The Genomes, Proteomes, and Structures of Three Novel Phages That Infect the Bacillus cereus Group and Carry Putative Virulence Factors

ABSTRACT

This article reports the results of studying three novel bacteriophages, JL, Shanette, and Basilisk, which infect the pathogen Bacillus cereus and carry genes that may contribute to its pathogenesis. We analyzed host range and superinfection ability, mapped their genomes, and characterized phage structure by mass spectrometry and transmission electron microscopy (TEM). The JL and Shanette genomes were 96% similar and contained 217 open reading frames (ORFs) and 220 ORFs, respectively, while Basilisk has an unrelated genome containing 138 ORFs. Mass spectrometry revealed 23 phage particle proteins for JL and 15 for Basilisk, while only 11 and 4, respectively, were predicted to be present by sequence analysis. Structural protein homology to well-characterized phages suggested that JL and Shanette were members of the family Myoviridae, which was confirmed by TEM. The third phage, Basilisk, was similar only to uncharacterized phages and is an unrelated siphovirus. Cryogenic electron microscopy of this novel phage revealed a T=9 icosahedral capsid structure with the major capsid protein (MCP) likely having the same fold as bacteriophage HK97 MCP despite the lack of sequence similarity. Several putative virulence factors were encoded by these phage genomes, including TerC and TerD involved in tellurium resistance. Host range analysis of all three phages supports genetic transfer of such factors within the B. cereus group, including B. cereus, B. anthracis, and B. thuringiensis. This study provides a basis for understanding these three phages and other related phages as well as their contributions to the pathogenicity of B. cereus group bacteria.

IMPORTANCE The Bacillus cereus group of bacteria contains several human and plant pathogens, including B. cereus, B. anthracis, and B. thuringiensis. Phages are intimately linked to the evolution of their bacterial hosts and often provide virulence factors, making the study of B. cereus phages important to understanding the evolution of pathogenic strains. Herein we provide the results of detailed study of three novel B. cereus phages, two highly related myoviruses (JL and Shanette) and an unrelated siphovirus (Basilisk). The detailed characterization of host range and superinfection, together with results of genomic, proteomic, and structural analyses, reveal several putative virulence factors as well as the ability of these phages to infect different pathogenic species.

INTRODUCTION

The evolutionary, environmental, and ecological importance of bacteriophages has recently come into focus. With at least 10³¹ bacteriophages in Earth's biosphere (1–5), these living entities are the most abundant on the planet and play a major role in the evolution of novel genes, horizontal gene transfer, and maintenance of environmental and ecological balance (for recent reviews, see references 6 to 12). However, little is known about the complexity and diversity of these organisms. The most studied bacteriophages are those that infect the genus Mycobacterium (671 fully sequenced genomes available at www.phagesdb.org) and those that infect the Enterobacteriaceae family (∼300 fully sequenced genomes available in GenBank). Although numerous Bacillus phages have been isolated (91 fully sequenced genomes available), few have been characterized in detail, including their genomic and proteomic composition, structure determination, and host range. This study focuses on the characterization of three recently isolated phages that infect Bacillus cereus and places them in the context of all Bacillus phage genomes reported thus far.

The Bacillus cereus group of bacteria, also termed B. cereus sensu lato, is an assemblage of six closely related endospore-producing Firmicutes, including B. cereus, B. anthracis, B. thuringiensis, B. weihenstephanensis, B. mycoides, and B. pseudomycoides (for a recent review, see reference 13). Results of molecular methods have recently highlighted the highly related nature of these species, and several investigators have suggested that they form a single species, with strains of different ecotypes and pathotypes (14, 15). These organisms are commonly found in soil but can also infect both mammalian and insect hosts. In humans, B. cereus can cause emetic and diarrheal forms of gastroenteritis (16, 17). B. anthracis is an obligate mammalian pathogen (18), and B. thuringiensis can infect insects (19–21). In contrast, B. weihenstephanensis, B. mycoides, and B. pseudomycoides are nonpathogenic. The study of bacteriophages that infect the B. cereus group is important in understanding their evolution, including the evolution of pathogenic strains because phages often carry or transfer virulence factors to their hosts (for a recent review, see reference 22). In addition, these phages may prove to be a valuable method for controlling pathogenic B. cereus, B. thuringiensis, and B. anthracis strains (6, 7, 9, 23, 24).

We recently reported the isolation and whole-genome sequence of three novel phages (Basilisk, JL, and Shanette) that infect B. cereus (25). Herein we provide the results of detailed study of these phages, including host range and superinfection ability, whole-genome analysis, and phage structure analysis by virion particle proteome identification and electron microscopy. In addition, we identify several genes carried on the JL and Shanette phages that may contribute to the pathogenicity of B. cereus hosts. This analysis presents a framework for understanding the basic properties of Bacillus phages and their ability to transfer virulence factors to their pathogenic hosts.

MATERIALS AND METHODS

Growth media and field isolates.

Bacteria were grown in LB broth (10 g tryptone, 5 g yeast extract, 5 g NaCl, 1 ml of 2 N NaOH) or on LB agar plates (12 g agar per liter). Bacillus cereus field isolates were obtained from apple orchard soil in Logan, UT, and were confirmed by 16S rRNA gene sequencing to be B. cereus species. Briefly, a single colony from each bacterial field isolate was boiled at 98°C for 5 min, and the lysate was used as the template in Taq PCR (New England BioLabs) with 16S RNA primers 27F (F stands for forward) and 907R (R stands for reverse) (26). Samples that produced an ∼1-kb band were submitted for BigDye sequencing (Applied Roche) to the Brigham Young University (BYU) DNA Sequencing Center. Strains were identified as B. cereus species (BC6132 and BC7003) through BLAST (27) analysis of the resulting 16S rRNA gene sequences. Other bacterial strains were B. anthracis Sterne 1043, B. subtilis ATCC 6033, B. thuringiensis kurstoki (HO-1), B. megaterium ATCC 13622, and B. mycoides ATCC 6462. Phages were isolated and sequenced by the method of Grose et al. (25).

Host range studies.

Host range studies were performed by both spot test and plaque-forming assays. For spot test assays, 500 μl of bacteria grown overnight was embedded in LB-top agar (LB with 0.75% agar), and then 5 μl of 10⁶ PFU/ml lysate was spotted in the center of the plate. For plaque-forming assays, 20 μl of 10³ PFU/ml lysate was incubated with 500 μl of bacteria grown overnight for 30 min before plating in top agar. The plates were incubated with the top agar facing up at 30°C overnight for both assays.

Phage superinfection studies.

Agar was removed from the center of an isolated plaque and streaked on LB agar plates, and the plates were incubated at 37°C for 24 h to facilitate lysogen growth. Single colonies were then tested for superinfection by plating overnight cultures in top agar and spotting 5 μl of each phage lysate on the plate. The plates were incubated with the top agar facing up at 30°C for 24 h.

Computational analysis and genomic comparison.

