Antirepression System Associated with the Life Cycle Switch in theTemperate Podoviridae Phage SPC32H

Minsik Kim, Sangryeol Ryu

Department of Food and Animal Biotechnology, Department of Agricultural Biotechnology, Research Institute for Agriculture and Life Sciences, and Center for Food andBioconvergence, Seoul National University, Seoul, South Korea

Prophages switch from lysogenic to lytic mode in response to the host SOS response. The primary factor that governs thisswitch is a phage repressor, which is typically a host RecA-dependent autocleavable protein. Here, in an effort to reveal themechanism underlying the phenotypic differences between the Salmonella temperate phages SPC32H and SPC32N, whosegenome sequences differ by only two nucleotides, we identified a new class of Podoviridae phage lytic switch antirepressorthat is structurally distinct from the previously reported Sipho- and Myoviridae phage antirepressors. The SPC32H repres-sor (Rep) is not cleaved by the SOS response but instead is inactivated by a small antirepressor (Ant), the expression ofwhich is negatively controlled by host LexA. A single nucleotide mutation in the consensus sequence of the LexA-bindingsite, which overlaps with the ant promoter, results in constitutive Ant synthesis and consequently induces SPC32N to enterthe lytic cycle. Numerous potential Ant homologues were identified in a variety of putative prophages and temperate Podo-viridae phages, indicating that antirepressors may be widespread among temperate phages in the order Caudovirales tomediate a prudent prophage induction.

Bacteriophages (phages), which are natural viral predators ofbacteria, multiply by infecting specific host bacteria. Although

there is an additional type of phage-host relationship called“steady-state infection,” which is exemplified by filamentousphages (1), phage genome replication generally occurs via twodifferent developmental paths: the lytic cycle and the lysogeniccycle. In contrast to the lytic cycle, which results in immediatebursting of the host bacteria and the release of bacteriophageprogeny, the lysogenic cycle involves the maintenance of thephage genome as a part of the host genome for several generations,typically by integrating into host chromosomes or, more rarely, byreplicating as low-copy-number phage plasmids (2–4). The ex-pression of genes necessary for progeny production and host celllysis is tightly repressed by a phage regulatory system, but somephysiological changes in the host induced by UV light irradiationor other DNA-damaging agents activate the lytic cycle by disablingthe phage repressor. Phages fall into two categories: virulentphages that replicate strictly by the lytic cycle and temperatephages that can enter both the lytic cycle and the lysogenic cycle.

The lytic switch following lysogenic development has been wellstudied in the temperate phage lambda. In the lambda lysogenicphase, phage CI repressors form dimers and bind to specific op-erators to prevent expression of lambda early genes and subse-quent late genes (5, 6). Upon host DNA damage, the activated hostRecA protein induces CI proteolysis in a manner similar to theinactivation of the host SOS response regulator LexA (7–9). CIproteolysis leads to the expression of early and late genes, resultingin lytic development. This mechanism illustrates how lambda andother similar phages exploit the host cell SOS response to escapequickly from a potentially damaged host using the RecA-depen-dent cleavable repressor. Alternatively, some phages in the fami-lies Sipho- and Myoviridae utilize the LexA-regulated antirepres-sors instead of the cleavable repressor to associate their lytic switchto the host SOS response (10–12).

Here, in an effort to identify the factor(s) that causes a pheno-typic difference between two very similar podoviral Salmonella

phages, SPC32H and SPC32N, we found a novel Podoviridaephage lytic switch antirepressor. We observed that a single nucle-otide change in the LexA-binding site, which overlaps with thepromoter of the phage antirepressor gene, causes constitutive ex-pression of the antirepressor Ant and consequent inhibition ofphage repressor function in SPC32N. As a result, SPC32N couldnot establish lysogeny as clear plaque mutants. A LexA-dependentlytic switch involving an antirepressor, rather than repressor pro-teolysis, has been found previously in only sipho- and myoviralphages (10–12), and the podoviral SPC32H/N Ant protein had nosignificant homology to these known antirepressors. A databasesearch identified many proteins with homology to Ant, suggestingthe extensive use of antirepressor-mediated lytic induction amongtemperate phages in the order Caudovirales.

MATERIALS AND METHODSBacterial strains, plasmids, and growth conditions. The bacterial strainsand plasmids used in this study are listed in Tables 1 and 2, respectively.All Salmonella mutants were derived from the prophage-cured Salmonellaenterica serovar Typhimurium strain LT2 [referred to as LT2(c)] and its�LT2gtrABC1 (SR5003) derivative to exclude the effect of prophages andspontaneous phage resistance via O-antigen glucosylation, respectively(13, 14). Standard cloning procedures were used to construct the recom-binant plasmids. Bacteria were grown aerobically at 37°C in LB mediumsupplemented with the following chemicals, as needed: ampicillin (Ap),50 �g ml�1; kanamycin (Km), 50 �g ml�1; chloramphenicol (Cm), 25 �gml�1; 5-bromo-4-chloro-3-indolyl-�-D-galactoside (X-Gal), 40 �g ml�1;

Received 6 August 2013 Accepted 19 August 2013

Published ahead of print 28 August 2013

Address correspondence to Sangryeol Ryu, sangryu@snu.ac.kr.

Supplemental material for this article may be found at http://dx.doi.org/10.1128/JVI.02173-13.

Copyright © 2013, American Society for Microbiology. All Rights Reserved.

doi:10.1128/JVI.02173-13

November 2013 Volume 87 Number 21 Journal of Virology p. 11775–11786 jvi.asm.org 11775

on February 12, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

L-arabinose, 0.2% (final concentration); isopropyl-�-D-thiogalactopyra-noside (IPTG), 1,000 �M (final concentration); and mitomycin C(MMC), 1 �g ml�1 (final concentration). For the disc diffusion assay,6-mm-diameter filter paper discs were soaked with 10 �l of arabinose,

antibiotics, or MMC at the indicated concentrations, placed on the sur-face of the bacterium-inoculated solidified soft top agar (LB supple-mented with 0.4% [wt/vol] agar and X-Gal, if necessary), and incubated at37°C for 8 h.

TABLE 1 Bacterial strains and bacteriophages used in this study

Strain or phage Relevant characteristicsa Reference or source

Bacterial strainsSalmonella enterica serovar

TyphimuriumLT2(c) Prophage-cured strain LT2; wild type; host for phage SPC32H and SPC32N 13SR5003 LT2(c) with �LT2gtrABC1 14SR5100 �LT2gtrABC1 (32H); SR5003 lysogenized by SPC32H This studySR5158 LT2(c) with �LT2gtrABC1 �sulA �lexA This studySR5176 LT2(c) with �LT2gtrABC1 �sulA lexA(G85D); non-cleavable LexA mutant This studySR5189 �LT2gtrABC1 (32H recET::lacZ); SR5003 lysogenized by SPC32H recET::lacZ This studySR5188 �LT2gtrABC1 �sulA lexA(G85D) (32H recET::lacZ); SR5176 lysogenized by SPC32H recET::lacZ This studySR5192 �LT2gtrABC1(32H rep-HA); SR5003 lysogenized by SPC32H rep-HA This studySR5197 �LT2gtrABC1(32H rep-HA ant-HA); SR5003 lysogenized by SPC32H rep-HA ant-HA This study

Escherichia coliBTH101 F� cya-99 araD139 galE15 galK16 rpsL1(Strr) hsdR2 mcrA1 mcrB1; reporter strain in bacterial

two-hybrid assay29

BL21(DE3) F� ompT hsdS (rB� mB

�) gal (DE3); protein overexpression Laboratory collection

BacteriophagesSPC32H Infect S. Typhimurium; O antigen specific This studySPC32N Infect S. Typhimurium; O antigen specific This studySPC32H �ant SPC32H derivative with �ant This studySPC32H m1 SPC32H derivative with one nucleotide substitution in tsp gene (C to T) This studySPC32H m2 SPC32H derivative with one nucleotide substitution in ant gene promoter (G to T) This studySPC32H m12 SPC32H derivative with two nucleotide substitutions in tsp gene (C to T) and ant gene

promoter (G to T)This study

a Strr, streptomycin resistant.

