. 3D). The same chromosomal segment was isolated by PCR from three vir stocks of independent origin. Combinations of X_V1 and X_V3 sites were identified in the mutants (Table 5). Strong expression of the ImmX-insensitive phenotype was linked to the double mutation X_V1 X_V3. A mutation in X_V2 enhanced the expressivity of X_V1 and X_V1 X_V3 double mutants (Table 4).

TABLE 5.

Mutations detected in ImmX-insensitive stocks

Phagea	Base change (5′ to 3′)b at site:			Level of ImmX insensitivityc
	XV1	XV2	XV3
16-3cti3 (for wt), 16-3avc17	AC	A	C	Sensitive
16-3v17-1, 16-3cti3x17-1	GC	C	T	High
16-3v17, 16-3cti3x17	GC	A	T	High
16-3vB1N1	TC	A	T	High
16-3v31, 16-3v27	AA	A	T	High medium
16-3cti3x377	GC	C	C	Below medium
16-3cti3x375	GC	A	C	Low
16-3v38	AC	A	T	Very low

^awt, wild type.

^bBoldface indicates mutation.

^cSee EOP values.

DISCUSSION

This study provided insight into a complex regulatory region of Rhizobium phage 16-3. The immX region contributes to the function of immunity to homoimmune phage superinfection. We dissected the immX region in two parts, X_U/L and X_V. The X_U/L region contains two overlapping cistrons, X_U and X_L, which code for proteins pX_U and pX_L, respectively. The X_V region functioned as a cognate target for the two proteins. To date, no DNA or peptide sequences homologous to the immX region have been found in the data banks (EMBL, SwissProt, PDB).

Mutations identified in the X_V region overcame the X_U/L repression. Generally, these mutations could define a cognate cis element(s). The sequence around the X_V3 site is comparable to known repressor binding sites (Fig. 4). A CG-to-TA base change at the mutant X_V3 site weakens the palindromic symmetry of the sequence box 5′-ATGGCCGGGCAT-3′ to 5′-ATGGCCGGGTAT-3′. Three copies of this box are separated by 11 and 9 bp on the genome. Consequently, proteins bound to the boxes should lie on the same face along the B-DNA, also providing an opportunity for cooperative interactions. The sequences for X_V3 (as well as for X_V1 [see below]) are not related to the operator sites O_R and O_L, which bind the 16-3 C repressor (i.e., 5′-ACAA-4/6 bp-TTGT-3′ [33]).

FIG. 4.

Putative functions in the X_V region: X_V1 protein binding site sequence (at the right), transcription terminator directing leftward (full sequence), X_V3 protein binding sites, and putative promoter (P) with −35 and −10 boxes. Boldfaced letters indicate mutant positions and mutational changes. The DNA logo (38) of the sites is emphasized for X_V3.

The sequence around the X_V1 site (i.e., 5′-CGACCGATCGCTGTCGTTTTATT-3′) can be evaluated in two ways (Fig. 4). The first 16-bp segment contains a strong palindromic symmetry, like X_V3, and may also be a protein binding site. However, the whole sequence, if transcribed in mRNA (bottom strand), could be folded into a rho-independent transcription termination structure. The mutations found at the 13th (G to T) and 14th (T to C or A) positions would weaken both the symmetry of a putative binding site and the stability of a putative stem-loop.

Phages with the sole mutation at the X_V2 site have not been isolated. The overlapping DNA sequence did not provide a hint for the function of X_V2. Phenotypically, the X_V2 mutation enhanced the expression of ImmX insensitivity when added to either a single mutation in X_V1 or a double mutation in X_V1 and X_V3 (Tables 4 and 5). The phenotypic change caused by mutant X_V2 was mild according to the EOP assays (Table 4), although its effect was very strong for plaque morphology. Mutants with an X_V2 mutation (X_V2 X_V1 versus X_V1, and X_V3 X_V2 X_V1 versus X_V3 X_V1) formed significantly bigger and more-transparent plaques in the lawn of ImmX-expressing bacteria.

Mutations in X_U or X_L destroyed the ImmX-controlled immunity function. The immunity function was restored in R. meliloti 41(16-3cti3Km^R-L469 16-3cti4Sp4Km^R-U463) double lysogens, i.e., the X_U and X_L mutant prophages complemented each other in trans. Our genetic analyses were compatible with the fact that both X_U and X_L cistrons were directly involved in the ImmX repression. Whether the cistrons act independently in a simple additive way or cooperatively by formation of complex structures remains open for further studies.

A number of temperate phages possess two repressor regions (see the introduction). One of the regions, often named immC (after the clear-plaque phenotype associated with it), shows structural similarity and sequence homology to the λcI region. The second immunity regions (immX in 16-3 and immI in P22) seem to be more diverse. The second immunity region of 16-3, immX, is unique. It does not show structural similarity or sequence homology to immunity regions of other phages with dual-control regions.

In this study we focused only on the immX region of 16-3 and did not investigate its functional connections with the immC region. Plaques of immX mutants are turbid (an indication of ImmC activity), while those of immC mutants and immC immX mutants are clear. This result is compatible with the ImmC function being epistatic to the ImmX function; that is, the rules of epistases may indicate that ImmC acts downstream of ImmX in the pathway toward the development of the lysogenic state. However, our results do not show how the roles for ImmX and ImmC are shared in establishing and maintaining lysogeny. The heat inducibility of 16-3cti3 lysogens (i.e., carrying immC temperature-sensitive [ts] mutant prophages) and the lack of immunity to superinfection of the lysogens carrying X_U/L mutant prophages (i.e., Km^r insertion mutants) show that the functions of immC and immX are involved in at least maintaining lysogeny.

The avirT locus may be a place where both the C and X repressors interact, since mutant 16-3c⁺avirT_1-9 escaped from the activity of immX and (to a lesser extent) from that of immC (12). To date there is no handle for the function of avirT. Mutant 16-3c⁺avirT_1-9 formed turbid plaques like the 16-3cti3 X_V mutants; however, unlike these, the former made very unstable R. meliloti 41 lysogens. The turbid plaques of 16-3c⁺avirT_1-9 cleared up after 2 to 3 days, and its lysogenic broth cultures lysed (6). From the genetic analyses of the lysogeny of 16-3, it can be deduced that in order for 16-3 to efficiently overcome the immunity of the lysogenic cells, at least the combination of two mutant elements is required: mutant immC combined with mutant X_V or with mutant avirT. Our next efforts will be directed at detailed functional dissection of these two regulatory regions.