Phylogenetic trees were constructed using a Muscle (28) alignment and the neighbor-joining method in Mega5 (29). Bootstrapping was set at 2,000, and the trees were collapsed at a less than 50% bootstrap value. Trees were rooted to Gram-negative bacteria, which did not alter the overall structure of the tree. Sequences for comparison were chosen by BLASTP alignment for closest relatives, BLASTP alignment to phages, and BLASTP alignment to homologues present in bacterial isolates known to be tellurium resistant. Bacillus phage sequences were retrieved from GenBank and the Bacillus Phage Database at PhagesDB.org and contact with the authors of this website. To ensure retrieval of all Bacillus phages from GenBank, the major capsid protein (MCP) from at least one phage in each cluster was used to retrieve all phages with similar MCP sequences via BLASTP (30). Conserved proteins were identified using the program CoreGenes 3.5 at the default BLASTP threshold of 75 (31, 32). Terminal repeats in Basilisk, JL, and Shanette were determined by processing the raw sequencing data using the PAUSE method (https://cpt.tamu.edu/pause/). Dot plots were generated using Gepard (33) and average nucleotide identity (ANI) was calculated with Kalign (34). The terminal repeats of the Pegasus and CampHawk genomes were removed for dot plot analysis. Genomic maps of all phages were prepared using Phamerator (35), an open-source program designed to compare phage genomes. Phamerator was also used to calculate percent G/C and the number of open reading frames (ORFs). The 101 Bacillus phages used for proteomic comparison by Phamerator were those available online by 15 June 2014 and included the following phages (host bacteria shown in parentheses and accession numbers shown in brackets; phagesdb refers to the Bacillus Phage Database available at phagesDB.org): vB_BceM_Bc431v3 (B. cereus) [NC_020873], 0305phi8-36 (B. thuringiensis) [NC_009760], 250 (B. cereus) [GU229986], Adelynn (Bacillus sp.) [phagesdb], AP50 (B. anthracis) [NC_011523], Andromeda (B. pumilus) [NC_020478], B4 (B. cereus) [JN790865], B5S (B. cereus) [JN797796], B103 (B. subtilis) [NC_004165], Bam35c (B. thuringiensis) [AY257527], BanS-Tsamsa (B. anthracis) [NC_023007], Bastille (B. cereus) [JF966203], BCD7 (B. cereus) [JN712910], BceA1(B. cereus) [HE614282], BCP78 (B. cereus) [JN797797], BCJA1C (Bacillus sp.) [NC_006557], BCP1 (B. cereus) [KJ451625], BCU4 (B. cereus) [JN797798], BigBertha (B. thuringiensis) [NC_022769], BMBtp2 (B. thuringiensis) [JX887877], Blastoid (B. pumilus) [NC_022773], BPS10C (B. cereus) [JN094431], BPS13 (B. cereus) [NC_018857], BtCS33 (B. thuringiensis) [NC_018085], CampHawk (B. subtilis) [NC_022761], CAM003 (Bacillus sp.) [phagesdb], Cherry (B. anthracis) [NC_007457], Curly (B. pumilus) [KC330679], Doofenshmirtz (Bacillus sp.) [phagesdb], Eoghan (B. pumilus) [KC330680], Evoli [NC_024207], Fah (B. anthracis) [NC_007814], Finn (B. pumilus) [KC330683], G (B. megaterium) [JN638751], GA-1 (Bacillus sp.) [NC_002649], Gamma (B. anthracis) [NC_007458], Gamma 51 (B. cereus) [DQ222853], Gamma 53 (B. anthracis) [DQ222855], Gamma isolate d'Herelle (B. cereus) [DQ289556], Gemini (B. pumilus) [KC330681], Gir1 (Bacillus sp.), GIL16c (B. thuringiensis) [NC_006945], Glittering (B. pumilus) [NC_022766], Grass (Bacillus sp.) [NC_022771], Hakuna (Bacillus sp.) [NC_024213], HoodyT (Bacillus sp.) [phagesdb], IEBH (B. thuringiensis) [NC_011167], JL (B. cereus) [KC595512], JPB9 (Bacillus sp.) [phagesdb], Megatron (Bacillus sp.) [phagesdb], MG-B1 (B. weihenstephanensis) [NC_021336], Nagalana (Bacillus sp.) [phagesdb], Nf (B. subtilis) [EU622808], Page (B. megaterium) [NC_022764], Pappano (Bacillus sp.) [phagesdb], PBC1 (B. cereus) [NC_017976], Pegasus (Bacillus sp.) [phagesdb], pGIL01 (B. thuringiensis) [AJ536073], phi105 (B. subtilis) [NC_004167], phi29 (B. Subtilis) [NC_011048], phiAGATE (B. pumilus) [JX238501], phiNIT1 (B. subtilis) [NC_021856], phIS3501 (B. thuringiensis) [NC_019502], Plant (Bacillus sp.), PM1 (B. subtilis) [NC_020883], Polaris (Bacillus sp.), Pony (B. megaterium) [NC_022770], Poppyseed (B. megaterium) [KF669657], PhiCM3 (B. thuringiensis) [NC_023599], ProCM3 (B. thuringiensis) [KF296717], PZA (B. subtilis) [M11813], Riggi (B. pumilus) [NC_022726], Riley (Bacillus sp.) [KJ489402], SIOphi (B. subtilis) [KC699836], Shanette (B. cereus) [KC595513], SP10 (B. subtilis) [NC_019487], SPO1 (B. subtilis) [NC_011421], SP10 (B. subtilis) [NC_019487], SPP1 (B. subtilis) [NC_004166], Slash (B. megaterium) [KF669661], SPBc2 (B. subtilis) [NC_001884], Spock (B. thuringiensis) [NC_022763], Staley (B. megaterium) [NC_022767], Stitch (Bacillus sp.) [phagesdb.org], Taylor (B. pumilus) [KC330682], TP21-L (B. cereus) [NC_011645], Troll (B. thuringiensis) [KF208639], W.Ph. (B. cereus) [NC_016563], WBeta (B. cereus) [NC_007734], Wip1 (B. anthracis) [KF188458], phBC6A51 (B. cereus) [NC_004820], phBC6A52 (B. cereus) [NC_004821], and Bacillus virus 1 (Bacillus sp. strain 6k512) [NC_009737].

Phage purification by CsCl gradient.

High-titer phage lysates were prepared by incubating 20 μl of phage with 500 μl of B. cereus BC7003 culture in exponential phase. The mixtures were plated in top agar after 30-min incubations, and the plates were incubated with the top agar facing up at 30°C for 24 h. The phages were harvested by layering 5 ml of LB on top of the plates, incubating 1 h, and then harvesting the supernatant. The supernatant was centrifuged twice (each time for 20 min at 8,000 × g) and then filtered through a 0.45-μm filter before being applied to a CsCl gradient composed of 1.3, 1.4, and 1.5 μg/ml (3 ml each), 1.6 μg/ml (2.5 ml), and 1.7 and 1.8 μg/ml CsCl (2 ml each).

Phage sample preparation and mass spectrometry analysis.

