Archives of Virology

, Volume 159, Issue 5, pp 871–884

Characterization and comparative genomic analysis of bacteriophages infecting members of the Bacillus cereus group

Authors

  • Ju-Hoon Lee
    • Department of Food Science and BiotechnologyKyung Hee University
  • Hakdong Shin
    • Department of Food and Animal Biotechnology, Research Institute for Agriculture and Life Sciences, and Center for Food and BioconvergenceSeoul National University
    • Department of Agricultural Biotechnology, Research Institute for Agriculture and Life Sciences, and Center for Food and BioconvergenceSeoul National University
    • Department of Food and Animal Biotechnology, Research Institute for Agriculture and Life Sciences, and Center for Food and BioconvergenceSeoul National University
    • Department of Agricultural Biotechnology, Research Institute for Agriculture and Life Sciences, and Center for Food and BioconvergenceSeoul National University
Brief Review

DOI: 10.1007/s00705-013-1920-3

Cite this article as:
Lee, J., Shin, H. & Ryu, S. Arch Virol (2014) 159: 871. doi:10.1007/s00705-013-1920-3

Abstract

The Bacillus cereus group phages infecting B. cereus, B. anthracis, and B. thuringiensis (Bt) have been studied at the molecular level and, recently, at the genomic level to control the pathogens B. cereus and B. anthracis and to prevent phage contamination of the natural insect pesticide Bt. A comparative phylogenetic analysis has revealed three different major phage groups with different morphologies (Myoviridae for group I, Siphoviridae for group II, and Tectiviridae for group III), genome size (group I > group II > group III), and lifestyle (virulent for group I and temperate for group II and III). A subsequent phage genome comparison using a dot plot analysis showed that phages in each group are highly homologous, substantiating the grouping of B. cereus phages. Endolysin is a host lysis protein that contains two conserved domains: a cell-wall-binding domain (CBD) and an enzymatic activity domain (EAD). In B. cereus sensu lato phage group I, four different endolysin groups have been detected, according to combinations of two types of CBD and four types of EAD. Group I phages have two copies of tail lysins and one copy of endolysin, but the functions of the tail lysins are still unknown. In the B. cereus sensu lato phage group II, the B. anthracis phages have been studied and applied for typing and rapid detection of pathogenic host strains. In the B. cereus sensu lato phage group III, the B. thuringiensis phages Bam35 and GIL01 have been studied to understand phage entry and lytic switch regulation mechanisms. In this review, we suggest that further study of the B. cereus group phages would be useful for various phage applications, such as biocontrol, typing, and rapid detection of the pathogens B. cereus and B. anthracis and for the prevention of phage contamination of the natural insect pesticide Bt.

Introduction

Since their discovery in 1915 [68], bacteriophages have been known to be viruses of prokaryotes that invade specific bacterial hosts, replicate using the host DNA replication and protein biosynthesis systems, and lyse the host for propagation [34, 43, 63]. The lifestyles of bacteriophages include a lysogenic cycle (for phage genome integration into the host chromosome) and a lytic cycle (for lysis the bacterial host because of bactericidal activity) [25, 34, 43]. Bacteriophages occur everywhere in the biosphere and are frequently found in the ocean and soil. The total biomass of phages and number of phage species in nature have been estimated at over 1030 particles and more than 106 species, respectively [10, 51]. Bacteriophages are classified by the International Committee on Taxonomy of Viruses (ICTV) according to phage morphology and nucleic acid type [30]. Approximately 96 % of all bacteriophages belong to the order Caudovirales (the tailed phages), and the rest either belong to the order Ligamenvirales (the linear bacterial viruses) or have not been assigned to an order (viruses with no formal head-tail structure), as is the case for members of the family Tectiviridae [3]. The bacteriophages in the order Caudovirales belong to three different families: Myoviridae, whose members have a contractile tail, Siphoviridae, whose members have a non-contractile tail, and Podoviridae, whose members have a short and non-contractile tail [30, 34].

The Bacillus cereus group consists of B. cereus, B. anthracis, B. thuringiensis, B. mycoides, B. pseudomycoides, and B. weihenstephanensis [27, 72]. Of these, B. cereus, B. anthracis, and B. thuringiensis have been suggested to be a single species of B. cereus sensu lato [14, 24]. This classified species is a pathogen that infects humans, animals, and even insects. B. cereus is a well-known food-borne pathogen that produces enterotoxins that causes diarrhea, vomiting, and nausea [9, 21]. In addition, B. anthracis is a category A bio-threat agent, causing fatal anthrax in humans and animals [22]. In 2001, B. anthracis endospores were used as a biological weapon, resulting in a mortality rate higher than 45 % in those exposed [7]. Generally, to control bacterial pathogens, various antibiotics have been widely used. However, B. cereus is generally insusceptible to penicillin-related antibiotics (because it produces β-lactamase) and also to other antibiotics, such as erythromycin and tetracycline [31, 53]. In addition, long-term antibiotic treatment with various antibiotics against B. anthracis results in rapid development of antibiotic resistance [5]. Therefore, because of the emergence of antibiotic-resistant strains, alternative biocontrol approaches against these pathogens are needed. Employing bacteriophages to infect B. cereus or B. anthracis could be a good strategy for controlling these pathogens. An additional member of the Bacillus cereus group, B. thuringiensis (Bt), has been thoroughly characterized and has been used as a biological pesticide for the biocontrol of insect pests. It produces insecticidal crystal proteins (ICPs), which are highly toxic to the pest larvae, but not to humans and other animals [4, 54]. Whereas Bt has been widely used for insect pest control, bacteriophage contamination causes damage to Bt production via fermentation. To overcome this problem, bacteriophages infecting B. thuringiensis should be studied to understand their infection mechanism and the mechanism by which they switch from the lytic to the lysogenic state.

