A Virulent Phage Infecting Lactococcus garvieae, with Homology to Lactococcus lactis Phages

Abstract

A new virulent phage belonging to the Siphoviridae family and able to infect Lactococcus garvieae strains was isolated from compost soil. Phage GE1 has a prolate capsid (56 by 38 nm) and a long noncontractile tail (123 nm). It had a burst size of 139 and a latent period of 31 min. Its host range was limited to only two L. garvieae strains out of 73 tested. Phage GE1 has a double-stranded DNA genome of 24,847 bp containing 48 predicted open reading frames (ORFs). Putative functions could be assigned to only 14 ORFs, and significant matches in public databases were found for only 17 ORFs, indicating that GE1 is a novel phage and its genome contains several new viral genes and encodes several new viral proteins. Of these 17 ORFs, 16 were homologous to deduced proteins of virulent phages infecting the dairy bacterium Lactococcus lactis, including previously characterized prolate-headed phages. Comparative genome analysis confirmed the relatedness of L. garvieae phage GE1 to L. lactis phages c2 (22,172 bp) and Q54 (26,537 bp), although its genome organization was closer to that of phage c2. Phage GE1 did not infect any of the 58 L. lactis strains tested. This study suggests that phages infecting different lactococcal species may have a common ancestor.

INTRODUCTION

Lactococcus garvieae is the etiological agent of lactococcosis, a pathology affecting a variety of fish species that is responsible for significant economic losses in both marine and freshwater aquacultures (1). In addition, L. garvieae has been associated with subclinical intramammary infections in cows (2). More recently, L. garvieae was associated with infections in humans and is now considered an opportunistic and potentially zoonotic pathogen (3, 4). In an effort to find its ecological niches, genetic investigations have revealed its presence, and sometimes dominance, in several habitats: river and sewage waters and raw and processed foods, including milk, meats, and vegetables (5–7).

Despite growing attention, genomic characterization of L. garvieae progresses slowly, yet the evolutionary history and global diversity of this emerging pathogen are particularly interesting. The population structure and diversity of L. garvieae strains have been studied using comparative genome analysis, which led to the development of a multilocus sequence type (MLST) scheme to identify specific strains (8). These studies also revealed that L. garvieae is related to Lactococcus lactis, a nonpathogenic bacterium used for the manufacture of several fermented dairy products (9). