1

Phage-borne factors and host LexA regulate the lytic switch in phage GIL01 1

Running title: Phage GIL01 is under LexA control 2

Nadine Fornelos1,2, Jaana K. H. Bamford2* and Jacques Mahillon1 3

4

5

1 Laboratory of Food and Environmental Microbiology 6

Université catholique de Louvain 7

Croix du Sud, 2, Box L7.05.12 8

B-1348 Louvain-la-Neuve 9

Belgium 10

2 Department of Biological and Environmental Science and Nanoscience Center 11

University of Jyväskylä 12

P. O. Box 35 13

40014 Jyväskylä 14

Finland 15

16

*Corresponding author. Mailing address: Department of Biological and Environmental Science 17

and Nanoscience Center, University of Jyväskylä, P. O. Box 35, 40014 Jyväskylä, Finland. 18

Phone: +358 14 260 4160. Fax: +358 14 260 2221. Email: jaana.bamford@jyu.fi 19

20

21

Keywords: prophage, lytic switch, clear plaque mutant, SOS response, LexA 22

Copyright © 2011, American Society for Microbiology and/or the Listed Authors/Institutions. All Rights Reserved.J. Bacteriol. doi:10.1128/JB.05618-11 JB Accepts, published online ahead of print on 2 September 2011

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

2

Abstract 23

The Bacillus thuringiensis temperate phage GIL01 does not integrate into the host chromosome 24

but exists stably as an independent linear replicon within the cell. Similar to the lambdoid 25

prophages, the lytic cycle of GIL01 is induced as part of the cellular SOS response to DNA 26

damage. However, no CI-like maintenance repressor has been detected in the phage genome, 27

suggesting that GIL01 uses a novel mechanism to maintain lysogeny. To gain insights into the 28

GIL01 regulatory circuit, we isolated and characterized a set of 17 clear plaque (cp) mutants that 29

are unable to lysogenize. Two phage-encoded proteins, gp1 and gp7, are required for stable 30

lysogen formation. Analysis of cp mutants also identified a 14-bp palindromic dinBox1 sequence 31

within the P1-P2 promoter region that resembles the known LexA binding site of Gram-positive 32

bacteria. Mutations at conserved positions in dinBox1 result in a cp phenotype. Genomic analysis 33

identified a total of three dinBox sites within GIL01 promoter regions. To investigate the 34

possibility that the host LexA regulates GIL01, phage induction was measured in a host carrying 35

a non-cleavable lexA (Ind-) mutation. GIL01 formed stable lysogens in this host but lytic growth 36

could not be induced by treatment with mitomycin C. Also, mitomycin C induced β-37

galactosidase expression from GIL01-lacZ promoter fusions and induction was similarly blocked 38

in the lexA (Ind-) host. These data support a model in which host LexA binds to dinBox 39

sequences in GIL01, repressing phage gene expression during lysogeny and providing the switch 40

necessary to enter lytic development. 41

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

3

Introduction 42

Temperate phages are defined by their ability to infect a susceptible host cell and opt between 43

either of two developmental pathways, the lytic or the lysogenic cycle. In the lytic cycle, the 44

phage genome is replicated and expressed to produce the structural components necessary to 45

assemble capsids. Once engaged, this pathway inevitably ends with the death of the host cell and 46

the release of a crop of newly formed virions in the environment. In contrast, the lysogenic cycle 47

is defined by the tight repression of lytic behavior and the phage genome replicates along with 48

that of the host cell, either physically integrated into the host genome or as an independent 49

plasmid prophage. The quiescent prophage is thereby efficiently inherited as cells divide, until 50

the lytic cycle is reactivated in response to alterations in host cell physiology. This transition 51

requires the efficient interaction of regulatory components that control the passage from a highly 52

stable stand-by mode to multiple rounds of genome replication, vigorous gene expression, virion 53

assembly and host lysis. 54

In the well-studied bacteriophage λ, establishment of lysogeny relies on the synthesis of CII, a 55

transcriptional regulator that directs the expression of the lytic cycle repressor CI (15). Once 56

produced, CI dimers bind cooperatively at operator sites in the so-called switch region to prevent 57

transcription from the early promoters PL and PR and subsequently, due to the lack of early gene 58

products, from the late lytic promoters as well. Expression of CI alone is sufficient to maintain 59

lysogeny once it has been established (26). Despite an intrinsically stable quiescent state, the λ 60

prophage can efficiently switch to an alternate lytic cycle. The λ CI repressor functions 61

analogously to the bacterial SOS response master repressor LexA, to which it shares significant 62

sequence similarity (21, 45). LexA binds as a dimer to specific operator sites and negatively 63

regulates the expression of a host of genes that are essentially involved in DNA repair and cell 64

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

4

survival (10, 18). Upon DNA damage, such as that inflicted by UV irradiation or mitomycin C 65

(MMC) treatment, the host recombinase RecA is converted to an activated form when it binds 66

single-stranded DNA. Activated RecA then acts as a co-protease to stimulate LexA auto-67

proteolysis and through a similar process, CI auto-proteolysis as well (43, 59). As a consequence, 68

LexA dissociates from its consensus sites to initiate the SOS response. Likewise, CI is no longer 69

able to bind its operator sites in the λ genome, leading to lytic gene expression and viral 70

proliferation. 71

Alternative mechanisms of regulation have been described in several SOS-inducible phages. The 72

repressors of coliphages 186 and N15, for example, are not RecA-sensitive but are instead 73

antagonized by phage-borne anti-repressors that are under LexA control (47, 61). In CTX�, the 74

filamentous phage that encodes cholera toxin in Vibrio cholerae, host LexA and phage repressor 75

function as both activators and repressors of gene expression to ensure the permanent production 76

and secretion of CTX� (36). 77

The Bacillus thuringiensis prophage GIL01 has been shown to be induced by DNA-damaging 78

treatments such as UV irradiation and MMC (69). The 15 kb linear prophage does not integrate 79

into the bacterial chromosome but instead, replicates independently within its host and 80

occasionally experiences induction of its lytic cycle. GIL01 shares a similar genome organization 81

and size with PRD1, the model tectivirus infecting Gram-negative hosts. However, unlike PRD1 82

and its close siblings, which are all lytic phages, GIL01 is able to enter a dormant state during 83

which its lytic functions are turned off. Two additional tectiviruses infecting B. thuringiensis 84

have also been described: Bam35, which is almost identical to GIL01 (55, 64) and has been the 85

object of many studies on the internal membrane that defines tectiviruses (20, 39-41), and 86

GIL16, which shares considerable sequence identity (83.6%) with GIL01 (67). A more distant 87

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

5

relative, pBClin15, has been sequenced alongside with the B. cereus reference strain ATCC 88

14579 genome but is not yet known to be a prophage (29). Recently, the sequence of AP50, a 89

temperate phage highly specific to B. anthracis strains, added a new member to the subgroup of 90

Tectiviridae that proliferate among members of the B. cereus group (62). 91

By analogy to other temperate phages, it has been assumed that GIL01 controls its lysogenic 92

cycle by expressing a lytic gene repressor. A potential candidate for this function was predicted 93

to be encoded by open reading frame 6 (ORF6), by virtue of its N-terminal LexA-like DNA-94

binding domain (67). However, unlike the phage λ CI repressor, ORF6 lacks any recognizable 95

protease linker required for repressor cleavage or C-terminal dimerization domain. Furthermore, 96

spontaneous GIL01 clear plaque (cp) mutants – that are unable to establish the lysogenic cycle – 97

were previously mapped to ORF7 but no such mutations were found to affect ORF6 (67). In this 98

study, we demonstrate that ORF7 is indeed involved in the regulation of the phage cycle and that 99

it does so along with ORF1. Importantly, we also show that GIL01 is induced in response to SOS 100

signals due to the direct control by the host LexA repressor. LexA likely binds to a conserved 101

operator, termed dinBox1 (for damage-inducible Box), that when mutated commits GIL01 to 102

propagate exclusively as a virulent phage. dinBox1 lies within two promoters, P1 and P2, that 103

direct the expression of regulation and replication functions that are likely important for 104

lysogenic development/maintenance. Two additional LexA binding sites, dinBox2 and dinBox3, 105

were identified at promoter P3, regulating the downstream structure and lysis module. Curiously, 106

no cp mutants were ever isolated that carried mutations within these operator sites, possibly 107

indicating that dinBox2 and dinBox3 are not required for lysogenic regulation but are important 108

for lytic growth. Altogether, we identified one cis-acting element (dinBox1), two phage-borne 109

genes (ORF1 and ORF7) and a host regulator (LexA) that determine the passage from the latent 110

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

6

state to lytic development in GIL01. The identification of these factors will ultimately help 111

elucidate the gene circuit that regulates GIL01 development. 112

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

7

Materials and Methods 113

Bacterial strains, phages and plasmids 114

The natural host of phage GIL01 is the insect pathogen B. thuringiensis subspecies israelensis 115

strain NB31 (2). The permissive plasmid-free strain GBJ002 (31), derived from strain 4Q2 and 116

related to NB31 (32), was used to propagate GIL01. In order to work in identical genetic 117

backgrounds, GBJ002 lysogens were generated as follows. GIL01 filter-sterilized suspensions 118

were spotted onto 0.7% top agar lawns seeded with GBJ002. Plates were incubated overnight at 119

37°C. The center of a lysed spot was then recovered with a sterile inoculation loop and streaked 120

on LB agar. After 16 hours at 37°C, individual colonies had grown and were screened by 121

polymerase chain reaction (PCR) with GIL01-specific primers to confirm lysogeny. A positive 122

colony was selected (GBJ338) and retained for the study. 123

GIL01 clear plaque (cp) mutants were obtained spontaneously by infecting GBJ002 with wild-124

type GIL01 phage suspensions. Briefly, phage preparations from serial 10-fold dilutions were 125

added to 200 µl of mid-log-phase GBJ002 and incubated at room temperature for 30 minutes. 126

The infected cells were then plated on LB agar with 4 ml of molten 0.7% top agar. The plates 127

were incubated overnight at 37°C and visually screened for the presence of clear plaques. Single 128

clear plaques, isolated from independent experiments, were picked with micropipette tips and 129

subjected to three rounds of single plaque purification on host GBJ002. The wild-type pGIL01 130

genome sequence is available from the NCBI GenBank database under the accession number 131

AJ536073. 132

Strain GBJ499 (recA-) was constructed by disrupting the recA gene of strain GBJ002 with the 133

integrative vector pMutin4 (Bacillus Genetic Stock Center). pMutin4 is unable to replicate in 134

Bacillus and is therefore particularly well-suited for insertional mutagenesis in this host (66). 135

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

8

Briefly, an internal fragment of recA (coordinates 3,719,072 through 3,720,009 in B. cereus 136

strain G9842, GenBank acc. number NC_011772) was amplified from strain GBJ002 with 137

primers RecAmutF (5’- GCGGCCGCTCTTCACCAATTCCGTGAT) and RecAmutR (5’- 138

GAATTCCAATTTGGTAAAGGTTCAATTA). The PCR product was digested with NotI and 139

EcoRI (restriction sites underlined), purified and cloned into NotI- and EcoRI-digested pMutin4 140

vector. The ligation was transformed into Escherichia coli (ER2925) and recombinant plasmids 141

were extracted and introduced into competent GBJ002. Bacillus transformants having integrated 142

the vector by homologous recombination were selected on 0.3 µg erythromycin ml-1. The 143

insertion of pMutin4 in the desired locus was confirmed by sequence analysis of recA. The 144

downstream gene, which encodes a phosphodiesterase (B1422), is separated from recA (B1421) 145

by a 500-bp non-coding region, including a Rho-independent terminator. It is therefore unlikely 146

that pMutin4 insertion in recA disrupts an operon. 147

Strain GBJ396 (lexAA96D) was constructed by substituting an Ala (GCT) residue at position 96 in 148

lexA with Asp (GAT). A 1,6-kb segment encompassing lexA (coordinates 3,623,569 through 149

