Constitutive Expression of a Transcription Termination Factor by a Repressed Prophage: Promoters for Transcribing the Phage HK022 nun Gene

Abstract

Lysogens of phage HK022 are resistant to infection by phage λ. Lambda resistance is caused by the action of the HK022 Nun protein, which prematurely terminates early λ transcripts. We report here that transcription of the nun gene initiates at a constitutive prophage promoter, P_Nun, located just upstream of the protein coding sequence. The 5′ end of the transcript was determined by primer extension analysis of RNA isolated from HK022 lysogens or RNA made in vitro by transcribing a template containing the promoter with purified Escherichia coli RNA polymerase. Inactivation of P_Nun by mutation greatly reduced Nun activity and Nun antigen in an HK022 lysogen. However, a low level of residual activity was detected, suggesting that a secondary promoter also contributes to nun expression. We found one possible secondary promoter, P_Nun′, just upstream of P_Nun. Neither promoter is likely to increase the expression of other phage genes in a lysogen because their transcripts should be terminated downstream of nun. We estimate that HK022 lysogens in stationary phase contain several hundred molecules of Nun per cell and that cells in exponential phase probably contain fewer.

Only a few prophage genes are expressed in lysogens. Some, such as the genes encoding prophage repressors, prevent the expression of genes that lead to cell death and prophage loss. Others, such as phage λ lom, bor, rexA, and rexB (2, 6), phage P22 sieA, sieB, and mnt (30, 31), phage P1 res and mod (35), phage P2 old and tin (9, 18), and prophage genes encoding virulence factors in a variety of pathogenic bacteria (see reference 32), have diverse functions. It is believed that the products of most of these genes confer a selective advantage on the lysogen, or, like repressors, prevent prophage curing and lytic phage growth. In some cases, for example λ rexA (16), the expressed genes are cotranscribed with the repressor gene. In others, for example, P2 old (9), transcription initiates at dedicated promoters. In both cases downstream transcription terminators can prevent cotranscription of other prophage genes that might harm the host or destabilize the prophage. In phages of the λ family, transcription antitermination mechanisms that function during lytic growth allow full expression of genes located downstream of the terminators only when they are needed for phage production (see reference 34).

Lysogens of temperate phage HK022 produce Nun, a protein that confers resistance to (or excludes) the related phage λ (24). Although the nun gene is located immediately downstream of p_L, a major HK022 early promoter, this promoter can be inactivated by mutation without preventing nun expression (3). A second promoter, p_RM, is located upstream of p_L (Fig. 1). Transcripts initiating at p_RM in a lysogen direct the synthesis of prophage repressor, which is encoded by the cI gene. Previous results showed that at least some p_RM-directed transcripts extend through the repressed p_L promoter, which suggested that nun is cotranscribed with cI (3).

FIG. 1.

Genetic map of the nun-cI region of HK022 (not to scale). Promoters and the direction of transcription are indicated by bent arrows. T_IS903 is a transcription termination site within an insertion element found in our original isolate of HK022 (21). In a lysogen, p_L is repressed and p_RM is activated by the product of the cI gene. The map underneath is an expanded view of the region that contains two new promoters identified in this study (P_Nun and P_Nun′). DNA fragments from positions 2 to 72, positions 2 to 174, and positions 151 to 247, relative to the HK022 p_L transcription start point, were fused to a promoterless lacZ gene. The corresponding β-galactosidase (β-gal) activities (numbers at left) were measured in exponentially growing cells carrying multicopy fusions. The relative strengths of P_Nun, P_Nun′, and derepressed p_L (but not p_RM) are suggested by the thickness of the corresponding arrows. The binding site for oligonucleotide PM1, which was used to map the P_Nun and P_Nun′ transcription start sites, is shown.

A potential problem with this idea is that nun transcripts initiating at p_RM will read through putL, a cis-acting transcription antitermination site that lies between p_RM and the beginning of nun (Fig. 1). Nascent putL transcripts modify elongating RNA polymerase molecules so that they read through downstream transcription termination sites (14). Indeed, the activity of putL is required for p_L-directed expression of phage genes located downstream of nun during lytic phage growth (20). Therefore, if nun were transcribed from p_RM in a lysogen, antitermination could lead to transcription of these downstream genes. This is likely to be harmful because some of the resulting proteins can reduce the fitness of the lysogen or destabilize the prophage (see Discussion). Accordingly, we looked for nun promoters located downstream of putL. In this article, we report that nun is transcribed from a dedicated promoter located between putL and the coding sequence, and that transcription from p_RM contributes little if anything to nun expression.

MATERIALS AND METHODS

Bacterial strains, plasmids, and bacteriophages.

Bacterial strains, plasmids, and bacteriophages are listed in Table 1.

TABLE 1.

