SP10 Infectivity Is Aborted after Bacteriophage SP10 Infection InducesnonA Transcription on the Prophage SP� Region of the Bacillussubtilis Genome

Tatsuya Yamamoto,a Nozomu Obana,b Lii Mien Yee,c,d Kei Asai,c Nobuhiko Nomura,b Kouji Nakamurab

Graduate School of Life and Environmental Sciences, University of Tsukuba, Tsukuba-Shi, Ibaraki, Japana; Faculty of Life and Environmental Sciences, University ofTsukuba, Tsukuba-Shi, Ibaraki, Japanb; Department of Biochemistry and Molecular Biology, Graduate School of Science and Engineering, Saitama University, Saitama-Shi,Saitama, Japanc; Biotechnology Research Center, The University of Tokyo, Bunkyo-ku, Tokyo, Japand

Bacteria have developed various strategies for phage resistance. Infection with phage induces the transcription of part of thephage resistance gene, but the regulatory mechanisms of such transcription remain largely unknown. The phage resistance genenonA is located on the SP� prophage region of the Bacillus subtilis Marburg strain genome. The nonA transcript was detected atthe late stage of SP10 infection but is undetectable in noninfected cells. The nonA transcript was detected after the induction ofthe sigma factor Orf199-Orf200 (�Orf199-200), when sigma factors encoded in the SP10 genome were expressed from a xylose-in-ducible plasmid. Thus, the SP10 sigma factor is an activator of a set of SP10 genes and nonA. The nonA gene encodes a 72-amino-acid protein with a transmembrane motif and has no significant homology with any protein in any database. NonA overexpres-sion halted cell growth and reduced the efficiency of B. subtilis colony formation and respiration activity. In addition, SP10virion protein synthesis was inhibited in the nonA� strain, and SP10 virion particles were scarce in it. These results indicate thatNonA is a novel protein that can abort SP10 infection, and its transcription was regulated by SP10 sigma factor.

Bacteriophages (phages) frequently attack bacteria in almost allenvironments. Phages outnumber bacteria by approximately

1- to 10-fold, and �1025 phages initiate infection per second on aglobal scale (1–3). Bacteria have evolved various systems to resistphage predation, and phages in turn have developed strategies toevade such resistance (4). Thus, constant infection with phages leadsto a coevolutionary arms race (5). Bacterial phage resistance systemsinclude mechanisms for the prevention of phage adsorption, the pre-vention of phage DNA entry, the cleavage of phage nucleic acids(such as by a restriction-modification system and the clustered regu-larly interspaced short palindromic repeats [CRISPR]/CRISPR-asso-ciated [Cas] system), and abortive infection (4). Abortive infection isalso called phage exclusion. The abortive infection genes affect crucialsteps in the phage infection cycle, such as genomic DNA replication(6–10), RNA transcription (11–13), translation (14, 15), or host celllysis (16), preventing the spread of phages. Toxin-antitoxin (TA) sys-tems are also associated with the abortive infection phenotype (17–22). These phage resistance genes are often found in mobile geneticelements such as plasmids or in the prophage region of the bacterialgenome, suggesting that bacteria acquired phage resistance genes byhorizontal gene transfer.

In 1930, Harold J. Conn described the Bacillus subtilis Marburgstrain and deposited it in the ATCC under accession number 6051(23). This strain is nonpermissive for the multiplication of B. sub-tilis phage SP10 (24). SP10 is a double-stranded DNA phage thatbelongs to the Myoviridae family, and it can multiply in B. subtilisstrain W23 and Bacillus amyloliquefaciens (25). The SP10 genomeis 143,986 bp in length and contains 236 predicted open readingframes (ORFs), and it has significant similarity to the gene orga-nization of SPO1, a virulent phage of B. subtilis that has beenstudied in detail (26). However, the B. subtilis Marburg strain ispermissive for SPO1 multiplication (25). A genetic study of a per-missive mutant revealed that both mutant alleles in the nonA andnonB genes are necessary for permissiveness (24). The nonA mu-

tant is cured of the SP� prophage, and the nonB mutant has anonsense mutation in ydiR, which is a component of the B. subtilisrestriction system, BsuM (27). The BsuM restriction (R) andmodification (M) system of B. subtilis Marburg comprises BsuMRencoded by the ydiR-ydiS-ydjA operon and BsuMM encoded bythe ydiO-ydiP operon. These genes are located in the prophage-like element pro3. The target sequence of BsuMR is identical tothat of XhoI (CTCGAG) (28, 29). Since 36 BsuMR recognitionsites are located in the SP10 genome, BsuMR cleaves SP10genomic DNA after DNA injection, thus restricting SP10 multi-plication (26, 30). In contrast to nonB, the function of nonA hasremained unknown. The nonA mutation has been mapped to theSP� prophage region (27). Partial deletion and complementationanalysis have shown that the 825-bp SP� prophage region, whichcontains parts of bnrdE and bnrdF, is necessary to resist the SP10phage (31). This suggests that the nonA gene is located in this825-bp region. The bnrdEF genes of SP� are homologues of B.subtilis nrdEF and xnrdEF of SP10. The nrdEF genes encode ribo-nucleotide reductase (RNR), which catalyzes the conversion ofnucleotides to deoxynucleotides and which is essential for thegrowth of B. subtilis (32). The xnrdEF genes complement a tem-perature-sensitive mutation of nrdE, and such complementationis inhibited by artificially induced transcription from the 3= end ofthe bnrdE antisense strand (31). This suggests that nonA is tran-

Received 18 October 2013 Accepted 19 November 2013

Published ahead of print 22 November 2013

Address correspondence to Kouji Nakamura, nakamura.kouji@biol.tsukuba.ac.jp.

Supplemental material for this article may be found at http://dx.doi.org/10.1128/JB.01240-13.

Copyright © 2014, American Society for Microbiology. All Rights Reserved.

doi:10.1128/JB.01240-13

February 2014 Volume 196 Number 3 Journal of Bacteriology p. 693–706 jb.asm.org 693

on October 16, 2017 by guest

http://jb.asm.org/

Dow

nloaded from

scribed from the antisense strand of bnrdE and that it inhibitsXnrdEF function (31).

Here, we found that nonA is transcribed from the bnrdEF in-tergenic region to the bnrdE 3= end region and that its transcrip-tion is activated by the SP10-encoded sigma factor Orf199-Orf200(�Orf199-200). The nonA gene encodes a small transmembrane pro-tein, which is necessary for SP10 to acquire resistance. On theother hand, the 3= untranslated region (UTR) of nonA corre-sponding to the bnrdE 3= region that is involved in the inhibitoryeffects of xnrdEF genes was not required. SP10 capsid proteinsynthesis was inhibited in the nonA� strain, and the isopropyl-�-D-thiogalactopyranoside (IPTG)-induced expression of NonAprotein resulted in the inhibition of B. subtilis growth. These re-sults demonstrate that nonA is a new abortive infection gene.

MATERIALS AND METHODSBacterial strains, phages, plasmids, and medium. Table 1 lists the bacte-rial strains, phages, and plasmids used in this study. B. subtilis and Esche-

richia coli were routinely grown in LB medium or on an LB plate at 37°C.LB medium (for SP10, SPO1, and SPP1) and LB medium supplementedwith 5 mM MgSO4 (for �29) were used in phage infection experiments,and 10 mM CaCl2 was added to the media before SPP1 infection. SP10,SPO1, �29, and SPP1 phages were propagated on B. subtilis RM125 asdescribed previously (31). The following antibiotics were included whenappropriate: ampicillin, kanamycin, erythromycin, spectinomycin, andtetracycline (50, 5, 0.5, 100, and 15 �g/ml, respectively).

Oligonucleotide primers. Table S1 in the supplemental material liststhe oligonucleotide primers used in this study.

Construction of nonA complementation strains. The E. coli rrnB ter-minator was fused to the region downstream of the amyE 5= region toterminate transcription from the amyE promoter. The amyE 5= region wasthen amplified by PCR using primers TYP2 and TAP6 from B. subtilisgenomic DNA as a template, and the rrnB terminator region was ampli-fied using primers TYP5 and TYP7 from pMutinNC (33). These two PCRproducts were ligated by fusion PCR with primers TYP1 and TYP7. Thekanamycin resistance gene (kan) was amplified by PCR using primersTYP8 and TYP10 from B. subtilis TMO 311 (34) genomic DNA, and theamyE 3= region was amplified using primers TYP5 and TYP4. These twoPCR products were ligated with primers TYP7 and TYP3 by fusion PCR.The wild-type nonA gene was amplified using primers TYP11 and TYP12.The nonA start codon mutation (GTG to GAG) was generated by two-stepPCR. The upstream and downstream regions of nonA start codon frag-ments were amplified using the primer sets TYP38/TYP14 and TYP13/TYP35, respectively, and the resultant PCR products were fused withprimers TYP11 and TYP12 by PCR. The same strategy was used to intro-duce the ribosome binding site (RBS) mutation of nonA (AGGA toAAAA). TYP38/TYP16 and TYP15/TYP35 primer sets were used for thisRBS mutation. The spoVG 3= UTR was amplified by PCR using primersTAP17 and TAP19, and fragments from the nonA upstream region to thestop codon of the nonA ORF were amplified using primers TAP11 andTAP18 from wild-type nonA and nonA with a mutated start codon or theRBS-mutated nonA gene as a template. The spoVG 3= UTR fragment andthe nonA gene fragments were fused with primers TAP11 and TAP20 byPCR. These products containing wild-type nonA or nonA variants weredigested by BamHI and EcoRI and then ligated with a BamHI-digestedamyE 5=-rrnB terminator fragment and an EcoRI-digested kan-amyE 3=fragment using the DNA Ligation Kit Mighty Mix (TaKaRa Bio). Eachligated product was amplified with primers TAP1 and TAP3 by PCR. ThePCR products were introduced into the amyE locus of the B. subtilisASK3002 strain by double-crossover recombination to generate theTAY3201, TAY3202, TAY3203, TAY3204, TAY3205, and TAY3206strains. Control strains were generated by amplifying the multicloning siteof pGEM-3Zf (�) (Promega) with primers TAP21 and TAP22 by PCR,and then this product was digested with EcoRI and BamHI and ligatedwith the BamHI-digested amyE 5= rrnB terminator fragment and theEcoRI-digested kan-amyE 3= fragment. This ligated product was intro-duced into the amyE locus of the B. subtilis ASK3000 and ASK3002 strainsto generate TAY3000 and TAY32000. Transformation was confirmed byPCR and DNA sequencing.

Construction of plasmids expressing SP10 sigma factor. orf120,orf183, and orf199-200 were amplified by PCR using the TAP27/TAP28,TAP23/TAP24, and TAP25/TAP26 primer sets, respectively. The PCRproducts were digested by KpnI and BamHI and then ligated into thesimilarly digested pWH1520 plasmid to generate pWSP120, pWSP183,and pWSP199-200, which were transformed into Escherichia coli JM109and amplified. The amplified plasmid was purified from the cells of thesestrains and used in the protoplast transformation of B. subtilis UOT1285.