3,625,169 in B. cereus strain G9842, GenBank acc. number NC_011772) was amplified from 150

strain GBJ002 with primers MutLex1 (5’- GGATCCTTGATATTCTTCAACAATAC) and 151

MutLex2 (5’- AAGCTTGTGTTAGAACACTTGCTGGA) and cloned into the pCR4-TOPO 152

vector (Invitrogen). The ligation was transformed into E. coli (TOP10) and recombinants were 153

extracted to undergo mutagenesis. A single nucleotide mutation was generated by using the 154

QuickChange II Site-Directed Mutagenesis kit (Stratagene) with primer pairs LexindF (5’- 155

GTCGGAAAAGTTACTGATGGTTTACCAATTACAGCG) and LexindR (5’- 156

CGCTGTAATTGGTAAACCATCAGTAACTTTTCCGAC) according to the manufacturer’s 157

instructions. DpnI-treated vector DNA was transformed into XL1-Blue E. coli cells and 158

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

9

transformants were selected on 50 µg kanamycin ml-1. Mutants were confirmed by DNA 159

sequence analysis with the universal sequencing primers M13. Mutated DNA from one selected 160

clone was restricted with BamHI and HindIII (restriction sites underlined in MutLex1 and 161

MutLex2) and the lexA-containing fragment was cloned into BamHI- and HindIII-digested 162

pMutin4 plasmid DNA. The ligation was transformed into E. coli (ER2925) and recombinant 163

plasmids were extracted and introduced into competent GBJ002. Similarly to above, Bacillus 164

transformants having integrated the vector by homologous recombination were selected on 0.3 165

µg erythromycin ml-1. One of these recombinants was grown for several consecutive cycles of 166

four hours in liquid LB and samples were recovered at each cycle and plated on LB agar at 167

different dilutions. Single colonies were subjected to UV irradiation (wavelength 254 nm) for 10 168

seconds and left at room temperature for two days. UV-sensitive colonies were amplified with 169

primers MutLex1 and MutLex2 and the resulting PCR products were sequenced in order to 170

screen for clones that had undergone a second recombination event resulting in the lexAA96D 171

genotype. Lysogens GBJ500 and GBJ406 were generated by infecting the mutant strains 172

GBJ499 and GBJ396 with GIL01 as explained above. 173

E. coli strain ER2925 (New England Biolabs), a dam- and dcm- K12 derivative, was used to 174

propagate un-methylated constructs prior to transformation into GBJ002. pCR4-TOPO-derived 175

constructs were transformed into E. coli strain TOP10 (Invitrogen) or XL1-Blue (Startagene). 176

ORF1 and ORF7 were cloned into the multicopy plasmid pDG148, an E. coli – Bacillus spp. 177

shuttle vector with an isopropyl-β-D-thiogalactopyranoside (IPTG)-inducible Pspac promoter 178

upstream of the multiple cloning site (63). The pCR4-TOPO (Invitrogen) positive selection 179

vector was used to clone 5’-RACE generated fragments for sequencing as well as the amplified 180

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

10

lexA gene and promoters P1-P2 segments for site-directed mutagenesis. pHT304-18Z is a low-181

copy-number shuttle vector used to generate lacZ fusions in B. cereus (1). 182

Growth conditions and phage induction 183

B. thuringiensis and E. coli strains were grown in modified Lysogeny Broth (LB) (6) (1% 184

tryptone, 0,5% yeast extract, 0,5% NaCl) at 37°C. B. thuringiensis liquid cultures were always 185

grown overnight with shaking at 200 rpm before being diluted 1:100 into LB and allowed to 186

grow exponentially. Transformation of B. thuringiensis and E. coli strains with the different 187

constructs was performed by electroporation (46, 57) and transformants were selected on LB 188

agar supplemented with 25 µg kanamycin ml-1 (pDG148-derived constructs), 50 µg kanamycin 189

ml-1 (pCR4-TOPO-derived constructs), 25 µg erythromycin ml-1 and 100 µg ampicillin ml-1 190

(pHT304-18Z-derived constructs in B. thuringiensis and E. coli, respectively) or 0.3 µg 191

erythromycin ml-1 and 100 µg ampicillin ml-1 (pMutin4-derived constructs in B. thuringiensis 192

and E. coli, respectively). 193

GIL01 was induced from GBJ338, GBJ500 and GBJ406 by UV irradiation and MMC treatment. 194

In both cases, exponential lysogenic cultures were pelleted and resuspended in 5 ml of 10 mM 195

MgSO4. For UV induction, the resuspended cells were irradiated at 254 nm for 10 seconds at a 196

distance of 10 cm. The irradiated cells were then incubated at 37°C and 200 rpm for 15 minutes, 197

pelleted and resuspended in liquid LB. Phage induction was allowed to proceed for 90 minutes at 198

37°C and 200 rpm before culture supernatants were recovered and filtered (0.45 µm). For MMC 199

induction, MMC was added to 5 ml of resuspended cells at a concentration of 50 ng ml-1. The 200

cells were then incubated at 37°C and 200 rpm for one hour before being pelleted and washed 201

twice with 10 mM MgSO4. The pellets were finally resuspended in liquid LB and phage 202

induction was allowed to proceed for 90 minutes at 37°C and 200 rpm before culture 203

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

11

supernatants were recovered and filtered (0.45 µm). Control samples that were not irradiated or 204

MMC-induced were otherwise treated equally. GIL01 titers from induced and un-induced 205

cultures were estimated by infecting GBJ002 in soft agar (0.7%). 206

DNA, enzymes and reagents 207

Nucleotides and DNA modification enzymes were purchased from Roche and Fermentas. The 208

Phusion Flash High-Fidelity PCR Master Mix (Finnzymes) and GoTaq Flexi DNA polymerase 209

(Promega) were used in PCR amplifications. Oligonucleotides, antibiotics, mitomycin C, 210

isopropyl-β-D-thiogalactopyranoside (IPTG) and GenElute Plasmid Miniprep Kits were 211

purchased from Sigma-Aldrich. QIAquick Gel Extraction and PCR Purification Kits were 212

purchased from Qiagen. 213

Primer extension assays 214

Total RNA was extracted from mid-log-phase GBJ338 cultures with the Ambion RiboPureTM-215

Bacteria Kit and treated with DNase DNA-freeTM solution (Ambion). Transcription start sites 216

were determined with the 5’ RACE System for Rapid Amplification of cDNA Ends (Invitrogen) 217

according to the manufacturer’s instructions. Phage-specific primers hybridizing downstream of 218

P1-P2 (5’- TGTCCACTGTCATAATATCATTT, at coordinates 619-641) and P3 (5’- 219

TTGGAGATTATTTATTGACAAGA, at coordinates 5,233-5,255) were used to synthesize 220

cDNA strands from GIL01 transcripts. cDNAs were tailed and PCR-amplified before being 221

cloned into pCR4-TOPO and sequenced. 222

lacZ fusions and β-galactosidase assays 223

The region encompassing promoters P1 and P2 as well as dinBox1 (coordinates 41 through 413) 224

was amplified using primers pHT1 (5’- AAGCTTGACATGTAATAACATTTATG) and pHT4 225

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

12

(5’- GGATCCTCTTTTTATGAACACGCA). The corresponding product was purified and 226

cloned into the selection vector pCR4-TOPO. The ligation was transformed in competent TOP10 227

cells and recombinants were extracted and restricted with enzymes HindIII and BamHI. The gel-228

purified fragments were then ligated into HindIII- and BamHI-digested pHT304-18Z to obtain 229

construct pDin1-2 in strain ER2925. A similar strategy was followed to generate a lacZ fusion 230

with promoter P3. The region encompassing the promoter sequence, dinBox2 and dinBox3 231

(coordinates 4,751 through 4,976) was amplified with primers pHT5 (5’- 232

AAGCTTAGACAAGCAAAACCGACA) and pHT6 (5’- 233

GTCGACTATGATTAGTTCCGTTAAAG), cloned into pCR4-TOPO, restricted with HindIII 234

and SalI and ligated to HindIII- and SalI-digested pHT304-18Z to obtain construct pDin3. Both 235

constructs pDin1-2 and pDin3 were confirmed by DNA sequence analysis and transformed to 236

competent GBJ338 cells. To generate pDin1-2mut, in which dinBox1 contains the same single 237

nucleotide mutation as in Bam35c (CGAACaagcGTTTT to CGAACaagcTTTTT), the 238

QuickChange II Site-Directed Mutagenesis kit (Stratagene) was used with primer pairs DinIndF 239

(5’- GTTAAAAAACGAACAAGCTTTTTATAAGTGTTCGG) and DinIndR (5’- 240

CCGAACACTTATAAAAAGCTTGTTCGTTTTTTAAC) on construct pCR4-TOPO containing 241

the Din1-2 insert. After verification of the desired mutation by sequence analysis, the same 242

cloning steps as for the generation of wild-type pDin1-2 were followed. 243

Bacillus transformants were grown in liquid cultures supplemented with 10 µg erythromycin ml-1 244

to exponential phase, at which point MMC 50 ng ml-1 was added to one-half of the cultures. 245

Cultures were further incubated for one additional hour and samples (20 µl) from each different 246

condition tested were taken. β-galactosidase activities were assayed as previously described (16) 247

with some modifications. Bacterial cells were permeabilized with 980 µl of buffer Z (containing 248

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

13

50 mM β-mercaptoethanol) and 10 µl of toluene. After 5 minutes at 37°C, 200 µl of buffer Z 249

supplemented with ONPG (4 mg ml-1) were added to each sample. The treated samples were 250

incubated at 37°C for a time period ranging from 20 to 30 minutes and reactions were stopped by 251

adding 500 µl of Na2CO3 1M. The samples were centrifuged for 7 minutes at 13,000 rpm and 252

supernatants were collected. Optical densities were measured at a wavelength of 420 nm. The β-253

galactosidase activity values are expressed as Miller units and were calculated with the formula: 254

1000*(A420/A595*time (minutes)*volume (0.02)). 255

pDG1, pDG7 and pDG1/7 plasmid constructions and complementation tests 256

ORF1 (coordinates 356-532) was amplified from GIL01 DNA with primers DG1F (5’- 257

GCTCTAGATGAGTAAACTACTGACTGCG) and DG1R (5’- 258

TGCGCATGCTTAATTGTCCAGTTTATTTT) and digested with XbaI and SphI (restriction 259

sites underlined). The restriction fragment was purified and ligated to XbaI- and SphI-digested 260

pDG148 plasmid DNA to generate pDG1. pDG7 was constructed similarly by using primer pairs 261

DG7F (5’- GCTCTAGATGCGTGACAAATTGCTCG) and DG7R (5’- 262

TGCGCATGCTCAGTCATCCTTCTTCCCC) (coordinates 4,564 through 4,716). An ORF1-263

ORF7 transcriptional unit was generated by amplifying ORF1 with primers DG1_rbsF (5’- 264