TABLE 2 Plasmids used in this study

Plasmid Relevant characteristicsa Reference or source

Gene overexpressionpBAD24 General expression vector with the PBAD promoter; Apr 44prep pBAD24-rep; Apr This studypant pBAD24-ant; Apr This studyPant* pBAD24-ant*; encoding frame-shifted ant; Apr This study

Luciferase reporter assaypBBRlux Derivative of broad-host-range cloning vector pBBR1MCS containing a promoterless luxCDABE; Cmr 28pPant_H::lux pBBRlux with ant promoter of SPC32H; Cmr This studypPant_N::lux pBBRlux with ant promoter of SPC32N; Cmr This study

Bacterial two-hybrid assaypKT25 Encodes T25 fragment of adenylate cyclase; Kmr 29pUT18C Encodes T18 fragment of adenylate cyclase; Apr 29pKT25-zip pKT25 with zip; Kmr 29pUT18C-zip pUT18C with zip; Apr 29pKT25-rep pKT25 with rep; Kmr This studypUT18C-ant pUT18C with ant; Apr This studypKT25-ant pKT25 with ant; Kmr This studypUT18C-rep pUT18C with rep; Apr This study

Protein expressionpHIS-parallel1 Protein expression vector; allowing N-terminal His6 tagging with a TEV cleavage site; Apr 45pHIS-LexA pHIS-parallel1 with lexA; Apr This studypHIS-Rep pHIS-parallel1 with rep; Apr This studypHIS-Ant pHIS-parallel1 with ant; Apr This study

a Kmr, kanamycin resistant; Apr, ampicillin resistant; Cmr, chloramphenicol resistant.

Kim and Ryu

11776 jvi.asm.org Journal of Virology

on February 12, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

Bacteriophage. The bacteriophages used in this study are listed inTable 1. All phage mutants were derived from the temperate phageSPC32H, which was previously isolated from chicken fecal samples ob-tained from a traditional marketplace in South Korea (15). Routine phagespotting and double-agar overlay assays were conducted to determine theefficiency of plating (EOP) in specific bacteria (14, 15). For the morpho-logical analysis, the phage stocks were negatively stained with 2% uranylacetate (pH 4.0) as previously described (15) and were examined by trans-mission electron microscopy (LEO 912AB TEM; Carl Zeiss, Jena, Ger-many) at 120-kV accelerating voltage. The images were scanned with aProscane 1,024 � 1,024-pixel charge-coupled device camera.

Bacteriophage genome sequencing and analysis. Phage nucleic acidsthat were extracted by the phenol-chloroform extraction method withprotease K/SDS treatment (16) were pyrosequenced using the GS FLXTitanium system by Macrogen, Seoul, South Korea. The quality-filteredreads were assembled using the GS de novo assembler (v. 2.60), and theopen reading frames (ORFs) that encode proteins of more than 35 aminoacids in size were predicted using the software programs GeneMarkS (17),Glimmer 3.02 (18), and FgenesB (Softberry, Inc., Mount Kisco, NY,USA). The predicted ORFs were annotated based on the results ofBLASTP (19), InterProScan (20), and NCBI Conserved Domain Database(21) analysis. tRNAscan-SE (22) and BPROM (Softberry, Inc.) were usedto predict the tRNA sequences and the putative promoter/transcriptionfactor-binding sites, respectively. Genomic comparison at the DNA levelwas visualized using the program Easyfig (23).

Construction of the Salmonella and phage mutants. The lambda redrecombination method was used for in-frame gene deletion (24). To con-struct the noncleavable LexA protein, a point mutation in lexA {resultingin a G85D mutation in the amino acid sequence [lexA(G85D)]} was gen-erated by lambda red recombination and double homologous recombi-nation-based counterselection, as previously described (14), using thesuicide vector pDS132 (25). The SPC32H lysogen [�LT2gtrABC1 (32H);SR5100] was isolated by sequential streaking of SPC32H-resistant clonesfrom a lawn of phage-treated �LT2gtrABC1 and was verified by PCR am-plification of the phage attachment (attR) site. The transcriptional recET::lacZ fusion was constructed using pCE70, as previously described (26, 27).Human influenza virus hemagglutinin (HA) epitope tagging of the spe-cific gene(s) was also accomplished by lambda red recombination usingoligonucleotides containing the HA tag sequence.

Phage mutants were induced from the SPC32H lysogen after the genemanipulations described above, with some modifications. Briefly, to gen-erate SPC32H m1, a truncated tailspike gene (tsp::Kmr), which was con-structed by lambda red recombination in the SPC32H lysogen, was re-placed with the m1-containing tsp gene by double homologousrecombination-based counterselection. The presence of the m1 mutationin the induced phage was confirmed by DNA sequencing. Similar meth-ods were used to construct SPC32H m2 and SPC32H m12. The oligonu-cleotides used in this study are listed in Table 3.

Bioluminescence reporter assay. The 197-bp fragment upstream ofthe ant gene in SPC32H (designated Pant_H) or SPC32N (designatedPant_N) was PCR amplified and cloned into pBBRlux (28), resulting in thetranscriptional fusion of the operon luxCDABE to the putative ant genepromoter. S. Typhimurium strains harboring this reporter plasmid werecultured in 200 �l of fresh LB broth supplemented with appropriate an-tibiotics in a 96-well plate. The cellular bioluminescence of the culture andthe absorbance at 600 nm (A600) were measured periodically using anInfinite 200 Pro plate reader (Tecan, Männedorf, Switzerland), and theresults were expressed in arbitrary relative light units (RLU). To triggerthe SOS responses, MMC was added to the culture after a 3-h incubation.The three independent assays with triple technical replications were per-formed.

Western blot analysis. At the mid-exponential phase, culture of theHA-tagged-gene(s)-containing Salmonella was treated by MMC, and por-tions of the culture were sampled at the indicated time points. Bacterialcells were harvested by centrifugation and were lysed with the B-Per re-

agent (Thermo Scientific, Illinois, USA). Soluble proteins (10 �g) fromcell lysates were separated by 15% SDS-PAGE and electrotransferred tothe polyvinylidene difluoride (PVDF) membrane. HA-tagged proteinsand DnaK were detected with anti-HA and anti-DnaK antibodies, respec-tively. The chemiluminescence signals were developed using the West-Zolplus Western blot detection system (iNtRON Biotechnology, Gyeonggi-do, South Korea) after the goat anti-mouse IgG-horseradish peroxidase(HRP) (Santa Cruz Biotechnology, CA, USA) treatment, and then X-rayfilm was exposed to chemiluminescent light to detect the signals.

Bacterial two-hybrid assay. Protein-protein interaction was deter-mined by the recovery of adenylate cyclase (CyaA) activity through het-erodimerization of fusion proteins in the Escherichia coli BTH101 reporterstrain (cyaA mutant) (29). The reporter strain harboring the fusion plas-mid pair (e.g., pKT25-rep and pUT18c-ant) was streaked on LB agar sup-plemented with Km, Ap. and X-Gal or subjected to the �-galactosidaseassay (30) to quantitatively measure the interaction.

Purification of proteins, rTEV protease treatment, and analyticalsize exclusion chromatography. Cultures of E. coli BL21(DE3) harboringpHIS-LexA, -Rep, or -Ant (optical density at 600 nm [OD600] � �0.15)were treated by 100 �M IPTG and incubated at 25°C for an additional 4 h.Cells were harvested by centrifugation and lysed in lysis buffer (20 mMTris [pH 8.0], 500 mM NaCl, and 20 mM imidazole) by sonication on ice.Centrifuged (16,000 � g, 4°C, for 30 min) and filtered (0.22-�m filter;Millipore, Ireland) cell lysate was subjected to nickel-nitrilotriacetic acid(Ni-NTA) affinity chromatography (Qiagen, California, USA) accordingto the manufacturer’s protocol with elution buffer (lysis buffer with 250mM imidazole). The eluted protein was concentrated using a Vivaspin 20instrument (3,000-molecular-weight cutoff [MWCO] polyethersulfone[PES]; Sartorius, Goettingen, Germany), and the buffer was changed (20mM [Tris pH 8.0], 500 mM NaCl, and 50% glycerol) using a PD MidiTrapG-25 column (GE Healthcare, Buckinghamshire, United Kingdom). Toremove the His6 tag from the purified proteins, recombinant tobacco etchvirus (rTEV) protease (1:5 ratios in concentration) was treated for 6 h at4°C in a cleavage buffer (10 mM Tris [pH 8.0], 150 mM NaCl, 0.5 mMEDTA, and 100 mM dithiothreitol [DTT]). For analytical size exclusionchromatography, a Superdex 200 10/300 GL column (GE Healthcare) wasused. The column was equilibrated with a buffer consisting of 500 mMNaCl and 20 mM Tris [pH 8.0], and then purified proteins (500 �l of 0.8�g �l�1) were loaded on to the column at a flow rate of 0.5 ml min�1.