Purified phage lysates were analyzed by SDS-polyacrylamide gel electrophoresis (PAGE) in duplicate, and proteins were stained with Coomassie brilliant blue. The SDS-polyacrylamide gel was cut into 11 slices for Basilisk and 9 slices for JL to fractionate the proteins. Thick bands, which indicate abundant protein, were placed in their own fraction. Phage protein gel slices were prepared for mass spectrometry using slightly modified versions of gel digestion methods (36). Briefly, protein gel slices were diced, destained with 50% acetonitrile (at least high-performance liquid chromatography [HPLC] grade) in 50 mM ammonium bicarbonate (Acros; 99%), reduced with dithiothreitol (Acros; 99%), alkylated with iodoacetamide (MP Biomedicals, LLC), and digested overnight with 1 μg trypsin (Pierce; mass spectrometry [MS] grade) in 25 mM ammonium bicarbonate (Acros; 99%) at 37°C. The peptides were extracted from the gel using 50% acetonitrile (at least HPLC grade) and vortexing. The resulting solution was decanted, speed vacuumed, and resuspended in 0.1% formic acid (Fisher Scientific; Optima grade). Samples were analyzed on the LTQ Orbitrap XL mass spectrometer coupled with the Eksigent Nano-LC Ultra system. The peptides were trapped on an Exp C18 stem trap (Optimize Technologies) and separated for 90 min using H₂O-acetonitrile acidified with 0.1% formic acid (Fisher Scientific; Optima grade) at 325 nl/min on a nanoACQUITY PST 25-cm C₁₈ column with 1.7-μm beads (Waters). Parent scans were preformed in the Orbitrap at 60,000 resolution at 400 m/z with a lock mass at 371.101230 m/z. The top 10 most-abundant species were fragmented and detected in the LTQ using collision-induced dissociation. Unassigned peptides and peptides with one charge state were excluded from fragmentation. Dynamic exclusion was set with a repeat count of 1 and a 180-s exclusion duration, but early expiration was allowed. Mass spectral data were searched against the annotated phage genomes (Basilisk [KC595511.1] and JL [KC595512], Uniprot) with and without B. cereus and Escherichia coli database backgrounds using the Thermo Proteome Discoverer 1.4 software package with Sequest HT and Percolator. To detect peptides that may be present but misannotated, each phage genome was also translated into all six frames (EMBOSS Sixpack [37]) and searched. Search parameters included Orbitrap tolerance of 10 ppm and tryptic digest LTQ tolerance of 0.6 Da, allowing at most two missed cleavages in the tryptic digestion, oxidation (M) as a dynamic modification and carbamidomethy (C) as a static modification. A high-confidence filter was applied with a false discovery rate of 1.0%.

TEM of phages.

High-titer phage lysates were prepared for transmission electron microscopy (TEM) on 200-mesh copper grids coated with Formvar and carbon (Ted Pella, Inc.) and stained with 2% phosphotungstic acid solution (pH 7) for 2 min or 1% uranyl acetate for 15 to 20 s. Residual liquid was wicked away with filter paper. Transmission electron micrographs were captured using a FEI Tecnai F20 TEM operated at 200 kV or a JEOL JEM-1400Plus TEM operated at 120 kV. The phages in electron micrographs were measured using ImageJ (38) with the average and standard deviation from a minimum of three separate measurements.

Cryogenic specimens of Basilisk phage were prepared by first applying 3.5 μl of purified virus to holey carbon grids. With the use of a FEI Vitrobot, the sample was then blotted with filter paper and plunge frozen in liquid ethane. Frozen-hydrated specimens were stored in liquid nitrogen and, during imaging, were maintained at approximately −180°C with a Gatan 626 Cryoholder (Pleasanton, CA, USA). Images were recorded on Kodak SO-163 film with a FEI Tecnai F20 TEM operated at 200 kV. The electron dose was less than 10 electrons/Ų. Images were digitized with a Nikon Super CoolScan 9000 ED scanner.

Parameter values for the contrast transfer function (CTF) were determined from scanned micrographs via the program CTFFIND3 (39). Basilisk particle images were selected and extracted using X3DPREPROCESS (40). With icosahedral symmetry enforced, two random-model methods were used to determine ab initio three-dimensional (3D) reconstructions from the imaged particles (41, 42). Both methods gave similar T=9 icosahedral structures. One of these structures was used as a starting model for model-based determination of the orientations and origins of the imaged particles. The package AUTO3DEM (43) was used to determine orientations and origins of imaged particles and to compute the three-dimensional reconstruction. During AUTO3DEM processing, CTF parameter values were refined for individual particle images. Bsoft (44) and UCSF Chimera (45) were used for graphical display of the 3D reconstruction. UCSF Chimera was also used to fit coordinates of the bacteriophage HK97 major capsid protein (46) into the Basilisk capsid density. Only the core of the HK97 protein was fitted (47). HK97 coordinates were first fit into the T=9 capsid of halophilic bacteriophage NS01 (D. Eng, P. Shen, M. Domek, D. Belnap, unpublished results), and this fit was used as a starting point for the fit into the Basilisk structure. The nine subunits in the NS01 asymmetric unit were adjusted as a rigid body by manual adjustment to better fit the Basilisk structure. The overall adjustment was small, between 3 and 4 Å per subunit. Magnification of the cryo-electron microscopy (cryo-EM) images was calibrated with bovine liver catalase as the calibration standard (48).

Accession numbers.

The reannotation of the three phages is available in GenBank with the following accession numbers: Basilisk, KC595511; JL, KC595512; and Shanette, KC595513. The 3D reconstruction of the Basilisk head is available as entry 6043 in the EM Data Bank.

RESULTS

Isolation and host range of Basilisk, JL, and Shanette.

Two B. cereus isolates (BC6132 and BC7003) were collected from soil samples from Logan, UT, and were confirmed as B. cereus by 16S rRNA gene sequencing which revealed more than 99% identity to B. cereus strains. Phages were subsequently isolated from enrichment cultures containing BC7003 and soil from various apple orchards along the Wasatch Front of Utah County in Utah as previously reported (25). Each phage was plaque purified at least three times. Three phages, Basilisk, Shanette, and JL, were cultured to moderately high titers (10⁶) for further characterization.

Cross-infectivity studies of Basilisk, JL, and Shanette were warranted because others have observed cross-infectivity of bacteriophages that infect the Bacillus genus (49–51). Basilisk, JL, and Shanette were capable of infecting both environmental B. cereus isolates along with B. thuringiensis and B. anthracis Sterne (Table 1). Infection was not detected with B. subtilis, B. mycoides, or B. megaterium. These results are consistent with the highly related nature of the B. cereus, B. anthracis, and B. thuringiensis species (14, 15, 52). In addition to host range, superinfection studies were performed in order to assess whether the phages were strictly lytic or temperate. Basilisk formed stable lysogens that were resistant to superinfection by the same phage. Both JL and Shanette were unable to form colonies resistant to superinfection, suggesting that they are strictly lytic.

TABLE 1.

Superinfection and Bacillus host range assays of B. cereus phages Basilisk, Shanette, and JL

Phage	Superinfection assayb	Bacillus host range assayb
	Stable lysogena	B. cereus BC6132	B. cereus BC7003	B. anthracis Sterne	B. thuringiensis	B. Subtilis	B. mycoides	B. megaterium
Basilisk	Yes	+	+	+	+	−	−	−
Shanette	No	+	+	+	+	−	−	−
JL	No	+	+	+	+	−	−	−

^aB. cereus strains BC7003 and BC6132 were used for superinfection studies (the ability of the phage to produce a stable lysogen is indicated by yes or no).

^bSymbols: +, plaque formation; −, failure to infect based on both plaque assay and spot tests.

Genome sequence analysis of Basilisk, JL, and Shanette show relationships to other Bacillus phages.