TGCAAGCTTGaaggtggtgCATATGTAAT) and DG1_rbsR (5’- 265

GCTCTAGATTAATTGTCCAGTTTATTTTTA) (coordinates 336 through 532, including the 266

ORF1 ribosome binding site shown in Fig. 1C in lower case), restricting the PCR product with 267

HindIII and XbaI, and cloning the purified product into HindIII- and XbaI-digested pDG148 268

plasmid resulting in construct pDG1_rbs. ORF7 was amplified with primers DG7_rbsF (5’- 269

GCTCTAGAaaggagggAAAAGGTATG) and DG7_rbsR (5’- 270

GCTCTAGATCAGTCATCCTTCTTCC) (coordinates 4,549 through 4,716, including the ORF7 271

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

14

ribosome binding site shown in lower case). After digestion with XbaI and gel purification, the 272

fragment was ligated into XbaI-digested pDG1_rbs plasmid, resulting in construct pDG1/7. 273

pDG1, pDG7 and pDG1/7 ligations were transformed into competent E. coli (ER2925) and 274

recombinant plasmids were extracted and confirmed by DNA sequence analysis. Competent 275

GBJ002 was transformed with un-methylated pDG1, pDG7 and pDG1/7 (containing ORF7 in 276

the correct orientation) to yield the clones GBJ250, GBJ223 and GBJ467, respectively. Protein 277

expression was induced by adding IPTG to a final concentration of 1 mM to one-half of liquid 278

cultures in early exponential growth phase. Control cultures were grown without addition of 279

IPTG. After an additional 3-hours incubation at 37°C with shaking, 200 µl of induced cultures 280

were infected with three 10-fold serial dilutions of cp mutant phage stocks and plated with 281

molten top agar (0.7%) on LB plates supplemented with IPTG to a final concentration of 1 mM. 282

Un-induced cultures followed the same treatment but were plated on LB plates devoid of IPTG. 283

The plates were incubated overnight at 37°C and plaque counts were determined. Plate pictures 284

were taken with the Bio-Rad Molecular Imager Gel Doc XR System using the Quantity One 285

version 4.6.3. imaging software. 286

Sequencing and Sequence analyses 287

Constructs were sequenced at the Flanders Interuniversity Institute for Biotechnology (VIB) 288

Genetic Service Facility (University of Antwerp, Belgium) by double strand primer walking and 289

with individual quality control of sequence reads included. Cp mutants were amplified with 290

GIL01-specific primer pairs designed to cover the 15-kb genome in segments of 1 kb with 100-291

bp overlaps. PCR products were purified and sequenced with a BigDye® Terminator v3.1 Cycle 292

Sequencing Kit on an ABI Prism 3130xl Genetic Analyzer (both from Applied Biosystems). Cp 293

genomes were sequenced on both strands to provide 3- to 4-fold sequence coverage. Sequences 294

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

15

were analyzed with the EMBOSS package at the Belgian EMBL node and the EMBL-EBI tools. 295

GBJ002 sequences were compared to the genome of strain G9842 (GenBank acc. number 296

NC_011772), to which GBJ002 shares most sequence identity. 297

Phylogenetic footprints discovery 298

To discover evolutionary conserved elements, we used the program footprint-discovery from 299

RSAT (Regulatory Sequence Analysis Tools) (65) and followed the protocol as described for the 300

study case 2 (17). This strategy has been systematically evaluated on E. coli and gives good 301

results at different taxonomical levels for well-conserved genes such as lexA (30). The algorithm 302

was run on the lexA and recA promoter sequences from the B. cereus reference strain ATCC 303

14579 (GenBank acc. number AE016877) and the GIL01 promoters. 304

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

16

Results 305

The GIL01 genome has two major functional domains 306

Like most phage genomes sequenced to date, GIL01 contains closely packed genes with few 307

intergenic spaces. All GIL01 genes are transcribed in the same rightward direction and the ORFs 308

are frequently overlapping, suggesting that GIL01 genes are organized into several operons (Fig. 309

1A). Interestingly, genes with similar functions cluster together, forming two functional 310

modules: the left arm of approximately 4.8 kbp contains replication and regulatory genes, such as 311

the predicted terminal protein (ORF4), the DNA polymerase (ORF5) and two proteins with 312

predicted DNA-binding domains (ORF1 and ORF6). The larger downstream module (10 kbp) 313

contains all the structural and assembly genes as well as two endolysins (ORF26 and ORF30) 314

involved in cell wall degradation (68). Two non-coding segments stand out by their length and 315

position in the genome as containing potential promoter/regulatory elements. The first one lies in 316

the left extremity, upstream of ORF1 (355 bp), while the second one separates ORF8 from ORF9 317

(118 bp) (Fig. 1A). Considering this genetic organization, both non-coding segments were 318

examined for the presence of promoters using RACE PCR, a rapid method for the amplification 319

of 5’-cDNA ends. The technique relies upon amplifying RNA transcripts between a defined 320

internal site and the unknown 5’-extremity of interest, thus allowing for the identification of the 321

transcription start site. Using RNA extracted from a lysogenic culture in the exponential growth 322

phase, two transcription start sites were identified upstream of ORF1, at positions 175 and 278, 323

and one initiation site was identified in the intergenic region between ORF8 and ORF9, at 324

coordinate 4,915 (Fig. 1B). Centered approximately -10 and -35 nucleotides upstream of each 325

transcription start site are conserved sigma A-dependent promoter sequences (Fig. 1A and 1C). 326

The two potential promoters on the left were named P1 and P2 and the internal promoter, P3. It 327

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

17

is interesting to note that the -10 hexamer in P1 is preceded by a TG dinucleotide, positioned 1 328

base upstream of the -10 sequence (Fig. 1C), a feature found in so-called extended -10 329

promoters. In E. coli and B. subtilis, the TGn extension allows for closer contacts between the 330

RNA polymerase sigma subunit and its promoters, especially when a well-conserved -35 box is 331

absent (5, 11, 34, 37, 58, 70). This could be the case for P1, where the -35 box does not begin 332

with T, the predominant base found at the beginning of the -35 hexamer in Gram-positive 333

bacteria (70). 334

In addition to the reported promoter sequences, potential ribosome binding sites (RBSs) were 335

identified 10 and 6 nucleotides upstream of the start codons of ORF1 and ORF9, respectively. 336

Also, Rho-independent transcription terminators were identified downstream of ORF8, where 337

transcription of the replication/regulation operon would be expected to terminate, and 338

downstream of ORF30, where transcription of the lysis/assembly operon is expected to terminate 339

(Fig. 1A). We confirmed that transcription does not span the regulatory region between ORF8 340

and ORF9 by performing RT-PCR with primer sets that span this region (data not shown). 341

Finally, lacZ reporter fusions were generated to the regions containing P1-P2 and P3 in order to 342

verify that these are biologically active promoters. As shown in Table 1, both promoter regions 343

directed the expression of β-galactosidase in the GIL01 B. thuringiensis host. Such 344

transcriptional organization and the corresponding gene distribution further strengthen the idea 345

that the GIL01 genome is transcribed as two major operons, one responsible for genome 346

replication and regulation, the other for virion structure, assembly and host cell lysis. 347

GIL01 clear plaque mutants map to three distinct loci 348

A chief objective of this study was to identify the components of the genetic circuit governing 349

GIL01 development. The only obvious candidates for transcription regulation were ORF1 and 350

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

18

ORF6. PSI-BLAST searches identified a potential helix-turn-helix (HTH) DNA-binding motif 351

in ORF1 similar to the HTH found in the MerR super-family of transcription regulators. In 352

ORF6, the same database searches revealed a potential winged HTH DNA-binding domain 353

similar to that found in the N-terminus of LexA, the master repressor of the SOS regulon. We 354

reasoned that since GIL01 is induced in DNA-damaging conditions, most likely by way of the 355

cellular SOS response, it is potentially under the transcriptional control of a cleavable repressor 356

analogous to CI in phage λ. However, neither ORF1/ORF6 nor any other protein-coding region 357

in GIL01 displays the typical structure of the CI/LexA class of repressors: an N-terminal DNA-358

binding domain, followed by an Ala-Gly or Cys-Gly cleavage site and a C-terminal protease 359

domain. Yet, we did not rule out the existence of a distinct type of SOS-dependent induction 360

mechanism, such as the LexA-regulated anti-repressors found in coliphages 186 (8, 38) and N15 361

(47) or that GIL01 induction relies on a yet unknown SOS pathway. 362

To investigate these possibilities, we took advantage of knowledge acquired from a previous 363

study of spontaneous GIL01 clear plaque (cp) mutants defective for the lysogenic cycle (67). 364

Since ORF6 was a strong candidate for encoding a repressor of the lytic cycle, the first cp 365

mutants had been sequenced within and in the vicinity of ORF6. Unexpectedly, no mutations 366

were detected in ORF6. Instead, mutations were found in the downstream ORF7, the function of 367

which was unknown. ORF7 cp mutants displayed either a deletion or insertion of 11 nucleotides 368

in an almost perfect direct repeat (DR) (5’-CAAGTCGTAaG CAAGTCGTAtG) at coordinates 369

4,615 through 4,636. 370

In this study, a more extensive collection of cp mutants was assembled from multiple 371

independent experiments and analyzed. GIL01 cp mutants arose at a frequency of approximately 372

10-4 and displayed otherwise normal growth properties. In addition to the expected DR mutations 373

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

19

in ORF7 (deletions in cp04 and cp18, insertions in cp25 and cp29), a mutant (cp08) carrying a 374

single nucleotide substitution in the second repeat, resulting in a premature terminator codon, 375

was also isolated. Another ORF7 mutant, cp27, has a single nucleotide insertion downstream of 376

the DR (Fig. 2A). In every instance, the mutation altered the predicted protein sequence, cp04, 377

cp08, cp18 , cp25 and cp29 all resulted in truncation of ORF7, while cp27 resulted in a modified 378

C-terminus (Fig. 2B). 379

Additionally, a class of GIL01 cp mutants was isolated that lacked mutations in ORF7, 380

suggesting that mutations in other regions of the GIL01 genome might account for their cp 381

phenotype. This class of mutants all contained single nucleotide substitutions, deletions or 382

insertions within ORF1 (Fig. 2A) or they contained mutations immediately upstream of ORF1 383

between promoters P1 and P2 (Fig. 1C and 3A). Interestingly, mutations in ORF1 cluster in two 384

regions: the predicted HTH DNA-binding motif (cp32, cp33 and cp36) or the most C-terminal 385

residues (cp06 and cp10) (Fig. 2B). Mutations in the non-coding region upstream of ORF1, 386

hereafter termed dinBox1, resulted in base substitutions (cp16, cp19, cp23 and cp26) or single 387

nucleotide deletions (cp17 and cp37) at either of two nucleotide positions within a 14-bp 388

sequence that bears strong similarity to the LexA binding site in Gram-positive bacteria (Fig. 389

3B). As outlined below, a single mutation in dinBox1 is sufficient to abolish the lysogenic 390

pathway of GIL01. 391

To rule out the possibility that additional mutations, outside the sequenced areas, contributed to, 392

or were responsible for, the observed cp phenotype, we sequenced the entire GIL01 genome of 393

cp26 (mutated in dinBox1), cp27/cp29 (mutated in ORF7) and cp32/cp33 (mutated in ORF1). 394

cp26 (dinBox1), cp27 (ORF7) and cp33 (ORF1) only displayed the initially identified nucleotide 395

substitutions (cp26 and cp33) and insertion (cp27). cp29 and cp32, however, each contained one 396