Recent developments in genome sequencing and bioinformatic technologies have allowed a large number of bacteriophage studies to be carried out at the genomic level, providing further opportunities for applications in novel biocontrol agents and phage therapy. In this review, we describe the general features of the B. cereus group phages and genomic insights resulting from comparative and functional genomic analysis. This genomic information is useful for extending our understanding of the general genomic characteristics of these phages and their various applications in the control of bacterial pathogens and in phage therapy.

The general genomic features and classification of B. cereus group bacteriophages based on comparative genomics

To date, 30 complete genome sequences of B. cereus group bacteriophages (18 for B. cereus phages, four for B. anthracis, and eight for B. thuringiensis) are available in the GenBank database. The general features of all 30 complete genome sequences are listed in Table 1. Members of three different morphological families are present among the 30 B. cereus group bacteriophages, including Myoviridae, Siphoviridae (order Caudovirales), and Tectiviridae (not assigned to an order). Interestingly, the phage genome size may be related to the family morphology of the B. cereus group bacteriophages. The ranges of phage genome sizes in the families Myoviridae, Siphoviridae, and Tectiviridae are 94 to 219 kb, 36 to 53 kb, and 14.3 to 14.9 kb, respectively (Table 1). Furthermore, the phage life cycle may also be related to the family morphology of the B. cereus group bacteriophages. Whereas all Myoviridae family bacteriophages are virulent, Siphoviridae and Tectiviridae family bacteriophages are temperate, supporting this hypothesis. Therefore, a close relationship may exist between the family morphology, genome size, and life cycle of the B. cereus group phages. Based on these results, there may be members of three potential major phage groups among the 30 B. cereus group bacteriophages. To further classify these 30 different B. cereus group bacteriophages, comparative phylogenetic analysis using the phage major capsid proteins (MCPs) and phage terminase large subunits were conducted, and this revealed that there are three major evolutionary groups: phage groups I, II and III (Fig. 1A and B). A subsequent comparative dot plot analysis of all 30 bacteriophage genomes at the DNA level also showed that the phage genomes in each phage group are similar, consistent with the phage classification (Fig. 2). In addition, tetranucleotide frequency comparisons of the 30 B. cereus group phage genomes showed that the phage genomes in each major phage group are closely related, supporting this phage grouping [48, 66, 73] (Fig. S1). As suggested recently by Lavigne et al. [36], the family Myoviridae is divided into the three subfamilies: Peduovirinae, Tevenvirinae, and Spounavirinae. Based on the phylogenetic tree in that report, phage group I in the family Myoviridae corresponds to the subfamily Spounavirinae. The subfamily Spounavirinae includes two genera: Spounalikevirus and Twortlikevirus. However, the members of phage group I do not belong to these two genera but most likely belong to another new genus, suggesting that the subfamily Spounavirinae may be more diverse [17].
Table 1

General genomic features of the B. cereus group bacteriophages

Phage

Host

Genome size (bp)

GC%

Predicted ORFs

No. Hypo. (%)a

Coding (%)

No. tRNAs

Morphology

Life style

Accession number

Group

Reference

B4

B. cereus

162,596

37.71

277

228 (82.3)

90.6

0

Myoviridae

Virulent

NC_018863

I

[38]

B5S

B. cereus

162,598

37.71

272

230 (84.6)

90.5

0

Myoviridae

Virulent

JN797796

I

This study

Bastille

B. cereus

153,962

38.14

273

231 (84.6)

92.6

7

Myoviridae

Virulent

NC_018856

I

[42]

W.Ph.

B. cereus

156,897

36.45

274

250 (91.2)

92.4

0

Myoviridae

Virulente

NC_016563

I

-

BPS10C

B. cereus

159,590

38.74

271

232 (85.6)

91.6

0

Myoviridae

Virulent

JN654439

I

-

BPS13

B. cereus

158,305

38.75

268

231 (86.2)

89.4

0

Myoviridae

Virulent

NC_018857

I

-

vB_BceM_Bc431v3

B. cereus

158,621

39.98

238

165 (69.3)

90.6

21

Myoviridae

Virulent

JX094431

I

[17]

BCP78

B. cereus

156,176

39.86

227

181 (79.7)

90.0

18

Myoviridae

Virulent

NC_018860

I

[37]

BCU4

B. cereus

154,371

39.86

223

171 (76.7)

89.9

19

Myoviridae

Virulent

JN797798

I

This study

BMBtp2

B. thuringiensis

36,932

37.79

53

39 (73.6)

86.9

0

Siphoviridae

Temperate

NC_019912

 

[16]

TP21-L

B. cereus

37,456

37.8

56

N/D

89.3

0

Siphoviridae

Temperatee

NC_011645

 