Electrophoretic mobility shift assay (EMSA). The purified PCR frag-ments of the ant gene promoter region (APR) was �-32P labeled using T4polynucleotide kinase (TaKaRa, Japan). The labeled DNA (approximately4 nM) was incubated with various concentrations of LexA for 30 min at37°C in 20 �l of reaction mixture containing 1� binding buffer (10 mMHEPES [pH 8.0], 10 mM Tris [pH 8.0], 50 mM KCl, 1 mM EDTA, 1 mMdithiothreitol, and 5% glycerol) and 1.1 �g of poly(dI-dC). For determi-nation of Rep binding, various amounts of Rep were incubated for 15 minat 20°C with the 4 nM labeled DNA in the 20-�l reaction mixture. Whenappropriate, Rep was preincubated with various concentrations of Ant for30 min at 20°C prior to incubation with labeled DNA. The samples wereresolved by 6% native PAGE in 0.5� TBE buffer (45 mM Tris-borate [pH8.3] and 1 mM EDTA). The gels were vacuum dried, and the radioactivitywas analyzed using a BAS2500 system (Fujifilm, Tokyo, Japan).

Nucleotide sequence accession numbers. The genome sequences ofSPC32H and SPC32N are available at GenBank under accession numbersKC911856 and KC911857, respectively.

RESULTSPhenotypic and genomic characterization of the two related S.Typhimurium phages SPC32H and SPC32N. Previously, we iso-lated nine phages specific for S. Typhimurium from chicken fecalsamples (15). Two of these phages, which originated from thesame sample collection, exhibited distinct plaque morphologieson a lawn of S. Typhimurium: one phage (SPC32H) formed tur-bid plaques surrounded by a halo, but the other phage (SPC32N)

Antirepression System in a Lytic Switch of Podophage

November 2013 Volume 87 Number 21 jvi.asm.org 11777

on February 12, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

formed clear plaques without a halo (Fig. 1A and B). Transmissionelectron microscopy (TEM) analysis revealed that both phagesbelonged to the family Podoviridae, since they had an isometrichead (�62.3 nm in diameter) and a short noncontractile tail(�15.4 nm in length) with tail shaft and tail spikes (Fig. 1C andD). These two phages infected identical repertories of Salmonellastrains using the O antigen (O-Ag) of Salmonella as the host re-ceptor (data not shown).

Sequencing of SPC32H and SPC32N revealed that both phagescontain 38,689 bp of double-stranded DNA with an identicalGC content of 50.16% and 51 predicted open reading frames(ORFs) with one Arg-tRNA. About half of the ORFs (24 ORFs)were annotated as hypothetical proteins, whereas the other anno-tated proteins were classified into the following modules: DNApackaging, virion structure morphogenesis, lysogenic conversion,host lysis, and DNA replication/recombination (Fig. 2A; see also

Table S1 in the supplemental material). The predicted proteinsincluded a phage integrase as well as a putative repressor, indicat-ing that both phages might be temperate phages. BLASTP searchesrevealed that the SPC32H and SPC32N genomes closely resem-bled those of Salmonella phage ε15 and other ε15-like phages (31,32). Indeed, whole-genome comparisons made at the DNA levelrevealed a significant degree of synteny between the genomes ofSPC32H, ε15, and the ε15-like phage phiV10 (Fig. 2A). In partic-ular, 34 out of 51 SPC32H gene products, including a small/largeterminase, a head-to-tail joining protein, a putative major coatprotein, a putative holin/endolysin, an integrase, a repressor, anda putative DNA replication protein, were highly similar (50 to�100% identity at the amino acid level) to those of ε15 (see TableS1). Genes for the putative SPC32H tail structure module (e.g.,SPC32H_016, 017, and 018) had higher similarity to those ofphiV10 than ε15 (Fig. 2A). Since these phages infect different

TABLE 3 Oligonucleotides used in this study

Oligonucleotide Sequence (5= ¡ 3=)a Purpose

32H-attP ATA CAG TTT GTC CTG CGG CTT GAG PCR for attR site32H-attB GAC AAA GTC AGT CAG GCG TTT ACC PCR for attR site32H-rep-CF2 AAT GGG CAA ATG AAT TCG CTA TGA AA prep construction32H-rep-CR1 TGG TAA TTG CGT GTC GAC TGA G prep construction32H-ant-CF1 TGT TTG CAT GGA GAA TTC GAG ATG C pant construction32H-ant-CR1 GCC GCA GAG TCG ACC TTT TAT TTT T Pant, Pant* construction32H-ant*-CF2 CAT GGA GAA TTC GAG ATG ACA ACG G Pant* construction32s-Pant-CF1 TCA GTT GAG CTC GTC ATG TAA G pPant_H::lux and pPant_N::

lux construction32s-Pant-CR1 GGT GAT ACT GCC ACT AGT TCT C pPant_H::lux and pPant_N::

lux construction32H-recE-lacZ-Red-F AGT AAG TCA CAA CTG GAT ATG GTG GCC AAG AAC CCT TCC

CTG TAG GCT GGA GCT GCT TCGSR5188 and SR5189

construction32H-recT-lacZ-Red-R TCT TTG TTG CAG GTG TGG CGC CGT GGC GCC ACG GTG

GTG A AT TCC GGG GAT CCG TCG ACCSR5188 and SR5189

construction32H-rep-HA-Red-F CTC ATT CAT AAA CTT CGT GTT TGA GCA GAA CAA AAG CAA

GTA TCC GTA TGA TGT TCC TGA TTA TGC TAG CCT CTAATG TAG GCT GGA GCT GCT TCG

SR5192 and SR5197construction

32H-rep-HA-Red-R ACC GCC ATC GGG CAG GTA AAG CGT CAG AAT GGC AGGGGA TAT TCC GGG GAT CCG TCG ACC

SR5192 and SR5197construction

32H-ant-HA-Red-F TCT CGA TTA CTG TGC AGA ACA GTT ACG AAA ACA AAC CACATA TCC GTA TGA TGT TCC TGA TTA TGC TAG CCT CTAATG TAG GCT GGA GCT GCT TCG

SR5197 construction

32H-ant-HA-Red-R TGT CAT AGC ATG AAT GTG ACA TGT CAC GAG GCC GCA GAAGAT TCC GGG GAT CCG TCG ACC

SR5197 construction

pKT25-32H-rep-CF1 AAC TGC AGG GAT GAA AAG TAT TTA TGA CAT pKT25-rep constructionpKT25-32H-rep-CR1 CGG GAT CCT TAC TTG CTT TTG TTC TG pKT25-rep and pUT18C-

rep constructionpUT18C-32H-ant-CF1 AAC TGC AGG ATG CAA CGG CAG TAT CA pUT18C-ant constructionpUT18C-32H-ant-CR1 CGG GAT CCT TAT GTG GTT TGT TTT CGT AAC pUT18C-ant and pKT25-

ant constructionpKT25-32H-ant-CF1 AAC TGC AGG GAT GCA ACG GCA GTA TCA pKT25-ant constructionpUT18C-32H-rep-CF1 AAC TGC AGG ATG AAA AGT ATT TAT GAC ATA AG pUT18C-rep constructionpHIS-LT2-lexA-CF2 TAT ATA CAC CCC ATG GGC GGA ATG AAA G pHIS-LexA constructionpHIS-LT2-lexA-CR2 ATT GCC GGA TCT CGA GTT ACA AGG AG pHIS-LexA constructionpHIS-32H-Rep-CF1 AAT ACC ATG GCT ATG AAA AGT ATT TAT GAC ATA AGA CGC pHIS-Rep constructionpHIS-32H-Rep-CR1 AAA GCT CGA GAA TGG CAG GGG ATT ACT TGC pHIS-Rep constructionpHIS-32H-Ant-CF4 CAT AAG CCA TGG GGA TGC AAC GGC AGT ATC AC pHIS-Ant constructionpHIS-32H-Ant-CR3 GCC GCA GAC TCG AGC TTT TAT TTT TCA TTA TGT GG pHIS-Ant construction32s-APR-CF1 CTT CAG TTG AGA CCG TCA TG PCR for APRH and APRN

32s-APR-CR1 TAA GAT GTG AGT CCT CCA CC PCR for APRH and APRN

a Restriction enzyme sites and HA tag coding sequences are underlined and italicized, respectively. Bold case indicates the artificially inserted redundant nucleotide to generate theframeshift in the ant gene.