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

20

additional mutation. In cp29, a single nucleotide substitution changed a Gln to Lys 397

(CAA→AAA) at residue 19 of ORF29, the function of which is currently unknown. cp32 398

contained an additional deletion of three nucleotides at codon 64 within ORF19 (function 399

unknown), resulting in the in-frame deletion of a single Ile residue. Yet, cp29 and cp32 display 400

the same infectious behavior as their cognate cp27 and cp33 mutants, which are solely mutated 401

in ORF7 and ORF1, respectively (see complementation experiments below and Table 2). This 402

strongly suggests that ORF18 and ORF29 are not involved in the regulatory network of GIL01. 403

Taken together, our results indicate that dinBox1, ORF1 and ORF7 are critical for the 404

establishment or maintenance of the GIL01 lysogenic cycle. 405

Ultimately, we wished to classify cp mutants according to their degree of virulence on lysogen 406

and non-lysogen hosts. However, none of the cp mutants tested produced plaques on lawns of the 407

lysogen strain GBJ338, indicating that GIL01 lysogens are immune to infection by cp mutants. 408

This could be due to the activity of phage- and/or host-encoded immunity factors that shift 409

GIL01 towards lysogenic development. Alternatively, we cannot rule out the possibility that 410

infection was barred by a yet unidentified phage-borne super-infection exclusion system. 411

gp1 and gp7 regulate lysogenic development 412

Since mutations in either ORF1 or ORF7 commit GIL01 to multiply exclusively as a lytic phage, 413

we investigated more closely the roles of their respective gene products, gp1 and gp7, in GIL01 414

development. The corresponding genes were cloned into the E. coli - B. subtilis shuttle vector 415

pDG148, placing them under the transcriptional control of the IPTG-inducible Pspac promoter. 416

Three different plasmids were generated, whereby gp1 and gp7 were expressed individually or 417

together as a single transcriptional unit. The permissive B. thuringiensis host GBJ002 was 418

transformed with the resulting plasmids, generating strains GBJ250 (gp1), GBJ223 (gp7) and 419

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

21

GBJ467 (gp1/gp7). Each of these strains were induced for protein expression and subjected to 420

infection with GIL01 cp mutants in dinBox1, ORF1 or ORF7. The results of these experiments 421

are summarized in Table 2. Expression of gp1 alone reduced the infective titers of ORF1 mutants 422

cp32 and cp33 approximately 10-fold. Surprisingly, gp1 expression also reduced the plating 423

efficiency of dinBox1 and ORF7 mutants, although to a lesser extent (2-fold and 5-fold PFU 424

reduction, respectively). This was not the case when gp7 was expressed in the host strain. In this 425

case, a reduction in phage titers was only observed when induced GBJ223 cells were infected 426

with the ORF7 mutants cp27 and cp29. Although titers only showed a 2-fold decrease, the 427

plaques were now turbid and considerably smaller than those seen on the un-induced strain 428

GBJ223 (Fig. 4). This observation was less remarkable when ORF1 mutants were plated on 429

induced cultures of GBJ250, as ORF1 mutants tend to produce smaller plaques in general. 430

Infection of these same induced cell lines with wild-type GIL01 yielded titer reductions that 431

varied between 2-fold and 10-fold when compared to infections of un-induced cultures, while 432

plaque morphology remained unaltered (data not shown). This noticeable reduction in titers, as 433

well as the considerable experiment-to-experiment variability in titers, could be explained by 434

variations in the cellular concentrations of gp1 and gp7 generated with our plasmid-borne, IPTG-435

inducible system. Wild-type GIL01 probably expresses balanced levels of gp1 and gp7 inside the 436

host cell and disrupting that balance is likely to affect phage behavior in many ways. These data 437

are consistent with the idea that gp1 and gp7, and possibly the proper expression rates of these 438

two factors, play important roles in the choice between lytic and lysogenic development. 439

Given the above results, we investigated whether there was any effect due to the simultaneous 440

expression of both gp1 and gp7 in the cell. There was indeed a substantial reduction in phage 441

titers when induced GBJ467 cells were infected with dinBox1 mutants and no plaques at all were 442

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

22

observed with ORF1 mutants (Table 2). In contrast, no difference in plating efficiency was 443

observed between infections of induced GBJ250 and GBJ467 with ORF7 mutants. Hence, co-444

expressing gp1 and gp7 in the same strain enhances complementation of dinBox1 and ORF1 445

mutants but does not result in enhanced rescue of ORF7 mutants. 446

Since complementation was associated with sharp reductions in infective titers, we wished to 447

verify that expression of gp1 and/or gp7 conferred immunity by preventing the infecting mutant 448

from propagating, rather than by blocking infection altogether. In order to investigate the fate of 449

the cp mutants post-infection, bacterial samples were recovered from infected spots and streaked 450

on fresh plates containing IPTG. PCR reactions with GIL01-specific primers were performed on 451

isolated colonies in which gp1 and/or gp7 were expressed and which had been infected with 452

mutants cp26 (dinBox1), cp27 (ORF7) and cp33 (ORF1). Induced GBJ223 colonies were only 453

found to be positive for GIL01 when infected with cp27, showing that expression of gp7 454

complements the missing immunity function of the corresponding cp mutant. GIL01-positive 455

colonies were also obtained from GBJ250 spot infections with cp33. Similar results were 456

obtained when induced GBJ467 was infected with each of the three cp mutants. These results are 457

in agreement with the infection patterns described above (Table 2) and altogether, they support 458

the idea that gp1 and gp7 act in immunity against infection and, together with the dinBox1 459

element, play important roles in the regulatory network of GIL01. 460

dinBox1 is a conserved LexA box 461

dinBox1 showed strong similarity to the consensus LexA binding site (CGAACn4GTTCG) in B. 462

subtilis (13, 30, 72). Its sequence, CGAACaagcGTTTT, is centered at positions +17 and -86 with 463

respect to the P1 and P2 transcription start sites, respectively (Fig. 1C). Cp mutants at dinBox1 464

were unable to enter lysogeny and mutations always involved changes to strongly conserved 465

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

23

positions within the SOS box. Interestingly, Bam35c, a cp mutant of phage Bam35 (55), only 466

differs from GIL01 at a few nucleotide positions, one of which is located within the critical bases 467

for LexA binding to dinBox1 (Fig. 3A). In agreement with this finding, experimental data for the 468

LexA box sequence requirements in B. subtilis shows that changes to any one of the five outer 469

nucleotides can considerably reduce or prevent LexA binding (24). In addition to conserved 470

dinBox1, two motifs with similar characteristics were found near promoter P3. dinBox2 and 471

dinBox3 are centered at -22 and +11 relative to the start of transcription (Fig. 1C). It is worth 472

noting that, while numerous mutations were isolated in dinBox1, none of the cp mutants isolated 473

in this study carried mutations in dinBox2 or dinBox3. 474

To further consolidate our observations, we searched for SOS-box sequences in the upstream 475

regions of SOS genes in B. cereus strain G9842, which is closest to strain GBJ002 in DNA 476

sequence. We compared the LexA operators found upstream of genes involved in transcriptional 477

regulation (lexA), recombination (recA and ruvAB) and DNA repair (uvrAB) to the consensus 478

sequence for LexA binding in B. subtilis and to the three dinBox sequences identified in GIL01. 479

As shown in Fig. 3B, the B. cereus SOS boxes and the three GIL01 dinBox sequences closely 480

match the B. subtilis consensus SOS box. Additionally, an in silico approach using the program 481

RSAT (17) identified the same three dinBox sites found by visual analysis of the GIL01 482

promoter sequences (Fig. 3B). Lastly, alignment of the B. cereus LexA protein sequence with the 483

Gram-positive consensus LexA sequence reveals strong conservation of the N-terminal regions 484

that make close contacts with DNA (Fig. 3C). Taken together, these observations strongly 485

suggest that the dinBox sequences identified in GIL01 function as LexA binding sites in the B. 486

cereus group of bacteria. 487

GIL01 is induced in DNA-damaging conditions in a LexA-dependent manner 488

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

24

In a previous study, we reported that GIL01 was induced by DNA-damaging agents such as 489

mitomycin C (MMC), nalidixic acid or exposure to UV irradiation (69). Such treatments are 490

known to elicit the cellular SOS response in a RecA-dependent manner (71). To investigate if 491

RecA is involved in the induction cycle of GIL01, we disrupted the B. thuringiensis recA gene 492

through insertional mutagenesis (66). The resulting strain, GBJ499, was particularly sensitive to 493

UV and MMC treatment and displayed a slower growth rate than the wild-type strain (data not 494

shown). GBJ499 was infected with GIL01 to generate the lysogen GBJ500. Phage titers 495

produced by spontaneous induction of GBJ500 were reduced approximately 10-fold compared to 496

the wild-type lysogen GBJ338 (Table 3). In addition, whereas GIL01 was strongly induced after 497

UV or MMC treatment of the wild-type lysogen strain, there was no increase in phage titers 498

when GBJ500 is similarly induced (Table 3, lines 2 and 5). These results indicate that GIL01 499

induction is RecA-dependent and potentially linked to the B. thuringiensis SOS response 500

pathway. To investigate this hypothesis, the role of cellular LexA in the lytic switch was 501

assessed. We were unable to isolate a lexA null mutant of B. thuringiensis. This result was not 502

entirely unexpected considering that LexA represses the expression of cell division inhibitors in 503

other bacterial groups (sulA in E. coli and yneA in B. subtilis). In E. coli, lexA null mutants are 504

not viable unless sulA is first inactivated (52). In B. subtilis, lexA-deficient cells grow into long 505

filaments (25) as a consequence of YneA upregulation (33). The B. cereus LexA is 70% identical 506

to its counterpart in B. subtilis and most of the shared residues lie in the N-terminal DNA-507

binding domain (Fig. 3C). Most importantly, B. cereus LexA contains an Ala96-Gly97 peptide 508

sequence that in B. subtilis and E. coli (at Ala91-Gly92 in both organisms) is the site of RecA-509

stimulated auto-proteolysis (28, 51). Certain amino acid changes at the Ala-Gly cleavage site are 510

known to abolish LexA proteolysis (19, 42). We used a similar approach to construct a non-511

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

25

cleavable lexA mutant by allele exchange in B. thuringiensis, replacing the Ala codon with an 512

Asp codon (lexAA96D). As expected, the resulting strain, GBJ396, was hyper-sensitive to DNA 513

damage conditions, displaying a 300-fold reduction in survival after treatment with MMC, 514

compared to a 9-fold reduction in survival of MMC-treated wild-type cells. Furthermore, a lexA-515

lacZ reporter fusion, when introduced into GBJ002 (lexA+), showed an expected 4-fold increase 516

in lacZ expression after treatment with MMC. However, there was no increase in lacZ expression 517

after SOS treatment of GBJ396 (lexAA96D) carrying the same lexA-lacZ reporter plasmid, 518

indicating that LexA was not proteolyzed and therefore not released from its operator sites by 519

SOS induction signals (data not shown). The lexAA96D mutant was infected with GIL01 to 520

generate the lysogen strain GBJ406. Similar to what we observed for the recA- lysogen, GIL01 521

induction from GBJ406 was strongly reduced in comparison to the wild-type host strain (Table 522