Construction of nonA or nonA-His expression plasmids. His-taggednonA and the spoVG 3=UTR were amplified from B. subtilis genomic DNAby PCR using the TYP11/TYP29 and TYP30/TYP19 primers sets, respec-tively. The PCR products were digested by PstI and ligated using a DNALigation Kit Mighty Mix. The ligated product was amplified by PCR withprimers TYP11 and TYP20, digested by BamHI and EcoRI, and ligated

TABLE 1 Bacterial strains, phages, and plasmids used in this study

Name Genotype or descriptionSource ofreference

B. subtilis strainsUOT1285 trpC2 lys1 �aprE3 nprE18 nprR2 31ASK3000 UOT1285 �ydiR-ydjA 31ASK3001 UOT1285 �yqpP-yodU (�SP�) 31ASK3002 UOT1285 �ydiR-ydjA �yqpP-yodU (�SP�) 31TAY3000 ASK3000 amyE::kan This studyTAY3200 ASK3002 amyE::kan This studyTAY3201 ASK3002 amyE::nonA kan This studyTAY3202 ASK3002 amyE::nonA(T27A) kana This studyTAY3203 ASK3002 amyE::nonA-spoVG 3= UTR kan This studyTAY3204 ASK3002 amyE::nonA (T27A)-spoVG 3= UTR kan This studyTAY3205 ASK3002 amyE::nonA(G13A G14A) kan This studyTAY3206 ASK3002 amyE::nonA(G13A G14A)-spoVG 3=

UTR kanThis study

TAY3221 ASK3002 amyE::lacI Pspac-nonA-His kan This studyTAY3227 ASK3002 amyE::lacI Pspac-nonA(T27A)-His kan This studyTMO310 trpC2 aprE::lacI Pspac-mazF spec 34TMO311 trpC2 aprE::lacI Pspac-mazF kan 34

E. coli strainJM109 recA1 endA1 gyrA96 thi-1 hsdR17(rK

mK�)

e14 (mcrA mutant) supE44 relA1 �(lac-proAB) F=[traD36 proAB� lacIq lacZ�M15]

TaKaRa Bio

PhagesSP10 Bacillus phage 31SPO1 Bacillus phage BGSC�29 Bacillus phage BGSCSPP1 Bacillus phage BGSC

PlasmidspHY300PLK E. coli-B. subtilis shuttle vector, tet TaKaRa BiopHY-nonA nonA in pHY300PLK This studypHY-nonA-HIs nonA-His in pHY300PLK This studypWH1520 Xylose-inducible expression vector, tet MoBiTecpWSP120 SP10 orf120 in pWH1520 This studypWSP183 SP10 orf183 in pWH1520 This studypWSP199-200 SP10 orf199-200 in pWH1520 This studypUC18 IPTG-inducible expression vector, amp TaKaRa BiopUC-nonA-His nonA-His in pUC18 This studypUC-nonAmut-His nonA(T27A)-His in pUC18 This studypMutinNC Integration vector, amp erm 33pGEM3Zf (�) Cloning vector with T7 and SP6 promoters, amp Promega

a nonA(T27A), the T-to-A change at position 27 encoded by nonA. Other mutations areshown similarly.

Yamamoto et al.

694 jb.asm.org Journal of Bacteriology

on October 16, 2017 by guest

http://jb.asm.org/

Dow

nloaded from

into the similarly digested pHY300PLK to generate the pHY-nonA-Hisplasmid. The pHY-nonA plasmid was similarly constructed. The nonA-spoVG 3= UTR was amplified using the primers TYP11 and TYP20 fromTAY3203 genomic DNA and cloned into the BamHI-EcoRI site ofpHY300PLK to generate pHY-nonA plasmids. These plasmids were trans-formed into E. coli JM109, amplified, purified from cultures of thesestrains, and used in the protoplast transformation of B. subtilis ASK3002.

Construction of a nonA-inducible strain. The lacI-Pspac cassette wasamplified by PCR using primers TYP33 and TYP32 from TMO310genomic DNA, and the nonA-His gene was amplified using the primersTYP31 and TYP20 from pHY-nonA-His. Both amplified products wereligated by PCR using primers TYP33 and TYP20. This PCR fragment wasdigested by EcoRI and BamHI and ligated with the BamHI-digested amyE5= rrnB terminator fragment and the EcoRI-digested kan-amyE 3= frag-ment using a DNA Ligation Kit Mighty Mix. This ligated product wasintroduced into the amyE locus of B. subtilis ASK3002 to generateTAY3221. The start codon mutation of the nonA-His gene (nonAmut-His)was generated by two-step PCR with the TYP31/TYP13 and TYP14/TYP20 primer sets as described above. The lacI-Pspac-nonAmut-His, amyE5= rrnB terminator, and kan-amyE 3= fragments were ligated and intro-duced into the amyE locus of B. subtilis ASK3002 as described above. Thistransformant was named TAY3227. The transformation was confirmedby PCR and DNA sequencing.

The nonA-His and nonAmut-His genes were amplified by primersTYP34 and TYP20. The PCR products were digested by EcoRI and BamHIand cloned into a similarly digested pUC18 plasmid to generate pUC-nonA-His and pUC1-nonAmut-His. The transformation of these plasmidsinto E. coli JM109 was confirmed by PCR and DNA sequence analysis.

RNA isolation. B. subtilis was grown to an optical density at 600 nm(OD600) of 0.5 and then left uninfected or infected with SP10, SPO1, �29,or SPP1 phage at a multiplicity of infection (MOI) 5 or 10. To assay theinduction of SP10 sigma factors, B. subtilis was grown in LB medium at37°C to an OD600 of 0.5, and then xylose was added to the medium at afinal concentration of 0.5% (wt/vol). Cells in 10-ml cultures were har-vested and washed once with an equal volume of 10 mM Tris-HCl buffer(pH 7.4), and then pellets were frozen at 80°C. Total RNA was isolatedfrom the cells as described previously (35). The RNA quality was con-firmed by agarose gel electrophoresis, and the amount of RNA was quan-tified using NanoDrop ND-1000 (NanoDrop Technologies).

Northern blot analysis. Total RNA was separated on a 1.4% agarosegel containing 2% formaldehyde and transferred to Gene Screen Plusnylon membranes (PerkinElmer) by capillary blotting. RNA was cross-linked under UV irradiation and then digoxigenin (DIG)-labeled DNA orRNA probes were hybridized and detected according to the DIG Applica-tion Manual provided by the manufacturer (Roche) (36).

Digoxigenin-labeled RNA probes were transcribed in vitro using T7RNA polymerase with DIG RNA labeling mix (Roche) according to themanufacturer’s instructions. To prepare the templates, DNA fragmentscorresponding to probes 1 and 3 and to probes 2 and 4 (Fig. 1) wereamplified by PCR using the TYP37/TYP38 and TYP25/TYP36 primerssets, respectively, and cloned into the EcoRI site of pGEM-3Zf (�) (Pro-mega). The direction of inserts was confirmed by PCR and DNA sequenc-ing. Fragments of DNA containing the T7 promoter and each insert wereamplified by PCR using TYP21 and a primer corresponding to the 3= endof the insert, and then the PCR product was used as the template fortranscription in vitro.

Digoxigenin-labeled DNA probes were generated using a DIG-HighPrime Kit (Roche) according to the manufacturer’s instructions. Tem-plates for DIG-labeled DNA probes were prepared by PCR using primersfrom TYP39 to TYP54 (see Table S1 in the supplemental material).

5=/3= Rapid amplification of cDNA ends (RACE). Total RNA (5 �g)prepared from the B. subtilis cells infected with SP10 for 30 min wasincubated with or without 30 units of tobacco acid pyrophosphatase(TAP) in 10 �l of reaction buffer at 37°C for 1 h, and then TAP-treated oruntreated RNA (1 �g) was circularized with 5 units of T4 RNA ligase at

37°C for 1 h. cDNA was synthesized using PrimeScript reverse transcrip-tase (TaKaRa Bio) and random nonamers from RNA (0.2 �g) treated withT4 RNA ligase. The ligated junction region was amplified from the cDNAby PCR using primers TYP55 and TYP56. The amplified DNA productswere digested with EcoRI and BamHI, cloned into the pUC18 plasmid,and sequenced.

Western blot analysis. B. subtilis was grown to an OD600 of 0.5 andthen infected with SP10 phage at an MOI of 10. Cells from 40-ml cultureswere harvested 30 min after SP10 infection and washed once with an equalvolume of 10 mM Tris-HCl buffer (pH 7.4). Pellets were frozen and storedat 80°C. The cells were resuspended in 4 ml of 50 mM sodium phosphatebuffer (pH 7.0) and passed three times through a French press. Cell debriswas removed by centrifugation at 4°C for 30 min at 20,000 g, and thesupernatant served as whole-cell lysate. Supernatant (3 ml) was separatedby centrifugation at 4°C for 1 h at 250,000 g, and the supernatant servedas the cytoplasmic fraction. The pellet was resuspended in 3 ml of 50 mMsodium phosphate buffer (pH 7.0) and then separated by centrifugation at4°C for 1 h at 250,000 g. The supernatant was discarded, and the pelletwas resuspended in 0.4 ml of 50 mM sodium phosphate buffer (pH 7.0)served as the membrane fraction.

An equal volume of 2 sample buffer (0.1 M Tris-HCl [pH 6.8], 4%SDS, 12% [vol/vol] 2-mercaptoethanol, 20% glycerol, and 0.004% [wt/vol] bromophenol blue [BPB]) was added to whole-cell lysates and cyto-plasmic and membrane fractions and incubated at 95°C for 10 min. Thesesamples (0.1 or 0.005 OD600 units) were resolved by SDS-PAGE andtransferred onto polyvinylidene fluoride (PVDF) membranes using asemidry electroblotter. The PVDF membrane was blocked with 2.5%skim milk in Tris-buffered saline containing 0.2% Tween 80. Mousemonoclonal anti-6His antibody (Wako) diluted 1:5,000 and rabbitpolyclonal anti-SigA diluted 1:5,000 were used as the primary antibodies,and horseradish peroxidase-conjugated anti-mouse IgG and anti-rabbitIgG (GE Healthcare) diluted 1:50,000 were used as the secondary antibod-ies. Primary and secondary antibodies were diluted in Can Get SignalImmunoreaction Enhancer Solution 1 and Solution 2 (Toyobo), respec-tively.

One-dimensional SDS-PAGE and matrix-assisted laser desorptionionization–time of flight mass spectrometry (MALDI-TOF MS) analy-sis. B. subtilis was cultured in 300-ml conical flasks containing 60 ml of LBmedium at 37°C to an OD600 of 0.5 and then harvested before infection (0min) and at 15, 30, 45, and 60 min after infection with SP10 at an MOI 10.The cells harvested from 1.0-ml cultures were washed once with an equalvolume of 10 mM Tris-HCl buffer (pH 7.4), pelleted, frozen, and stored at80°C. The cells were resuspended in 50 �l of Solution 1 (25 mM Tris-HCl [pH 8.0], 10 mM EDTA, and 50 mM glucose) with 0.3 mg/ml ly-sozyme and incubated at 37°C for 10 min. An equal volume of 2 samplebuffer was then added, and the cells were incubated at 95°C for 10 min.The SP10 phage was precipitated from 5 ml of SP10 phage lysates mixedwith 5 ml of 20% (wt/vol) polyethylene glycol 6000 containing 2 M NaCland incubated at 4°C overnight. SP10 phage was harvested by centrifuga-tion at 4°C for 10 min at 20,000 g. The pellet was rinsed with 99.5%ethanol, resuspended in 200 �l of 1 sample buffer (2 sample bufferdiluted with an equal volume of Solution 1), and incubated at 95°C for 10min. These B. subtilis and SP10 samples were resolved by 13% or 7%SDS-PAGE and stained with Coomassie brilliant blue (CBB).