Bacterial strains, plasmids, and bacteriophages

Strain, plasmid, or bacteriophage	Relevant genotype or descriptiona	Comment and/or source or reference
Strains
CF881	rna	Lacks RNase I (36)
JO350	MC1000 (HK022 nun::Tn10)1	21
LM86	MC1000 (HK022::kan PNun′[−10])1	This work
LM87	MC1000 (HK022::kan)1	The prophage is strain O295 (this work)
LM89	MC1000 (HK022::kan Δ[PNun-PNun′])1	This work
MC1000	lacΔX74	Host for HK022 prophages (27)
N5947	lacZΔ21 (λ int2 xis1 Δ[S-X] Nam7 Nam53 cI857 ΔH1[cro-bio-uvrB])	Used as a host for pL(λ)-lacZ fusions (22)
RK422	MC1000 (HK022::kan PNun[−10])1	This work
RK444	MC1000 (λRS88 PNun-lacZ)1	This work
RK445	MC1000 (λRS88 PNun[−10]-lacZ)1	This work
RW4055	W3110 lacZΔM15 lacIq/pNUN	pNUN carries a PLac-nun fusion (1)
RW4072	(argF-lac)ΔU169 recA56 (λ imm-HK022 cIts12 nun::lacZ sam7)n
RW3177	N5947 (λ imm21 pL[λ]-lacZ) (HK022 nun::Tn10)	This work
RW3181	N5947 (λ imm21 pL[λ]-nutL-lacZ) (HK022 nun::Tn10)	This work
RW4089	N5947 (λ imm21 pL[λ]-lacZ) (HK022::kan PNun′[−10])1	This work
RW4091	N5947 (λ imm21 pL[λ]-lacZ) (HK022::kan)1	This work
RW4093	N5947 (λ imm21 pL[λ]-lacZ) (HK022::kan Δ[PNun-PNun′])1	This work
RW4095	N5947 (λ imm21 pL[λ]-lacZ) (HK022::kan PNun[−10])1	This work
RW4097	N5947 (λ imm21 pL[λ]-nutL-lacZ) (HK022::kan PNun′[−10])1	This work
RW4099	N5947 (λ imm21 pL[λ]-nutL-lacZ) (HK022::kan)1	This work
RW4101	N5947 (λ imm21 pL[λ]-nutL-lacZ) (HK022::kan Δ[PNun-PNun′])1	This work
RW4103	N5947 (λ imm21 pL[λ]-nutL-lacZ) (HK022::kan PNun[−10])1	This work
SK38	pK1 in XL1	3
SK37	pJO9.18 in XL1	3
SK11	CF881 (HK022)	3
XL1	recA	Host for plasmids (Stratagene)
Plasmids
pJO9.18	pEMBL-18 derivative that carries the pR-cI-pL-nun segment of HK022 DNA	21; see reference 7 for a description of pEMBL-18
pK1	pJO9.18 with a mutation inactivating pL	3
pLM11	pJO9.18 PNun′(−10)	This work
pLM12	pJO9.18 Δ(PNun-PNun′)	This work
pMOC192	pRS415 containing HK022 sequence from positions +2 to +174 relative to the start site of pL transcription fused to lacZ	M. Clerget, personal communication
pMOC194	pRS415 containing HK022 sequence from positions +2 to +72 relative to the start site of pL transcription fused to lacZ	M. Clerget, personal communication
pNGC25	pGB2 containing the HK022 cI and nun genes	J. Little, personal communication
pRK395	pRS415 with a PNun-lacZ fusion	This work
pRK412	pJO9.18 with the PNun(−10) mutation	This work
pRK438	pRS415 with a PNun(−10)-lacZ fusion	This work
pRS415	Promoterless lacZ expression vector	28
Bacteriophages
O295	HK022 cIts12 with a kanamycin resistance determinant inserted in the pL operon	3
RS88	λ imm434 cIind−	Parent of single-copy PNun-lacZ fusions (28)
W30	λ b2 cI−	NIH collectionb

^aSubscripts indicate the number of tandem prophages, as follows: 1, single lysogen; and n, polylysogen. p_L(λ)-lacZ (fusion HA25) and p_L(λ)-nutL-lacZ (fusion HA22) are located in the b2 region of a modified λ imm21 prophage (24). Fusion HA25, which has 22 bp less of the λ p_L operon than does fusion HA22, lacks the boxB element of nutL. 

^bNIH, National Institutes of Health. 

Phage and bacterial growth.

Bacteria were grown in tryptone or Luria-Bertani (LB) broth and supplemented with ampicillin (100 μg/ml) or kanamycin (50 μg/ml) where appropriate. HK022 was grown and assayed as described in reference 21.

Cloning of fragments with promoter activity.