3, lines 3 and 6). The data obtained with both recA and lexA mutants show that GIL01 is induced 523

as part of the host SOS response pathway. 524

In order to assess if promoters P1-P2 are regulated by LexA, we determined whether a 373-bp 525

region encompassing both promoters and dinBox1 could direct transcription of a promoter-less 526

lacZ gene upon SOS induction. The 373-bp segment was cloned into the low-copy-number 527

pHT304-18Z plasmid (1), transcriptionally fusing the 5’ end of ORF1 to the lacZ reporter gene. 528

The construct was transformed to GBJ338 (lexA+) and GBJ338 (lexAA96D) and β-galactosidase 529

expression from the P1.P2-lacZ fusion was measured with and without SOS induction by MMC. 530

Table 1 shows that expression from promoters P1-P2 was induced approximately 4-fold in the 531

wild-type host in SOS conditions whereas no increase in β-galactosidase levels was observed in 532

the lexAA96D strain. Moreover, β-galactosidase expression from a P1.P2-lacZ fusion in which 533

dinBox1 was mutated at a critical residue was no longer repressed in either GBJ338 (lexA+) or 534

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

26

GBJ338 (lexAA96D) and MMC treatment did not further increase LacZ expression. Taken 535

together, these data strongly indicate that LexA regulates P1-P2 and that dinBox1 is necessary 536

for this LexA-mediated regulation. Parallel experiments with promoter P3, in which a 226-bp 537

fragment including dinBox2 and dinBox3 was fused to lacZ, yielded a 9-fold increase in β-538

galactosidase activity in the wild-type host after MMC treatment. Once again, the increase in 539

lacZ expression levels after MMC treatment was eliminated when the same experiment was 540

performed in a host carrying the lexAA96D mutation. These experiments demonstrate that host 541

LexA regulates promoters P1-P2 and P3, most likely by binding to the dinBox sequences and 542

preventing transcription initiation. 543

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

27

Discussion 544

In this study we used a genetic approach to identify central components of the system regulating 545

the temperate state in phage GIL01. We provide evidence that the host SOS repressor LexA is 546

directly involved in the switch but that phage-borne factors gp1 and gp7 are also necessary for 547

lysogeny. LexA is likely to act by binding to dinBox1, a conserved SOS box lying between two 548

newly identified promoters upstream of lysogeny and replication genes. Two additional 549

conserved LexA binding sites were also identified upstream of structure and lysis genes, further 550

suggesting that LexA plays a central role in GIL01 development. Our results lead to the 551

conclusion that GIL01 uses a combination of host- and phage-derived regulators to control its 552

cycle. 553

At first glance, GIL01 does not appear to display the characteristic modular organization usually 554

encountered in temperate phages. Its 15 kbp genome is smaller than most sequenced prophage 555

genomes and all 30 ORFs are transcribed in the same rightward direction, a rarely observed 556

feature. In addition, only a few genes have attributed functions, making it difficult to define the 557

boundaries of gene clusters involved in related functions. Most notably, a lysogeny module is not 558

readily apparent as there are only two gene candidates, ORF1 and ORF6, predicted to code for 559

DNA-binding proteins that could function as transcription regulators. Nevertheless, the 560

identification of promoters in large non-coding segments and Rho-independent terminators at the 561

end of transcriptional units revealed a functional organization into two major modules (Fig. 1A). 562

Two adjacent promoters, P1 and P2, at the left extremity of the phage genome, control 563

transcription of lysogeny and replication functions whereas a probable single internal promoter, 564

P3, directs the expression of structural and lytic genes. Such transcriptional organization 565

typically allows for complex, multi-layered regulation of gene transcription. It would, for 566

instance, permit the expression of regulatory factors and phage DNA polymerase to levels that 567

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

28

are required for lysogeny while keeping lytic genes fully repressed. Conversely, upon perception 568

of an inducing signal, synthesis of terminal protein and DNA polymerase would be significantly 569

enhanced along with induction of structural and lytic genes. 570

A similar overall genome organization can be found in the B. subtilis lytic phage φ29, with the 571

notable difference that early and late genes are divergently transcribed. Tandem promoters A2b 572

and A2c jointly direct the expression of the leftward regulation and replication genes, which are 573

similarly distributed in GIL01 (49). Both promoters are strong and their repression by two early 574

factors coincides with late gene transcription activation from the flanking divergent promoter 575

A3, a determining step in the switch from early to late transcription (11, 50). The coupling of 576

tandem promoters to the same operator is frequently observed in bacteria (examples are the 577

rRNA operons (23, 73), crp gene (22) and fis operon (53)) and offers a certain flexibility for 578

differential gene regulation. In the autonomously replicating coliphage N15 for example, two 579

adjacent promoters direct gene expression from the anti-immunity locus immA and their 580

differential control is important for the choice between lysis and lysogeny (56). Similarly 581

arranged promoters regulate anti-immunity operons in distinct phages such as P1, P4 and P7 582

(56), suggesting that this regulatory strategy is widespread among temperate phages. It is 583

therefore conceivable that in GIL01, promoters P1 and P2 display distinct strengths, are subject 584

to variable levels of regulation or give rise to different transcripts. 585

GIL01 lytic development is induced by DNA-damaging treatments that elicit the cellular SOS 586

response and yet, GIL01 does not code for a cleavable CI-like repressor. Instead, we found that 587

regulation is mediated by host LexA since the phage lytic pathway was no longer induced in a 588

non-cleavable lexA (Ind-) background. Accordingly, three putative LexA operators that bore a 589

strong resemblance to the consensus LexA binding site in Gram-positive bacteria were identified 590

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

29

between promoters P1 and P2 (dinBox1) and at P3 (dinBox2 and dinBox3). We showed that the 591

SOS induction of both promoter regions was blocked by a non-cleavable lexA (Ind-) host 592

mutation. Moreover, a single nucleotide mutation at a strongly conserved position within 593

dinBox1 completely eliminates transcription repression at P1-P2. Although additional 594

experiments will be needed in order to confirm LexA binding to dinBox1, these data lend 595

credence to the idea that host LexA is directly involved in the GIL01 control circuit. 596

Whereas numerous cp mutants were isolated in dinBox1, no mutations were ever detected in the 597

downstream dinBox2 or dinBox3. This suggests that dinBox2 and dinBox3 are either essential for 598

phage development or, if such mutants are indeed viable, they cannot be detected as clear plaque 599

formers. Also dinBox2 and dinBox3 could perform an additional function in the regulatory 600

system such as in LexA-mediated DNA looping between dinBox2/3 and the dinBox1 promoter 601

region. In this case, a cp mutant might not be obtained unless both boxes were simultaneously 602

mutated, which is likely to occur at too low a frequency to be detected. In any event, LexA 603

binding to both sites provides the basis for a plausible model in which a cleavable repressor 604

would be immediately responsible for lytic gene expression. However, it is unlikely that LexA is 605

the sole regulator of the switch as this would lead to inefficient GIL01 induction. Due to the 606

negative auto-regulation of lexA transcription, normal LexA levels are restored once DNA 607

damage inside the cell has been repaired (7, 44). LexA would then be able to bind its operators in 608

GIL01 again and interrupt lytic development. In phage lambda, for instance, this is prevented by 609

strong repression of CI synthesis once lytic development commences (60). It is therefore 610

conceivable that LexA only plays an ancillary role in the regulatory network by controlling the 611

expression of phage-borne repressor(s) and activator(s), these in turn being responsible for the 612

lytic switch per se. This hypothesis is strongly supported by the isolation of GIL01 cp mutants 613

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

30

deficient for gp1 or gp7. Both proteins are required for lysogeny and their absence can be 614

complemented by expressing the missing functions in trans. Interestingly, cross-615

complementation is observed with gp1 whereas it is not seen with gp7. The observed differences 616

could be attributed to distinct and/or independent roles of gp1 and gp7, gp1 potentially having a 617

broader activity range. In addition, since the expression system used in this study does not 618

provide any insight into the protein concentrations generated, it is quite probable that they differ 619

from the intracellular levels produced by GIL01 during infection. This could also explain why 620

wild-type GIL01 is not fully repressed when infecting cells expressing both gp1 and gp7. One 621

would expect the wild-type phage cycle to be hindered by the two factors known to be involved 622

in lysogeny and yet, titers are only slightly reduced. This observation suggests there is the need 623

for precise levels of both factors and probably, a determined timing of expression, too. 624

gp1 displays a DNA-binding domain similar to the one found in the MerR (repressor of mercury 625

resistance operons) super-family of transcription regulators. MerR itself is known to mediate 626

both repression and activation of transcription by virtue of its ability to distort the DNA helix (3, 627

4). A role for gp1 as a modulator of DNA structure is an attractive idea as it is a common theme 628

among transcription regulators. gp7, on the other hand, has no predicted function or recognizable 629

sequence motif. It might cooperate with gp1 in the same regulatory process or it could act 630

independently, at a completely different level of immunity regulation. 631

Temperate phages that are under direct control of host LexA utilize a variety of regulatory 632

mechanisms. Coliphages 186 and N15 code for LexA-regulated anti-repressors (8, 47) and gene 633

homologues that are preceded by LexA binding sites have been identified in other phages (9, 12, 634

27). A unique example where LexA exerts a dual function is provided by the Vibrio cholerae 635

filamentous phage CTX. The phage-encoded repressor RstR acts together with host LexA to 636

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

31

activate and repress gene transcription (36, 54). Such a mechanism allows the phage to respond 637

to small variations in RstR-LexA levels and switch from a low particle formation mode to higher 638

yields and vice versa. This regulatory system appears to be well suited for a filamentous phage, 639

the production of which is maintained throughout the cell cycle. 640

GIL01 cp mutants were isolated in three distinct loci – dinBox1, ORF1 and ORF7 – each of them 641

necessary for the lysogenic cycle. Surprisingly, none of the isolated mutants produced plaques on 642

lysogens, a result that we expected to see with at least the operator mutants in dinBox1. There are 643

many possible explanations for the observed immunity effect. One is that GIL01 codes for an as 644

yet unidentified super-infection exclusion system that blocks entry of newly infecting phage. A 645

more attractive explanation, however, would be that dinBox1 is not the only operator needed for 646

lysogeny. If other regulatory sites are present on the phage genome, potentially responding to 647

gp1 and/or gp7, these sites could constitute themselves a LexA-independent pathway for 648

regulation and compensate for the loss of the dinBox1 pathway. In this case, a mutant super-649

infecting phage would still be sensitive to repression upon infection of a lysogen, provided that 650

the appropriate immunity factors are readily available in the cell. It is noteworthy that even 651

though the dinBox1, ORF1 and ORF7 alleles were independently isolated multiple times, they 652

probably do not represent a saturated screen of all possible mutations leading to the cp 653

phenotype. We therefore do not exclude the possibility that yet unidentified factors also 654

participate in GIL01 regulation. 655

GIL01 is regulated by a multi-component system composed of at least host LexA, gp1 and gp7. 656

The combination of host- and phage-encoded functions in the lytic switch has already been seen 657

in other phages and enables a fine regulation of lytic functions. LexA is used as an internal 658

sensor of stress and allows for the rapid escape from potentially deleterious damage to the host. 659

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

32

The addition of phage-borne functions could modulate the switch by conferring more precision 660

to the system or by adding a further layer of control. GIL01 has developed an elaborate network 661

to control its development and future studies will address how the different components 662

identified in this study intervene in the switch. 663

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

33

Acknowledgements 664

We are grateful to B. Clément and S. Degand for their contribution to the early stages of this 665

work and to A. Toussaint and A. Aertsen for stimulating discussions. We thank R. Janky for 666

continuous encouragement and skilful advice with RSAT and R. Leplae for assistance with 667

protein analysis. We are particularly indebted to H. Kimsey for insightful observations and 668

comments on the manuscript. We wish to thank C. Condon for kindly providing the pDG148 669

vector. This work was supported by grants from the National Fund for Scientific Research and 670

the Université catholique de Louvain (grants to NF and JM). Part of this work was also 671

supported by the Finnish Centre of Excellence Program of the Academy of Finland (grant 672