The genome sizes for Basilisk, JL, and Shanette were 81,790, 137,918, and 138,877 bp, respectively. Sequencing was accomplished with 454 pyrosequencing, and putative open reading frames (ORFs) of each genome were predicted using GeneMark and Glimmer within the DNA Master phage annotation software (http://cobamide2.bio.pitt.edu) as previously described (25). The data from coding potential (GeneMark [53]), BLASTX (54), and Shine-Dalgarno prediction (http://cobamide2.bio.pitt.edu) were used to determine the best call for annotation, with GeneMark and BLASTX given the most weight. Functions were assigned on the basis of BLASTP hits with an E value of at least 10⁻³. The ends of the phages were recently determined by processing the raw sequencing data using the PAUSE method (https://cpt.tamu.edu/pause/) which resulted in a new bp 1 call than previously reported (25), and GenBank files were updated accordingly.

Nucleotide sequence revealed JL and Shanette as highly similar phages (96% average nucleotide identity [ANI]), while Basilisk was unrelated. Whole-genome maps of JL, Shanette, and Basilisk are presented in Fig. 1, and their basic genomic properties are given in Table 2. Annotation of the JL and Shanette genomes confirmed their similar nature and revealed 217 ORFs with four tRNAs for JL and 220 ORFs with three tRNAs for Shanette. The genome of JL encodes four unique putative proteins of unknown function (gp21, gp77, gp188, and gp196) and one tRNA that were not encoded by Shanette, while Shanette encodes eight unique putative proteins of unknown function (gp10, gp17, gp43, gp45, gp75, gp77, gp187, and gp221). Each of these 12 unique putative proteins are less than 44 amino acids and are uncharacterized; therefore, we are uncertain whether they are expressed, but all have coding potential based on GeneMark (53). All putative proteins with a predicted function were encoded by genes on both genomes. From genome annotation and nucleotide analysis, Basilisk appeared to be unrelated to JL or Shanette and harbors 138 ORFs and two tRNAs. The presence of tRNAs in all three genomes is common for phages and is expected to compensate for differences in codon usage or the obliteration of host tRNAs (55, 56).

FIG 1.

Genomic maps of B. cereus phages illustrate the high similarity of Shanette and JL contrasted to the unrelated phage Basilisk. Phages were mapped using Phamerator (19). Purple lines between phages denote regions of high nucleotide similarity, and the ruler shows genome base pairs. Boxes for gene products are labeled with predicted function, occasionally numbered, and colored to indicate similarity between the phages (E value < 1e⁻⁴). Predicted tRNAs are shown with black boxes and are labeled. DHFR, dihydrofolate reductase; MCP, major capsid protein; MurNAc-LAA, N-acetylmuramolyl-alanine amidase.

TABLE 2.

Characteristics of Basilik, Shanette, and JL, as well as related Bacillus phages

Phage	Hosta	Size (bp)	GC%	No. of ORFs	No. of tRNAs	Accession no.	Familyb	Reference
Basilisk	C	81,790	33.9	140	2	KC595511	S	25
Staley	M	81,656	35.35	113	0	NC_022767	S	91
Slash	M	80,382	35.23	111	0	KF669661	S	92
JL	C	137,918	40.8	222	4	KC595512	M	25
Shanette	C	138,877	40.8	223	3	KC595513	M	25
SPO1	S	132,562	39.97	204	5	NC_011421	M	93
Pegasus	B	146,685	40.3	236	3
CampHawk	S	146,193	40.2	231	2	NC_022761	M	89

^aThe hosts that phages were isolated from are shown using abbreviations as follows: C, Bacillus cereus; M, Bacillus megaterium; S, Bacillus subtilis; B, Bacillus sp.

^bS, Siphoviridae; M, Myoviridae.

Despite the inability of all three phages to infect B. subtilis, BLAST hits for proteins encoded by JL and Shanette revealed similarity to the well-characterized lytic B. subtilis phage SPO1 and the SPO1-like isolate CampHawk, which is consistent with the strictly lytic nature of JL and Shanette in the above superinfection studies (57). In contrast, proteins encoded by Basilisk were similar to those found in newly isolated, uncharacterized B. megaterium phages Staley and Slash. In a search for phages related to these seven phages, we were also able to retrieve the nucleotide sequence of a recently isolated SPO1-like phage Pegasus online (http://bacillus.phagesdb.org). The basic characteristics of these eight phages are provided in Table 2.

To determine the overall genomic and proteomic similarity of these phages, nucleotide and gene product comparisons were made using Gepard dot plot (Fig. 2) and CoreGenes (Table 3). Nucleotide dot plot analysis confirmed a weak nucleotide relationship between JL and Shanette and the highly related phages SPO1, CampHawk, and Pegasus (Fig. 2). Dot plot analysis consists of aligning the phage nucleotide sequences along both the x and y axes. A dot is placed where the sequences are identical; hence, a single sequence aligned with itself results in a diagonal line down the center of the plot. JL and Shanette had a strong diagonal line across their comparison, consistent with 96% ANI. In comparison, there was a weak diagonal line running the length of the genomes when JL or Shanette was compared to SPO1, CampHawk, or Pegasus (JL and SPO1 have 53.51% ANI, while Shanette and SPO1 have 53.39% ANI). Basilisk displayed an even weaker relationship to its closest relatives, the uncharacterized phages Staley and Slash (no significant ANI). No relationship could be seen by dot plot between JL/Shanette and Basilisk.

FIG 2.

Whole-genome nucleotide (A) and amino acid (B) dot plot analysis of JL, Shanette. and Basilisk reveals relationships to other Bacillus phages and a clear division between JL/Shanette and Basilisk. Nucleotide dot plots were produced using Gepard (33) at a word length of 10 and amino acid at a word of 5. Whole-genome amino acid sequences were obtained from Phamerator (35).

TABLE 3.

Percentage of the predicted proteome conserved between JL, Shanette, Basilisk, and related phages SPO1, CampHawk, Staley, and Slash

Reference genome	% predicted proteome conserved between the reference genome and the following phagea:
	JL (218 gp)	Shanette (220 gp)	SPO1 (204 gp)	CampHawk (231 gp)	Basilisk (138 gp)	Staley (138 gp)	Slash (138 gp)
JL	100
Shanette	96.33*	100
SP01	39.45*	39.55*	100
CampHawk	38.99*	39.09*	92.65*	100
Basilisk	4.59	4.09	5.39	4.76	100
Staley	3.21	2.73	5.88	5.63	35.51	100
Slash	3.67	3.18	4.90	4.76	34.78	94.69*	100

^aThe number of total gene products for each phage listed is given. CoreGenes scores of ∼40% are typically used to indicate phage families (54, 58, 59). The CoreGenes scores of ∼40% are indicated by an asterisk.

CoreGenes analysis confirmed the relationship between JL, Shanette, SPO1, and CampHawk, with JL sharing 96.33% of its proteome with Shanette, 39.45% with SPO1, 38.99% with CampHawk, and less than 5% with Basilisk, Staley, and Slash (Table 3). In contrast, Basilisk shares about 35% of its proteome with Staley and Slash. These proteomic similarities are right on the border of the 40% cutoff that is typically used to group phages by relatedness (54, 58, 59), highlighting the novel nature of our three isolates.