Oligonucleotides were purchased from BioServe Biotechnologies (Laurel, Md.). Plasmid DNAs were isolated by using a Qiagen spin miniprep kit. Restriction enzymes and ligase were purchased from New England Biolabs (NEB). A 97-bp DNA fragment that immediately precedes the nun gene and contains the P_Nun promoter was amplified from plasmid pNGC25 by PCR with primers RK99 (5′-GGCGGATCCAATAAGCACCGTACGG-3′) and RK100 (5′-CAGCGAATTCAAGTATTTATTGCAAAGATTC-3′). The amplified DNA was digested with EcoRI and BamHI (underlined sequences) and cloned into the promoterless lacZ vector pRS415.

Primer extension.

Total RNA was isolated with an RNA isolation kit (Totally RNA; Ambion) from log-phase cultures of the following strains: SK37, SK38, and SK11 (Table 1). Strains SK37 and SK38 carry plasmids with inserts of HK022 DNA, and strain SK11 is an HK022 lysogen. Seven milliliters of culture was added directly to 10 ml of lysing buffer and then processed according to the Ambion protocol. RNA was suspended in diethyl pyrocarbonate-treated water supplemented with 1 mM EDTA. The purity and integrity of the RNA was analyzed on a 1.2% agarose gel, followed by staining with ethidium bromide. Primer extension reactions (20-μl volumes) were performed by using the Promega reverse transcription system. Oligonucleotide PM1 (5′-CTCGAGATGTAAGACCTC-3′) is complementary to the nun mRNA and hybridizes approximately 90 bp downstream of the P_Nun start site. PM1 was 5′ end-labeled in a 30-μl reaction mixture containing 50 μCi of [γ-³²P]ATP (Redivue [3,000 Ci/mmol]; Amersham) 10 U of T4 polynucleotide kinase (NEB) and 1× kinase buffer (NEB). Reaction mixtures were incubated for 30 min at 37°C. Primer extension reaction mixtures were incubated at 42°C for 1 h, and the reactions were stopped by adding 10 μl of Stop solution (95% formamide, 20 mM EDTA, 0.05% bromophenol blue, 0.02% xylene cyanol). Reaction mixtures were heated at 94°C for 3 min, chilled immediately on ice, and then loaded onto an 8% denaturing sequencing gel. Sequencing reactions were performed with the Perkin-Elmer Amplicycle sequencing kit. Sequence was obtained directly from purified pK1 plasmid and PCR-amplified template from wild-type P_Nun and P_Nun(−10) mutant clones by using 5′ end-labeled oligonucleotide PM1 as the primer.

In vitro transcription.

Multiple round in vitro transcription reactions were performed on templates amplified from P_Nun⁺ and P_Nun(−10) mutant clones. Reaction mixtures consisted of approximately 50 nM PCR template, 100 nM RNA polymerase holoenzyme (Epicentre, Madison, Wis.), 1 mM concentrations of each nucleoside triphosphate (NTP), 20 mM Tris-glutamate, 50 mM K-glutamate, 10 mM Mg-glutamate, and 0.001 μg of bovine serum albumin per ml. The reaction mixtures were incubated for approximately 1 h at 37°C. After ethanol precipitation and washing, the pelleted material was suspended in diethyl pyrocarbonate-treated water supplemented with 1 mM EDTA. The RNA products were then used directly in a primer extension reaction.

P_Nun and P_Nun′ mutations.

To inactivate the P_Nun promoter, plasmid pJO9.18 was digested with BsrGI, and the overhangs were filled in with Klenow fragment. The resulting blunt-ended plasmid was ligated and electroporated into MC1000 cells. Transformants were screened by PCR with primers PM1 and RK100, and the products were tested for their susceptibility to BsrGI digestion. The P_Nun′ promoter was inactivated by replacing the sequence in the predicted −10 hexamer with a BamHI restriction site. The replacement was accomplished by overlap extension PCR with primers PM2 (5′-GCGGATAACAATTTCACACAGG-3′), PM3 (5′-CTTCTATTTTTTCGAACGACTTCTGGATCCAAAAGCTGCTTTGCTTTTTGTGAC-3′), PM4 (5′-CACAAAAAGCAAAGCAGCTTTTTGGATCCAGAAGTCGTTCGAAAAAATAG-3′), and PM6 (5′-CGCCAGGGTTTTCCCAGTCACGAC-3′). The BamHI site that replaces the predicted −10 sequence is underlined in the relevant oligonucleotide sequences. Amplifications were performed with Vent polymerase with pJO9.18 used as the template. The final amplified product was digested with HindIII and EcoRI and ligated into pJO9.18. The desired clones were identified by restriction analysis of purified plasmid DNA. The mutation was transferred onto HK022 by recombination, and the presence of the BamHI site was confirmed by PCR amplification with primers PM1 and PM5 (5′-CCTCATTAGGCAGTCAATCG-3′) followed by digestion of the products with BamHI. Inactivation of both the P_Nun′ and P_Nun promoters was accomplished by digesting the P_Nun′ mutant construct with BamHI and BsrGI. The DNA ends were filled in by Klenow fragment, ligated, and electroporated into MC1000 cells. The desired deletion mutant was identified by analyzing the size of DNA amplified by PCR with primers PM1 and PM5.