Kim and Ryu

11778 jvi.asm.org Journal of Virology

on February 12, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

hosts (i.e., S. Typhimurium for SPC32H, Salmonella enterica se-rovar Anatum for ε15, and E. coli O157:H7 for phiV10), differ-ences were observed in the genes encoding tailspike proteins andthe flanked lysogenic conversion module (which converts O-Ag toprevent superinfection). Taken together, these results suggest that

SPC32H and SPC32N should be assigned to the class of ε15-likephages.

A single nucleotide change is responsible for the phenotypicdifference between the two phages. Interestingly, a comparisonof the full genome sequences of SPC32H and SPC32N revealed

FIG 1 Two similar S. Typhimurium-specific Podoviridae phages, SPC32H and SPC32N, produce morphologically distinct plaques. (A and B) Plaque morphol-ogy of SPC32H (A) or SPC32N (B). Dilutions (10 �l) of each phage stock were spotted onto a lawn of the S. Typhimurium LT2(c) �LT2gtrABC1 strain (SR5003).(C and D) TEM image of SPC32H (C) or SPC32N (D). Inset at the bottom left of each panel shows the enlarged virion morphology with a black scale bar (50 nm).The white arrow and arrowheads indicate the tail shaft and tail spikes, respectively.

FIG 2 There are two single nucleotide differences between the genomes of ε15-like phage SPC32H and SPC32N. (A) DNA alignment of the genomes of phageε15 (NC_004775.1), SPC32H, and phiV10 (NC_007804.2) using Easyfig. High sequence similarity between the genomes is indicated by the gray regions. SPC32HORFs are indicated by numbered or annotated arrows. Phage functional modules are indicated under the arrows. ant, antirepressor; tsp, tailspike; oac, o-acetyltransferase; hol, holin; end, endolysin; int, integrase; rep, repressor. Note that the SPC32N genome is identical to that of SPC32H with the exception of twosingle nucleotide differences (see panel B). (B) Schematic representation of the location of the two single nucleotide differences, m1 and m2. The partial SPC32Hgenome sequence surrounding the two single nucleotide differences is shown. m1 (located within the tsp gene) and m2 (located in the intergenic region betweenSPC32H_020 and tsp) are indicated in bold, uppercase letters. The predicted �10 and �35 sites of the putative promoter for SPC32H_020 gene are boxed. Theputative LexA-binding site (SOS box) and the putative repressor-binding site are underlined and doubly underlined, respectively. (C) Consensus sequence of theLexA-binding site from E. coli (8, 33, 34) and the putative LexA-binding sites from phage SPC32H and SPC32N. m2 is indicated with a gray background. Notethat the LexA-binding site sequences for SPC32H and SPC32N shown here are reverse complements of the sequence shown in panel A.

Antirepression System in a Lytic Switch of Podophage

November 2013 Volume 87 Number 21 jvi.asm.org 11779

on February 12, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

only two nucleotide differences. One nucleotide difference, desig-nated m1, is located within the tsp gene (SPC32H_021), whichencodes a phage tailspike, and the other nucleotide difference, m2,is located in the intergenic region between a gene (SPC32H_020)encoding a hypothetical protein and the tsp gene (Fig. 2B). Toverify whether m1, m2, or both single nucleotide differences wereresponsible for the differences between SPC32H and SPC32N, wemutated the SPC32H sequence to match that of SPC32N. Asshown in Fig. 3A, we observed no significant changes in the tur-bidity of the lysis zone when the SPC32H m1 sequence waschanged to that of SPC32N. However, the turbidity decreased dra-matically when the SPC32H m2 sequence was replaced by that ofSPC32N. We therefore investigated in detail how the single nucle-otide difference at the m2 locus leads to this phenotypic difference.

Supplementation with the repressor induces the lysogenicdevelopment of the lytic cycle-biased phage SPC32N. Becauselysogen formation is normally associated with plaque morphol-ogy, we investigated the ability of SPC32H and SPC32N to lysog-enize. The SPC32H and ε15 integrases have 93% identity at theamino acid level, and both phages contain the highly conservedcommon core regions and arm-type binding sequences that arerequired for phage genome integration (31). This suggests thatboth phages may integrate their genome into the same attachmentsite, near the end of the Salmonella guaA gene. Therefore, to detectlysogenization by SPC32H or SPC32N, we PCR amplified theright end of the phage genome attachment site (attR site) using aprimer pair that specifically anneals to the upstream region of thephage integrase gene (int) and within the guaA gene. The specificattR band was amplified from DNA isolated from the SPC32Hlysis zone, whereas no band was detected using DNA isolated fromthe SPC32N lysis zone (Fig. 3B). The specific attR band was am-plified from both colony and genomic DNA from the putative

SPC32H lysogen [�LT2gtrABC1 (32H)] but not from DNA iso-lated from the parental Salmonella strain, SPC32H, or SPC32N(Fig. 3B). Furthermore, the SPC32H lysogen spontaneously pro-duced phages that formed halo plaques during prolonged incuba-tion (data not shown). These results clearly demonstrate that Sal-monella can be lysogenized by SPC32H but not by SPC32N. Thespecific attR band was amplified from DNA isolated from the lysiszone of SPC32H m1 but not SPC32H m2 or m12 (Fig. 3C), con-firming that m2 is the reason for phenotypic differences betweenSPC32H and SPC32N.

Because the phage repressor plays a critical role in the mainte-nance of the lysogenic state by repressing the expression of lyticgenes and both phages have a putative repressor gene (rep), weassessed the deficiency of repression in SPC32N. When SPC32Hand SPC32N were spotted onto lawns of a Salmonella strain over-expressing rep (�LT2gtrABC1 prep), the EOP of both phages wassignificantly reduced (10�5 for SPC32H and 10�2 forSPC32N), and both strains exhibited a more turbid lysis zone thanthe control (Fig. 4A). Moreover, the specific attR band was ampli-fied from DNA isolated from the lysis zone of both phages (datanot shown), suggesting that supplementation with the repressorcan promote lysogenic development in SPC32N. These resultssuggest that SPC32N is defective in maintaining lysogeny, mostlikely due to an insufficient amount of the active repressor.

FIG 3 Introducing the m2 sequence from SPC32N induces SPC32H to enter thelytic cycle. (A) High-titer phage stocks (�107 PFU ml�1; 10 �l) of SPC32H,SPC32N, and three mutant phages derived from SPC32H were spotted onto alawn of S. Typhimurium LT2(c) �LT2gtrABC1. (B) SPC32H can lysogenize hostSalmonella, whereas SPC32N cannot. Various template samples were PCR ampli-fied with an attR-specific primer pair. M, DNA marker 1 Kb (Invitrogen); i, innerpart of the lysis zone; e, edge of the lysis zone; gDNA, genomic DNA;�LT2gtrABC1(32H), SPC32H lysogen (SR5100). (C) DNA isolated from the lysiszones shown in panel A was PCR amplified using primers specific for the attR siteto determine the lysogenization of each phage. Lanes 1 to 5 correspond to each lysiszone shown in panel A. Note that the introduction of m2 resulted in a disappear-ance of the lysogen-specific attR band (lanes 4 and 5).