Protein bands were excised from gels and reduced with 1.5 mg/mldithiothreitol (DTT) in 100 mM NH4HCO3 at room temperature for 15min. The DTT solutions were removed, and the proteins in the gels werealkylated with 10 mg/ml of iodoacetamide in 100 mM NH4HCO3 at roomtemperature for 15 min. The iodoacetamide solutions were removed, andCBB was washed out from the gels twice with 50 mM NH4HCO3 in 50%methanol. The gels were dehydrated by incubation with 100% acetonitrile(ACN) at room temperature for 10 min. The ACN was removed, and thegels were dried by centrifugal evaporation. The proteins in the gels weredigested with sequencing-grade modified trypsin (Promega) at 37°C over-night and then eluted from the gels with 50 �l of 50% (vol/vol) ACN–

Abortive Infection in B. subtilis Marburg Strain

February 2014 Volume 196 Number 3 jb.asm.org 695

on October 16, 2017 by guest

http://jb.asm.org/

Dow

nloaded from

0.1% (wt/vol) trifluoroacetic acid (TFA) at room temperature for 20 min.The eluates were collected, and peptide was eluted once with 70 �l of 50%ACN– 0.1% TFA. These peptide solutions were condensed by centrifugalevaporation, and then 2-�l samples were mixed with 2 �l of matrix (3mg/ml of �-cyano-4-hydroxycinnamic acid in 50% ACN– 0.1% TFA),and 0.5 �l of this mixture was spotted onto plates.

We measured MALDI-TOF MS spectra using a tandem TOF (TOF/TOF) 5800 system (AB SCIEX) and analyzed them using mMass software(37). The proteins were identified by a Mascot peptide mass fingerprint

search (Matrix Science, Boston, MA) using the NCBI nonredundant(NCBInr) database (taxonomy, all entries).

Transmission electron microscopy. B. subtilis TAY3000, TAY3200,and TAY3201 were grown in 300-ml conical flasks containing 60 ml of LBmedium at 37°C to an OD600 of 0.5 and infected with SP10 at an MOI 10.Cultures (1 ml) were harvested 45 min after SP10 infection and washedonce with an equal volume of 10 mM Tris-HCl buffer (pH 7.4). These cellswere resuspended in 1 ml of fixative solution (2% glutaraldehyde– 0.1 Mphosphate buffer, pH 7.4) and incubated at 4°C for 30 min. The cells were

bnrdF bnrdE

100 bpProbe 1 Probe 2

Probe 3 Probe 4RNAprobes

4,7276,948

2,6611,8211,5171,049

575483310

- + - +0 5 15304560 0 5 15304560 0 5 15304560 0 5 15304560

ASK3000 ASK3002

Time (min)SP10 infection

nt

23S rRNA

4,7276,948

2,6611,8211,5171,049

575483310

- + - +0 5 15304560 0 5 15304560 0 5 15304560 0 5 15304560

ASK3000 ASK3002

Time (min)SP10 infection

nt

23S rRNA

4,7276,948

2,6611,8211,5171,049

575483310

- + - +0 5 15304560 0 5 15304560 0 5 15304560 0 5 15304560

ASK3000 ASK3002

Time (min)SP10 infection

nt

23S rRNA

4.1 knt3.4 knt

- + - +0 5 15304560 0 5 15304560 0 5 15304560 0 5 15304560

ASK3000 ASK3002

4,7276,948

2,6611,8211,5171,049

575483310

ntTime (min)

SP10 infection

23S rRNA

350 nt

A

B

C

Probe1 Probe2

Probe3 Probe4

1 kbp

nrdF nrdE nrdIymaB ymzA

bnrdF bnrdF bnrdIintein

yosQyosR yosL

xnrdIxnrdExnrdF orf174orf178

B. subtilis

SPβ

SP10

1 2 3 4 5 6 7 8 9 10 1112 1314 1516 1718 19 20 2122 23 24 1 2 3 4 5 6 7 8 9 10 1112 1314 1516 1718 19 20 2122 23 24

1 2 3 4 5 6 7 8 9 10 1112 1314 1516 17 18 19 20 2122 2324 1 2 3 4 5 6 7 8 9 101112 1314 1516 171819 202122 23 24

Lane Lane

Lane Lane

*

*

FIG 1 Detection of nonA transcripts. (A) Schematic of the intergenic region between bnrdE and bnrdF genes in the prophage SP� region of B. subtilis genome.Thin black arrows represent the directions and locations of RNA probes. Probes 1 and 2 and probes 3 and 4 were designed to hybridize to bnrdEF sense and bnrdEFantisense transcripts, respectively. (B) Northern blot of ASK3000 (�ydiR-ydjA) and ASK3002 (�SP� �ydiR-ydjA) infected with or without SP10 phage. TotalRNA (1 �g) isolated from cultures at the indicated time points after SP10 infection was used for Northern blot analysis against probes 1, 2, 3, and 4. The loadingcontrol is 23S rRNA on membrane stained with methylene blue. The size marker is RNA molecular weight marker I (Roche). Asterisks represent the locations ofthe 23S and 16S rRNAs. (C) Schematic of locus encoding the ribonucleotide reductase in the B. subtilis, SP�, and SP10 genome. Gray, striped, and black arrowsrepresent nrdI, nrdE, and nrdF homologs, respectively. White arrows represent nonconserved genes.

Yamamoto et al.

696 jb.asm.org Journal of Bacteriology

on October 16, 2017 by guest

http://jb.asm.org/

Dow

nloaded from

collected by centrifugation, the supernatant was decanted, and the cellswere resuspended in 1 ml of fixative solution and stored at 4°C. Theelectron microscope study was carried out at Hanaichi UltraStructureResearch Institute. The cells were fixed in osmium tetroxide, dehydratedusing a graded ethanol series (50% to 100%), and embedded in Epon812. Ultrathin sections (80 to 90 nm) were cut using an ultrami-crotome, stained with uranyl acetate and lead, and then examinedusing a JEM1200EX transmission electron microscope (TEM; JEOL,Ltd.) at 80 kV.

Phage assays. The number of PFU was measured as described previ-ously (31). B. subtilis grown in 5 ml of LB medium in test tubes at 37°C toan OD610 of 0.5 was infected with SP10 at an MOI of 10 to assess lysis bymeasuring the OD610 for 5 h after infection.

Bacterial cell toxicity assays. B. subtilis TAY3221 and TAY3227 werecultured in LB medium at 37°C to an OD600 of 0.35 and then NonA-6His was induced with 1 mM IPTG. The OD600s of controls and ofIPTG-induced samples were measured for 5 h after induction. Cells in1-ml cultures before and 1 h after of IPTG induction were harvested andresuspended in an equal volume of LB medium and serially diluted 10-fold, and then 2-�l portions were spotted onto LB plates and incubated at37°C overnight to measure cell viability. NonA toxicity in E. coli wasanalyzed in a similar manner to the method used for B. subtilis, but E. coliJM109 harboring pUC-nonA-His and pUC-nonAmut-His was cultured inLB medium with 50 �g/ml of ampicillin. E. coli was cultured at 30°C toreduce the copy numbers of these plasmids. To repress basal level tran-scription of the nonA gene from the lac promoter of pUC-nonA-His orpUC-nonAmut-His, 1% (vol/wt) glucose was added to LB medium (forpreculture) and LB plates (for colony-forming assays).

Fluorescence microscopy. Strains TAY3221 and TAY3227 harborednonA, which was expressed as described above. TAY3221 cells were killedby incubation at 80°C for 20 min for use as dead cell controls. Cells in 1-mlcultures after IPTG induction for 1 h were harvested and washed oncewith 1 ml of 0.85% (vol/wt) NaCl. The cells were stained with 4=,6=-diamidino-2-phenylindole (DAPI; Dojindo), propidium iodide (PI; Bio-Vision), or CTC (5-cyano-2,3-ditolyl tetrazolium chloride) (BacstainCTC Rapid Staining Kit for microscopy; Dojindo) at 37°C for 30 min.Cells stained with CTC were subsequently stained with DAPI at roomtemperature for 5 min, and then 5 �l of these cells was placed on agarose-coated glass slides and viewed using an LSM710 (Carl Zeiss) microscope.

RESULTSThe region between the bnrdE and bnrdF genes is the site ofnonA RNA transcription induced by SP10 phage infection. The825-bp region of prophage SP� encompassing the end of bnrdE(245 bp), the intergenic region between bnrdE and bnrdF (332 bp),and the start of bnrdF (248 bp) is required for wild-type nonA genefunction (31), which suggests that the nonA transcript is tran-scribed from this region. We prepared four RNA probes (probes 1,2, 3, and 4) corresponding to each strand of each half of the 825-bpregion (Fig. 1A) to determine the location and direction of nonAgene transcription by Northern blotting. Probes 1 and 2 were de-signed to hybridize to the bnrdEF sense transcript, and probes 3and 4 were designed to hybridize to the bnrdEF antisense tran-script. Total RNAs were prepared from the ASK3000 (�ydiO-ydjA) and ASK3002 (�SP� �ydiO-ydjA) strains with or withoutSP10 infection for Northern blotting. Probe 1 detected transcriptsof 4.1 and 3.4 kilonucleotides (knt) in both ASK3000 andASK3002 with and without SP10 infection (Fig. 1B, probe 1).Since the ASK3002 strain has an SP� deletion, the two transcriptswere not bnrdEF mRNAs. In addition to the bnrdIEF operon in theprophage SP� region, the B. subtilis genome harbors another ri-bonucleotide reductase operon, nrdIEF (Fig. 1C). Furthermore,phage SP10 has the ribonucleotide reductase genes xnrdIEF,

which are homologs of nrdIEF and bnrdIEF (Fig. 1C). The 4.1- and3.4-knt transcripts were detected regardless of SP10 infection (Fig.1B, probe 1), indicating that these transcripts were not transcribedfrom the xnrdEF locus. The nucleotide sequence of the first 248 bpof the 5= end of the SP� bnrdF genes is very similar to that of the B.subtilis nrdF gene (99% identity), and the nrdF gene is transcribedas 4.1- and 3.4-knt transcripts corresponding to nrdIEF-ymaB andthe nrdIEF operon, respectively (32). These data indicated thatprobe 1 cross-hybridized with, and thus detected, the nrdIEF-ymaB and the nrdIEF transcripts. Probes 2 and 3 did not detectsignals in either strain at any time (Fig. 1B, probes 2 and 3). On theother hand, probe 4 detected a transcript of about 350 nt inASK3000 from 15 to 60 min after SP10 infection (Fig. 1B, probe 4,lanes 9 to 12), but this transcript was not detected in ASK3002(Fig. 1B, probe 4, lanes 13 to 24), suggesting that its transcriptionwas occurred from bnrdEF in the SP� region. Probe 2 did notdetect this band, indicating that nonA is likely transcribed in theopposite direction to bnrdEF. We concluded that the �350-nttranscript detected by probe 4 corresponds to the nonA transcript.This nonA transcript was not detected when SP10 phage did notinfect with ASK3002 (Fig. 1B, probe 4, lanes 1 to 6), suggestingthat SP10 phage induces the nonA transcription.