Proteins of predicted function encoded within the three B. cereus phage genomes revealed putative virulence factors and the novel nature of Basilisk.

The three phages encoded a variety of gene products with predicted roles in virion structure, virion assembly, DNA replication/metabolism enzyme, gene expression control, cell lysis, or host pathogenesis (Table 4).

TABLE 4.

Proteins with predicted function encoded by JL, Shanette, and Basilisk^a

^aVirion components identified by LC/MS/MS are indicated by an asterisk, with lightest to darkest shading indicating functional categories for virion component (no shading), virion assembly (lightest gray), DNA replication/metabolism, gene expression control, cell lysis, and host function/pathogenesis (darkest gray). MurNAc-LAA, N-Acetylmuramolyl-alanine amidase.

Virion components.

The JL and Shanette genomes encoded 11 homologues of proteins known to be virion components of Bacillus phages. JL and Shanette contain phage structural proteins that are highly similar to those encoded by Bacillus myoviruses SPO1 and CampHawk (percent identity to a homologue in SPO1 is given), including the major capsid protein (78%), portal protein (68%), tail sheath protein (50%), tube protein (73%), two tail fiber proteins (gp105 is 53% identical to SPO1, while gp113 is 35% identical to B. subtilis phage SP10), absorption tail protein (46%), and two baseplate proteins (gp108 and gp109 are both 43%). In addition, the JL genome encodes a putative tail lysin (gp104, 50%) that has similarity to cell wall-associated hydrolases as well as a protein containing an N-acetylmuramolyl domain (JL gp117) found in several Bacillus genomes and also encoded by phage SPBc2 (59% identity). Tailed double-stranded DNA (dsDNA) bacteriophages frequently harbor structural proteins displaying peptidoglycan hydrolases known to hydrolyze the amide bond between N-acetylmuramoyl and l-amino acids in certain cell wall glycopeptides (60–62). This suggests that both of these proteins aid in degradation of the Gram-positive cell wall, allowing phage access to the cell membrane.

In contrast to JL and Shanette, the Basilisk genome encoded only four putative virion proteins, revealing its novel nature. The first two proteins are the putative tape measure protein (TMP, gp47) and the minor structural protein (gp49) and are similar to proteins encoded by the recently isolated, related phages Slash and Staley (32% and 50% identity for Basilisk gp47 and gp49, respectively). The putative TMP harbors a peptide domain common in zinc metallopeptidases that are encoded by prophages in several Bacillus species. The third protein is Basilisk's tail fiber protein (gp51), which is one of only two proteins also encoded by JL and Shanette (38% identity). Since all three phages display the same host range (Table 1), these results are consistent with tail fibers being a highly recombinogenic host determinant (63, 64). The tail fiber has homology to carbohydrate binding proteins (CenC from B. cereus is 43% identical over 103 amino acids), which suggests that carbohydrates are involved in host recognition. The fourth protein (gp52) may aid Basilisk by allowing access to the cell, since it has significant similarity (41% identity over 251 amino acids) to a chitinase found in Paenibacillus sp. (65). A putative major capsid protein (MCP) was not identified by sequence comparison.

Virion assembly.

The genomes of JL and Shanette harbor a putative procapsid protease as well as tail assembly chaperones that are required for the assembly of other Myoviridae phages, and all three genomes encode a recognizable large terminase for DNA packaging. The predicted tail assembly chaperones (JL gp101 and gp102) result from a predicted +1 frameshift that is based on the well-characterized phage SPO1. The SPO1 phage encodes two chaperones as the result of a +1 ribosomal slippage (SPO1_86 and SPO1_87). Although frameshift prediction software such as FrameD was unable to detect this frameshift in JL and the nucleotide similarity between JL and SPO1 is weak, amino acid conservation exists between the JL (JL_102) and SPO1 (SPO1_87) frameshift gene products (58% similarity).

The terminase for JL (gp30) and Shanette is also highly similar (77% identity) to that of SPO1. This observation suggests that they share similar DNA packaging strategies that result in nonpermutated DNA with long terminal repeats (57, 66). Processing the raw sequencing data using the PAUSE method (https://cpt.tamu.edu/pause/) confirmed that JL and Shanette do indeed have long terminal repeats of 16,292 bp and 16,480 bp, respectively. In contrast, Basilisk's terminase is similar to those of uncharacterized phages Staley and Slash (67% identity to both), and PAUSE revealed short terminal repeats of 218 bp.

DNA replication/metabolism.

All three phage genomes encode proteins with putative DNA replication functions, including DNA helicase and primase. JL and Shanette also encode their own DNA polymerase, ligase, and topoisomerase as well as two proteins that appear to be involved in DNA partitioning (JL gp67 and gp154). In addition, all three phage genomes encode putative enzymes for nucleotide metabolism, including thymidylate synthase and deoxyuridine (dUTP) pyrophosphatase, while Basilisk encodes a putative nucleotide reduction system consisting of dihydrofolate reductase, class I ribonucleotide reductase, ribonucleotide reductase stimulatory protein, a glutaredoxin-like protein, and thioredoxin.

Several putative proteins for DNA recombination are evident in our three phages, including an HNH domain endonuclease (JL and Shanette), exonucleases (all three phages), a Holliday junction resolvase (Basilisk), and a putative integrase (Basilisk). This putative integrase (gp109) is consistent with the results of our superinfection studies in which Basilisk behaved as a temperate phage (Table 1). High rates of recombination and the presence of recombination systems are a hallmark of tailed bacteriophages (36, 67–69), and such systems have proven useful for genetic recombineering in both bacteria and phages (70, 71). The JL and Shanette genomes also encode a recombination nuclease (SbcCD) that is found in other Bacillus phages, including SPO1 and CampHawk. This ATP-dependent double-strand DNA exonuclease is present in many bacteria and has been shown to eliminate or repair DNA secondary structures, preventing propagation of palindromic phage sequences (72–74). Phage Lula/phi80 encodes an inhibitor of the host SbcCD that facilitates its replication (75), suggesting the SbcCD encoded by JL/Shanette may inhibit replication of other phages or may act as a decoy/inhibitor complex to allow replication of JL/Shanette.

Gene expression control.

In addition to encoding their own DNA replication machinery, JL and Shanette also harbor genes for transcriptional control, including those that encode putative sigma factors or accessory proteins (JL gp144, gp175, and gp176) as well as putative transcriptional regulators (JL gp134 and gp145). These proteins may be used to regulate gene expression of either phage or host targets. Basilisk contains two putative transcription factors (gp22 and gp103); gp103 contains a KilA domain that is common in phage transcription factors (76).

Cell lysis.

All three genomes encode proteins that appear to function in cell wall degradation and host cell lysis, allowing phage progeny to enter or exit the cell. JL, Shanette, and Basilisk all contain a putative hydrolase/lysin (JL gp117 and Basilisk gp53) with homology to N-acetylmuramoyl-l-alanine amidases found in a wide variety of Bacillus species. JL and Shanette also contain a lysin conserved in SPO1 (JL gp116, a putative holin) and an autolysin (gp103), while Basilisk contains a lysin homologue (gp54).

Host functions/pathogenesis.