FIG 4 The novel antirepressor, encoded by SPC32H_020 (ant), induces thelytic development of SPC32H. (A) Supplementation with the putative repres-sor leads to the lysogenic development of the lytic cycle-biased phage SPC32N,while supplementation with the putative antirepressor results in the lytic de-velopment of SPC32H. Salmonella strains transformed with a control plasmid(pBAD24), a putative repressor-overexpressing plasmid (prep), or anSPC32H_020-overexpressing plasmid (pant) were infected with serially di-luted (10-fold) stocks of SPC32H or SPC32N. L-Arabinose (0.2%, final con-centration) was added to induce SPC32H_020 expression from pant. (B) Theexpression of the SPC32H_020 protein promotes the switch from lysogenic tolytic development. The SPC32H lysogen [�LT2gtrABC1 (32H); SR5100] andnonlysogen (�LT2gtrABC1; SR5003) strains were transformed with pant or acontrol plasmid (pBAD24), and the resulting strains were subjected to a discdiffusion assay with 10 �l of 15% L-arabinose. pant* indicates the plasmidencoding a frame-shifted ant gene. Arabinose-induced bacterial lysis was ob-served only in the SPC32H lysogen harboring pant.

Kim and Ryu

11780 jvi.asm.org Journal of Virology

on February 12, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

A novel antirepressor encoded by SPC32H_020 governs thelytic switch. The m2 mutation is located 24 bp upstream of thestart codon of the hypothetical protein SPC32H_020, suggestingthat m2 may cause the observed phenotypic differences by affect-ing the expression of this protein. SPC32H_020 is a small, 86-amino-acid protein with no known conserved domain or motif. ABLASTP search identified 40 hypothetical proteins with morethan 64% identity with SPC32H_020 but did not identify anyprotein with a known function. The proteins exhibiting high ho-mology to SPC32H_020 were from members of the Enterobacte-riaceae, such as E. coli, Salmonella spp., Klebsiella spp., Citrobacterspp., and Cronobacter spp., and from ε15-like phages, includingε15, TL-2011b, phiV10, SPN1S, and SPN9TCW (see Table S2 inthe supplemental material). Interestingly, some larger proteins(�218 amino acids) with a relatively low identity (42%) wereannotated as putative antirepressors, suggesting the possibility ofan antirepressor role for SPC32H_020.

To determine whether SPC32H_020 functions as an antire-pressor, we measured the prophage induction efficiency from aSalmonella strain harboring a SPC32H mutant which lacks thegene SPC32H_020 [�LT2gtrABC1 (32H �ant) strain]. Comparedwith results for the WT phage lysogen, the spontaneous inductionrate of the mutant phage lysogen was significantly lower (ca. 6 �10�6-fold lower than that of the WT phage) (Table 4), indicatingthe critical role of SPC32H_020 in normal prophage induction.Furthermore, MMC treatment did not notably enhance inductionof the mutant phage (1.11-fold increase) but did cause a 78.65-fold increase in induction of the WT phage (Table 4), suggesting apotential network between the SPC32H_020 gene and the hostSOS response. Because the EOPs in Salmonella of the WT andmutant phage were similar (1.8 � 107 and 4.0 � 107 PFU ml�1,respectively), these results suggest that SPC32H_020 might act asan antirepressor. The function of the SPC32H_020 gene was fur-ther tested by a phage spotting assay using Salmonella harboring aplasmid overexpressing SPC32H_020 from a pBAD promoter(pant). As expected, both SPC32H and SPC32N generated clearlysis zones/plaques (Fig. 4A), and the lysogen-specific attR bandwas not PCR amplified from DNA isolated from the SPC32H lysiszone (data not shown). To verify the function of SPC32H_020 inlytic switching and prophage induction, an arabinose disc diffu-sion assay was conducted using Salmonella (�LT2gtrABC1) and theSPC32H Salmonella lysogen [�LT2gtrABC1 (32H)], both harbor-ing pant. The SPC32H lysogen carrying pant underwent lysis inthe presence of 15% arabinose, but no lysis was observed in theabsence of pant (Fig. 4B). When the arabinose-inducible plasmidcontained a frame-shifted ant gene (ant*), which was generated byinserting an additional adenine directly downstream from theSPC32H_020 start codon, the arabinose treatment did not inducelysis (Fig. 4B), suggesting that lytic switching is induced by theSPC32H_020 protein rather than the RNA. Taken together, ourresults strongly suggest that SPC32H_020 encodes a novel antire-

pressor protein that plays a significant role in the switch from thelysogenic cycle to the lytic cycle. We have annotated theSPC32H_020 gene as ant (antirepressor) and its gene product asAnt.

The m2 sequence in the SOS box causes constitutive expres-sion of the antirepressor by SPC32N. The results described aboveindicate that the m2 mutation may allow the overexpression of antin SPC32N. Using the BPROM software program, the �10 and�35 sites of the putative ant promoter and one LexA-binding site(SOS box), which overlaps the predicted �10 site, were predictedin the upstream region of the ant gene (Fig. 2B). Intriguingly, m2is located in the consensus LexA-binding site sequence (8, 33, 34)(Fig. 2C). LexA is a transcriptional repressor that represses variousSOS regulons, including LexA itself and the RecA protein, viabinding to the SOS box. DNA damage induces the formation ofactivated RecA nucleoprotein filaments that promote autocleav-age of LexA and consequent derepression of SOS regulons (8).Therefore, we hypothesized that the ant gene is an SOS reguloncontrolled by LexA and that m2 in the consensus LexA-bindingsite sequence might prevent the LexA-mediated repression of theant gene in SPC32N.

To test this hypothesis, we first examined the promoter activityof the ant gene in both SPC32H and SPC32N via a biolumines-cence reporter assay using luxCDABE. In contrast to the low num-ber of RLU (relative light units) detected using the ant promoterfrom SPC32H (Pant_H), the promoter from SPC32N (Pant_N) ex-hibited approximately 2-log higher values (Fig. 5A and B). Treat-ment with MMC significantly increased the RLU produced from aclone harboring pPant_H::lux but not from a clone harboringpPant_N::lux (Fig. 5A and B), suggesting that Pant_H expression wasactivated by DNA damage, whereas Pant_N was expressed consti-tutively and independent of DNA damage. To elucidate whetherthese responses were associated with LexA, we constructed Salmo-nella mutants without the lexA gene or expressing a noncleavableform of LexA [lexA(G85D)] and measured the bioluminescencefrom the reporter plasmid pPant_H::lux. Both mutants were con-structed in a �sulA background to suppress the lethality of thelexA deletion (35). This sulA deletion did not affect reporter geneexpression (data not shown). In the absence of lexA, Pant_H activitywas comparable to that observed in the lexA background in thepresence of MMC, and the Pant_H activity was not affected byMMC treatment (Fig. 5B, �lexA). In addition, the lexA pheno-type was partially rescued by in trans complementation of lexA(data not shown). In contrast, replacing LexA with the noncleav-able form of LexA prevented promoter activation by MMC [Fig.5B, lexA(G85D)], indicating that DNA damage activates the antpromoter through LexA proteolysis. The results were similar re-gardless of the presence of the SPC32H prophage [Fig. 5B, lexA

versus lexA(32H)], suggesting that no other factors, includingthe SPC32H repressor, are involved in ant gene regulation, despitethe presence of a repressor-binding site immediately upstreamof the ant promoter (Fig. 2B; also see below).

We next performed an electrophoretic mobility shift assay(EMSA) to show the binding of LexA to the SOS box within Pant_H

or Pant_N. When the radiolabeled DNA fragment APRH* (ant genepromoter region from SPC32H) was incubated with an increasingamount of purified Salmonella LexA, a specific mobility shift wasobserved, and the APRH* fragment was released by the addition ofthe unlabeled competing cold probe APRH (Fig. 5C, lanes 1 to 8).In contrast, the unlabeled cold probe APRN (ant gene promoter

TABLE 4 Comparison of prophage induction efficiencies

Strain description

Titer of phage (PFU ml�1)

Foldchange� MMC

MMC(1 �g ml�1)

�LT2gtrABC1 (32H) 4.45 � 108 3.50 � 1010 78.65�LT2gtrABC1 (32H �ant) 2.65 � 103 2.95 � 103 1.11

Antirepression System in a Lytic Switch of Podophage

November 2013 Volume 87 Number 21 jvi.asm.org 11781

on February 12, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

region from SPC32N) was unable to compete with APRH* forLexA, and the APRN* fragment was not shifted in the presence ofLexA (Fig. 5C, lanes 9 to14), confirming that LexA cannot repressant expression in SPC32N due to an inability to bind to the SOSbox containing m2.