We investigated the transcriptional start and end sites of nonAusing 5=/3= RACE analysis (data not shown). Total RNA was pre-pared from ASK3000 cells at 30 min after infection with SP10.Total RNA with or without tobacco acid pyrophosphate (TAP)treatment was circularized using T4 RNA ligase, and reverse tran-scription-PCR (RT-PCR) amplification across the 5=/3= junctionof nonA RNA proceeded using custom-designed primers. Sincethe 5= ends of primary transcripts were triphosphorylated, pri-mary transcripts were circularized by T4 RNA ligase whentriphosphorylated RNA 5= ends were converted to monophos-phorylated 5= ends by TAP. Therefore, the 5= end site mappedfrom TAP-treated RNA represents a transcriptional start site. AnRT-PCR product of �320 bp was generated from RNA treatedwith TAP but not from that RNA without TAP treatment (datanot shown). Cloning and sequencing of the � 320-bp RT-PCRproduct (data not shown) showed that the 5= and 3= ends of theprimary nonA transcript comprised an A and a T residue located243 bases downstream and 127 bases upstream of the bnrdE trans-lational stop codon, respectively (Fig. 2B). The length of nonARNA is 370 nt, and the region located 130 nt from its 3= endoverlaps with the coding sequence of the bnrdE gene (Fig. 2A andB). The consensus sequences of any sigma factors of B. subtiliswere not found upstream of the nonA transcriptional start sites.The transcriptional regulator of nonA is discussed later in thisarticle. An inverted repeat followed by T-rich sequences, whichcould act as an intrinsic (rho-independent) terminator, was iden-tified in the 3= end of nonA (Fig. 2B).

Taken together, these findings indicated that the nonA gene islocated in the bnrdEF intergenic region and that it is transcribed inthe opposite direction to the bnrdEF genes. Moreover, SP10 infec-tion apparently induces nonA transcription.

The wild-type nonA phenotype requires a 72-amino-acidprotein encoded by the nonA gene. We found a 72-amino-acidORF in the nonA transcriptional region and a canonical RBS(AGGA) 10 bp upstream of its start codon (Fig. 2B and C). Todetermine whether or not the nonA ORF is responsible for thephage SP10 resistance phenotype, mutations were introduced intoits start codon or the RBS, and wild-type or mutated nonA genes

Abortive Infection in B. subtilis Marburg Strain

February 2014 Volume 196 Number 3 jb.asm.org 697

on October 16, 2017 by guest

http://jb.asm.org/

Dow

nloaded from

were integrated into the amyE locus of the phage SP10-sensitivestrain, ASK3002 (Fig. 3A). The resultant strains were infected withphage SP10 at an MOI of 10, and then cell lysis was determined bymeasuring reductions in the OD610 (Fig. 3B). The nonA� strain,TAY3000 (�ydiO-ydjA amyE::kan), was not lysed at 5 h after SP10infection (Fig. 3B). On the other hand, the nonA strain, TAY3200(�SP� �ydiO-ydjA amyE::kan), started to become lysed at 1 hafter SP10 infection (Fig. 3B). The OD610 of TAY3201 that ex-pressed the wild-type nonA gene from its native promoter at theamyE locus did not become reduced after SP10 infection (Fig. 3B),whereas strains harboring the untranslatable nonA mutation,TAY3202 (start codon mutant) and TAY3205 (RBS mutant), werelysed at 1 h after infection. These results indicated that the 72-amino-acid protein encoded by the nonA gene is required fornonA-dependent resistance to the SP10 phage in B. subtilis.

The nucleotide sequences at the 3= ends of nrdE, bnrdE, andxnrdE are similar. To exclude the possibility that the 3=UTR of thenonA transcript, which overlaps the 3= end of the nrdE codingsequence, inhibits the expression of nrdE and/or xnrdE by baseparing to inhibit SP10 multiplication, we replaced the 3=UTRs ofthe wild-type and untranslatable nonA mutants with that of thespoVG gene of B. subtilis and integrated them into the amyE locusof the ASK3002 chromosome (Fig. 3C). SP10 phage-induced lysisof the resultant strains was then analyzed as described above. TheOD610 of TAY3204, TAY3206, TAY3202, and TAY3205 becamereduced after SP10 infection, whereas that of TAY3203 andTAY3201 did not (Fig. 3D). These findings suggested that the 3=UTR of nonA is not responsible for SP10 phage resistance.

We measured the efficiency of SP10 plaque formation in thestrains shown in Fig. 3 (Table 2). Similar to the results shown inFig. 3, TAY3201 and TAY3203 were resistant to the SP10 phage(efficiency of plaquing [EOP] of 1.2 105 and 7.4 106, re-

spectively), whereas TAY3202 and TAY3204 were sensitive (EOPof 8.5 101 and 8.2 101, respectively). The EOPs and plaquesizes did not significantly differ among nonA 3= UTR and corre-sponding spoVG 3= UTR strains (Table 2). Taken together, theseresults indicate that NonA protein is necessary for the nonA-de-pendent SP10 phage resistance phenotype, whereas the 3=UTR ofnonA is not.

NonA localizes to the membrane. NonA consists of 72 aminoacids and has a calculated molecular mass of 8,465 Da, a C termi-nus containing many basic amino acids, and a calculated isoelec-tric point (pI) of 9.64. A BLASTP and a TBLASTN search of NonAdid not find any protein similar to it. We predicted the domainstructure of NonA using the InterProScan program (38) andfound a predicted transmembrane domain between amino acids25 and 47 (Fig. 2C). To determine whether NonA localizes to themembrane, we constructed the pHY-nonA-His plasmid that ex-presses NonA with a His tag at the C terminus, which comple-mented the nonA-deficient phenotype, as well as untagged NonA(data not shown). Extracts of ASK3002 cells harboring pHY-nonAor pHY-nonA-His infected with SP10 phage were fractionated byultracentrifugation, and His-tagged NonA and SigA were detectedby Western blotting against mouse anti-His-tagged antiserum andrabbit anti-SigA antiserum, respectively. Both His-tagged NonAand SigA were detected in whole-cell lysates (Fig. 4, lanes 1 and 4).SigA, which is a cytosolic protein, was detected in the cytosolicfraction but not in the membrane faction (Fig. 4, lanes 2, 3, 5, and6). In contrast, the detection of His-tagged NonA exclusively inthe membrane fraction (Fig. 4, lanes 5 and 6) indicated that itlocalizes to the membrane.

The nonA gene is transcribed at the late stage of the SP10infection. We analyzed transcription of the SP10 genes, nonA, andmodification and restriction genes after SP10 infection in wild-

10 20 30 40 50 60 70MNFLDNLISL LESTVISTGW LCDRLLSCLI EYILKLIFGL LCNALMRKFN KIKCKRRPGD PVNFEPKRKR TY

5’ AAATTTTTGT CAATTAATCA TGTCCTTTCA ATTGAAATAT TTTTTTCCGA AAAAAATGCA GTGAATTTAA AAGCTGAATT GACAGACTCC TATATTACTG 3’3’ TTTAAAAACA GTTAATTAGT ACAGGAAAGT TAACTTTATA AAAAAAGGCT TTTTTTACGT CACTTAAATT TTCGACTTAA CTGTCTGAGG ATATAATGAC 5’

5’ TCAAACATTC GAAAGGAATG AAAGACAGTG AATTTTTTAG ATAACCTTAT ATCTTTGTTA GAGAGTACAG TCATTTCAAC TGGATGGTTG TGTGATAGGC 3’3’ AGTTTGTAAG CTTTCCTTAC TTTCTGTCAC TTAAAAAATC TATTGGAATA TAGAAACAAT CTCTCATGTC AGTAAAGTTG ACCTACCAAC ACACTATCCG 5’

5’ TTTTATCATG TTTAATCGAA TATATTTTGA AGCTTATATT TGGTTTACTT TGTAATGCTC TTATGAGAAA ATTTAACAAA ATTAAGTGCA AACGACGTCC 3’3’ AAAATAGTAC AAATTAGCTT ATATAAAACT TCGAATATAA ACCAAATGAA ACATTACGAG AATACTCTTT TAAATTGTTT TAATTCACGT TTGCTGCAGG 5’

5’ GGGAGATCCT GTTAATTTTG AACCAAAACG TAAACGTACC TATTAAACTA CACAGGACAA GCATCCCTCT TGGGTCGTAT CTTTAGTTCT GGCATAATAT 3’3’ CCCTCTAGGA CAATTAAAAC TTGGTTTTGC ATTTGCATGG ATAATTTGAT GTGTCCTGTT CGTAGGGAGA ACCCAGCATA GAAATCAAGA CCGTATTATA 5’

5’ AAAGTTTTAA TTCCACGATG ATGTGCGTAG AGGTCTATTC TATTTAGGTC TCTCGTCGTC ATCGTATCCT TTAAGAACAA CGTAAATGAA ATCCCTTGGT 3’3’ TTTCAAAATT AAGGTGCTAC TACACGCATC TCCAGATAAG ATAAATCCAG AGAGCAGCAG TAGCATAGGA AATTCTTGTT GCATTTACTT TAGGGAACCA 5’

bnrdF start codonnonA RBS nonA start codon

nonA stop codon

bnrdE stop codonnonA 3’ end

nonA TSS

A

B

bnrdF bnrdE100 bp

nonA

C - - + - + ++ + + +++ - ++++

FIG 2 Determination of nonA transcriptional start and end sites and prediction of the nonA ORF. (A) Schematic of the nonA locus in the SP� region of the B.subtilis genome. The bent arrow and lollipop structure, respectively, indicate the transcriptional start site and end site of nonA. White arrows represent the bnrdEand bnrdF genes, and the gray arrow represents the predicted nonA ORF. (B) DNA sequence around the nonA gene. The bent arrow and inverted arrows representthe transcriptional start site and terminator structure of nonA, respectively. Boldface and shaded areas indicate transcriptional end site, predicted ribosomalbinding site (RBS), and predicted start and stop codons of nonA. Boldface and shaded areas also indicate bnrdF start and bnrdE stop codons. (C) Amino acidsequence of the predicted NonA protein. Boldface and shaded areas indicate the transmembrane domain predicted by InterProScan (38). Plus and minus symbolsindicate positively and negatively charged amino acids, respectively.

Yamamoto et al.