Crossing P_Nun and P_Nun′ mutations onto HK022 phage.

Cells carrying pJO9.18 or various mutant derivatives were infected with HK022::Tn10(kan) cIts12 (phage O295), and the resulting lysate was used to infect MC1000 cells. This plasmid carries the wild-type HK022 cI gene but, for unknown reasons, does not prevent HK022 growth (21). The Tn10(kan) insertion is in the p_L operon, downstream from nun. Phage recombinants carrying the plasmid cI⁺ gene were selected by infection of a sensitive host and isolation of Km^r colonies at 42°C, the nonpermissive temperature for cIts12. The presence of the mutation on the prophage was confirmed by the size of PCR-amplified products and the ability to digest the products with the enzymes BsrGI or BamHI. Lysates of recombinant phage were made by induction of the lysogens with UV light. These lysates were used to lysogenize cells containing λp_L-nutL-lacZ fusions (24).

Crossing lacZ fusions onto λRS88.

Promoter fusions in the lacZ reporter vector, pRS415, were transferred onto λRS88 by recombination as described previously (14, 28). Lambda prophage copy number was determined as described previously (23).

Beta-galactosidase activities.

Strains containing thermosensitive λ repressor and a p_L(λ)-lacZ fusion were grown overnight at 32°C in tryptone broth. The overnight cultures were diluted 1:50 in LB broth, grown to approximately 1 × 10⁸ to 2 × 10⁸ cells/ml, and shifted to 42°C for 60 min to allow derepression of the λ p_L promoter. Beta-galactosidase activity was measured as described previously (17). It was important to use toluene to permeabilize the cells because β-galactosidase produced by these fusions was destabilized by the alternative treatment with chloroform and detergent. Cells containing single-copy lacZ fusions were grown overnight at 37°C in tryptone broth. The overnight cultures were diluted 1:1,000 in LB broth and grown at 37°C to an optical density at 650 nm of approximately 0.5 × 10⁸ to 1 × 10⁸ cells/ml. Thereafter, aliquots were removed and placed on ice at various intervals and the β-galactosidase activities were determined as described above.

Determination of HK022 prophage copy number.

The copy number of HK022 prophages was determined essentially as described in reference 23 by using primers specific for the HK022 int and attP and the host attB sites, as follows: RK111 (int oligonucleotide [5′-AGCAATGCAGGGAGGCCAGC-3′]), RK112 (attP oligonucleotide [5′-AATAATCTTGCGGTAGCCAGAC-3′]), and RK113 (attB oligonucleotide [5′-GGAGATCTCCTGCTCCTGTTG-3′]). Single-copy lysogens were identified by the amplification of an approximately 747-bp DNA generated from the attB and int primers. Polylysogens produced the 747-bp fragment in addition to an approximately 613-bp fragment.

Immunoblots (10).

Cells were grown in LB broth at 37°C, chilled, concentrated, diluted into loading buffer, boiled for 5 min, and frozen at −70°C. Aliquots were loaded directly onto sodium dodecyl sulfate Tris-tricine–16.5% acrylamide gels (Bio-Rad), electrophoresed, and electroblotted to Immobilon polyvinylidene difluoride membranes. We found that Nun comigrated with prestained marker proteins (Bio-Rad) in the 14- to 18-kDa range. The membranes were blocked with 5% milk powder and allowed to react overnight with antiserum against partially purified Nun raised in rabbits (antiserum kindly supplied to us by David Friedman [11]). We treated the antiserum before use with an acetone powder prepared from strain MC1000 to remove antibodies to bacterial proteins (10). It was diluted 1:50 into a solution of 5% milk powder for use. Bands were visualized by treating the membranes with peroxidase-coupled anti-rabbit serum for 1 h, allowing them to react with ECL Western blotting detecting reagent (Amersham), and placing them in contact with X-ray film. To quantitatively estimate the amount of Nun, we mixed serial dilutions of purified Nun (a gift of Randy Watnick) with extracts of strain MC1000 and fractionated the mixtures in separate lanes of a gel (see Fig. 3).

FIG. 3.