1129648 to JKHB), an EMBO short-term fellowship and the National Research Fund in 673

Luxembourg, co-funded under the Marie Curie Actions of the European Commission (FP7-674

COFUND) (grants to NF). 675

676

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

34

References 677

1. Agaisse, H. and D. Lereclus. 1994. Structural and functional analysis of the promoter region 678 involved in full expression of the cryIIIA toxin gene of Bacillus thuringiensis. Mol. Microbiol. 679 13:97-107. 680

2. Andrup, L., J. Damgaard, and K. Wassermann. 1993. Mobilization of small plasmids in 681 Bacillus thuringiensis subsp. israelensis is accompanied by specific aggregation. J. Bacteriol. 682 175:6530-6536. 683

3. Ansari, A. Z., J. E. Bradner, and T. V. O'Halloran. 1995. DNA-bend modulation in a 684 repressor-to-activator switching mechanism. Nature 374:371-375. 685

4. Ansari, A. Z., M. L. Chael, and T. V. O'Halloran. 1992. Allosteric underwinding of DNA is 686 a critical step in positive control of transcription by Hg-MerR. Nature 355:87-89. 687

5. Barne, K. A., J. A. Bown, S. J. Busby, and S. D. Minchin. 1997. Region 2.5 of the 688 Escherichia coli RNA polymerase sigma70 subunit is responsible for the recognition of the 689 'extended-10' motif at promoters. EMBO J. 16:4034-4040. 690

6. Bertani, G. 1951. Studies on lysogenesis. I. The mode of phage liberation by lysogenic 691 Escherichia coli. J. Bacteriol. 62:293-300. 692

7. Brent, R. and M. Ptashne. 1980. The lexA gene product represses its own promoter. Proc. 693 Natl. Acad. Sci. U. S. A. 77:1932-1936. 694

8. Brumby, A. M., I. Lamont, I. B. Dodd, and J. B. Egan. 1996. Defining the SOS operon of 695 coliphage 186. Virology 219:105-114. 696

9. Bunny, K., J. Liu, and J. Roth. 2002. Phenotypes of lexA mutations in Salmonella enterica: 697 evidence for a lethal lexA null phenotype due to the Fels-2 prophage. J. Bacteriol. 184:6235-698 6249. 699

10. Butala, M., D. Zgur-Bertok, and S. J. Busby. 2009. The bacterial LexA transcriptional 700 repressor. Cell Mol. Life Sci. 66:82-93. 701

11. Camacho, A. and M. Salas. 2001. Mechanism for the switch of phi29 DNA early to late 702 transcription by regulatory protein p4 and histone-like protein p6. EMBO J. 20:6060-6070. 703

12. Casjens, S. R., E. B. Gilcrease, W. M. Huang, K. L. Bunny, M. L. Pedulla, M. E. Ford, 704 J. M. Houtz, G. F. Hatfull, and R. W. Hendrix. 2004. The pKO2 linear plasmid prophage of 705 Klebsiella oxytoca. J. Bacteriol. 186:1818-1832. 706

13. Cheo, D. L., K. W. Bayles, and R. E. Yasbin. 1991. Cloning and characterization of DNA 707 damage-inducible promoter regions from Bacillus subtilis. J. Bacteriol. 173:1696-1703. 708

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

35

14. Chu, W. Y., Y. F. Huang, C. C. Huang, Y. S. Cheng, C. K. Huang, and Y. J. Oyang. 709 2009. ProteDNA: a sequence-based predictor of sequence-specific DNA-binding residues in 710 transcription factors. Nucleic Acids Res. 37:W396-401. 711

15. Court, D., L. Green, and H. Echols. 1975. Positive and negative regulation by the cII and 712 cIII gene products of bacteriophage lambda. Virology 63:484-491. 713

16. Cutting, S. M. and P. B. Vander Horn. 1990. Genetic Analysis, p. 27-74. In C. R. Harwood 714 and S. M. Cutting (ed.), Molecular Biological Methods for Bacillus. John Wiley, Chichester. 715

17. Defrance, M., R. Janky, O. Sand, and J. van Helden. 2008. Using RSAT oligo-analysis 716 and dyad-analysis tools to discover regulatory signals in nucleic sequences. Nat. Protoc. 3:1589-717 1603. 718

18. Erill, I., S. Campoy, and J. Barbe. 2007. Aeons of distress: an evolutionary perspective on 719 the bacterial SOS response. FEMS Microbiol. Rev. 31:637-656. 720

19. Fabret, C., S. D. Ehrlich, and P. Noirot. 2002. A new mutation delivery system for 721 genome-scale approaches in Bacillus subtilis. Mol. Microbiol. 46:25-36. 722

20. Gaidelyte, A., V. Cvirkaite-Krupovic, R. Daugelavicius, J. K. Bamford, and D. H. 723 Bamford. 2006. The entry mechanism of membrane-containing phage Bam35 infecting Bacillus 724 thuringiensis. J. Bacteriol. 188:5925-5934. 725

21. Gicquel-Sanzey, B. and P. Cossart. 1982. Homologies between different procaryotic DNA-726 binding regulatory proteins and between their sites of action. EMBO J. 1:591-595. 727

22. Gonzalez-Gil, G., R. Kahmann, and G. Muskhelishvili. 1998. Regulation of crp 728 transcription by oscillation between distinct nucleoprotein complexes. EMBO J. 17:2877-2885. 729

23. Gourse, R. L., H. A. de Boer, and M. Nomura. 1986. DNA determinants of rRNA 730 synthesis in E. coli: growth rate dependent regulation, feedback inhibition, upstream activation, 731 antitermination. Cell 44:197-205. 732

24. Groban, E. S., M. B. Johnson, P. Banky, P. G. Burnett, G. L. Calderon, E. C. Dwyer, S. 733 N. Fuller, B. Gebre, L. M. King, I. N. Sheren, L. D. Von Mutius, T. M. O'Gara, and C. M. 734 Lovett. 2005. Binding of the Bacillus subtilis LexA protein to the SOS operator. Nucleic Acids 735 Res. 33:6287-6295. 736

25. Haijema, B. J., D. van Sinderen, K. Winterling, J. Kooistra, G. Venema, and L. W. 737 Hamoen. 1996. Regulated expression of the dinR and recA genes during competence 738 development and SOS induction in Bacillus subtilis. Mol. Microbiol. 22:75-85. 739

26. Hendrix, R. W., J. W. Roberts, F. W. Stahl, and R. A. Weisberg (eds.). 1983. Lambda II. 740 Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 741

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

36

27. Hertwig, S., I. Klein, V. Schmidt, S. Beck, J. A. Hammerl, and B. Appel. 2003. Sequence 742 analysis of the genome of the temperate Yersinia enterocolitica phage PY54. J. Mol. Biol. 743 331:605-622. 744

28. Horii, T., T. Ogawa, T. Nakatani, T. Hase, H. Matsubara, and H. Ogawa. 1981. 745 Regulation of SOS functions: purification of E. coli LexA protein and determination of its 746 specific site cleaved by the RecA protein. Cell 27:515-522. 747

29. Ivanova, N., A. Sorokin, I. Anderson, N. Galleron, B. Candelon, V. Kapatral, A. 748 Bhattacharyya, G. Reznik, N. Mikhailova, A. Lapidus, L. Chu, M. Mazur, E. Goltsman, N. 749 Larsen, M. D'Souza, T. Walunas, Y. Grechkin, G. Pusch, R. Haselkorn, M. Fonstein, S. D. 750 Ehrlich, R. Overbeek, and N. Kyrpides. 2003. Genome sequence of Bacillus cereus and 751 comparative analysis with Bacillus anthracis. Nature 423:87-91. 752

30. Janky, R. and J. van Helden. 2008. Evaluation of phylogenetic footprint discovery for 753 predicting bacterial cis-regulatory elements and revealing their evolution. BMC Bioinformatics 754 9:37. 755

31. Jensen, G. B., L. Andrup, A. Wilcks, L. Smidt, and O. M. Poulsen. 1996. The 756 aggregation-mediated conjugation system of Bacillus thuringiensis subsp. israelensis: host range 757 and kinetics of transfer. Curr. Microbiol. 33:228-236. 758

32. Jensen, G. B., A. Wilcks, S. S. Petersen, J. Damgaard, J. A. Baum, and L. Andrup. 759 1995. The genetic basis of the aggregation system in Bacillus thuringiensis subsp. israelensis is 760 located on the large conjugative plasmid pXO16. J. Bacteriol. 177:2914-2917. 761

33. Kawai, Y., S. Moriya, and N. Ogasawara. 2003. Identification of a protein, YneA, 762 responsible for cell division suppression during the SOS response in Bacillus subtilis. Mol. 763 Microbiol. 47:1113-1122. 764

34. Keilty, S. and M. Rosenberg. 1987. Constitutive function of a positively regulated promoter 765 reveals new sequences essential for activity. J. Biol. Chem. 262:6389-6395. 766

35. Kelley, L. A. and M. J. Sternberg. 2009. Protein structure prediction on the Web: a case 767 study using the Phyre server. Nat. Protoc. 4:363-371. 768

36. Kimsey, H. H. and M. K. Waldor. 2009. Vibrio cholerae LexA coordinates CTX prophage 769 gene expression. J. Bacteriol. 191:6788-6795. 770

37. Kumar, A., R. A. Malloch, N. Fujita, D. A. Smillie, A. Ishihama, and R. S. Hayward. 771 1993. The minus 35-recognition region of Escherichia coli sigma 70 is inessential for initiation 772 of transcription at an "extended minus 10" promoter. J. Mol. Biol. 232:406-418. 773

38. Lamont, I., A. M. Brumby, and J. B. Egan. 1989. UV induction of coliphage 186: 774 prophage induction as an SOS function. Proc. Natl. Acad. Sci. U. S. A. 86:5492-5496. 775

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

37

39. Laurinavicius, S., D. H. Bamford, and P. Somerharju. 2007. Transbilayer distribution of 776 phospholipids in bacteriophage membranes. Biochim. Biophys. Acta 1768:2568-2577. 777

40. Laurinavicius, S., R. Kakela, P. Somerharju, and D. H. Bamford. 2004. Phospholipid 778 molecular species profiles of tectiviruses infecting Gram-negative and Gram-positive hosts. 779 Virology 322:328-336. 780

41. Laurinmaki, P. A., J. T. Huiskonen, D. H. Bamford, and S. J. Butcher. 2005. Membrane 781 proteins modulate the bilayer curvature in the bacterial virus Bam35. Structure 13:1819-1828. 782

42. Lin, L. L. and J. W. Little. 1988. Isolation and characterization of noncleavable (Ind-) 783 mutants of the LexA repressor of Escherichia coli K-12. J. Bacteriol. 170:2163-2173. 784

43. Little, J. W. 1984. Autodigestion of lexA and phage lambda repressors. Proc. Natl. Acad. 785 Sci. U. S. A. 81:1375-1379. 786

44. Little, J. W. and J. E. Harper. 1979. Identification of the lexA gene product of Escherichia 787 coli K-12. Proc. Natl. Acad. Sci. U. S. A. 76:6147-6151. 788