Based on these results, we investigated the overall cascade ofSPC32H induction using Salmonella strains lysogenized by a de-rivative of SPC32H containing lacZ transcriptionally fused to theputative recET genes. Because the phage recE and recT gene prod-ucts, a 5= ¡ 3= exonuclease and a single-strand DNA binding/annealing protein, respectively, promote homologous recombi-nation to mediate the integration/excision of phage genome to/from the host chromosome (36–38), the expression of recET (andits orthologous genes) can be used as a reporter for prophageinduction. As shown in Fig. 6, treatment with MMC but not otherantibiotics activated the recET::lacZ fusion in the lexA back-ground but not in the lexA(G85D) background, indicating thatthe DNA damage generated by MMC induces SPC32H inductiondependent on LexA proteolysis. The expression of Ant clearly re-sulted in lysogen-specific lysis (Fig. 4B), supporting the idea thatderepression of the ant gene via MMC-induced LexA proteolysis

FIG 5 The ant promoter of SPC32H is activated by DNA damage via LexA proteolysis, whereas the SPC32N ant promoter is constitutively active, due to theinability of LexA to bind to the m2-containing consensus LexA-binding site. The RLU (relative light units) were calculated by dividing the measured biolumi-nescence by the A600 value. The mean and SD for three independent assays are shown on a log scale on the y axis (A and B). (A) Time course observation of antpromoter activity in the presence or absence of DNA damage. Salmonella strains harboring the bioluminescence reporter plasmid pPant_H::lux (luxCDABE fusedto the putative ant promoter of SPC32H) or pPant_N::lux (luxCDABE fused to the putative ant promoter of SPC32N) were incubated at 37°C, and thebioluminescence, as well as the A600 of the culture, was measured every half-hour. The vertical arrows indicate MMC treatment (1 �g ml�1, final concentration;3 h after incubation). (B) ant promoter activities of the various Salmonella strains at an A600 of �0.6, harboring the bioluminescence plasmid. MMC (1 �g ml�1,final concentration) was added after 3 h of incubation. lexA, �LT2gtrABC1, SR5003; lexA(32H), �LT2gtrABC1(32H), SR5100; �lexA, �LT2gtrABC1 �sulA�lexA, SR5158; lexA(G85D), �LT2gtrABC1 �sulA lexA(G85D), SR5176. ���, P 0.001. (C) LexA specifically binds to the putative ant gene promoter region ofSPC32H but not to that of SPC32N, which contains m2. The �-32P-labeled DNA fragment of the ant gene promoter region from SPC32H (APRH*) or fromSPC32N (APRN*) was incubated with the indicated amounts of purified Salmonella LexA and was subjected to an electrophoretic mobility shift assay (EMSA).Corresponding unlabeled DNA fragments (APRH and APRN) were used for the competition analysis. The position of the unbound fragments (F) and fragmentsretarded by LexA binding (B) are indicated.

FIG 6 DNA damage-induced LexA proteolysis followed by SPC32H ant ex-pression induces the switch to lytic development. The lacZ gene, transcription-ally fused to the putative recET genes, was introduced into the SPC32H lyso-gens harboring an intact (lexA) or noncleavable [lexA(G85D)] LexA, and theresulting strains were subjected to a disc diffusion assay with the followingsolutions: MMC, 0.5 mg ml�1 mitomycin C; Cm, 2.5 mg ml�1 chloramphen-icol; Ap, 10 mg ml�1 ampicillin; D.W., distilled water. Note that the blue zoneappears to surround the MMC disc in the lexA background only.

Kim and Ryu

11782 jvi.asm.org Journal of Virology

on February 12, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

leads to phage induction. Taken together, these results demon-strate that the ant gene of SPC32H is negatively regulated by LexAand that m2 in the SOS box causes the dramatic phenotype differ-ences between SPC32H and SPC32N by influencing ant expres-sion.

The antirepressor Ant interacts directly with the cognate re-pressor Rep. The putative repressor from SPC32H (designatedRep), encoded by SPC32H_041, is a 198-amino-acid protein thatcontains a helix-turn-helix motif. The RecA-mediated autocleav-age site (Ala-Gly or Cys-Gly), a highly conserved site in cleavablerepressors such as lambda CI (39), is not present in SPC32H Rep,strongly supporting the notion that SPC32H prophage inductioninvolves the inhibition of Rep through means other than auto-cleavage assisted by RecA nucleofilaments. The immunodetectionof HA epitope-tagged Rep demonstrated that the expression levelof Rep remained virtually constant (i.e., was not cleaved) through-out a 1-h treatment with MMC (Fig. 7A). The lytic switch wasactivated by MMC in this experiment, as shown by the fact thatHA-tagged Ant was expressed and accumulated after treatmentwith MMC (Fig. 7A, lower panel). The HA-tagged versions of theRep and Ant proteins are fully functional (data not shown).

To explore the possible interaction between Rep and Ant, weperformed a bacterial two-hybrid assay based on the restoration of�-galactosidase activity in E. coli cyaA mutant strain BTH101 (29).The reporter strain E. coli BTH101 expressing the combination ofhybrid proteins (i.e., T25-Rep/T18-Ant or T25-Ant/T18-Rep)produced blue colonies on X-Gal plates (data not shown) and

exhibited a significantly higher (approximately 10- to 50-fold)level of �-galactosidase activity than the negative control (i.e., E.coli BTH101 expressing the unfused T18 and T25 peptides) (Fig.7B), indicating a heterodimerization of the hybrid proteins viainteraction between Rep and Ant. Strong Rep-Rep and Ant-Antinteractions were also observed, implying the possibility of multi-merization by each protein. Indeed, the results of analytical sizeexclusion chromatography demonstrated that Rep and Ant wereable to dimerize and tetramerize, respectively (data not shown).Taken together, these results suggest that Rep and Ant can interactwith themselves and each other.

EMSA using purified Rep and Ant revealed that Ant inhibitsRep target site binding. The high homology (97% identity) be-tween SPC32H Rep and the ε15 repressor suggests that Rep mayrecognize the same DNA sequence (5=-ATTACCnnnnGGTAAT�3=) as the ε15 repressor. Radiolabeled APRH*, which includesthe putative repressor-binding site as well as the SOS box, was alsoused in this assay. Two DNA-protein complex bands with differ-ent mobilities appeared when purified Rep was incubated withAPRH* (Fig. 7C, lanes 1 to 4), suggesting that APR may have twoRep-binding sites with different affinities for Rep. A competitionassay using a nonlabeled cold probe demonstrated the specificityof Rep-binding for APR (Fig. 7C. lanes 5 and 6). Notably, prein-cubation of Rep with purified Ant prevents the mobility shift ofthe APRH* fragment in an Ant concentration-dependent manner(Fig. 7C, lanes 7 to 9), suggesting that the specific interaction be-tween Ant and its cognate repressor Rep interferes with Rep bind-

FIG 7 DNA damage induces Ant accumulation but not Rep cleavage, and the consequent binding of Ant to Rep inhibits the binding of Rep to specific operators.(A) Salmonella strains lysogenized by SPC32H expressing HA-tagged Rep (upper panel; �LT2gtrABC1 [32H rep-HA], SR5192) or both HA-tagged Rep andHA-tagged Ant (lower panel; �LT2gtrABC1 [32H rep-HA ant-HA], SR5197) were exposed to MMC for 1 or 2 h, respectively. The MMC-treated bacterial cultureswere sampled at the indicated time points and subjected to the Western blotting to immunodetect the HA-tagged proteins. DnaK was used as an internal control.(B) Bacterial two-hybrid assays revealed the direct binding of Ant to Rep. The �-galactosidase activity of E. coli BTH101 reporter strains harboring the indicatedplasmid pairs were measured. The activities are presented in Miller units. B, a backbone plasmid. (C) EMSA with purified Rep and Ant demonstrates theAnt-mediated inhibition of Rep binding to its operators. Mixtures of APRH* and the indicated amounts of Rep were incubated at 20°C for 15 min in 1� bindingbuffer supplemented with 1.1 �g of poly(dI-dC) and then electrophoresed on a 6% native acrylamide slab gel for EMSA. For competition analysis, unlabeledAPRH fragments were added as cold probes to the mixture. When appropriate, Rep was preincubated with the indicated amounts of Ant at 20°C for 30 min andfurther incubated with APRH* as described above. The positions of the unbound fragments (F) and fragments retarded by Rep binding (B1 and B2) are indicated.