Sequence of predicted promoters and promoter mutants. The bent arrows indicate the observed 5′ ends of P_Nun and P_Nun′ transcripts (top and middle lines, respectively), as indicated by primer extension analysis (Fig. 2). The predicted −10 and −35 hexamers for these promoters are shown in bold type. The E. coli ς70 consensus hexamers are TATAAT and TTGACA, respectively. The straight arrows indicate the sequence changes in the P_Nun(−10) and P_Nun′(−10) mutants (see Materials and Methods). The sequence Δ[P_Nun-P_Nun′] is shown in the third line (see Materials and Methods). The slashes indicate the fusion points between P_Nun′(−10) sequence (left) and P_Nun sequence (right). Possible alternative −10 and −35 hexamers are shown in bold type.

RESULTS

Identification of promoters located between putL and nun.

We have discovered two promoters located between putL and nun. These were found by extending an oligonucleotide primer that was hybridized to RNA isolated from cells that carry plasmid clones of HK022 DNA. Extension of the primer, which is complementary to the beginning of the nun coding sequence, produced a strong signal corresponding to a 5′ RNA end close to the start of the nun coding region and a weak signal corresponding to a 5′ RNA end further upstream, but still downstream of putL (Fig. 2). We also saw the strong primer extension signal with RNA isolated from an HK022 lysogen, although the amount of product was much less in this case (data not shown) (we did not see the weak signal, as would be expected if its intensity were correspondingly reduced). We call the putative promoter corresponding to the strong downstream signal P_Nun, and we called the putative promoter corresponding to the weak upstream signal P_Nun′.

FIG. 2.

Mapping of the transcription start sites of the P_Nun and P_Nun′ promoters. Oligonucleotide PM1, complementary to the nun mRNA, was used for reverse transcriptase primer extension and sequencing. The DNA sequence of the promoter −10 regions is shown on the left. The arrows mark the position of the extension products (lane 1). The predominant initiation sites are marked by asterisks beside the DNA sequence (lanes G, A, T, and C).

Short DNA segments containing either of these putative promoters had promoter activity when fused to a promoterless lacZ reporter gene in a multicopy plasmid (Fig. 1). A 97-bp DNA fragment containing the presumed start point of P_Nun-directed transcription accumulated about 3,000 U of β-galactosidase during exponential growth of a culture carrying the plasmid-borne fusion, suggesting that this interval contains a functional promoter (plasmid pRK395, the segment from positions +151 to +247 relative to the p_L start). A 173-bp DNA fragment containing the presumed start point of P_Nun′-directed transcription accumulated about 600 U of β-galactosidase under the same conditions (plasmid pMOC192, the segment from positions +2 to +174 relative to the p_L start). A similar plasmid clone lacking the P_Nun′ start accumulated about 17 U of β-galactosidase (pMOC194, the segment from positions +2 to +72 relative to the p_L start), suggesting that the segment from positions +73 to +174 contains at least part of a promoter.

A computer-assisted search (19) identified a candidate promoter in the P_Nun region and a second one in the P_Nun′ region. The sequences, predicted −10 and −35 hexamers, and observed start sites are shown in Fig. 3. Although the match to E. coli consensus ς70 promoters is poor, the location of the predicted promoters is completely consistent with the functional analysis described in the preceding paragraph.

P_Nun and P_Nun′ mutations.

To see if these promoters contribute to Nun production in an HK022 lysogen, we mutated them individually or together and measured the activity and intracellular concentration of Nun. In mutant P_Nun(−10), the predicted −10 hexamer of P_Nun was changed, in mutant P_Nun′(−10), the predicted −10 hexamer of P_Nun′ was changed, and in mutant Δ[P_Nun-P_Nun′], the DNA segment extending from the (altered) −10 hexamer of P_Nun′(−10) to the −10 hexamer of P_Nun was deleted (Fig. 3). We showed that P_Nun(−10) had substantially reduced promoter activity: the steady-state level of β-galactosidase activity in cells carrying a single copy of a P_Nun(−10)-lacZ fusion was 1 to 3% of that of a comparable P_Nun-lacZ fusion (strains RK445 and RK444, respectively) (data not shown). The P_Nun′(−10) and Δ[P_Nun-P_Nun′] mutations were not checked in this way, and we have no experimental evidence that they inactivate the respective promoters. Indeed, inspection of the sequence revealed Δ[P_Nun-P_Nun′] could have created a new promoter (Fig. 3). The three mutations were crossed from the plasmids in which they were constructed onto HK022 as described in Materials and Methods, and single lysogens of the mutant phages were isolated. The Nun phenotype of these lysogens was checked in several ways.

Plaque formation by λ on lawns of HK022 lysogens (Table 2).

TABLE 2.