[42]

IEBH

B. cereus

53,104

36.42

86

60 (69.8)

85.8

0

Siphoviridae

Temperate

NC_011167

 

[59]

phBC6A51

B. cereus

61,395

37.69

75

56 (74.7)

83.0

0

N/D

Temperatee

NC_004820

 

[26]

BCD7

B. cereus

93,839

38.04

140

107 (76.4)

90.2

0

Myoviridae

Virulent

NC_019515

 

-

PBC1

B. cereus

41,164

41.68

50

29 (58.0)

92.0

0

Siphoviridae

Virulent

NC_017976

 

[32]

BceA1

B. cereus

42,932

35.66

63

22 (34.9)

88.1

0

Siphoviridae

Temperate

HE614282

 

[65]

MZTP02

B. thuringiensis

15,717

37.55

20

11 (55.0)

78.8

0

N/D

Temperate

AY894696

 

[41]

AP50

B. anthracis

14,398

38.65

31

14 (45.2)

96.2

0

Tectiviridae

Temperate

NC_011523

III

[61]

GIL16c

B. thuringiensis

14,844

40.07

31

26 (83.9)

98.8

0

Tectiviridae

Virulentb

NC_006945

III

[71]

Bam35c

B. thuringiensis

14,935

39.72

32

30 (93.8)

99.7

0

Tectiviridae

Virulentc

NC_005258

III

[62]

GIL01

B. thuringiensis

14,931

39.73

30

24(80.0)

96.7

0

Tectiviridae

Temperate

AJ536073

III

[69]

0305phi8-36

B. thuringiensis

218,948

41.8

246

142 (57.7)

95

0

Myoviridae

Virulent

NC_009760

 

[67]

phBC6A52

B. cereus

38,472

34.72

49

31 (63.3)

80.4

0

N/D

Temperatee

NC_004821

 

[26]

11143

B. cereus

39,077

34.96

49

23 (46.9)

85.3

0

Siphoviridae

Temperate

GU233956

 

[40]

BtCS33

B. thuringiensis

41,992

35.22

57

29 (50.9)

85.1

0

Siphoviridae

Temperate

NC_018085

II

[74]

phIS3501

B. thuringiensis

44,401

34.86

53

25 (47.2)

75.7

1

Siphoviridae

Temperate

NC_019502

II

[46]

Cherry

B. anthracis

36,615

35.26

51

29 (56.9)

91.5

0

Siphoviridae

Virulentd

NC_007457

II

[19]

Fah

B. anthracis

37,974

34.94

50

18 (36.0)

89.6

0

Siphoviridae

Virulentd

NC_007814

II

[45]

Gamma

B. anthracis

37,253

35.22

53

30 (56.6)

90.7

0

Siphoviridae

Virulentd

NC_007458

II

[19, 57]

Wbeta

B. cereus

40,867

35.26

53

27 (50.9)

90.2

0

Siphoviridae

Temperate

NC_007734

II

[57]

aThe number of hypothetical proteins

bClear plaque mutant of the phage Bam35

cClear plaque mutant of GIL16

dLytic variants of the phage Wbeta

ePhage Classification Tool Set (PHACTS) prediction result [44]

https://static-content.springer.com/image/art%3A10.1007%2Fs00705-013-1920-3/MediaObjects/705_2013_1920_Fig1_HTML.gif
Fig. 1

Comparative phylogenetic analysis of the major capsid proteins (A) and the terminase large subunits (B) using the MEGA5 [33] and ClustalW [35] programs by the neighbor-joining method. The numbers associated with the branches represent bootstrap values

https://static-content.springer.com/image/art%3A10.1007%2Fs00705-013-1920-3/MediaObjects/705_2013_1920_Fig2_HTML.gif
Fig. 2

A comparative dot plot analysis of all 30 bacteriophage genomes using the JDotter program [11] with a maximum plot size for 700. 1, B4; 2, B5S; 3, Bastille; 4, BPS10C; 5, BPS13; 6, W.Ph.; 7, vB_BceM_Bc431v3; 8, BCP78; 9, BCU4; 10. BceA1; 11, MZTP02; 12, BtCS33; 13, phIS3501; 14, Cherry; 15, Wbeta; 16, Fah; 17, Gamma; 18, 0305phi8-36; 19, IEBH; 20, BMBtp2; 21, TP21-L; 22, PBC1; 23, phBC6A52; 24, BCD7; 25, phBC6A51; 26, 11143; 27, AP50; 28, GIL16c; 29, Bam35c; 30, pGIL01

Although 30 complete genome sequences of the B. cereus group bacteriophages are available, 34.9 to 93.5 % of the annotated open reading frames (ORFs) are annotated as hypothetical proteins, most likely because of insufficient phage gene annotation data in the sequence databases (Table 1). Therefore, a genomic study of B. cereus group bacteriophages is required to extend our understanding to allow further applications in the development of biocontrol agents and phage therapy.