698 jb.asm.org Journal of Bacteriology

on October 16, 2017 by guest

http://jb.asm.org/

Dow

nloaded from

type UOT1285 and SP� and/or ydiR-ydjA-deficient mutants byNorthern blotting. The orf057 (superinfection exclusion protein),orf191 (DNA polymerase), and orf148 (tail sheath protein precur-sor) are regarded as early, middle, and late genes, respectively, ofSP10. These SP10 phage genes were similarly expressed inASK3000 and ASK3002 (Fig. 5, lanes 7 to 12 and 19 to 24), sug-gesting that NonA does not affect their transcription. These find-ings are consistent with our previous data obtained by quantita-tive RT-PCR (qRT-PCR) (31). On the other hand, the amounts ofSP10 gene transcripts were obviously lower in UOT1285 andASK3001 than in ASK3000 and ASK3002 (Fig. 5). UOT1285 andASK3001 both have restriction (BsuMR) and modification(BsuMM) enzymes encoded by ydiR-ydiS-ydjA and ydiO-ydiP,

respectively, and these genes were transcribed at all time points(Fig. 5, lanes 1 to 6 and 13 to 18). Both ydiR-ydiS-ydjA and ydiO-ydiP are transcribed independently of SP10 infection during thevegetative growth phase (39). These results showed that the re-striction enzyme BsuMR was constitutively expressed before SP10infection and that it can immediately cleave the SP10 genomeupon infection. In contrast, the nonA transcript was detected onlyin ASK3000 (Fig. 5, lanes 9 to 12) although UOT1285 andASK3000 both carry the nonA gene. The transcription of nonAstarted in ASK3000 from 15 to 60 min after SP10 infection when thelate gene orf148 was expressed (Fig. 5, lanes 9 to 12), indicating thatthe nonA gene is expressed during the late stage of the SP10 infection.The SP10 genes were transcribed in ASK3000 but not in UOT1285(Fig. 5, lanes 1 to 12), and nonA was transcribed in ASK3000 only after

nonA bnrdEbnrdFAGGA/GTG

TAY3201

nonA bnrdEbnrdFAGGA/GAG

TAY3202

nonAbnrdFAGGA/GTG

TAY3203

nonA bnrdEbnrdFAAAA/GAG

TAY3205

RBS/Start codon

nonAbnrdFAGGA/GAG

TAY3204

nonAbnrdFAAAA/GTG

TAY3206

spoVG 3′ UTR

TAY3000

TAY3200

TAY3201

TAY3202TAY3205

TAY3000

TAY3200

TAY3203

TAY3204TAY3206

A B

C D

0 1 2 3 4 50.01

0.10

1.00

OD

610

Time after SP10 infection (hours)

0 1 2 3 4 50.01

0.10

1.00

OD

610

Time after SP10 infection (hours)

spoVG 3′ UTR

spoVG 3′ UTR

FIG 3 SP10 phage resistance phenotype requires NonA protein. (A and C) Schematic of the nonA gene and various mutants included in nonA complementationanalysis. Each gene was integrated into the amyE locus of ASK3002 (�SP� �ydiR-ydjA). Boldface and underlining indicate mutated nucleotides. (B and D) OD610

of indicated strains after infection with SP10. Strains were cultured in LB medium at 37°C to an OD610 0.5 and then infected with SP10 at an MOI of 10. Resultsare means of three independent experiments, and bars represent standard errors. (B and D) Results from same set of experiments.

TABLE 2 EOP and plaque size of SP10

Strain Genotype EOPa,b

Plaquediameter(mm)b

TAY3000 �ydiR-ydjA amyE::kan �8.7 109 NDc

TAY3200 �SP� �ydiR-ydjA amyE::kan 1.0 0.4 0.83 0.22TAY3201 �SP� �ydiR-ydjA amyE::nonA (1.2 0.6) 105 0.31 0.11TAY3202 �SP� �ydiR-ydjA amyE::nonA(T27A) (8.5 1.4) 101 0.75 0.21TAY3203 �SP� �ydiR-ydjA amyE::nonA-spoVG

3= UTR(7.4 4.1) 106 0.3 0.09

TAY3204 �SP� �ydiR-ydjA amyE::nonA(T27A)-spoVG 3= UTR

(8.2 2.3) 101 0.75 0.21

a The EOP of SP10 phages is 1.0 on TAY3200.b EOP, plaque size, and standard error values were calculated from three independentexperiments.c ND, not detected.

W C M W C MpHY-NonA pHY-NonA-His

-6 His

-SigA

1 2 3 4 65Lane

FIG 4 NonA protein localizes to cell membrane. Western blots of ASK3002harboring pHY-NonA and pHY-NonA-His. Cells harvested at 30 min afterSP10 infection were disrupted using a French press. Whole-cell lysates (W)were separated into cytosol (C) and membrane (M) fractions by ultracentrif-ugation. Lanes were loaded with cell lysate corresponding to 0.1 (top panel) or0.005 OD600 units (bottom panel) of cultures. �, anti.

Abortive Infection in B. subtilis Marburg Strain

February 2014 Volume 196 Number 3 jb.asm.org 699

on October 16, 2017 by guest

http://jb.asm.org/

Dow

nloaded from

SP10 infection (Fig. 1, lanes 1 to 12), suggesting that products of theSP10 genes are involved in nonA expression.

The sigma factor Orf199-200 encoded by SP10 phage is re-quired for nonA transcription. The SP10 phage has three sigmafactor homologs, Orf120, Orf183, and Orf200, and they may di-rect the transcription of middle and late genes (26). This suggeststhat one of these sigma factors regulates nonA transcription. Weinvestigated this notion by expressing Orf120, Orf183, and Orf200under the xylose-inducible promoter in cells that were not in-fected and detected nonA mRNA by Northern blotting. Theorf199, annotated as a sigma factor accessory protein in GenBank,is located upstream of orf200. The Orf199 and Orf200 are ho-mologs of SPO1 Gp33 and Gp34, respectively (26). Gp33 is alsoannotated as a sigma factor accessory protein. Gp33 binds RNApolymerase and is required for Gp34-dependent transcriptionalactivation (40, 41). Based on these facts, we assumed that Orf199 isrelated to Orf200-dependent transcription and therefore coex-

pressed Orf199 with Orf200. Each of genes orf120, orf183, andorf199-200 was cloned under the control of the xylose-induciblepromoter in pWH1520, including the native RBS. UOT1285 har-boring plasmids expressing each sigma factor or pWH1520 weregrown to the mid-log phase, and then the sigma factors were in-duced by xylose. Total RNAs isolated from these cells were sub-jected to Northern blot analyses. These sigma factors were tran-scribed after xylose induction (Fig. 6A, lanes 6 to 8, 10 to 12, and14 to 16). The nonA transcript was detectable after the inductionof Orf199-200 (Fig. 6A, lanes 10 to 12) but not Orf120 or Orf183(Fig. 6A, lanes 6 to 8 and 14 to 16). Thus, the SP10 phage-encodedsigma factor, Orf200, and the sigma factor accessory protein,Orf199, regulated nonA transcription.

Infection with the B. subtilis phages SPO1, �29, and SPP1that can grow on B. subtilis Marburg does not induce nonA tran-scription. The B. subtilis Marburg strain is resistant to the SP10phage but susceptible to some other B. subtilis phages, such asSPO1, �29, and SPP1 (25). The genomes of SPO1, �29, and SPP1phages do not contain any BsuMR target sites, thus allowing thesephages to escape the BsuMR restriction system. However, whythey can grow normally in the nonA� strain remains unknown.Since the SP10 sigma factor regulated nonA transcription, we ex-amined whether the nonA gene is transcribed during the infectionof B. subtilis by SPO1, �29, and SPP1 phages. Total RNAs isolatedfrom ASK3000 cells infected in parallel with SP10, SPO1, �29, andSPP1 phages were used for Northern blot analysis. SP10 infectioninduced nonA gene transcription (Fig. 6B, upper panel, lanes 3 to6). In contrast, nonA transcription was undetectable in cells in-fected with SPO1, �29, and SPP1 (Fig. 6B, upper panel, lanes 7 to24), whereas their late genes were transcribed (Fig. 6, lower panel).This indicates that nonA was not transcribed in ASK3000 duringSPO1, �29, and SPP1 infection.

SP10 protein synthesis is inhibited in strains expressingnonA. SP10 genes were transcribed in the nonA� strain, ASK3000,during SP10 infection, suggesting that NonA does not affect tran-scription of SP10 genes. To test whether NonA affects SP10 pro-tein synthesis, whole-cell extracts prepared from TAY3000,TAY3200, and TAY3201 cells infected with SP10 and SP10 virions

0 5 15 30 45 60 0 5 15 30 45 60 0 5 15 30 45 60 0 5 15 30 45 60UOT1285 ASK3000 ASK3001 ASK3002

Time (min)

orf057

orf191

orf148

nonA

ydiO

ydiR

23S rRNA

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24Lane

FIG 5 Transcription of nonA occurs at a later stage of SP10 phage develop-ment. Northern blots of UOT1285 (wild type), ASK3000 (�ydiR-ydjA),ASK3001 (�SP�), and ASK3002 (�ydiR-ydjA �SP�) strains infected withSP10. Total RNA (1 �g for nonA; 3 �g for orf057, orf191, orf148, ydiO, andydiR) isolated from cultures at indicated times after SP10 infection was usedfor Northern blot analysis against the indicated probes. The 23S rRNA on themembrane stained with methylene blue is shown as a loading control.

0 1 2 3 0 1 2 3 0 1 2 3 0 1 2 3pWH1520 pWSP183 pWSP199-200 pWSP120

nonA

orf183

orf199-200

orf120

Time (h)

23S rRNA

0 5 15 30 45 60 0 5 15 30 45 60 0 5 15 30 45 60 0 5 15 30 45 60SP10 SPO1 φ29 SPP1

nonA

23S rRNA

PhagesTime (min)

0 5 15 30 45 60 0 5 15 30 45 60 0 5 15 30 45 60 0 5 15 30 45 60SP10 orf148 SPO1 gp16.1 φ29 gp8 SPP1 gp8

47276948

2661182115171049575483313

Time (min)nt

A B

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1617 18 19 20 2122 23 24

1 2 3 4 5 6 7 8 9 1011 12 1314 15 16 17 18 19 20 2122 23 24

Lane

Lane

Lane

FIG 6 Transcription of nonA is regulated by SP10-encoded sigma factor Orf199-200. (A) Northern blot of UOT1285 harboring pWH1520, pWSP183,pWSP199-200, and pWSP120. Total RNA (3 �g) isolated from cultures at the indicated times after xylose addition was used for Northern blot analysis againstnonA and SP10-encoded sigma factor probes. (B) Northern blots of total RNA (1 and 3 �g, top and bottom panels, respectively) from ASK3000 at the indicatedtimes after infection with SP10, SPO1 �29, or SPP1 phage. The size marker is RNA Molecular Weight Marker I. The 23S rRNA on the membrane stained withmethylene blue is shown as a loading control in both panels.

Yamamoto et al.