Phenotypes of P_Nun and P_Nun′ mutations in single-copy HK022 lysogens^a

Phenotype			λ Plaque formation (EOP)b	β-Galactosidase activity
PNun	PNun′	nun		pL(λ)-nutL-lacZc	pL(λ)-lacZd
+	+	+	<10−8	20	1,349
+	(−10)	+	<10−8	37	1,690
(−10)	+	+	∼10−2e	2,069	1,445
Δ	Δ	+	∼10−3e	408	1,543
+	+	Tn10	1.0	1,735	1,281

^aThe nun::Tn10 insertion mutation and the two p_L(λ)-lacZ fusions have been described previously (21, 24). The phage used for measuring the efficiency of plating (EOP) was λ b2 c (strain W30), and efficiencies are relative to that on lawns of MC1000, the nonlysogenic host. Beta-galactosidase activity was measured 1 h after thermal derepression of the λ p_L promoter, and values are averages obtained from at least two independent determinations. The standard error of the mean was ≤14%. 

^bThe bacterial strains used were (from top to bottom) LM87, LM86, RK422, LM89, and JO350. 

^cThe bacterial strains used were (from top to bottom) RW4099, RW4097, RW4103, RW4101, and RW3181. 

^dThe bacterial strains used were (from top to bottom) RW4091, RW4089, RW4095, RW4093, and RW3177. 

^eThere was a background of many tiny plaques, which were not included in the estimate of plating efficiency. 

A wild-type HK022 prophage reduced the number of plaques formed by superinfecting λ by more than 8 orders of magnitude. No λ mutants able to form plaques were found, probably because they cannot arise in a single step (25). However, the HK022 P_Nun(−10) mutant prophage excluded λ much less efficiently than did its parent. We observed a large number of tiny, uncountable plaques at a frequency of ≥10⁻² plaques per added λ phage. These are probably formed by limited growth of wild-type λ on this host. We also found plaques of normal or near-normal size at a frequency of about 10⁻² plaques per added phage. Most or all of these were formed by mutants that probably arose during limited λ growth on the lawn (see below). The residual λ exclusion by the P_Nun(−10) lysogen is clearly due to Nun activity, since a lysogen of a nun::Tn10 insertion mutant plated λ with unit efficiency, and the plaques were of near-normal size. The phenotype of a Δ[P_Nun-P_Nun′] lysogen was similar to that of the P_Nun single mutant, although, surprisingly, the frequency of normal-sized (i.e., mutant; see below) λ plaques was somewhat lower for the former (ca. 10⁻³ plaques per added phage). These results suggest that P_Nun makes a major but not the only contribution to nun transcription in a lysogen. The quantitative difference in λ exclusion between P_Nun(−10) and Δ[P_Nun-P_Nun′] can be accounted for by assuming that the deletion mutation creates a new promoter that is weaker than P_Nun but stronger than P_Nun(−10) (see above). The P_Nun′(−10) lysogen excluded λ as efficiently as did its wild-type parent, but this result cannot be unambiguously interpreted because we have no experimental evidence that the mutation inactivated the promoter.

Most or all of the normal-sized λ plaques that arose on lawns of HK022 P_Nun(−10) and Δ[P_Nun-P_Nun′] lysogens were formed by mutants. Phage from two such plaques recovered from lawns of each lysogen were subjected to several cycles of single-plaque purification on lawns of a nonlysogenic host and then amplified on the same host. We found that each of these phage lines formed plaques with unit efficiency on lawns of both the P_Nun(−10) and the Δ[P_Nun-P_Nun′] lysogens. However, they did not form plaques on wild-type or P_Nun′(−10) lysogens (efficiency of plating, ≤10⁻⁶), and they have not been further characterized.

Effect of promoter mutations on the expression of a p_L(λ)-nutL-lacZ fusion.

Nun blocks λ growth by binding to the boxB element of nascent nut transcripts and prematurely terminating early λ transcription (5, 24) (see Discussion). Thus, measurement of β-galactosidase production in a cell containing a nut site fused to a reporter gene gives a direct and quantitative estimate of Nun activity. As shown in Table 2, the wild-type Nun concentration greatly reduced β-galactosidase accumulation by the p_L(λ)-nutL-lacZ fusion but had no effect on a similar fusion that lacks a complete boxB. This agrees with previously published results (24). The P_Nun(−10) mutation restored a high level of β-galactosidase activity, similar to that observed in a nun insertion mutant. These results confirm that P_Nun makes a major contribution to nun expression. In contrast to measurement of phage exclusion, we saw no evidence of any residual Nun activity in this mutant. We propose an explanation for this difference in the Discussion. The Δ[P_Nun-P_Nun′] mutation gave an intermediate level of lacZ expression compared to the wild-type and the P_Nun mutant lysogens, a result that is qualitatively consistent with its intermediate effect on λ exclusion. The P_Nun′(−10) mutation caused a small but reproducible increase in the expression of the nutL-lacZ fusion.