B. cereus sensu lato phage group I

According to the phage classification system, B. cereus sensu lato phage group I includes all bacteriophages belonging to the family Myoviridae except for BCD7 and 0305phi8-36. While these two distinct bacteriophages belong to the family Myoviridae, their genome sizes are different from the phages in group I. However, all phages in the family Myoviridae are virulent and predominantly infect B. cereus. A comparative genomic analysis of the phages in group I showed that functional gene clusters are located at identical positions, such as those for packaging, host lysis, DNA manipulation, phage structure, and additional functions, suggesting that the genome arrangement is identical in all members of phage group I (Fig. 3). In addition, a dot plot analysis also supports this similarity in the genome arrangements (Fig. 2). To further understand their functional characteristics, predicted functional genes were categorized into seven functional groups, including packaging, host lysis, regulation, host interaction, DNA manipulation, phage structure, and additional functions (Table 2). All predicted functional genes are shared by all bacteriophages in phage group I, except for the genes encoding thioredoxin, flavodoxin, and a transcriptional regulator in the phage Bastille (Table 2). In addition, the putative ribose-phosphate pyrophosphokinase was also missing in phage W.Ph. (Table 2). A comparative protein domain analysis of endolysins in the phage group I revealed the existence of four different homologous endolysin groups: endolysin group I, containing N-acetylmuramoyl-L-alanine amidase and an SH3-like domain (phages W.Ph., BPS10C, and BPS13); endolysin group II, containing cell wall hydrolysis/autolysin and an SH3-like domain (phages Bc431v3, BCP78, and BCU4); endolysin group III, containing peptidase M15B/M15C and an SH3-like domain (phages B4 and B5S); and endolysin group IV, containing glycoside hydrolase family 25 and N-acetylmuramoyl-L-alanine amidase (phage Bastille) (Fig. 4). All endolysin groups, except for group IV, share the SH3-like domain (PF08460), which has been predicted to be a cell-wall-binding domain. However, the phage Bastille maintains a different type of N-acetylmuramoyl-L-alanine amidase (PF12123) from other phages in endolysin group I (PF01510). Furthermore, the endolysin of the phage Bastille does not have a SH3-like cell-wall-binding domain, indicating that the phage Bastille may have a different type of cell-wall-binding domain in the endolysin. Loessner et al. suggested that the C-terminus of endolysin in the phage Bastille has 77-amino-acid repeat sequences (located at amino acids 211-287 and 288-364) and may be involved in cell wall binding, like the SH3-like domain [42]. Endolysins in endolysin group I have an N-acetylmuramoyl-L-alanine amidase domain, which is probably involved in cell wall lysis. To characterize this amidase domain (PF01510), Park et al. purified the endolysin of the phage BPS13 and tested the cell wall lysis mechanism [49]. This endolysin cleaves the bond between N-acetylmuramic acid and L-alanine in the cell wall through its amidase activity. In addition, Son et al. purified endolysin from the phage B4 and tested its cell wall lysis activity to characterize the peptidase M15B/M15C domain (PF02557), revealing that this peptidase domain acts like a L-alanoyl-D-glutamate endopeptidase by cleaving the peptide bond between L-alanine and D-glutamate [60]. Therefore, these two endolysin domains, N-acetylmuramoyl-L-alanine amidase and peptidase M15B/M15C, performs host cell wall lysis. However, the other two endolysin domains, the cell wall hydrolase/autolysin (PF01520) and glycoside hydrolase family 25 (PF01183) domains, are predicted to perform host cell lysis, but their cell wall lysis mechanisms should be confirmed experimentally (Fig. 4).
https://static-content.springer.com/image/art%3A10.1007%2Fs00705-013-1920-3/MediaObjects/705_2013_1920_Fig3_HTML.gif
Fig. 3

Comparative genomic analysis of phages in the B. cereus sensu lato phage group I at the DNA level using the Easyfig program [64]

Table 2

Core gene analysis of bacteriophage group I

Functional group

Predicted function

W.Ph.

BPS10C/BPS13

Bastille

B4/B5S

Bc431v3

BCP78/BCU4

Packaging

Terminase large subunit

P

P

P

P

P

P

 

Portal protein

P

P

P

P

P

P

Lysis

Endolysin*

I

I

II

II

III

VI

 

Putative holin

P

P

P

P

P

P

Regulation

DNA-binding protein

P

P

P

P

P

P

 

Transcriptional regulator 1

P

P

P

P

P

P

 

Transcriptional regulator 2

P

P

-

P

P

P

Host interaction

Sporulation sigma factor SigF-like protein

P

P

P

P

P

P

 

Putative RNA polymerase sigma factor

P

P

P

P

P

P

DNA manipulation

DNA polymerase 1

P

P

P

P

P

P

 

DNA recombination/repair protein

P

P

P

P

P

P

 

DNA polymerase 2

P

P

P

P

P

P

 

DNA primase

P

P

P

P

P

P

 

Exonuclease

P

P

P

P

P

P

 

Helicase 1

P

P

P

P

P

P

 

Helicase 2

P

P

P

P

P

P

Structure

Adsorption associated tail protein/ tail fiber

P

P

P

P

P

P

 

Baseplate J protein

P

P

P

P

P

P

 

Baseplate protein

P

P

P

P

P

P

 

Minor structural protein

P

P

P

P

P

P

 

Minor structural protein/ putative tail fiber

P

P

P

P

P

P

 

Tail lysin 1

P

P

P

P

P

P

 

Tail lysin 2

P

P

P

P

P

P

 