45. Luo, Y., R. A. Pfuetzner, S. Mosimann, M. Paetzel, E. A. Frey, M. Cherney, B. Kim, J. 789 W. Little, and N. C. Strynadka. 2001. Crystal structure of LexA: a conformational switch for 790 regulation of self-cleavage. Cell 106:585-594. 791

46. Mahillon, J. and D. Lereclus. 2000. Electroporation of Bacillus thuringiensis and Bacillus 792 cereus, p. 242-252. In J. Teissié and N. Eynard (ed.), Electrotransformation of Bacteria. Springer 793 Lab Manual, Springer-Verlag, Germany. 794

47. Mardanov, A. V. and N. V. Ravin. 2007. The antirepressor needed for induction of linear 795 plasmid-prophage N15 belongs to the SOS regulon. J. Bacteriol. 189:6333-6338. 796

48. Mazon, G., I. Erill, S. Campoy, P. Cortes, E. Forano, and J. Barbe. 2004. Reconstruction 797 of the evolutionary history of the LexA-binding sequence. Microbiology 150:3783-3795. 798

49. Meijer, W. J., J. A. Horcajadas, and M. Salas. 2001. Phi29 family of phages. Microbiol. 799 Mol. Biol. Rev. 65:261-87 ; second page, table of contents. 800

50. Mencia, M., M. Monsalve, M. Salas, and F. Rojo. 1996. Transcriptional activator of phage 801 phi 29 late promoter: mapping of residues involved in interaction with RNA polymerase and in 802 DNA bending. Mol. Microbiol. 20:273-282. 803

51. Miller, M. C., J. B. Resnick, B. T. Smith, and C. M. Lovett Jr. 1996. The Bacillus subtilis 804 dinR gene codes for the analogue of Escherichia coli LexA. Purification and characterization of 805 the DinR protein. J. Biol. Chem. 271:33502-33508. 806

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

38

52. Mount, D. W. 1977. A mutant of Escherichia coli showing constitutive expression of the 807 lysogenic induction and error-prone DNA repair pathways. Proc. Natl. Acad. Sci. U. S. A. 808 74:300-304. 809

53. Nasser, W., M. Rochman, and G. Muskhelishvili. 2002. Transcriptional regulation of fis 810 operon involves a module of multiple coupled promoters. EMBO J. 21:715-724. 811

54. Quinones, M., H. H. Kimsey, and M. K. Waldor. 2005. LexA cleavage is required for 812 CTX prophage induction. Mol. Cell 17:291-300. 813

55. Ravantti, J. J., A. Gaidelyte, D. H. Bamford, and J. K. Bamford. 2003. Comparative 814 analysis of bacterial viruses Bam35, infecting a gram-positive host, and PRD1, infecting gram-815 negative hosts, demonstrates a viral lineage. Virology 313:401-414. 816

56. Ravin, N. V., A. N. Svarchevsky, and G. Deho. 1999. The anti-immunity system of phage-817 plasmid N15: identification of the antirepressor gene and its control by a small processed RNA. 818 Mol. Microbiol. 34:980-994. 819

57. Sambrook, J. and D. Russell. 2001. Molecular Cloning: a Laboratory Manual. 3rd ed. Cold 820 Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 821

58. Sanderson, A., J. E. Mitchell, S. D. Minchin, and S. J. Busby. 2003. Substitutions in the 822 Escherichia coli RNA polymerase sigma70 factor that affect recognition of extended -10 823 elements at promoters. FEBS Lett. 544:199-205. 824

59. Sauer, R. T., M. J. Ross, and M. Ptashne. 1982. Cleavage of the lambda and P22 825 repressors by recA protein. J. Biol. Chem. 257:4458-4462. 826

60. Schubert, R. A., I. B. Dodd, J. B. Egan, and K. E. Shearwin. 2007. Cro's role in the CI 827 Cro bistable switch is critical for {lambda}'s transition from lysogeny to lytic development. 828 Genes Dev. 21:2461-2472. 829

61. Shearwin, K. E., A. M. Brumby, and J. B. Egan. 1998. The Tum protein of coliphage 186 830 is an antirepressor. J. Biol. Chem. 273:5708-5715. 831

62. Sozhamannan, S., M. McKinstry, S. M. Lentz, M. Jalasvuori, F. McAfee, A. Smith, J. 832 Dabbs, H. W. Ackermann, J. K. Bamford, A. Mateczun, and T. D. Read. 2008. Molecular 833 characterization of a variant of Bacillus anthracis-specific phage AP50 with improved 834 bacteriolytic activity. Appl. Environ. Microbiol. 74:6792-6796. 835

63. Stragier, P., C. Bonamy, and C. Karmazyn-Campelli. 1988. Processing of a sporulation 836 sigma factor in Bacillus subtilis: how morphological structure could control gene expression. 837 Cell 52:697-704. 838

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

39

64. Stromsten, N. J., S. D. Benson, R. M. Burnett, D. H. Bamford, and J. K. Bamford. 2003. 839 The Bacillus thuringiensis linear double-stranded DNA phage Bam35, which is highly similar to 840 the Bacillus cereus linear plasmid pBClin15, has a prophage state. J. Bacteriol. 185:6985-6989. 841

65. Thomas-Chollier, M., O. Sand, J. V. Turatsinze, R. Janky, M. Defrance, E. Vervisch, S. 842 Brohee, and J. van Helden. 2008. RSAT: regulatory sequence analysis tools. Nucleic Acids 843 Res. 36:W119-27. 844

66. Vagner, V., E. Dervyn, and S. D. Ehrlich. 1998. A vector for systematic gene inactivation 845 in Bacillus subtilis. Microbiology 144 ( Pt 11):3097-3104. 846

67. Verheust, C., N. Fornelos, and J. Mahillon. 2005. GIL16, a new gram-positive tectiviral 847 phage related to the Bacillus thuringiensis GIL01 and the Bacillus cereus pBClin15 elements. J. 848 Bacteriol. 187:1966-1973. 849

68. Verheust, C., N. Fornelos, and J. Mahillon. 2004. The Bacillus thuringiensis phage GIL01 850 encodes two enzymes with peptidoglycan hydrolase activity. FEMS Microbiol. Lett. 237:289-851 295. 852

69. Verheust, C., G. Jensen, and J. Mahillon. 2003. pGIL01, a linear tectiviral plasmid 853 prophage originating from Bacillus thuringiensis serovar israelensis. Microbiology 149:2083-854 2092. 855

70. Voskuil, M. I. and G. H. Chambliss. 1998. The -16 region of Bacillus subtilis and other 856 gram-positive bacterial promoters. Nucleic Acids Res. 26:3584-3590. 857

71. Walker, G. C. 1984. Mutagenesis and inducible responses to deoxyribonucleic acid damage 858 in Escherichia coli. Microbiol. Rev. 48:60-93. 859

72. Winterling, K. W., D. Chafin, J. J. Hayes, J. Sun, A. S. Levine, R. E. Yasbin, and R. 860 Woodgate. 1998. The Bacillus subtilis DinR binding site: redefinition of the consensus 861 sequence. J. Bacteriol. 180:2201-2211. 862

73. Young, R. A. and J. A. Steitz. 1979. Tandem promoters direct E. coli ribosomal RNA 863 synthesis. Cell 17:225-234. 864

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

40

Figure legends 865

Fig. 1. Partial physical map of phage GIL01 with SOS-regulated promoters. (A) Predicted 866

open reading frames (ORFs) are depicted as open boxes and left-to-right arrows indicate the 867

direction of transcription of the left and right functional domains. Probable gene functions are 868

indicated above certain ORFs. Putative Rho-independent transcription terminators are shown as 869

open stem-loops downstream of ORF8 and ORF30. Promoters are shown as filled arrowheads 870

and each promoter region is enlarged below the map. -35 and -10 boxes are depicted as black 871

rectangles and angled arrows represent transcription start sites. Grey boxes show the location of 872

conserved LexA binding sites with regard to the three promoters. (B) Transcription start site 873

mapping of GIL01 ORF1 and ORF9 by 5’-RACE. cDNAs were prepared from GIL01 mRNA 874

and were tailed with dCTP and PCR-amplified. The purified products were cloned into pCR4-875

TOPO and sequenced using the universal primers M13. The electropherograms show the 876

transcription start sites (indicated with an arrow; sequences correspond to the GIL01 negative 877

strand) fused to their oligo-dC tails. The respective promoters from which transcription is 878

initiated are indicated for each electropherogram. Note that for the transcription start at P3, an 879

additional thymine residue that is not found in the GIL01 sequence has been inserted upstream of 880

the dC tail. (C) Detailed nucleotide sequence of promoters depicted in (A). Transcription start 881

sites detected in (B) are indicated with angled arrows (P1, P2 and P3). Putative ribosome 882

binding sites (RBSs) are shown in lower case, start codons in bold and SOS boxes are shaded 883

grey. A TG dinucleotide motif characteristic of -10 extended promoters is double-underlined 884

within P1. Genome coordinates are indicated above each sequence. 885

Fig. 2. Clear plaque (cp) mutants in ORF1 and ORF7. Gene (A) and protein (B) sequences of 886

cp mutants were aligned and compared to their respective wild-type sequences. Nucleotide 887

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

41

coordinates (A) and the corresponding protein sizes (B) are indicated for each sequence. Start 888

and stop codons are shown in bold and mutations appear on a light grey background. (A) The 889

direct repeat (DR) in ORF7 is highlighted against a black background in the wild-type sequence. 890

(B) ORF1 harbors a putative helix-turn-helix DNA-binding motif in its N-terminal domain and 891

the two α helices, α1 and α2, are shown boxed in black in the wild-type sequence. α1 and α2 892

were predicted using the softwares Phyre (35) and ProteDNA (14). 893

Fig. 3. dinBox1 is a conserved LexA operator. (A) A class of GIL01 cp mutants harbor single 894

nucleotide mutations (shaded grey) in dinBox1, a 14-bp predicted SOS box located between 895

promoters P1 and P2 (see Fig. 1A and 1C). The Bam35c operator differs from GIL01 by one 896

single nucleotide and this difference is believed to account for the opposed propagation styles of 897

both phages (see text for details). The GIL01 nucleotide coordinates are indicated above the 898

wild-type sequence. (B) Putative LexA operators were detected upstream of SOS genes in strain 899

B. cereus G9842 (GenBank acc. number NC_011772) and compared to the consensus sequence 900

for LexA binding in B. subtilis (24). In addition to dinBox1, two more potential SOS boxes, 901

dinBox2 and dinBox3 were identified within promoter P3 (see Fig. 1A and 1C). The outer 902

nucleotides, shown in capital letters, are highly conserved in SOS boxes of Gram-positive 903

bacteria and particularly important for LexA binding (24). (C) The N-terminal DNA binding 904

region of the B. cereus LexA protein is compared to the consensus sequence determined for 905

Gram-positive bacteria (48). The three α helices directly involved in DNA binding are 906

underlined and amino acid differences are shaded grey (24). 907

Fig. 4. gp7 is involved in immunity against infection. GBJ223 was grown exponentially and 908

gp7 expression was induced by adding IPTG 1 mM to one-half of the culture. Un-induced and 909