Antirepression System in a Lytic Switch of Podophage

November 2013 Volume 87 Number 21 jvi.asm.org 11783

on February 12, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

ing to its target DNA. Note that the protein concentrations indi-cated were calculated based on the assumption that the Rep andAnt stocks consisted entirely of active dimers and tetramers, re-spectively. The APRH* fragment clearly did not exhibit a mobilityshift when incubated with Ant alone (Fig. 7C, lane 10), excludingthe possibility that Ant inhibits Rep activity by competing for theRep-binding site with Rep.

DISCUSSION

The goal of this study was to determine the cause for the pheno-typic differences between two highly similar podoviral ε15-likephages, SPC32H and SPC32N. We detected two nucleotide differ-ences between the two phage genomes, but only one, locatedwithin a noncoding region, was responsible for the phenotypicdifferences. This nucleotide polymorphism, m2, was locatedwithin a consensus LexA-binding site sequence that overlaps the�10 site of the promoter for SPC32H_020, which encodes a hy-pothetical protein (Fig. 2). This sequence difference preventsLexA from binding to its binding site, allowing the constitutiveexpression of a small hypothetical protein (Fig. 5), which we haveidentified as a novel antirepressor of the family Podoviridae. Thisantirepressor inhibits the binding of its cognate phage repressor toregulatory regions (Fig. 7C), resulting in a switch of the phage lifecycle from lysogenic to lytic.

To date, at least two categories of lytic switch antirepressionsystems have been identified in temperate phages. The first systemis represented by the Cro protein of several lambdoid phages, suchas phage lambda, HK022, and HK97. In this system, the binding ofthe Cro protein to target operator sites prevents expression of thecI gene, which encodes the phage repressor CI (40). In contrast,the second system controls repressor activity at the protein level.For example, the antirepressor Tum from myoviral coliphage 186binds directly to the phage repressor CI, preventing CI from bind-ing to its operator sites (12). Notably, the latter system has beenfound in only a few temperate phages, including siphoviral co-liphage N15 (11) and the siphoviral prophages Gifsy-1 and Gifsy-3identified in S. Typhimurium strain 14028 (10). In the presentstudy, we have elucidated the mechanism and regulation of anantirepressor, which belongs to the second category of lytic switch

antirepression systems and is the first example of this type of sys-tem in the Podoviridae family of the order Caudovirales. A notablecommon feature of this second system type is the LexA-regulatedinitiation of antirepressor expression. Although the phage P22also produces an antirepressor that inactivates the c2 repressorand prevents RecA-dependent c2 proteolysis (41, 42), it is un-known whether LexA regulates the expression of the antirepres-sor. However, the presence of a consensus LexA-binding sequence38 bp upstream of the start codon of the antirepressor proteinsuggests that LexA may be involved in the regulation of the P22antirepressor and consequently that the P22 antirepressor may bea member of the second category of antirepression systems.

Linking the host SOS response to the lytic switch is a funda-mental strategy used by prophages to escape from damaged hostcells. Compared to the RecA-dependent cleavable repressor sys-tem, such as the lambda CI (7, 40), the antirepression systemappears to be more advantageous to the prophages. If host bacteriaare able to repair DNA damage before prophage induction andsurvive (43), it would be more beneficial for the prophages toremain in the host cell. If lysis occurred, the induced phages wouldneed to reestablish the prophage state in new host bacterial cells tostably maintain their genome as a part of a host genome. Thissuperfluous step could easily be prevented by expressing the anti-repressor in a LexA cleavage-dependent manner. Since antirepres-sor levels are reduced by the replenished LexA pool, lysogenicdevelopment could resume because inactivated, rather than de-graded, repressors can be restored to function by dissociatingfrom the antirepressor. Although we did not demonstrate the re-versible binding of Ant and the recovery of Rep activity after Antdissociation in the present study, Rep levels were stably main-tained without degradation during MMC treatment (Fig. 6), sug-gesting that recycled Rep could be used in the resumed SPC32Hlysogenic development. Indeed, the antirepressor Tum/repressorCI pair from coliphage 186 exhibits reversible Tum binding andthe recovery of CI activity after dissociation from Tum (12). Weare currently attempting to elucidate this issue by investigating thestructure of the Rep-Ant complex as well as the individual pro-teins.

Considering the advantages of rapid resumption of the regula-

FIG 8 Amino acid alignment of the phage antirepressors. The amino acid sequences of Tum (from coliphage 186), AntC (from coliphage N15), GfoA (fromGifsy-1) and Ant (from SPC32H) were aligned using ClustalW2. There are no noticeable consensus residues, demonstrating the diversity of phage antirepressorsin the order Caudovirales.

Kim and Ryu

11784 jvi.asm.org Journal of Virology

on February 12, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

tory circuit, it is possible that this type of repressor/antirepressorsystem is widespread among the temperate phages. Remarkably,several homologues of SPC32H Ant (38 to 100% amino acid sim-ilarities) were identified in other Podoviridae phages and variousbacteria in the family Enterobacteriaceae (see Table S2 in the sup-plemental material), most likely as a gene product of unknownfunction of prophages. Moreover, the phage antirepressors Tum(from Myoviridae coliphage 186), AntC (from Siphoviridae co-liphage N15), GfoA (from Siphoviridae phage Gifsy-1), and Ant(from Podoviridae phage SPC32H) are distinct from each other atthe amino acid sequence level (Fig. 8), suggesting that diverserepressor/antirepressor pairs are present in the order Caudoviralesto allow for more prudent control of lytic/lysogenic switching.Therefore, as suggested by Mardanov and Ravin, the cleavablerepressor system may not be the exclusive mechanism for lytic/lysogenic regulation in temperate phages (11). As recently illus-trated by Lemire et al., antirepressors can mediate cross talk be-tween prophages in polylysogenic strains (10). Thus, furtherstudies regarding the trans activity of diverse phage antirepressors,including SPC32H Ant, would provide insight into the coordi-nated behavior of temperate phage subversion of their bacterialprey.

ACKNOWLEDGMENT

This work was supported by a National Research Foundation of Korea(NRF) grant funded by the Ministry of Education, Science and Technol-ogy (no. 20090078983).

REFERENCES1. Rakonjac J, Bennett NJ, Spagnuolo J, Gagic D, Russel M. 2011. Fila-

mentous bacteriophage: biology, phage display and nanotechnology ap-plications. Curr. Issues Mol. Biol. 13:51–76.

2. Girons IS, Bourhy P, Ottone C, Picardeau M, Yelton D, Hendrix RW,Glaser P, Charon N. 2000. The LE1 bacteriophage replicates as a plasmidwithin Leptospira biflexa: construction of an L. biflexa-Escherichia colishuttle vector. J. Bacteriol. 182:5700 –5705.

3. Lobocka MB, Rose DJ, Plunkett G, III, Rusin M, Samojedny A, Lehn-herr H, Yarmolinsky MB, Blattner FR. 2004. Genome of bacteriophageP1. J. Bacteriol. 186:7032–7068.

4. Ravin NV. 2011. N15: the linear phage-plasmid. Plasmid 65:102–109.5. Echols H, Green L. 1971. Establishment and maintenance of repression

by bacteriophage lambda: the role of the cI, cII, and c3 proteins. Proc. Natl.Acad. Sci. U. S. A. 68:2190 –2194.

6. Reichardt L, Kaiser AD. 1971. Control of lambda repressor synthesis.Proc. Natl. Acad. Sci. U. S. A. 68:2185–2189.

7. Little JW. 1984. Autodigestion of LexA and phage lambda repressors.Proc. Natl. Acad. Sci. U. S. A. 81:1375–1379.

8. Little JW, Mount DW. 1982. The SOS regulatory system of Escherichiacoli. Cell 29:11–22.

9. Sauer RT, Ross MJ, Ptashne M. 1982. Cleavage of the lambda and P22repressors by recA protein. J. Biol. Chem. 257:4458 – 4462.

10. Lemire S, Figueroa-Bossi N, Bossi L. 2011. Bacteriophage crosstalk:coordination of prophage induction by trans-acting antirepressors. PLoSGenet. 7:e1002149. doi:10.1371/journal.pgen.1002149.

11. Mardanov AV, Ravin NV. 2007. The antirepressor needed for inductionof linear plasmid-prophage N15 belongs to the SOS regulon. J. Bacteriol.189:6333– 6338.