Tail sheath protein

P

P

P

P

P

P

 

Major capsid protein

P

P

P

P

P

P

 

Prohead protease

P

P

P

P

P

P

Additional function

3D domain-containing protein

P

P

P

P

P

P

 

Thymidylate synthase

P

P

P

P

P

P

 

Dephospho-CoA kinase

P

P

P

P

P

P

 

Dihydrofolate reductase

P

P

P

P

P

P

 

Metal-dependent hydrolase

P

P

P

P

P

P

 

Thioredoxin

P

P

-

P

P

P

 

Flavodoxin

P

P

-

P

P

P

 

Putative ribose-phosphate pyrophosphokinase

-

P

P

P

P

P

 

Ribonucleotide-diphosphate reductase subunit beta

P

P

P

P

P

P

 

Ribonucleotide-diphosphate reductase subunit alpha

P

P

P

P

P

P

 

Deoxyuridine 5’-triphosphate nucleotidohydrolase

P

P

P

P

P

P

P, present; -, not available

* B. cereus sensu lato phage group I have different types of endolysins

https://static-content.springer.com/image/art%3A10.1007%2Fs00705-013-1920-3/MediaObjects/705_2013_1920_Fig4_HTML.gif
Fig. 4

A comparative functional domain analysis of endolysins in the B. cereus sensu lato phage group I phage using the InterProScan program [76]

Two B. cereus phages, BPS10C and BPS13, were characterized and compared. The host range analysis for these two phages showed that they completely inhibit the growth of B. cereus group strains over 6 h (H. Shin, J.-H. Lee, J. Park, and S. Ryu, unpublished). These two phages are stable over a wide range of pH (5 to 8) and temperature (<50 °C). While these two phages showed very similar characteristics, phage BPS13 has a higher pH and temperature stability range than phage BPS10C. As discussed above, the endolysin of phage BPS13, LysBPS13, belongs to endolysin phage group I. In the presence of glycerol, lytic activity of LysBPS13 was displayed at temperatures up to 100 °C, suggesting high temperature stability [49]. A comparative genome sequence analysis of these two phages revealed that they have an endolysin and two copies of tail lysins. This gene set encoding an endolysin and two tail lysins is present in all nine phage genomes of the B. cereus sensu lato phage group I (Table 2). Although a few endolysins were characterized previously, the characteristics of these two tail lysins are still unknown. Therefore, further characterization of these tail lysins would extend our understanding of phage infection and host lysis mechanisms by members of the B. cereus sensu lato phage group I. The host range of phage Bc431v3 showed a relatively broad growth inhibition spectrum, including B. cereus group strains and even B. licheniformis, B. psychrosaccharolyticus, and B. megaterium, but not B. subtilis [17]. Bc431v3 has a long latent period (85 min) with a large burst size (>300 PFU). A comparative genome analysis of the phages BCP78 and BCU4 revealed that their genomes are nearly identical (>96 % DNA sequence identity) (Fig. 3). The endolysin group II phages (Bc431v3, BCP78, and BCU4) and endolysin group IV (Bastille) have different numbers of tRNAs in their genomes (Table 2). Comparative codon usage analysis of B. cereus strains (AH187 and ATCC 14597) and phage BCP78 showed different codon preferences for asparagine, phenylalanine, and serine, suggesting that extra tRNAs in the phage genome may play a role in translation of phage genes, but not of host genes [37]. Experimental characterization and a complete genome sequence analysis of the B. cereus phage B4 showed that it has a short eclipse/latent period (10/15 min) and a large burst size (>200 PFU) [38]. Moreover, a bacterial challenge assay showed complete growth inhibition of B. cereus up to 20 h, and its endolysin, LysB4, also showed efficient host cell lysis in 15 min [60], suggesting that phage B4 could be a good candidate as a novel biocontrol agent against B. cereus. Furthermore, a host range analysis of phage B4 revealed that it can inhibit the growth of B. cereus group strains and B. subtilis (Shin et al. 2013. Arch. Virol.). The genome sequence of the B. cereus phage B5S is 99 % identical to that of the phage B4. Further experimental characterization of phage B5S also showed it has almost identical host cell lysis activity. In addition, characterization of the endolysin of phage Bastille, PlyBa, showed efficient lysis of B. cereus and B. thuringiensis [42].

The B. cereus sensu lato phage group I contains all virulent phages with high host lysis activity against B. cereus. Therefore, the phages in this group may be good candidates for various applications as biocontrol agents against B. cereus. However, most of the phages (except for the phage Bastille) in this group were isolated only recently, and their genomic annotation information is not sufficient. Further research on the phages in this group is needed to increase our understanding and to develop applications for inhibiting the growth of B. cereus.

B. cereus sensu lato phage group II

According to the phage classification, B. cereus sensu lato phage group II includes the phages BtCS33, phIS3501, and Wbeta (Wβ) with virulent variants of Cherry, Fah, and Gamma (Wγ) in the family Siphoviridae. Their genome sizes (36 to 53 kb) are smaller than those of the B. cereus sensu lato phage group I, but larger than those of the B. cereus sensu lato phage group III. The phages BtCS33 and phIS3501 were isolated from induced strains of B. thuringiensis, and phage Wbeta was isolated from induced strains of B. cereus [46, 74].