700 jb.asm.org Journal of Bacteriology

on October 16, 2017 by guest

http://jb.asm.org/

Dow

nloaded from

were resolved by SDS-PAGE, and proteins were visualized by CBBstaining. Global protein profiles after SP10 infection did not differat any time point in these strains (Fig. 7A). However, bands cor-responding to 61 (Fig. 7A, lower panel, lanes 8 to 10), 51, and 28kDa (Fig. 7A, upper panel, lanes 8 to 10) appeared in TAY3200from 30 to 60 min after SP10 infection. The intensity of thesebands was significantly lower in TAY3000 and TAY3201 (Fig. 7A,lanes 1 to 5 and 11 to 15). The bands corresponding to 61 and 51kDa were also detected in protein samples prepared from SP10virions (Fig. 7A, lane 16). These results suggest that the 61- and51-kDa proteins detected in ASK3002 are derived from SP10 vi-rion proteins. The 61- and 51-kDa protein bands detected in SP10virion samples (Fig. 7, lane 16) were excised and digested withtrypsin, and their mass spectra were obtained by MALDI-TOF MSas described in Materials and Methods. Peptide mass fingerprint-ing with Mascot software identified these 61- and 51-kDa proteinsas SP10 Orf148 (tail sheath protein precursor) and Orf141 (majorcapsid protein precursor) with E values of 8.1 1018 and 8.1 1010, respectively. We also excised the 51-kDa band detected inTAY3200 at 45 min after SP10 infection and the same regions ofthe gels in TAY3000 and TAY3201 and analyzed them by peptidemass fingerprinting as described above. Orf141 was detected in gelslices of TAY3200 samples, with an E value of 4 106, but not inthose of TAY3000 and TAY3201 samples (data not shown). Thus,synthesis of the major capsid protein precursor, Orf141, was in-hibited during SP10 infection of the nonA� strain.

Both Orf148 and Orf141 are components of SP10 virions. Weused transmission electron microscopy (TEM) to determinewhether or not SP10 virions are formed in B. subtilis cells.TAY3000, TAY3200, and TAY3201 cells infected with SP10 for 45min were prepared for TEM observation. Phage particles weredetected throughout the cytoplasm of TAY3200 (Fig. 7B, middlepanel), whereas particles were scant or absent in the nonA�

strains, TAY3000 and TAY3201 (Fig. 7B, upper and lower pan-els). Thus, the number of intracellular SP10 particles correlateswith the amounts of the 61-, 51-, and 28-kDa proteins shown inFig. 7A. Taken together, these results suggest that NonA disturbsthe synthesis of a set of SP10 proteins, which diminishes SP10virion formation in nonA� strains.

Growth of B. subtilis and E. coli is halted by overexpressionof NonA protein. The nonA gene was transcribed in the cells onlyduring SP10 infection. We analyzed the effect of NonA expressionin noninfected cells. To artificially control NonA expression, theHis-tagged nonA gene under the xylose-inducible promoter(PxylA) was transformed into B. subtilis ASK3002. A few transfor-mants were obtained, but PxylA was obviously mutated or deletedin all of them (data not shown). This suggests that NonA protein istoxic to B. subtilis even when it is expressed at a basal level withoutan inducer. To control the expression of NonA more strictly, weused the lacI and IPTG-inducible spac promoter (Pspac) cassettein B. subtilis TMO310 (34). This lacI-Pspac cassette was used totightly control expression of the counterselection marker MazF

BTAY3000

TAY3201

TAY3200

A

200

11697.4

66.2

45.0

31.0

21.5

0 15 30 45 60 0 15 30 45 60 0 15 30 45 60 TAY3000 TAY3200 TAY3201

Time (min)

61

51

28

SP10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16Lane

610 15 30 45 60 0 15 30 45 60 0 15 30 45 60

TAY3000 TAY3200 TAY3201Time (min) SP10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16Lane

FIG 7 SP10 phage protein synthesis is inhibited in nonA� strains. (A) Lysates (corresponding to 0.05 OD600 units) prepared from TAY3000 (�ydiR-ydjAamyE::kan), TAY3200 (�ydiR-ydjA �SP� amyE::kan), and TAY3201 (�ydiR-ydjA �SP� amyE::nonA kan) at indicated times after SP10 infection resolved by 10%(top panel) and 7% (bottom panel) SDS-PAGE. Proteins were stained with CBB. Size markers are SDS-PAGE broad-range standards (Bio-Rad). (B) Typicaltransmission electron microscope images (magnification, 10,000) of TAY3000, TAY3200, and TAY3201 infected with SP10 for 45 min. Scale bar, 500 nm.

Abortive Infection in B. subtilis Marburg Strain

February 2014 Volume 196 Number 3 jb.asm.org 701

on October 16, 2017 by guest

http://jb.asm.org/

Dow

nloaded from

toxin in that strain. The His-tagged nonA gene or its untranslat-able mutant (first codon mutant) was fused with a lacI-Pspac cas-sette and integrated into the amyE locus of the ASK3002 chromo-some. These strains were grown to an OD600 of 0.35 and incubatedwith or without 1 mM IPTG, and then the OD600 was measuredfor 5 h. Figure 8A shows that cell growth halted immediately afterthe NonA-6His protein was expressed, and the OD600 graduallyfell. Furthermore, the efficiency of TAY3221 colony formationwas reduced 106-fold when NonA-6His protein was overex-pressed (Fig. 8B). In contrast, induction of the untranslatablenonA-6His gene did not affect TAY3227 cell growth and colonyformation (Fig. 8A and B). These results indicate that NonA pro-tein overexpression is harmful to B. subtilis cells.

The overexpression of NonA-6His protein from the pUC18plasmid in the Gram-negative bacteria E. coli JM109 also haltedcell growth but only slightly reduced the efficiency of colony for-mation (Fig. 8C and D). As with B. subtilis, overexpression of theuntranslatable 6His-tagged nonA gene did not affect E. coli cellgrowth and colony formation. In the absence of IPTG as an in-ducer, the growth rate of JM109 cells harboring pUC18-NonAbecame lower than that of a strain harboring pUC18-NonAmut

after the late log phase (Fig. 8C). This is probably due to the leakyexpression of NonA-6His protein. Taken together, these results

show that NonA protein is similarly harmful to both Gram-posi-tive B. subtilis and Gram-negative E. coli.

Overexpression of NonA reduced B. subtilis respiration ac-tivity. NonA is localized to the membrane, and its overexpressionseverely affects cell growth and colony formation. We examinedwhether NonA overexpression results in membrane damage thatis involved in the inhibition of growth and colony formation. Weinduced NonA and control cells and then visualized them by flu-orescence microscopy using 4=,6-diamidino-2-phenylindole(DAPI) that stains DNA in cells with both intact and injuredmembranes (live and dead cells) and with propidium iodide (PI)that stains DNA and double-stranded RNA in cells with injuredmembranes (dead cells). Most TAY3221 cells killed by heating(dead cell control) were stained with PI (Fig. 9A). Few cells ex-pressing NonA were stained by PI (Fig. 9A). These results sug-gested that NonA overexpression does not seriously damage themembrane. We then stained these cells with DAPI and 5-cyano-2,3-ditolyl tetrazolium chloride (CTC), the reduced form (form-azan) of which emits red fluorescence. Thus, CTC specificallystains cells with respiratory activity. Although CTC stained mostcells that did not express NonA and the untranslatable mutants ofcells either expressing or not expressing nonA, CTC stained fewcells expressing NonA and heat-killed cells (Fig. 9B). These results

T0T1

T0 T1

100 10-1 10-2 10-3 10-4 10-5 10-6

TAY 3221 IPTG -100 10-1 10-2 10-3 10-4 10-5 10-6

T0 T1

TAY 3221 IPTG +

TAY 3227 IPTG -

TAY 3227 IPTG +

100 10-1 10-2 10-3 10-4 10-5 100 10-1 10-2 10-3 10-4 10-5T0 T1

pUCnonA IPTG -

pUCnonA IPTG +

pUCnonAmut IPTG -

pUCnonAmut IPTG +

A

C

B

D

0 2 4 6 80.01

0.1

1

10

OD

600

Time (hours)

0 2 4 6 8 100.01

0.1

1

10

OD

600

Time (hours)

FIG 8 Expression of NonA protein is toxic for B. subtilis and E. coli. (A) OD600 of B. subtilis TAY3221 (�SP� �ydiR-ydjA amyE::lacI Pspac-nonA-His) andTAY3227 (�SP� �ydiR-ydjA amyE::lacI Pspac-nonAmut-His). These strains were grown in LB medium at 37°C to an OD600 of 0.35 (time zero, T0) and thenincubated with or without 1 mM IPTG. Results are means of three individual experiments; bars represent standard deviations. Filled circles, TAY3221 withoutIPTG; open circles, TAY3221 with 1 mM IPTG; filled triangles, TAY3227 without IPTG; open triangles, TAY3227 with 1 mM IPTG. (B) Assay of TAY3221 andTAY3227 cell viability. Cells from 1-ml cultures at IPTG addition (as described for panel A) (time zero, T0) and at 1 h later (T1) were harvested by centrifugation,resuspended in LB medium (1 ml), serially 10-fold diluted, spotted (2 �l) on LB plates, and incubated at 37°C overnight. (C) OD600 of E. coli JM109 harboringpUC-nonA and pUC-nonAmut. These strains were cultured as described for panel A but with ampicillin (50 �g/ml) at 30°C. Filled circles, plasmid pUC-nonAwithout IPTG; open circles, pUC-nonA with 1 mM IPTG; filled triangles, pUC-nonAmut without IPTG; open triangles, pUC-nonAmut with 1 mM IPTG. (D) Cellviability of JM109 carrying pUC-nonA and pUC-nonAmut was assayed as described for panel B, but cells were cultured in LB plates containing ampicillin (50�g/ml) and glucose (1%, wt/vol) at 30°C overnight.

Yamamoto et al.

702 jb.asm.org Journal of Bacteriology

on October 16, 2017 by guest

http://jb.asm.org/

Dow

nloaded from

suggest that NonA overexpression reduces cellular respiratory ac-tivities and thus inhibits cell growth and colony formation.

DISCUSSION

The two SP10 resistance systems, nonA and nonB, have been iden-tified in the B. subtilis wild-type Marburg strain (24). The nonB

gene encodes the restriction enzyme BsuMR, which cleaves phagegenomic DNA and thus allows nonB to inhibit phage growth (27).The nonA gene is located in the intergenic region between bnrdEand bnrdF (Fig. 1 and 2), and it encodes a small membrane proteinof 72 amino acids (Fig. 2 and 4). Complementation analysisshowed that the NonA protein confers a nonpermissive pheno-

TAY3221IPTG-

TAY3221IPTG+

TAY3227IPTG-

TAY3227IPTG+

TAY3221Heat

TAY3221IPTG-

TAY3221IPTG+

TAY3227IPTG-

TAY3227IPTG+

TAY3221Heat

DAPI PI Merge

DAPI CTC Merge

PI stained cell

3.5 ± 1.2 %

6.8 ± 2.9 %

5.6 ± 1.8 %

7.6 ± 4.1 %

94.2 ± 3.8 %

CTC stained cell

89.0 ± 4.3 %

2.6 ± 3.1 %

91.8 ± 2.0 %

88.9 ± 2.1 %

0.0 ± 0.0 %

A

B

FIG 9 Overexpression of NonA reduces respiration activity. Fluorescence microscopy images (magnification, 640) of TAY3221 and TAY3227 with or without1 mM IPTG for 1 h. The dead-cell control is TAY3221 incubated at 80°C for 30 min. Cells were stained with DAPI (blue) and PI (red) (A) or with DAPI (blue)and CTC (red) (B). Scale bar, 10 �m. Ratios (percent) of cells stained with PI or CTC were calculated from three independent experiments. Standard error valuesare included.