All three phage genomes possess proteins that may affect host pathogenesis. They all encode a putative dUTPase; dUTPases are common in many bacteriophages and have been shown to function as G-protein-like regulators required for the transfer of staphylococcal virulence factors (77, 78). In addition, they all encode a PhoH-like protein that is most similar to PhoH from the bacteria Agrobacterium vistis and Rhizobium sp. In addition, JL and Shanette encode the tellurium resistance proteins TerD and TerC (JL gp140 and gp142 and Shanette gp141 and gp143).

Tellurium oxyanion (TeO₃²⁻) has been used in the treatment of mycobacterial infections, and resistance is a feature of many pathogenic bacteria (79). A BLASTP search for TerC and TerD revealed homologues in many bacteria, including Bacillus strains, and a few phages that infect Gram-positive hosts. Since bacterial virulence factors are commonly carried by phages, we produced neighbor-joining alignments of the TerC and TerD proteins in order to observe phylogenetic relationships for both of these proteins (Fig. 3). All known phage homologues were included in the alignments; however, due to the large number of homologues from bacteria, the bacterial sequences included in the alignment were a few with the highest E values and those from Bacillus species as well as a few to represent the Gram-negative species known to be tellurium resistant. Phylogenetic trees suggest that both the JL TerC and TerD proteins are more similar to proteins found in Bacillus bacteria than to those encoded by Gram-negative bacteria or other phages. Since the only other phages found to carry these genes are phages that infect Gram-negative hosts, the predicted phylogeny is not unexpected and suggests genetic exchange between JL/Shanette and their Bacillus hosts. The only outlier to a Gram-negative/Gram-positive division is the TerD protein from B. weihenstephanensis, which appears to be more closely related to proteins encoded by members of the Enterobacteriaceae family than members of the Bacillus genus. There is no reported phage capable of infecting both Gram-negative and Gram-positive hosts, but the close relationship could be facilitated by another genetic method such as a plasmid. The TerD protein from Pseudomonas phage phiPsa374 and Cronobacter phage vP_CsaP_GAP52 also appear to be more closely related to Gram-positive hosts; however, the bootstrap values for those associations are low (55% and 78%).

FIG 3.

Phylogenetic trees of TerC (A) and TerD (B) produced by neighbor joining suggest that the JL and Shanette gene products are related to proteins encoded in Bacillus strains. All TerC and TerD proteins encoded by phages as well as a few representative Gram-positive and Gram-negative strains are shown (all have a BLASTP E value of greater than 10⁻⁷). Phylogenetic trees were constructed using a Muscle (28) alignment and the neighbor-joining method in Mega5 (29). Bootstrapping was set at 2000, and trees were collapsed at a less than 50% bootstrap value. str., strain.

Mass spectrometry analysis detected novel proteins as virion components and verified annotation.

As stated above, the genome of phage Basilisk was similar only to uncharacterized Siphoviridae phages, and Basilisk contained only four putative virion proteins predicted by sequence homology. In order to identify virion components and check the quality of our annotation, phages JL and Basilisk were purified by CsCl gradient, fractionated by SDS-polyacrylamide gel electrophoresis (Fig. 4A), and subjected to liquid chromatography-tandem mass spectrometry (LC/MS/MS). Shanette was not analyzed by LC/MS/MS due to the high genomic/proteomic similarity with JL discussed above. For JL, 23 proteins were detected with high confidence, while 15 were detected for Basilisk (Table 4). Comparison of these proteins with those of putative function indicates that all but one (gp108) of the predicted virion component proteins for JL was detected, including the putative portal protein, major capsid protein (MCP), tail sheath protein, tail tube subunit, tail assembly chaperone, tail lysin, tail fiber proteins, baseplate J family protein, and N-acetylmuramoyl-l-alanine amidase. In addition, we identified a putative nuclease with homology to the plasmid partitioning protein ParB (gp67).

FIG 4.

Structural characterization of Basilisk, JL, and Shanette. (A) CsCl-purified phages run on SDS-polyacrylamide gels for virion component analysis. Basilisk (left lane) and JL (right lane) were run (the middle lane is empty). Lanes were cut into slices and analyzed by LC/MS/MS. Major proteins identified by high spectral count score and good coverage are indicated with their predicted molecular masses (in kilodaltons) at the sides of the gel. (B) Transmission electron microscopy confirms Basilisk as a Siphoviridae and JL and Shanette as Myoviridae. Plaque-purified phages were negatively stained and imaged.

The four virion component proteins predicted above for Basilisk (tape measure protein, minor structural protein, tail fiber, and chitinase) were all detected. We were unable to identify the MCP for Basilisk by sequence analysis, but two proteins were most abundant on the SDS-polyacrylamide gel (Fig. 4A) and gave extremely high spectral counts in the LC/MS/MS analysis. The most abundant protein identified by SDS-PAGE and LC/MS/MS was an ∼35-kDa protein that matched the sequence of gp36. The high abundance of this protein suggests that gp36 is the MCP; however, BLAST analysis of this protein yields a single hit to a hypothetical protein of unknown function (Staley gp36). The second most abundant protein was an ∼16-kDa protein matching the sequence of gp35, an uncharacterized protein. The long Basilisk tail (Fig. 4B) would suggest that gp35 is a tail component.

In addition to detecting proteins of putative structural function, 12 proteins with homology to putative uncharacterized proteins were identified for JL and 11 for Basilisk. We also found spectral evidence for the presence of uncharacterized protein Basilisk gp46 in virions when searching against the phage proteome alone, but gp46 fell below our confidence threshold when the data were searched in a more stringent but less sensitive manner (using both B. cereus and E. coli proteomes as background in the proteome database). Support for these uncharacterized proteins as virion components comes from the genomic organization of JL, Shanette, and Basilisk because all of these clustered with known structural genes near the beginning of the genome, except four (JL gp147, gp215, and gp217 and Basilisk gp123).

In addition to identifying virion components, the mass spectrometry analysis verified the annotation of all three phages. Although the spectral data were searched for peptides that could result from translation of all six reading frames of the phage genomes, only peptides corresponding to annotated gene products were recovered.

Identification of genes conserved in other Bacillus phages confirmed the novel nature of Basilisk.

We used Phamerator (35) to determine the number of gene products contained within our three phages that are conserved within the extant 101 Bacillus phages available on NCBI or at www.phagesdb.org as of 15 June 2014 (Table 5). The structural and assembly proteins of the virion are the most highly conserved gene products compared with other Bacillus phages, with the tail fiber of JL (gp113), Shanette (gp115), and Basilisk (gp51) being found in 43 phages. Other well-conserved gene products of JL and Shanette include the structural proteins terminase (JL gp30), tape measure protein (JL gp103), major capsid protein (gp92), baseplate proteins (gp108 and 109), tail sheath (gp99), portal protein (gp89), and procapsid protease (gp90), which were each identified in at least 22 other phages. In addition, several of the DNA replication/recombination enzymes were well conserved with a DNA polymerase (JL gp162) being the most conserved gene product (found in 52 other phages). The most highly conserved gene products in the novel phage Basilisk are the DNA metabolism enzymes ribonucleotide reductase (gp91), an uncharacterized ATPase (gp50), and dihydrofolate reductase (gp62), which are each found in at least 32 other Bacillus phages.

TABLE 5.