Effect of promoter mutations on the level of Nun antigen.

We measured the effects of our promoter mutants on the intracellular concentration of Nun protein in HK022 lysogens by immunoblots. We found that the P_Nun(−10) and the Δ[P_Nun-P_Nun′] mutations reduced Nun protein concentrations to undetectable levels (Fig. 4). We estimate that we would have been able to see as little as 25% of the level of Nun found in a wild-type lysogen. By contrast, the P_Nun′(−10) mutation had no detectable effect on Nun level. The simplest conclusion from these results is that P_Nun is the major promoter directing nun transcription in a lysogen. We consider the source of the residual phage exclusion activity by the P_Nun(−10) prophage in the Discussion.

FIG. 4.

Detection of Nun in HK022 lysogens by immunoblotting. Cell extracts of strains lysogenic for wild-type or mutant HK022 prophages, of a nonlysogen, or of a strain carrying an induced P_Lac-nun fusion (as indicated below the gel image) were fractionated by electrophoresis in a denaturing gel. The Nun phenotype refers to the λ exclusion data of Table 1. Nun was detected with antibody as described in Materials and Methods. The position of Nun is indicated by the arrow. Its mobility relative to protein standards is consistent with a previous report (21), and the identification is confirmed by the large amount of protein with identical mobility extracted from a Nun-overproducing strain (lane 7). Lanes 1 to 6 each contained 0.02 to 0.025 mg of protein, and lane 7 contained about half that amount. Lane 1, P_Nun′(−10) (strain LM86); lane 2, P_Nun⁺ (strain LM87); lane 3, Δ[P_Nun-P_Nun′] (strain LM89); lane 4, P_Nun(−10) (strain RK422); lane 5, nun::Tn10 (strain JO350); lane 6, strain MC1000 (the nonlysogenic parent); lane 7, P_Lac-nun (strain RW4055). The first six strains were grown to 1 × 10⁹ to 2 × 10⁹ cells per ml and the seventh strain was grown to 4 × 10⁸ to 6 × 10⁸ cells per ml in LB broth at 37°C before extraction. Isopropyl-β-d-thiogalactopyranoside (IPTG) (1 mM) was added to the culture of RW4055, which contains a plasmid with a P_Lac-nun fusion, 50 min before the cells were harvested. The production of Nun by this strain was much reduced if IPTG was omitted (data not shown).

Further analysis of P_Nun.

We estimated the amount of Nun per cell by comparing immunoblots of lysogens to those of known quantities of purified Nun protein (Fig. 3). There are 120 to 360 molecules of Nun per cell in a culture grown to 1 × 10⁹ to 2 × 10⁹ cells/ml in LB broth. This number is roughly consistent with the activity of β-galactosidase per cell in comparable cultures of an HK022 lysogen containing a nun::lacZ translational fusion (strain RW4072) (the calculation assumes that the fusion protein is as active as native β-galactosidase [26]). The specific activity in this strain and another strain with a P_Nun-lacZ transcriptional fusion (RK444) increased from three- to sixfold as the cultures proceeded from mid-log phase (0.5 × 10⁸ to 1.5 × 10⁸ cells/ml) into stationary phase (1 × 10⁹ to 2 × 10⁹ cells/ml). The mechanism and biological significance of this increase are unknown.

To determine if any factors are required for P_Nun activity, we analyzed transcripts generated from P_Nun and P_Nun(−10) templates using purified E. coli RNA polymerase holoenzyme (E-ς70). In a multiround transcription reaction that included linear wild-type template and radioactive substrate, a runoff transcript of the predicted size was observed (data not shown). The corresponding transcript from the P_Nun(−10) mutant template was not detected. We were also unable to see a transcript in similar reactions with a template that contained P_Nun′. This result is consistent with our primer extension experiments and with results from lacZ fusions that show that P_Nun′ is less active than P_Nun. Finally, we determined that purified E-ς70 utilizes the same transcription start site for P_Nun in vitro as it does in vivo by performing primer extension analysis on RNA generated in multiround transcription reactions (data not shown). These results suggest that no additional factor besides E. coli polymerase containing the ς70 subunit is absolutely required for transcription from P_Nun.

DISCUSSION

Our experiments show that P_Nun is the major promoter for expression of the nun gene in an HK022 lysogen. This promoter thus joins a fairly short list of identified promoters that are active in repressed prophages. We estimate that there are several hundred molecules of Nun per cell in early-stationary-phase cells, and perhaps one-sixth to one-third that number in exponentially growing cultures. The mechanism and biological significance of this growth phase regulation are unknown. Since Nun is stable in vivo (21), its low level probably reflects, at least in part, the weakness of P_Nun. Measurements of the activity of comparable promoter-lacZ fusions suggest that P_Nun is about 1% as strong as fully derepressed HK022 p_L (R. A. King and R. A. Weisberg, unpublished results). This result is qualitatively consistent with the relative strengths of the two promoters in vitro. The ς70 consensus is not well conserved in P_Nun, and the transcript is unusual in that it starts with a U residue. These factors may contribute to the low activity of this promoter.