Abortive Infection in B. subtilis Marburg Strain

February 2014 Volume 196 Number 3 jb.asm.org 703

on October 16, 2017 by guest

http://jb.asm.org/

Dow

nloaded from

type upon SP10, whereas nonA RNA does not (Fig. 3). SP10 phagecan normally adsorb to ASK3000 that harbors the nonA, but notthe nonB gene, and inject its genome into the cells, and its genesare transcribed in the cells (Fig. 5). However, capsid protein ex-pression and phage particle formation were inhibited in these cells(Fig. 7). We therefore classified nonA as an abortive infection gene.

We previously suggested that an artificially induced RNAproduct corresponding to the region from �53 to �463 (relativeto the first nonA codon) might inhibit the expression of B. subtilisnrdE and SP10 xnrdE (31). Since this induced RNA lacks the nonA5= end region containing its RBS and start codon, the NonA pro-tein does not seem to be involved in the inhibitory effects of nrdEand xnrdE. The nucleotide sequences of the 3= ends of nrdE, bnrdE,and xnrdE are similar. On the other hand, the nucleotide se-quences and the lengths of the nrdEF, bnrdEF, and xnrdEF inter-genic regions distinctly differed. These findings suggested that the3= UTR of the nonA transcript, which overlaps the 3= end of thenrdE coding sequence, inhibits the expression of nrdE and xnrdEby base paring. However, we discovered here that the 3= UTR ofnonA was not required to abort SP10 infection (Fig. 3). Further-more, the amounts of the nrdE transcript during SP10 infectiondid not differ between nonA� (ASK3000) and nonA (ASK3002)strains (Fig. 1B, probe 1). These results suggest that inhibitingnrdE and xnrdE expression does not result in the abortion of SP10infection.

Some phages have several sigma factors that regulate phagegene expression (42, 43). SP10 phage has the sigma factors Orf183,Orf120, and Orf199-200 (26). Both Orf183 and Orf120 were ex-pressed from the early to middle stage, and Orf199-200 was ex-pressed only at the middle stage of SP10 infection (data notshown). These findings suggest that gene expression is regulatedby Orf183 and Orf120 at the middle stage and by Orf199-200 atthe late stage of SP10 infection. The consensus sequences of�Orf199-200-dependent promoters remain unknown. The putativepromoter sequences of the nonA gene were TTAAAA (35) andTATATT (10), and they were positioned at 18-bp intervals.These sequences distinctly differed from the consensus sequencesof B. subtilis or any sigma factor-dependent promoters. The arti-ficial expression of Orf199-200 induced the transcription of nonAin cells that were not infected with SP10 (Fig. 6A). These findingssuggest that �Orf199-200 regulates the transcription of nonA duringphage infection. Indeed, nonA was transcribed in cells at the latestage of SP10 infection but not in uninfected cells (Fig. 1 and 5).The nonA gene is located in the SP� prophage region of the B.subtilis genome. Nevertheless, its transcription is regulated by thesigma factor of SP10 but not by that of B. subtilis and SP�. Thefunctions of NonA for SP� are obscure. The present findings im-ply that SP� gains the nonA gene from the SP10 phage with itspromoter and that B. subtilis Marburg acquires the nonA-medi-ated ability to abort infection with virulent phages by lysogeniza-tion with SP�. The activation or expression of proteins that canabort infection depends on the products of phage genes. Lactococ-cus lactis AbiD1 protein aborts infection with Lactococcus phagebIL66. The translation of abiD1 is repressed under normal growthconditions, but the phage bIL66 Orf1 protein activates abiD1translation by binding to abiD1 mRNA during bIL66 infection(44). The phage T4 Gol and Stp proteins regulate the activity of E.coli Lit protease and PrrC anticodon nuclease, respectively, thatabort infection (14). Gol promotes the Lit-mediated hydrolysis ofelongation factor (EF)-Tu by binding to EF-Tu domains II and III

(45). The activity of PrrC is neutralized by EcoprrI; Stp altersinteraction between PrrC and EcoprrI, and activated PrrC is re-leased and aborts infection (14). In contrast to the genes that abortinfection, nonA was regulated at the transcriptional level. To thebest of our knowledge, this is the first indication of a virulentphage sigma factor regulating the expression of a gene in the hostgenome that can abort infection.

Each of the phages SPO1, �29, and SPP1 can grow in the wild-type B. subtilis Marburg strain. Neither �29 nor SPP1 has proteinshomologous to sigma factors, and nonA was not transcribed dur-ing infection with these phages (Fig. 6B), which is consistent withthe finding that SP10 sigma factor directs nonA transcription. Onthe other hand, SPO1 phage has the sigma factors Gp28, Gp2.21,and Gp33-Gp34 (Gp33-34) (46). A BLAST search revealed thatGp33 and Gp34 are putative homologs of the SP10 sigma factors,Orf199 and Orf200, that direct the transcription of nonA. Gp33shares 22.2% identity and 47.9% similarity with Orf199, and Gp24shares 35.5% identity and 54.8% similarity with Orf200. However,the nonA transcript was undetectable during infection with thisphage (Fig. 6B), indicating that �Gp33-34 does not regulate thetranscription of nonA. We did not find any other Orf199-200 ho-mologs in the NCBInr databases except for SPO1 Gp33-34. Ourfindings support the notion that Orf199-200 is required for nonAtranscription. Although Saito et al. reported that NonA aborts�NR2 infection, whether or not Orf199-200 homologs exist in the�NR2 genome remains unknown because this genome has not yetbeen sequenced (24). Because nonA is not transcribed during in-fection with SPO1, �29, and SPP1, whether or not NonA proteinaborts infection with these phages also remains unknown.

The amounts of early, middle, and late gene transcripts of theSP10 phage did not differ between nonA� and nonA strains (Fig.5). However, proteins such as SP10 capsid protein that is ex-pressed at the late stage of SP10 infection (26) did not accumulatein the nonA strain at this stage (Fig. 7A). Gene transcription at themiddle and late stages of infection is regulated by sigma factorsthat are expressed at early and middle stages, suggesting that earlyand middle genes are translated in nonA� cells. Thus, NonA mightinhibit SP10 gene expression at the level of translation during thelate stage. Some genes that abort phage infection by inhibitingtranslation have been identified. The E. coli Lit protein possesses azinc metalloproteinase domain, and it is activated by T4 Gol pro-tein at the late stage of infection (14). Activated Lit terminatesprotein synthesis by cleaving EF-Tu, which causes bacterial deathand aborts infection (47). The E. coli gene prrC encodes an anti-codon nuclease. PrrC cleaves the anti-codon loop of tRNALys,causing protein synthesis to stop. Lactococcus lactis AbiV proteininteracts with p2 SaV protein and inhibits p2 phage early proteinsynthesis, but details of the mechanisms remain unknown (48).NonA has no protease or nuclease domain, and its amino acidsequence is not similar to the sequences of the proteins that abortinfection described above. The artificial induction of NonA invegetative cells disturbed B. subtilis and E. coli cell growth. Fur-thermore, NonA-overexpressing cells of B. subtilis were stainedwith CTC but not with PI (Fig. 9), indicating that NonA overex-pression did not damage the cell membrane but, rather, reducedrespiration activity. We compared the amount of NonA betweenIPTG-induced and SP10-infected cells by Western blotting. Theamounts of NonA in TAY3210 (amyE::PnonA-nonA-His) cells in-fected with SP10 and in IPTG-induced TAY3221 (amyE::lacIPspac-nonA-His) cells were almost identical (data not shown).

Yamamoto et al.

704 jb.asm.org Journal of Bacteriology

on October 16, 2017 by guest

http://jb.asm.org/

Dow

nloaded from

This suggests that respiratory activity was reduced in SP10-in-fected and in IPTG-induced cells. We infer that the reduction ofrespiratory activity induced by NonA is related to abortive infec-tion of SP10. Based on these findings, we speculate that NonA doesnot directly inhibit SP10 gene translation but reduces the respira-tion activity of infected cells, which results in the inhibition oftranslation and in abortive infection.

NonA was not expressed in the B. subtilis Marburg strain be-cause the nonB restriction enzyme BsuMR cleaved SP10 phagegenomic DNA immediately after infection (Fig. 5, lanes 1 to 6).This means that NonA is not involved in the resistance of thisstrain to infection with SP10 phage. However, NonA was ex-pressed and inhibited SP10 phage proliferation in the Marburgstrain instead of BsuMR when these cells lost the nonB gene (Fig. 5,lanes 13 to 18). Therefore, NonA might be a backup system forSP10 phage resistance in the Marburg strain. Phages can easilyevolve to escape resistance by either acquiring genes from otherorganisms or mutating their own genes. However, even if SP10acquires the ability to evade BsuMR, the Marburg strain will re-main resistant to SP10 due to having a NonA-mediated systemthat aborts infection. Thus, NonA acts as a second layer of protec-tion against the SP10 phage in the Marburg strain.

ACKNOWLEDGMENTS

We gratefully thank Hideko Urushihara for continuous guidance, consid-erable encouragement, and valuable discussions during this study. Wethank Motoo Utsumi for helpful advice on the fluorescence microscopyanalysis. We also thank Norma Foster for a critical reading of the manu-script.

The MALDI-TOF MS analysis proceeded using a TOF/TOF 5800 sys-tem at the Chemical Analysis Division and Open Facility, Research Facil-ity Center for Science and Technology, University of Tsukuba.

This work was supported by the Grant-in-Aid for Scientific Research(C) (grant number 25450092 to K.N.) from the Japan Society for thePromotion of Science and by the Advanced Low Carbon Technology Re-search and Development Program to N.N. from the Japan Science andTechnology Agency.

REFERENCES1. Chibani-Chennoufi S, Bruttin A, Dillmann M, Brüssow H. 2004. Phage-

host interaction: an ecological perspective. J. Bacteriol. 186:3677–3686.http://dx.doi.org/10.1128/JB.186.12.3677-3686.2004.

2. Wommack KE, Colwell RR. 2000. Virioplankton: viruses in aquatic eco-systems. Microbiol. Mol. Biol. Rev. 64:69 –114. http://dx.doi.org/10.1128/MMBR.64.1.69-114.2000.

3. Lima-Mendez G, Toussaint A, Leplae R. 2007. Analysis of the phagesequence space: the benefit of structured information. Virology 365:241–249. http://dx.doi.org/10.1016/j.virol.2007.03.047.

4. Labrie SJ, Samson JE, Moineau S. 2010. Bacteriophage resistancemechanisms. Nat. Rev. Microbiol. 8:317–327. http://dx.doi.org/10.1038/nrmicro2315.

5. Stern A, Sorek R. 2011. The phage-host arms race: shaping the evolution ofmicrobes. Bioessays 33:43–51. http://dx.doi.org/10.1002/bies.201000071.

6. Bouchard JD, Dion E, Bissonnette F, Moineau S. 2002. Characterizationof the two-component abortive phage infection mechanism AbiT fromLactococcus lactis. J. Bacteriol. 184:6325– 6332. http://dx.doi.org/10.1128/JB.184.22.6325-6332.2002.

7. Domingues S, Chopin A, Ehrlich SD, Chopin M. 2004. The Lactococcalabortive phage infection system AbiP prevents both phage DNA replica-tion and temporal transcription switch. J. Bacteriol. 186:713–721. http://dx.doi.org/10.1128/JB.186.3.713-721.2004.

8. Emond E, Holler BJ, Boucher I, Vandenbergh PA, Vedamuthu ER,Kondo JK, Moineau S. 1997. Phenotypic and genetic characterization ofthe bacteriophage abortive infection mechanism AbiK from Lactococcuslactis. Appl. Environ. Microbiol. 63:1274 –1283.