The phage Wbeta and its variants inhibit the growth of B. anthracis (Table 1). The phages in this group have been widely used for phage typing. Furthermore, the phage Wbeta was used to construct a reporter phage containing luxAB for rapid detection of B. anthracis within 1 h after infection [55]. Phage Gamma is a variant of phage Wbeta, and it can infect the B. cereus W strain with phage Wbeta in the genome as a prophage, whereas phage Wbeta cannot re-infect B. cereus W [12, 57]. Therefore, phage Gamma was obtained by re-infection of the host strain, B. cereus W, with phage Wbeta lysate. Whereas phage Wbeta cannot infect capsulated B. anthracis, phage Gamma infects both encapsulated and non-encapsulated B. anthracis, indicating that phage Gamma can infect a broad range of B. anthracis strains [1]. Because of this broad host range, phage Gamma has been widely used for identification and indirect detection of B. anthracis [1, 50]. In addition, phage Gamma has been used to inhibit the growth of B. anthracis. Endolysin of the phage Gamma, PlyG was characterized, showing that it has two conserved protein domains: an N-terminal T7 lysozyme-like amidase domain for host cell lysis and a C-terminal cell-wall-binding domain [28, 29, 52, 56]. Notably, deletion of the binding domain abolished the host cell lysis activity of PlyG, suggesting that this binding domain may be important for specific host cell lysis by the N-terminal catalytic domain [29]. To further understand this binding domain, several motifs of the binding domain were chemically synthesized, and their binding activities were investigated [52]. This study revealed that a short 10-amino-acid sequence (LKMTADFILQ) is a key motif for host cell wall binding. This short sequence was coupled with Qdot nanocrystals and used for rapid detection of B. anthracis [52]. A study of the host receptor of phage Gamma showed that the host receptor of B. anthracis for phage Gamma is a surface-anchored protein containing an LPXTG motif, designated GamR [15]. However, the genome sequence of B. cereus ATCC 14579 does not have the host receptor protein GamR, and this explains its narrow host specificity [15]. Phage Cherry has also been used for typing of B. anthracis, but this phage is almost identical to phage Gamma in many aspects, such as phenotype, morphology and genome sequence [19]. In addition, genome sequencing and gene expression studies of phage Fah have shown the presence of unique viral promoters and a unique sigma factor that is most likely involved in host transcription events [45].

Phage BtCS33 was obtained by the induction of B. thuringiensis subsp. kurstaki CS33 [74]. This phage efficiently lyses B. thuringiensis because of its endolysin, PlyBt33, which contains a conserved protein domain of N-terminal glycoside hydrolase family 23 and a C-terminal amidase02_C domain. Notably, the C-terminal amindase02_C domain binds to the cell wall of B. thuringiensis and B. subtilis [75]. Phage phIS3501 was isolated from B. thuringiensis var. israelensis ATCC 35646 after induction [46]. The integration site of this phage is in hlyII, encoding hemolysin II. Therefore, this toxic hemolysin II may be activated by the induction of the prophage phIS3501.

The phages in B. cereus sensu lato phage group II generally infect B. anthracis and B. thuringiensis, and they have generally been used for typing of these species. The phage Gamma or its endolysin PlyG is a good candidate to control B. anthracis, thereby protecting against fatal anthrax disease. Therefore, these B. anthracis-infecting phages in this group should be studied further as a means to control the bio-threat posed by B. anthracis. In addition, the study of B. thuringiensis-infecting phages would help to protect the Bt production via fermentation.

B. cereus sensu lato phage group III

According to the phage classification, B. cereus sensu lato phage group III includes phages AP50, Bam35, GIL16C, and GIL10 in the family Tectiviridae. The family Tectiviridae has not yet been assigned to an order. Phages in the family Tectiviridae have no head-tail structure, but they instead possess tail-like tubes and viral membranes consisting of a lipid bilayer with inner and outer capsids. Their genome sizes (14.3 to 14.9 kb) are the smallest among all B. cereus phages in this report. In 1972, the phage AP50 was originally isolated from a soil sample with B. anthracis Sterne as a host strain [2, 47]. The host range analysis of this phage showed high host specificity for B. anthracis, indicating that it has a narrow host range. In addition, whereas phage AP50 makes turbid plaques, the variant AP50c produced clear plaques, and the CsaB cell-surface anchoring protein of B. anthracis may be involved in phage adsorption [8]. Phage Bam35 was initially isolated from B. thuringiensis [2]. Notably, this phage has a discrete entry mechanism. N-acetyl-muramic acid in the host membrane is essential for binding of phage Bam35 to the host, and the Bam35 virion has peptidoglycan hydrolysis activity. In addition, phage Bam35 requires the divalent cations Ca2+ and Mg2+ for penetration [20]. The temperate phage GIL01 was isolated from B. thuringiensis, and the phage genome was observed to be present in the host strain as a linear form of a tectiviral plasmid [69]. Notably, the induction of the lytic cycle of phage GIL01 has been reported to be associated with the host cellular SOS response to DNA damage [18]. Unlike endolysins of other phages, two ORFs, encoding Mur1 and Mur2, were expressed, and these proteins were experimentally confirmed to have peptidoglycan hydrolase activity during host lysis [70]. In addition, the genome sequence and characteristics of the phage GIL16 were similar to those of phage GIL01 [71].