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

42

induced cultures were then allowed to grow for three more hours. Cells were then infected with 910

cp29, a mutant in ORF7 (see Fig. 2), and plated. 911

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

Table 1. P1-P2 and P3 promoter activities in lexA+ and lexAA96D backgrounds. 1 2

Promoter fusion lexA allele Mean β-galactosidase activity (Miller units ± SD)a

-MMC +MMC

P1-P2 lexA+

lexAA96D 75 ± 13 91 ± 5

303 ± 106 78 ± 5

P1-P2 (dinBox1mut) lexA+

lexAA96D 345 ± 28 339 ± 19

351 ± 10 326 ± 12

P3 lexA+

lexAA96D 3 ± 1 3 ± 1

26 ± 10 6 ± 2

3 alacZ fusions to P1-P2, P1-P2 mutated in dinBox1 (CGAACaagcTTTTT) and P3 4

promoters were generated and β-galactosidase activity was measured in both lexA+ and 5

lexAA96D lysogens in normal growth (-MMC) and in SOS-inducing (+MMC) conditions. 6

Cultures were grown for 3 hours at 37°C when MMC50 was added to one-half of the 7

cells and growth was allowed to proceed for one additional hour. Cells were then 8

permeabilized and assayed for β-galactosidase. Activity values are the average of three 9

independent experiments. 10

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

Table 2: Functions of defective gp1 and gp7 can be complemented 1 2

cp mutants (relevant mutation)

Host (expressed protein(s))

cp23/cp26 (dinBox1)

cp32/cp33 (ORF1)

cp27/cp29 (ORF7)

Ratio of plating efficiencies (PFU on host +IPTG/

PFU on host -IPTG)a

GBJ250 (gp1) 0.5 0.1b 0.2

GBJ223 (gp7) 1 1b 0.5c

GBJ467 (gp1/gp7) 0.2 <10-8 0.2c

aThe GIL01-cured host strain (GBJ002) was transformed with plasmids expressing either 3

gp1 (GBJ250), gp7 (GBJ223) or gp1 and gp7 (GBJ467). The resulting strains were 4

grown exponentially before IPTG 1mM was added to one-half of the culture and growth 5

was allowed to proceed for three additional hours. Un-induced and induced cultures 6

were then infected with GIL01 cp mutants in dinBox1 (cp23/cp26), ORF1 (cp32/cp33) or 7

ORF7 (cp27/cp29). Two different cp mutants at each locus were tested independently 8

and similar infection patterns were obtained for each pair of mutants. The results shown 9

are representative of four independent experiments and give the ratio between titers 10

produced on un-induced cells and titers counted on the same induced strains. 11

bcp32 and cp33 form small clear plaques on the cured host strain. There was no 12

systematic change in plaque morphology between induced and un-induced cells infected 13

with ORF1 mutants. 14

cThe plaques observed on induced lawns were turbid and significantly smaller than 15

those produced on the un-induced lawns (see Fig. 4). 16

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

Table 3: GIL01 induction is recA- and lexA-dependent 1 2 Phage titera

Strain - MMC + MMCb Ratio

GBJ338 (GBJ002 (GIL01)) 3.1 x 107 1.9 x 109 61.3

GBJ500 (GBJ338 recA::pMutin4) 4.0 x 106 3.0 x 106 0.8

GBJ406 (GBJ338 lexAA96D) 1.5 x 105 1.8 x 105 1.2

-UV +UVc

GBJ338 (GBJ002 (GIL01)) 4.0 x 107 2.7 x 109 67.5

GBJ500 (GBJ338 recA::pMutin4) 1.1 x 106 2.0 x 106 1.8

GBJ406 (GBJ338 lexAA96D) 6.7 x 104 5.3 x 104 0.8

aPhage titers were determined by plating serial dilutions of GIL01 on the cured host 3

GBJ002. Results are the average from three independent experiments. 4

bThe MMC concentration used was 50 ng ml-1 and cultures were induced for one hour. 5

cCells were irradiated at 254 nm for 10 seconds. 6

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

0.3 kb 4.8 kb 15 kb

1 2 4 5

P1 P2 P3

6 93 8 107

replication & regulation structure & lysis

11 27 2826 29 30

-35 -10 -35 -10 -35 -10

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

P1

P2

P3

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

GAAAAGTGTGACAAGATATTCCAAGGGTGCTATAATAGGAAATGAGTTAAAAAACGAACAAGCGTTTTAT -35 -10 dinBox1 AAGTGTTCGGTTTTTGTAACATAACTGTAATATAAAAACATTTGCATTACTGGTAACTTTCGGGTAATAT

P1 131

AAGTGTTCGGTTTTTGTAACATAACTGTAATATAAAAACATTTGCATTACTGGTAACTTTCGGGTAATAT -35 -10 TGGTAGCAGAGAGGGACAAACGGGACAAATCGAGACACGATACATACTCTCTATACAAGTGAAAAGaagg tggtgCATATGTAATATG

P2

358 (ORF1)

CACACGTGTGACGTTATGCGGAACACTCGTTCGTATTATAGTATACATATAGCGAACAAACATTCGATaagg -35 dinBox2 -10 dinBox3

P3 4867

4952 (ORF9)agtgtGACAAAGTG

4952 (ORF9)

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

(A) ORF1

wt atgagtaaactactgactgcgcaagaggttgcagacttattgcgtgttcataaaaagactatcc cp06 atgagtaaactactgactgcgcaagaggttgcagacttattgcgtgttcataaaaagactatcc

356

cp06 atgagtaaactactgactgcgcaagaggttgcagacttattgcgtgttcataaaaagactatcccp10 atgagtaaactactgactgcgcaagaggttgcagacttattgcgtgttcataaaaagactatcc cp32 atgagtaaactactgactgcgcaagaggttgcagacttattgcgtattcataaaaagactatcc cp33 atgagtaaactactgactgcgcaagaggttgtagacttattgcgtgttcataaaaagactatcc cp36 atgagtaaactactgactgcgcaagaggttgcagacttattgcgtgttcataaaaagactatcc

wt atagaatgatacattctggaaagttggacgcttcaaaagtagcaaacaaattcctcattaaagag g gg g gg g g g gcp06 atagaatgatacattctggaaagttggacgcttcaaaagtagcaaacaaattcctcattaaaga cp10 atagaatgatacattctggaaagttggacgcttcaaaagtagcaaacaaattcctcattaaaga cp32 atagaatgatacattctggaaagttggacgcttcaaaagtagcaaacaaattcctcattaaaga cp33 atagaatgatacattctggaaagttggacgcttcaaaagtagcaaacaaattcctcattaaaga cp36 atagaatgatacattttggaaagttggacgcttcaaaagtagcaaacaaattcctcattaaaga

wt ggaagatgcaaaa-ggactattggaaaataaaaataaactggacaattaa--------------cp06 ggaagatgcaaaa-ggactattggaaaataaaa-taaactggacaattaataggaggaatattc cp10 ggaagatgcaaaaaggactattggaaaataa---------------------------------cp32 ggaagatgcaaaa-ggactattggaaaataaaaataaactggacaattaa--------------cp33 ggaagatgcaaaa-ggactattggaaaataaaaataaactggacaattaa--------------cp36 ggaagatgcaaaa-ggactattggaaaataaaaataaactggacaattaa-------------- wt -------------------------- 532 cp06 aatggaaaacaatactaaacaattaa 571 cp10 -------------------------- 514 cp32 -------------------------- 532 cp33 -------------------------- 532 cp36 -------------------------- 532cp36 532

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

ORF7

wt atgcgtgacaaattgctcgactttatcatcgaactatcacaatctagcaaacaagtcgtaag-- cp08 atgcgtgacaaattgctcgactttatcatcgaactatcacaatctagcaaacaagtcgtaag-- cp27 atgcgtgacaaattgctcgactttatcatcgaactatcacaatctagcaaacaagtcgtaag-- cp04/18 atgcgtgacaaattgctcgactttatcatcgaactatcacaatctagcaaa------------- cp25/29 atgcgtgacaaattgctcgactttatcatcgaactatcacaatctagcaaacaagtcgtaagca

4564

cp25/29 atgcgtgacaaattgctcgactttatcatcgaactatcacaatctagcaaacaagtcgtaagca

wt --------caagtcgtatgtcattgatagactgatgcaagtaacaaaa-gaagactacaaggaa cp08 --------caagtcgtaa----------------------------------------------cp27 --------caagtcgtatgtcattgatagactgatgcaagtaacaaaaagaagactacaaggaa cp04/18 --------caagtcgtatgtcattga--------------------------------------cp25/29 agtcgtaa--------------------------------------------------------cp25/29 agtcgtaa

wt ctagaaaagaatgtggaggggaagaaggatgactga 4716 cp08 ------------------------------------ 4635 cp27 ctagaaaagaatgtggaggggaagaaggatga---- 4713 cp04/18 ------------------------------------ 4632 cp25/29 ------------------------------------ 4635 cp 5/ 9 635

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

(B) gp1 wt MSKLLTAQEVADLLRVHKKTIHRMIHSGKLDASKVANKFLIKEEDAKGLLENKNKLDN------------- 58 cp06 MSKLLTAQEVADLLRVHKKTIHRMIHSGKLDASKVANKFLIKEEDAKGLLENKINWTINRRNIQWKTILNN 71

α1 α2

cp10 MSKLLTAQEVADLLRVHKKTIHRMIHSGKLDASKVANKFLIKEEDAKRTIGK------------------- 52cp32 MSKLLTAQEVADLLRIHKKTIHRMIHSGKLDASKVANKFLIKEEDAKGLLENKNKLDN------------- 58 cp33 MSKLLTAQEVVDLLRVHKKTIHRMIHSGKLDASKVANKFLIKEEDAKGLLENKNKLDN------------- 58 cp36 MSKLLTAQEVADLLRVHKKTIHRMIHFGKLDASKVANKFLIKEEDAKGLLENKNKLDN------------- 58

gp7gp7 wt MRDKLLDFIIELSQSSKQVVSKSYVIDRLMQVTKEDYKELEKNVEGKKDD 50 cp08 MRDKLLDFIIELSQSSKQVVSKS--------------------------- 23 cp27 MRDKLLDFIIELSQSSKQVVSKSYVIDRLMQVTKRRLQGTRKECGGEEG- 49 cp04/18 MRDKLLDFIIELSQSSKQVVCH---------------------------- 22 cp25/29 MRDKLLDFIIELSQSSKQVVSKS--------------------------- 23

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

(A) dinBox1 wt CGAACaagcGTTTT

185 198 g

cp16 CCAACaagcGTTTT cp17 CGAACaagcG-TTT cp19 CGAACaagcGATTT cp23 CAAACaagcGTTTT cp26 CGAACaagcGCTTT cp37 CGAACaagcG TTTcp37 CGAACaagcG-TTT Bam35c CGAACaagcTTTTT (B) B. subtilis cons. CGAACatatGTTCG lexA CGAACttatGTTTGrecA CGAACatttATTCG uvrAB CGAACagatATTCG ruvAB AGAACatttGTTGG dinBox1 CGAACaagcGTTTT di B 2 GG C G CGdinBox2 GGAACactcGTTCGdinBox3 CGAACaaacATTCG

(C) α1 α2 α3 B. cereus G9842 5 MEKLTKRQQDILDFIKLKVQEKGYPPSVREIGQAVGLASSSTVHGHLSRLEEKGYIRRDP 64 Gram-positive cons. 1 MSKLTKRQREILDVIKASVQSKGYPPSVREIGQAVGLASSSTVHGHLSRLERKGYIRRDP 60

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

-IPTG +IPTG

on May 17, 2018 by guest

http://jb.asm.org/

Dow

nloaded from