9. Garvey P, Fitzgerald GF, Hill C. 1995. Cloning and DNA sequence

analysis of two abortive infection phage resistance determinants from thelactococcal plasmid pNP40. Appl. Environ. Microbiol. 61:4321– 4328.

10. Hill C, Miller L a, Klaenhammer TR. 1990. Nucleotide sequence anddistribution of the pTR2030 resistance determinant (hsp) which abortsbacteriophage infection in lactococci. Appl. Environ. Microbiol. 56:2255–2258.

11. Cluzel PJ, Chopin A, Ehrlich SD, Chopin MC. 1991. Phage abortiveinfection mechanism from Lactococcus lactis subsp. lactis, expression ofwhich is mediated by an Iso-ISS1 element. Appl. Environ. Microbiol. 57:3547–3551.

12. Dai G, Su P, Allison GE, Geller BL, Zhu P, Kim WS, Dunn NW. 2001.Molecular characterization of a new abortive infection system (AbiU)from Lactococcus lactis LL51-1. Appl. Environ. Microbiol. 67:5225–5232.http://dx.doi.org/10.1128/AEM.67.11.5225-5232.2001.

13. O’Connor L, Tangney M, Fitzgerald G. 1999. Expression, regulation, andmode of action of the AbiG abortive infection system of lactococcus lactissubsp. cremoris UC653. Appl. Environ. Microbiol. 65:330 –335.

14. Snyder L. 1995. Phage-exclusion enzymes: a bonanza of biochemical andcell biology reagents? Mol. Microbiol. 15:415– 420. http://dx.doi.org/10.1111/j.1365-2958.1995.tb02255.x.

15. Haaber J, Samson JE, Labrie SJ, Campanacci V, Cambillau C, MoineauS, Hammer K. 2010. Lactococcal abortive infection protein AbiV interactsdirectly with the phage protein SaV and prevents translation of phageproteins. Appl. Environ. Microbiol. 76:7085–7092. http://dx.doi.org/10.1128/AEM.00093-10.

16. Durmaz E, Klaenhammer TR. 2007. Abortive phage resistance mecha-nism AbiZ speeds the lysis clock to cause premature lysis of phage-infectedLactococcus lactis. J. Bacteriol. 189:1417–1425. http://dx.doi.org/10.1128/JB.00904-06.

17. Fineran PC, Blower TR, Foulds IJ, Humphreys DP, Lilley KS, SalmondGPC. 2009. The phage abortive infection system, ToxIN, functions as aprotein-RNA toxin-antitoxin pair. Proc. Natl. Acad. Sci. U. S. A. 106:894 –899. http://dx.doi.org/10.1073/pnas.0808832106.

18. Pecota DC, Wood TK. 1996. Exclusion of T4 phage by the hok/sok killerlocus from plasmid R1. J. Bacteriol. 178:2044 –2050.

19. Samson JE, Spinelli S, Cambillau C, Moineau S. 2013. Structure andactivity of AbiQ, a lactococcal endoribonuclease belonging to the type IIItoxin-antitoxin system. Mol. Microbiol. 87:756 –768. http://dx.doi.org/10.1111/mmi.12129.

20. Kai T, Ueno H, Yonesaki T. 1998. Involvement of other bacteriophage T4genes in the blockade of protein synthesis and mRNA destabilization by amutation of gene 61.5. Virology 248:148 –155. http://dx.doi.org/10.1006/viro.1998.9270.

21. Otsuka Y, Yonesaki T. 2005. A novel endoribonuclease, RNase LS, inEscherichia coli. Genetics 169:13–20. http://dx.doi.org/10.1534/genetics.104.033290.

22. Otsuka Y, Yonesaki T. 2012. Dmd of bacteriophage T4 functions as anantitoxin against Escherichia coli LsoA and RnlA toxins. Mol. Microbiol.83:669 – 681. http://dx.doi.org/10.1111/j.1365-2958.2012.07975.x.

23. Conn HJ. 1930. The identity of Bacillus subtilis. J. Infect. Dis. 46:341–350.http://dx.doi.org/10.1093/infdis/46.4.341.

24. Saito H, Shibata T, Ando T. 1979. Mapping of genes determining non-permissiveness and host-specific restriction to bacteriophages in Bacillussubtilis Marburg. Mol. Gen. Genet. 170:117–122. http://dx.doi.org/10.1007/BF00337785.

25. Hemphill HE, Whiteley HR. 1975. Bacteriophages of Bacillus subtilis.Bacteriol. Rev. 39:257–315.

26. Yee LM, Matsumoto T, Yano K, Matsuoka S, Sadaie Y, Yoshikawa H,Asai K. 2011. The genome of Bacillus subtilis phage SP10: a comparativeanalysis with phage SPO1. Biosci. Biotechnol. Biochem. 75:944 –952. http://dx.doi.org/10.1271/bbb.100921.

27. Matsuoka S, Arai T, Murayama R, Kawamura F, Asai K, Sadaie Y. 2004.Identification of the nonA and nonB loci of Bacillus subtilis Marburg per-mitting the growth of SP10 phage. Genes Genet. Syst. 79:311–317. http://dx.doi.org/10.1266/ggs.79.311.

28. Jentsch S. 1983. Restriction and modification in Bacillus subtilis: sequencespecificities of restriction/modification systems BsuM, BsuE, and BsuF. J.Bacteriol. 156:800 – 808.

29. Bron S, Jannière L, Ehrlich SD. 1988. Restriction and modificationin Bacillus subtilis Marburg 168: target sites and effects on plasmid trans-formation. Mol. Gen. Genet. 211:186 –189. http://dx.doi.org/10.1007/BF00338412.

30. Matsuoka S, Asai K, Sadaie Y. 2005. Restriction and modification of SP10

Abortive Infection in B. subtilis Marburg Strain

February 2014 Volume 196 Number 3 jb.asm.org 705

on October 16, 2017 by guest

http://jb.asm.org/

Dow

nloaded from

phage by BsuM of Bacillus subtilis Marburg. FEMS Microbiol. Lett. 244:335–339. http://dx.doi.org/10.1016/j.femsle.2005.02.006.

31. Yee LM, Matsuoka S, Yano K, Sadaie Y, Asai K. 2011. Inhibitory effectof prophage SP� fragments on phage SP10 ribonucleotide reductase func-tion and its multiplication in Bacillus subtilis. Genes Genet. Syst. 86:7–18.http://dx.doi.org/10.1266/ggs.86.7.

32. Härtig E, Hartmann A, Schätzle M, Albertini AM, Jahn D. 2006. TheBacillus subtilis nrdEF genes, encoding a class Ib ribonucleotide reductase,are essential for aerobic and anaerobic growth. Appl. Environ. Microbiol.72:5260 –5265. http://dx.doi.org/10.1128/AEM.00599-06.

33. Morimoto T, Loh PC, Hirai T, Asai K, Kobayashi K, Moriya S,Ogasawara N. 2002. Six GTP-binding proteins of the Era/Obg family areessential for cell growth in Bacillus subtilis. Microbiology 148:3539 –3552.

34. Morimoto T, Ara K, Ozaki K, Ogasawara N. 2009. A new simple methodto introduce marker-free deletions in the Bacillus subtilis genome. GenesGenet. Syst. 84:315–318. http://dx.doi.org/10.1266/ggs.84.315.

35. Obana N, Shirahama Y, Abe K, Nakamura K. 2010. Stabilization ofClostridium perfringens collagenase mRNA by VR-RNA-dependent cleav-age in 5= leader sequence. Mol. Microbiol. 77:1416 –1428. http://dx.doi.org/10.1111/j.1365-2958.2010.07258.x.

36. Roche Diagnostics GmbH. 2008. DIG application manual for filter hy-bridization. Roche Diagnostics GmbH, Mannheim, Germany.

37. Strohalm M, Kavan D, Novák P, Volný M, Havlícek V. 2010. mMass3: a cross-platform software environment for precise analysis of massspectrometric data. Anal. Chem. 82:4648 – 4651. http://dx.doi.org/10.1021/ac100818g.

38. Quevillon E, Silventoinen V, Pillai S, Harte N, Mulder N, Apweiler R,Lopez R. 2005. InterProScan: protein domains identifier. Nucleic AcidsRes. 33:W116 –W120. http://dx.doi.org/10.1093/nar/gki442.

39. Ohshima H, Matsuoka S, Asai K, Sadaie Y. 2002. Molecular organizationof intrinsic restriction and modification genes BsuM of Bacillus subtilis

Marburg. J. Bacteriol. 184:381–389. http://dx.doi.org/10.1128/JB.184.2.381-389.2002.

40. Tijan R, Pero J. 1976. Bacteriophage SP01 regulatory proteins directinglate gene transcription in vitro. Nature 262:753–757. http://dx.doi.org/10.1038/262753a0.

41. Costanzo M, Brzustowicz L, Hannett N, Pero J. 1984. BacteriophageSPO1 genes 33 and 34. Location and primary structure of genes encodingregulatory subunits of Bacillus subtilis RNA polymerase. J. Mol. Biol. 180:533–547.

42. Geiduschek EP, Kassavetis GA. 2010. Transcription of the T4 late genes.Virol. J. 7:288. http://dx.doi.org/10.1186/1743-422X-7-288.

43. Reznikoff WS, Siegele DA, Cowing DW, Gross CA. 1985. The regulationof transcription initiation in bacteria. Annu. Rev. Genet. 19:355–387. http://dx.doi.org/10.1146/annurev.ge.19.120185.002035.

44. Bidnenko E, Chopin A, Ehrlich SD, Chopin M-C. 2009. Activation ofmRNA translation by phage protein and low temperature: the case ofLactococcus lactis abortive infection system AbiD1. BMC Mol. Biol. 10:4.http://dx.doi.org/10.1186/1471-2199-10-4.

45. Copeland NA, Kleanthous C. 2005. The role of an activating peptide inprotease-mediated suicide of Escherichia coli K12. J. Biol. Chem. 280:112–117. http://dx.doi.org/10.1074/jbc.M411280200.

46. Stewart CR, Casjens SR, Cresawn SG, Houtz JM, Smith AL, Ford ME,Peebles CL, Hatfull GF, Hendrix RW, Huang WM, Pedulla ML. 2009.The genome of Bacillus subtilis bacteriophage SPO1. J. Mol. Biol. 388:48 –70. http://dx.doi.org/10.1016/j.jmb.2009.03.009.

47. Yu YT, Snyder L. 1994. Translation elongation factor Tu cleaved by aphage-exclusion system. Proc. Natl. Acad. Sci. U. S. A. 91:802– 806. http://dx.doi.org/10.1073/pnas.91.2.802.

48. Haaber J, Rousseau GM, Hammer K, Moineau S. 2009. Identificationand characterization of the phage gene sav, involved in sensitivity to thelactococcal abortive infection mechanism AbiV. Appl. Environ. Micro-biol. 75:2484 –2494. http://dx.doi.org/10.1128/AEM.02093-08.

Yamamoto et al.

706 jb.asm.org Journal of Bacteriology

on October 16, 2017 by guest

http://jb.asm.org/

Dow

nloaded from