The phages in the B. cereus sensu lato phage group III infect B. anthracis and B. thuringiensis. In this group, the B. anthracis-infecting phage AP50 may be useful for B. anthracis typing and biocontrol. For example, the application of the endolysin from AP50 may serve as a host-specific biocontrol. However, B. thuringiensis-infecting phages negatively affect Bt production via fermentation [41]. Therefore, studies of host-phage interactions, infection mechanisms, and lytic/lysogenic decision mechanisms may be important to provide extended information about the characteristics of B. thuringiensis-infecting phages (or prophages), ultimately describing a possible technique to prevent phage contamination of Bt.

Other B. cereus bacteriophages

As discussed above, we categorized the B. cereus group-infecting phages into three major phage groups. We discussed these three major groups in terms of their general and genomic features. However, eleven phages are not categorized into any of these groups when using comparative phylogenetic analysis, because the ungrouped phages are evolutionarily distant and quite different from the grouped phages (Fig. 1). Grouping of these ungrouped phages was difficult due to a lack of genome information about other similar phages in the databases. Nevertheless, two phages, BMBtp2 and TP21-L, are marginally related to each other evolutionarily (Table 1; Fig. 1, and Fig. 2), but it is not enough to make a new major phage group, because only two similar phages are available in the sequence databases to date. In addition, two other phages, BceA1 and MZTP02, are also highly similar to each other despite their different genome lengths. Interestingly, their genome sequence alignment revealed that the genome sequence of phage MZTP02 is more than 99 % identical to a region of the genome sequence of phage BceA1 at the DNA level (Fig. 2), suggesting that the genome of phage MZTP02 is a part of the genome of phage BceA1 [65]. However, it is still not enough to make a new major phage group, because only two similar phage genomes are available in the sequence databases. Seven other phage genomes are evolutionarily distant and differ significantly from each other, so their grouping may be inappropriate until more similar phage genomes are available in the sequence databases.

A few phages have their own distinct characteristics, such as the B. cereus-infecting phages BCD7 and PBC1 and the B. thuringiensis-infecting phage 0305phi8-36. Phage BCD7 has the smallest genome and phage 0305phi8-36 has the largest genome of the phages of the family Myoviridae that infect members of the B. cereus group (Table 1). Phage PBC1 belongs to the family Siphoviridae, and the other two phages, BCD7 and 0305phi8-36, belong to the family Myoviridae. The virulent phage PBC1 in the B. cereus group is the first Siphoviridae phage for which a complete genome sequence was reported. A genome analysis of phage PBC1 showed the absence of lysogeny-related genes, suggesting that it is a virulent phage [32]. Phage BCD7 was isolated from a soybean sample, and it displayed strong host lysis and growth inhibition activity in broth culture over 20 h (H. Shin, J.-H. Lee, and S. Ryu, unpublished). Its genome has two copies of host-cell-wall hydrolases with a holin. However, a comparative dot plot analysis of the phage BCD7 showed that its genome sequence is not homologous to those of the other B. cereus group–infecting phages (Fig. 2). In addition, the genome sequence of phage 0305phi8-36 also showed no homology to other phage genomes [23, 67], suggesting that phages BCD7 and 0305phi8-36 may have evolved from different ancestors than the other Myoviridae phages in the B. cereus sensu lato phage group I. The different genome sizes of these phages (94 kb for BCD7 and 219 kb for 0305phi8-36) support this hypothesis (Table 1).

Conclusion

The pathogens B. cereus and B. anthracis and the insect pathogen B. thuringiensis are designated as a single species of B. cereus sensu lato [14, 24]. B. cereus is a well-known food-borne pathogen, and B. anthracis causes anthrax [7, 9, 21, 22]. Therefore, control of these pathogens is important in the prevention of food poisoning and bio-threats. Antibiotics have been widely used to control them. However, penicillin-related antibiotics are ineffectual because of the production of β-lactamase by these bacteria [31, 53]. Therefore, an alternative phage-based approach has been suggested for their control. B. thuringiensis (Bt) has been widely used as a natural pesticide to kill harmful insect pests [4, 54]. However, phage contamination negatively affects the production of Bt via fermentation. Therefore, a Bt phage study is required to investigate a process to protect Bt from phage contamination.