Gene products of phages JL, Shanette, and Basilisk with predicted function that are conserved among Bacillus phages

Function and proteina	Gene product of the following phage:			No. of membersb	Closest relativec
	Basilisk	JL	Shanette
Virion structural and assembly proteins
Tail fiber protein*	gp51	gp113	gp115	43	B. thuringiensis, B. gaemokensis
Terminase large subunit		gp30	gp29	41	SPO1_54
Tape measure protein		gp103	gp105	35	CampHawk_86
Major capsid protein*		gp92	gp94	35	SPO1_77
Baseplate protein		gp108	gp110	34	SPO1_93
Tail sheath protein*		gp99	gp101	34	SPO1_84
Portal protein		gp89	gp91	34	SPO1_73
Baseplate J family *		gp109	gp111	34	CampHawk_92
Procapsid protease		gp90	gp92	33	CampHawk_73
Putative tail lysin*		gp104	gp106	6	CampHawk_87
Virion component*		gp81	gp82	5	SPO1_67
Tape measure protein*	gp47			5	Slash_53
MurNAc-LAA*		gp117	gp119	4	B. massilioanorexius
Terminase large subunit	gp31			3	Staley_33
Terminase	gp32			3	Staley_33
Virion component*		gp84	gp86	3	SP10_134
Virion component*		gp215		3	BanS-Tsamsa_213
Virion component*	gp33			3	Staley_34
Virion component*	gp35			3	Staley_36
Virion component*	gp36			3	Staley_37
Virion component*	gp38			3	Staley_39
Virion component*	gp40			3	Staley_41
Virion component*	gp41			3	Staley_48
Virion component*	gp44			3	Staley_51
Virion component*	gp46			3	Staley_53
Minor structural protein*	gp49			3	Staley_57
Cell lysis
Autolysin/holin protein	gp54			7	B. cereus
Lysin (MurNAc-LAA)	gp53			6	Staley_59
Lysin/holin protein		gp116	gp118	5	CampHawk_103
Gene regulation
Transcriptional regulator		gp134	gp136	6	CampHawk_118
Sigma factor		gp144	gp146	5	SPO1_125
Late gene expression		gp145	gp147	5	CampHawk_124
Antirepressor KilA	gp103			3	B. cereus
DNA replication/metabolism
DNA polymerase		gp162	gp164	52	SPO1_31
Ribonucleotide reductase	gp91			47	B. thuringiensis
RNA helicase		gp119	gp121	36	SPO1_109
SbcC-like nuclease		gp130	gp132	34	SPO1_119
SbcD-like nuclease		gp126	gp128	33	CampHawk_111
Dihydrofolate reductase	gp62			32	L. sacchari
NrdE-like glutaredoxin	gp90			13	B. cereus
NrdI stimulatory protein	gp92			12	Paenibacillus sp.
DNA polymerase		gp143		12	SPO1_123
DNA polymerase		gp5		12	SPO1_51
		gp143			CampHawk_122
DNA ligase		gp174	gp176	7	SPO1_157
DNA helicase (UvrD-like)		gp198	gp199	7	SPO1_173
Thymidylate synthase		gp158	gp160	6	SPO1_29
Integration host factor		gp161	gp163	6	CampHawk_145
Exonuclease		gp171	gp173	6	CampHawk_154
Sigma factor		gp176	gp178	6	SPO1_34
HNH endonuclease	gp126	gp59	gp60	6	Paenibacillus sp.
Integration host factor		gp160	gp162	6	SPO1_30
ParB-like nuclease*		gp67	gp68	5	CampHawk_55
Replicative helicase		gp120	gp122	5	SPO1_110
ATPase (replication relax)		gp148	gp150	5	CampHawk
ATPase		gp149	gp151	5	CampHawk_129
Plasmid stability (StbA)		gp154	gp156	5	CampHawk_136
Histone		gp161	gp163	5	SPO1_145
Nucleosome binding		gp167	gp169	5	SPO1_150
Thioredoxin	gp11			4	MG-B1_26
Thymidylate synthase	gp64			4	L. sacchari
Nucleoside hydrolase	gp102			4	Staley_98
dUTPase		gp135	gp137	3	SP10_72
Lectin/gluconase		gp217		3	Bans-Tsamsa_218
dUTPase	gp21			3	B. cereus
Resolvase	gp61			3	Slash_62
Primase	gp65			3	B. cereus
Exonuclease	gp68			3	Slash_68
DNA primase	gp74			3	Staley_76
DNA helicase	gp75			3	Slash_76
XerD recombinase	gp109			3	Slash_98
Host functions/pathogenesis
PhoH-like		gp80	gp81	5	SPO1_66
PhoH-like	gp85			3	Bcp1_50

^aProteins identified or confirmed by GC/MS/MS spectrometry (this study) are indicated by an asterisk.

^bThe number of members of Bacillus phage gene products belonging to this protein pham as identified by Phamerator.

^cThe closest relative of the JL or Basilisk gene product, whether it be a Bacillus phage or bacterial homologue, identified by BLASTP.

Electron microscopy confirmed predicted morphology of the JL and Shanette phages and identified Basilisk as a siphovirus.

Transmission electron microscopy (TEM) was used to confirm the predicted morphological properties of JL and Shanette and to establish Basilisk as a member of the Siphoviridae (Fig. 4B). JL and Shanette had an icosahedral head with a contractile tail consistent with their structural protein homology to the Myoviridae family. Basilisk had an icosahedral head with a noncontractile tail, which suggests that it is a novel member of the Siphoviridae family. From images of negatively stained particles, the heads of myoviruses JL and Shanette were 96.3 (±4.3) nm and 95.9 (±4.2) nm in diameter, respectively. The tails were also nearly identical in size, 152 (±8) and 153 (±12) nm, respectively. Basilisk heads were measured at 72.2 (±3.2) nm and tails at 420 (±9) nm. Basilisk tails were striated, which is indicative of a helical arrangement of the structural proteins (Fig. 5A). At the tail end distal to the head, two types of ends were observed (Fig. 5A). In one type, a thin, needle-like structure was seen. In other particles, a rounded, donut-shaped end was observed. The functions of the distinct types of ends are unknown.

FIG 5.

Transmission electron microscopy reveals two unique types of distal tail ends for Basilisk. (A) In a negative-stained preparation, tail striations were observed as well as two types of distal ends: a needle-like structure (black arrowhead) and a closed, rounded structure (white arrowhead). (B) In a cryogenic, unstained preparation, the tail striations are less visible because of the thin tails, unstained nature of the specimen, and low-electron-dose procedure used to record images. Partially empty (black arrowhead) and empty (white arrowhead) particles were also observed. Bars, 50 nm.

Cryo-EM reconstruction of Basilisk, a phage with novel, uncharacterized structural proteins.

The uncharacterized nature of phage Basilisk prompted us to perform three-dimensional (3D) structural reconstruction of this phage. This reconstruction should also be valuable in understanding similar recently isolated phages, such as Staley and Slash, as well as phages yet to be discovered.