12. Shearwin KE, Brumby AM, Egan JB. 1998. The Tum protein of coliphage186 is an antirepressor. J. Biol. Chem. 273:5708 –5715.

13. Erickson M, Newman D, Helm RA, Dino A, Calcutt M, French W,Eisenstark A. 2009. Competition among isolates of Salmonella entericassp. enterica serovar Typhimurium: role of prophage/phage in archivedcultures. FEMS Microbiol. Lett. 294:37– 44.

14. Kim M, Ryu S. 2012. Spontaneous and transient defence against bacte-riophage by phase-variable glucosylation of O-antigen in Salmonella en-terica serovar Typhimurium. Mol. Microbiol. 86:411– 425.

15. Kim M, Ryu S. 2011. Characterization of a T5-like coliphage, SPC35, and

differential development of resistance to SPC35 in Salmonella entericaserovar Typhimurium and Escherichia coli. Appl. Environ. Microbiol. 77:2042–2050.

16. Sambrook J, Russell DW. 2001. Molecular cloning: a laboratory manual,3rd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

17. Besemer J, Lomsadze A, Borodovsky M. 2001. GeneMarkS: a self-training method for prediction of gene starts in microbial genomes. Im-plications for finding sequence motifs in regulatory regions. Nucleic AcidsRes. 29:2607–2618.

18. Delcher AL, Bratke KA, Powers EC, Salzberg SL. 2007. Identifyingbacterial genes and endosymbiont DNA with Glimmer. Bioinformatics23:673– 679.

19. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. 1990. BasicLocal Alignment Search Tool. J. Mol. Biol. 215:403– 410.

20. Zdobnov EM, Apweiler R. 2001. InterProScan—an integration platformfor the signature-recognition methods in InterPro. Bioinformatics 17:847– 848.

21. Marchler-Bauer A, Anderson JB, Derbyshire MK, DeWeese-Scott C,Gonzales NR, Gwadz M, Hao L, He S, Hurwitz DI, Jackson JD, Ke Z,Krylov D, Lanczycki CJ, Liebert CA, Liu C, Lu F, Lu S, Marchler GH,Mullokandov M, Song JS, Thanki N, Yamashita RA, Yin JJ, Zhang D,Bryant SH. 2007. CDD: a conserved domain database for interactive do-main family analysis. Nucleic Acids Res. 35:D237–D240. doi:10.1093/nar/gkl951.

22. Lowe TM, Eddy SR. 1997. tRNAscan-SE: a program for improved detec-tion of transfer RNA genes in genomic sequence. Nucleic Acids Res. 25:955–964.

23. Sullivan MJ, Petty NK, Beatson SA. 2011. Easyfig: a genome comparisonvisualizer. Bioinformatics 27:1009 –1010.

24. Datsenko KA, Wanner BL. 2000. One-step inactivation of chromosomalgenes in Escherichia coli K-12 using PCR products. Proc. Natl. Acad. Sci.U. S. A. 97:6640 – 6645.

25. Philippe N, Alcaraz JP, Coursange E, Geiselmann J, Schneider D. 2004.Improvement of pCVD442, a suicide plasmid for gene allele exchange inbacteria. Plasmid 51:246 –255.

26. Ellermeier CD, Janakiraman A, Slauch JM. 2002. Construction of tar-geted single copy lac fusions using lambda Red and FLP-mediated site-specific recombination in bacteria. Gene 290:153–161.

27. Merighi M, Ellermeier CD, Slauch JM, Gunn JS. 2005. Resolvase-in vivoexpression technology analysis of the Salmonella enterica serovar Typhi-murium PhoP and PmrA regulons in BALB/c mice. J. Bacteriol. 187:7407–7416.

28. Lenz DH, Mok KC, Lilley BN, Kulkarni RV, Wingreen NS, Bassler BL.2004. The small RNA chaperone Hfq and multiple small RNAs controlquorum sensing in Vibrio harveyi and Vibrio cholerae. Cell 118:69 – 82.

29. Karimova G, Pidoux J, Ullmann A, Ladant D. 1998. A bacterial two-hybrid system based on a reconstituted signal transduction pathway. Proc.Natl. Acad. Sci. U. S. A. 95:5752–5756.

30. Miller JH. 1972. Assay of �-galactosidase, p 352–355. In Experiments inmolecular genetics. Cold Spring Harbor Laboratory Press, Cold SpringHarbor, NY.

31. Kropinski AM, Kovalyova IV, Billington SJ, Patrick AN, Butts BD,Guichard JA, Pitcher TJ, Guthrie CC, Sydlaske AD, Barnhill LM,Havens KA, Day KR, Falk DR, McConnell MR. 2007. The genome ofε15, a serotype-converting, group E1 Salmonella enterica-specific bacte-riophage. Virology. 369:234 –244.

32. Perry LL, SanMiguel P, Minocha U, Terekhov AI, Shroyer ML, FarrisLA, Bright N, Reuhs BL, Applegate BM. 2009. Sequence analysis ofEscherichia coli O157:H7 bacteriophage PhiV10 and identification of aphage-encoded immunity protein that modifies the O157 antigen. FEMSMicrobiol. Lett. 292:182–186.

33. Fernandez de Henestrosa AR, Ogi T, Aoyagi S, Chafin D, Hayes JJ,Ohmori H, Woodgate R. 2000. Identification of additional genes belong-ing to the LexA regulon in Escherichia coli. Mol. Microbiol. 35:1560 –1572.

34. Lewis LK, Harlow GR, Gregg-Jolly LA, Mount DW. 1994. Identificationof high affinity binding sites for LexA which define new DNA damage-inducible genes in Escherichia coli. J. Mol. Biol. 241:507–523.

35. Bunny K, Liu J, Roth J. 2002. Phenotypes of lexA mutations in Salmonellaenterica: evidence for a lethal lexA null phenotype due to the Fels-2 pro-phage. J. Bacteriol. 184:6235– 6249.

36. Carter DM, Radding CM. 1971. The role of exonuclease and beta proteinof phage lambda in genetic recombination. II. Substrate specificity and themode of action of lambda exonuclease. J. Biol. Chem. 246:2502–2512.

Antirepression System in a Lytic Switch of Podophage

November 2013 Volume 87 Number 21 jvi.asm.org 11785

on February 12, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

37. Kmiec E, Holloman WK. 1981. Beta protein of bacteriophage lambdapromotes renaturation of DNA. J. Biol. Chem. 256:12636 –12639.

38. Li Z, Karakousis G, Chiu SK, Reddy G, Radding CM. 1998. The betaprotein of phage lambda promotes strand exchange. J. Mol. Biol. 276:733–744.

39. Daniels DL, Schroeder JL, Szybalski W, Sanger F, Coulson AR, HongGF, Hill DF, Petersen GF, Blattner FR. 1983. Complete annotatedlambda sequence, p 519 – 676. In Hendrix RW, Roberts JW, Stahl RW,Weisberg RA (ed), Lambda II. Cold Spring Harbor Laboratory Press,Cold Spring Harbor, NY.

40. Friedman DI, Gottesman M. 1983. Lytic mode of lambda develop-ment, p 21–51. In Hendrix RW, Roberts JW, Stahl RW, Weisberg RA(ed), Lambda II. Cold Spring Harbor Laboratory Press, Cold SpringHarbor, NY.

41. Prell HH, Harvey AM. 1983. P22 antirepressor protein prevents in vivorecA-dependent proteolysis of P22 repressor. Mol. Gen. Genet. 190:427–431.

42. Susskind MM, Botstein D. 1975. Mechanism of action of Salmonellaphage P22 antirepressor. J. Mol. Biol. 98:413– 424.

43. Kuzminov A. 1999. Recombinational repair of DNA damage in Esche-richia coli and bacteriophage lambda. Microbiol. Mol. Biol. Rev. 63:751–813.

44. Guzman LM, Belin D, Carson MJ, Beckwith J. 1995. Tight regulation,modulation, and high-level expression by vectors containing the arabi-nose PBAD promoter. J. Bacteriol. 177:4121– 4130.

45. Sheffield P, Garrard S, Derewenda Z. 1999. Overcoming expression andpurification problems of RhoGDI using a family of “parallel” expressionvectors. Protein Expr. Purif. 15:34 –39.

Kim and Ryu

11786 jvi.asm.org Journal of Virology

on February 12, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from