Bacteriophages are bacterial viruses that specifically invade and kill the host bacteria [34, 43, 63]. Therefore, phages have been applied to control the pathogenic B. cereus and B. anthracis. To control these pathogens, two approaches have been studied: direct growth inhibition of these pathogens and host cell lysis using phage endolysins. As an example, two B. cereus phages, FWLBc1 and FWLBc2, were isolated from a soil sample and used to treat mashed potatoes, resulting in a more than 5-log reduction in the cell numbers of B. cereus. This suggests that applying phages to foods may be useful in the control of food-borne pathogens, including B. cereus [39]. In addition, the B. cereus phage BCP1-1 showed high host specificity and inhibited only B. cereus, and not other fermentative bacteria such as B. subtilis in Korean fermented soybean food, suggesting selective growth inhibition of only the target bacterium [6]. In investigating the lysis of the host by phages, several endolysins, such as LysB4, LysBPS13, Ply12, Ply21, PlyBa (for B. cereus), and PlyG (for B. anthracis), have been characterized [42, 49, 56, 60]. Generally, endolysins have the two conserved protein domains of a host-cell-wall-binding domain (CBD) and an enzymatic activity domain for peptidoglycan lysis (EAD). The CBD may provide host specificity, transfer of endolysin, and cell wall binding to the specific membrane. The EAD lyses the peptidoglycan in the host membrane, but the cleavage site of each endolysin depends on the type of EAD domain. To control B. anthracis, PlyG was characterized. This endolysin recognizes and cleaves a neutral polysaccharide (NPS) comprised of galactose (Gal), N-acetylglucosamine (GlcNAc), and N-acetylmannosamine (Man-NAc) [13, 58]. The application of PlyG (20 U) resulted in nearly complete elimination of B. anthracis in 15 min, suggesting that purified endolysin may be useful in the control of pathogens [56]. In addition to the phage applications, the B. anthracis phages have been used for the typing of this pathogen. As an example, the phage Gamma is a virulent phage that specifically infects and lyses B. anthracis with >95 % accuracy [1]. In addition, B. anthracis can be detected using real-time PCR with the phage Gamma in 5 h with a detection limit of 207 CFU/ml, suggesting that the Gamma phage is a good detection tool for B. anthracis [50]. Furthermore, the phage Wbeta was used to construct a reporter phage containing the luxAB genes. This reporter phage was able to produce light in the B. anthracis host in 1 h [55]. The detection limit of B. anthracis using this reporter phage system was 103 CFU/ml.

Because of recent developments in genomic technologies and bioinformatics, phage genomics is becoming more popular. Genomics provides further information about the physiology, genetics, and host-phage infection/interaction mechanisms of the phage. Therefore, 30 complete phage genome sequences of the B. cereus group in the families Myoviridae, Siphoviridae, and Tectiviridae are currently available (Table 1). The categorization of the phages into three groups revealed that morphology, genome size, and lifestyle may be associated (Table 1). Three B. cereus phage groups showed different genome size (group I > group II > group III), phage family association (Myoviridae for group I, Siphoviridae for group II, and Tectiviridae for group III), and lifestyle (virulent phenotype for group I and temperate phenotype for group II and III). Although 30 complete genome sequences for phages of the B. cereus group are available in the GenBank database, the functions of the proteins encoded by 34.9 to 93.8 % of the genes in the phage genomes are still unknown (Table 1). More than 69 % of the genes in the Myoviridae phage genomes are hypothetical, most likely because of insufficient annotation information for their genomes. They were recently isolated and their genome sequences were reported. The insufficient annotation of the genomes highlighted the shortage of available information in the GenBank database (Table 1). However, annotation information about core genes of the Myoviridae phages, which are generally involved in host infection/interaction and phage replication/reconstruction, is available, and the core genes are shared in all phages in group I, suggesting that they may have evolved from a common ancestor (Table 2). Notably, two phages, BCD7 and 0305phi8-36, belong to the family Myoviridae, but they are not in phage group I. A comparative phylogenetic and comparative dot plot analysis of phage genomes showed that the genome sequences of phages BCD7 and 0305phi8-36 are not homologous to those of other phages in the phage group I (Fig. 1 and 2). This finding suggested that their ancestors are different from one another and even different from the common ancestor of phage group I, even though they belong to the family Myoviridae and have virulent phenotypes (Table 1). The different genome sizes of phages BCD7 (94 kb) and 0305phi8-36 (219 kb) support this hypothesis (Table 1).

In this review, we compared B. cereus group phages at the genomic level. A comparative genomic analysis of these phages showed that the B. cereus group phages can be categorized into three different major groups with each group maintaining its own set of specific features. Whereas B. anthracis phages have been applied in the biocontrol, typing, and rapid detection of B. anthracis, the recently isolated and analyzed B. cereus phages were not well suited for the biocontrol of the food-borne pathogen B. cereus. Nevertheless, members of phage group I generally inhibited the growth of B. cereus, and all of them are virulent phages. This group of phages may be useful for the efficient biocontrol of B. cereus, either via infection with these phages or the application of purified endolysins. However, more than 36% of the B. cereus group phages are not assigned in this grouping, based on the comparative phylogenetic and dot plot analyses. When more phage genome sequences are available in the GenBank database, new phage groups could be generated from a further comparative phylogenetic analysis, and the phages that are not assigned to a group may belong to these new phage groups. As discussed previously, insufficient genome analysis has been conducted for the B. cereus group phages. Therefore, further genome sequencing and bioinformatic analysis should be performed to overcome the lack of genome annotation information, to extend our understanding of these phages, and to successfully apply them in the biocontrol of the pathogens B. cereus and B. anthracis.

Acknowledgements

This research was supported by the Public Welfare & Safety research program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT and Future Planning (NRF-2012M3A2A1051684) and the "Cooperative Research Program for Agriculture Science & Technology Development (Project No. PJ009842)" Rural Development Administration, Republic of Korea.

Supplementary material

705_2013_1920_MOESM1_ESM.docx (1.3 mb)
Supplementary material 1 (DOCX 1331 kb)

Copyright information

© Springer-Verlag Wien 2013