Frozen-hydrated Basilisk particles were prepared and imaged. Basilisk particles full of DNA, partially full, and without DNA were observed in the cryo-EM images (Fig. 5B). The vast majority of particles were filled with DNA. A 3D image reconstruction was computed from the full particles (Fig. 6). The structure is solved to 18-Å resolution and was computed from 2,085 individual-particle images (underfocus range was 0.0 to 3.3 μm). As viewed from the outside, the Basilisk phage head is distinguished by pentameric capsomeres, two types of hexameric capsomeres, and trimeric protrusions between capsomeres. In a line between every two adjacent 5-fold vertices lie at the centers of two hexavalent capsomeres (hexameric assemblies of the capsid protein). This pattern indicates that the Basilisk head has a triangulation number (T) of nine. The thin-shelled protein capsid suggested that the capsid proteins have the bacteriophage HK97 fold (46, 80). Indeed, the core domain of the HK97 capsid protein fits well within the Basilisk capsid density (Fig. 6C), suggesting that Basilisk MCP also has this fold. Interestingly, the Basilisk MCP could not be identified from a TBLASTN search using the HK97 MCP protein sequence as the query, suggesting broad evolutionary sequence divergence while maintaining a conserved structural fold.

FIG 6.

Three-dimensional reconstruction of the head of Basilisk phage from cryo-TEM images reveals an HK97-like fold. (A) A central slice perpendicular to an icosahedral 2-fold axis of symmetry (black oval). Protein and nucleic acid densities are represented with lighter intensities. The thin, corrugated outer ring is the capsid. Inside, a few layers of dsDNA are observed as concentric rings. Bar, 25 nm. (B) Surface rendering of the capsid viewed from the outside centered on an icosahedral 2-fold axis. (The contour selected corresponds to a sigma level of approximately 0.4.) Darker hues represent density closer to the center of the head. Icosahedral 5-fold axes lie in the center of the white five-pointed, star-shaped capsomeres (diameter of 78 nm). Between 5-fold and 2-fold (e.g., center of image) axes are hexamers with the appearance of a six- pointed star. A second type of hexameric capsomere is centered on the icosahedral 3-fold axes, which lie in the center of three adjacent 5-fold vertices. A lower-density feature was observed on 3-fold axes (see black arrowheads in panels A and B), which is likely not present at all 3-fold positions or is disordered. Between adjacent hexameric capsomeres and between hexameric and pentameric capsomeres are extensions from the capsid that have the appearance of donut-shaped trimers. (C) Fit of bacteriophage HK97 capsid protein core domain into cryo-EM reconstruction of the Basilisk capsid. The top view is from outside the capsid and is centered on one hexamer that is adjacent to the 5-fold symmetry axis. Threefold (triangle), 2-fold (oval), and 5-fold (pentagon) symmetry axes are marked. The bottom two views show slices through the capsid. The cryo-EM density is shown as an orange mesh. Black ribbons designate the nine subunits that form one asymmetric unit of the capsid. Blue ribbons show symmetrically related subunits. (Inset) HK97 capsid protein monomer (46) shown as a ribbon diagram. The N-terminal domain (yellow), cross-linking arm (green), A domain (blue), and P domain (red) are shown. As in a previous study (47), only the A and P domains were used in the fitting into the Basilisk capsid density. Coordinates omitted from the CW02 fitting (yellow, green, light red, and light blue) are indicated.

DISCUSSION

Although more than 100 Bacillus phages have been isolated and fully sequenced, few have been characterized well. Phages are intimately linked to the ecology and evolution of their hosts, making their characterization vital to understanding the diversity and evolution of the Bacillus genus. Herein we described the characterization of two recently isolated Myoviridae phages, JL and Shanette, and a unique siphovirus, Basilisk.

While JL and Shanette were similar to the well-characterized Bacillus phage SPO1, Basilisk was similar only to uncharacterized phages. Our results support the utility of mass spectrometry to identify virion components in such uncharacterized phages, since we were able to detect 15 for Basilisk when only 4 were identified by sequence similarity. Although basic structural components were unrecognized in the Basilisk proteome (including the MCP), cryo-EM was used to reconstruct the capsid of Basilisk, which indicated a T=9 icosahedral capsid structure containing protein subunits with a fold similar to the fold in the HK97 MCP structure (Fig. 6). TEM imaging of Basilisk also revealed two unique tail ends, a thin, needle-like structure or a rounded, donut-shaped end (Fig. 5A). This structural analysis will be helpful as more Basilisk-like phages are identified and characterized, including Staley and Slash, two recently isolated phages that are weakly similar to Basilisk.

Even for JL, a phage similar to the well-characterized phage SPO1, we were able to identify 12 hypothetical, uncharacterized proteins as virion components in need of further characterization (in addition to the 11 with putative function identified from sequence analysis). Future experiments are necessary to determine whether these uncharacterized virion proteins are structural components or virion components with nonstructural roles (such as proteins to facilitate phage entry, protect phage DNA, initiate transcription of phage DNA, or take over host functions).

Several gene products were identified that may be involved in host functions or pathogenesis, including a PhoH-like protein and two that have been shown to contribute to tellurium resistance in other bacteria (TerC and TerD). PhoH is part of the Pho regulon and is expressed in response to phosphate starvation (81). Little is known about its function, but it is similar to the N-terminal domain of superfamily I helicases (81, 82). Interestingly, Pho regulon genes have been found in more than 40% of marine phages but in only 4% of nonmarine phages, with phoH being the most prevalent of these genes (83). A PhoH-like homologue was identified in JL, Shanette, and Basilisk. Further research is necessary to understand the role of this important conserved gene in both bacteria and phages. TerC is required for tellurite resistance in E. coli, and several tellurite resistance genes are known; however, the mechanisms by which these proteins function is largely unknown (84–86). Tellurium oxyanion (TeO₃²⁻) has been used in the treatment of mycobacterial infections, and resistance is a feature of many pathogenic bacteria (79). Additionally, tellurite resistance is commonly used for the identification and isolation of Shiga toxin-producing E. coli (79). This resistance is plasmid-borne in E. coli strains (86, 87), but we were able to find TerC and TerD homologues in a few phages that infect Enterobacteriaceae hosts. A neighbor-joining phylogenetic tree suggests a close relationship between the JL and Shanette TerC and TerD proteins and Bacillus strains, supporting genetic exchange. This identification of tellurite resistance genes in different phages that infect Gram-positive and Gram-negative bacteria suggests that phages are widely shared vehicles for genetic transfer of these virulence factors.

Transfer of virulence genes from phage to host has been repeatedly shown to drive pathogenesis in both Gram-positive and Gram-negative bacteria, including E. coli, Salmonella, Vibrio cholerae and Corynebacterium diphtheriae (for recent reviews, see references 22 and 88). The identification of virulence factors in these three newly isolated B. cereus phages also supports this trend in the genus Bacillus. These virulence factors could easily be carried to not only B. cereus but B. anthracis and B. thuringiensis as well, since all three of our phages are able to infect all three of these hosts.

In addition to being factors that drive B. cereus pathogenesis, phage may provide a solution for the treatment of pathogenic bacteria. Recently, several B. cereus-infecting phages have been characterized as candidates for treatment of food (89, 90); however, some of these are broad-host-range phages, capable of infecting Bacillus subtilis as well (89). The isolation and characterization of narrow-range phages, such as our newly isolated phages, may provide a more targeted approach. In addition, our results urge caution and careful characterization of phages for treatment due to their ability to carry and transfer putative pathogenic genes to Bacillus bacteria, as well as to other phages.