Nun acts by binding stoichiometrically to a stem-loop formed by the transcript of the boxB element in the nut sites of the nascent λ p_L and p_R transcripts (5). It then interacts with and arrests elongating RNA polymerase (12, 13, 33). The arrested transcription elongation complexes are stable in vitro, but the arrested polymerase probably disassociates from the template in vivo (24, 25, 29). The efficiency of Nun action is remarkable: it completely blocks λ lytic growth and cell killing and reduces to about 1% the expression of reporter genes that are fused to λ p_L or p_R and preceded by either nutL or nutR, respectively. The reduction persists for more than 1 h after derepression (24) (Table 1). λ p_L directs the synthesis of 10 to 20 transcripts/min in vivo (15), and the strength of λ p_R is likely to be similar. It is unlikely that a few hundred molecules of Nun could arrest such a large number of transcription elongation complexes if Nun were consumed during the reaction. Therefore, we suggest that Nun bound to arrested elongation complexes is released in active form after termination and recycles to newly synthesized transcripts where it acts again.

RNA polymerase molecules that initiate at P_Nun do not transcribe the putL antitermination site (Fig. 1) and should therefore efficiently terminate at downstream transcription terminators. One of these terminators is within the IS903 insertion element located immediately downstream of nun (8, 21). Termination of nun transcripts is probably advantageous to HK022 and its lysogens because the prophage kil, cIII, xis, and int genes lie downstream of transcription terminators in the p_L operon, and the products of these genes are likely to destabilize the prophage or harm the lysogen. As expected, nun can also be expressed from the p_L promoter: we found that more than 10 times the amount of β-galactosidase was produced by a p_L-P_Nun-nun::lacZ gene fusion in the absence than in the presence of repressor (unpublished results). However, lytic growth of HK022 does not require Nun (21).

Mutation of P_Nun depressed but did not completely prevent Nun production. Nun was undetectable by immunoblot (≤25% of wild type) and did not reduce expression of a p_L(λ)-nutL-lacZ fusion in a P_Nun(−10) lysogen. Nevertheless, it did reduce growth of superinfecting λ (Table 2). Lambda growth could be more severely affected by low Nun levels than is expression of a reporter gene fused to nutL because the phage also has a nutR site, and this could amplify its sensitivity. The source of the residual Nun in the P_Nun mutant is unknown, but several possibilities are open. First, P_Nun(−10) retains 1 to 2% of P_Nun⁺ promoter activity, as judged by expression of a reporter gene fused to the promoter, and this could be enough to impede the growth of superinfecting λ. Second, P_Nun′ could be a source of nun transcripts. We have no independent evidence that the P_Nun′(−10) mutation inactivates P_Nun′, and we were surprised to find that the Δ[P_Nun-P_Nun′] mutant retained considerable Nun activity. An inspection of the fused sequence suggests that this deletion creates a new promoter. Third, p_L could contribute to residual nun transcription if repression is incomplete. Finally, p_RM could contribute to residual nun transcription.

With respect to the last possibility, Cam et al. (3) detected low-intensity transcription initiating upstream of p_L and extending into putL in HK022 lysogens. They also found that a high level of HK022 repressor supplied in trans to a prophage by a plasmid reduced both Nun activity and the expression of a nun::lacZ gene fusion. Cam et al. (3) suggested that transcripts originating at p_RM are the principal source of nun expression in a lysogen and that excess repressor reduces nun expression by repressing p_RM. An alternative explanation of these results is that P_Nun responds to repressor. Although inspection of the DNA sequence revealed no obvious HK022 operator site in the vicinity of P_Nun, high repressor concentrations might relax the specificity of binding. Indeed, Carlson and Little (4) demonstrated binding of repressor to sequences adjacent to an isolated operator under such conditions.

Bacteriophage HK022 resembles other phages of the λ family with respect to its genetic organization. However, in striking contrast to λ, the first gene in the HK022 p_L operon is expressed in the presence of the prophage repressor. Our identification of a dedicated, constitutive promoter for nun explains in a simple way the expression of the phage exclusion function of a repressed HK022 prophage. The advantage that Nun expression confers on HK022 is less clear. Nun excludes phages with nut sites of λ specificity; the growth of several other phages, including some with related nut sites, is unaffected (15). Perhaps phage strains with nut sites of λ specificity are important natural competitors of HK022. Alternatively, perhaps Nun has functions that are yet undiscovered.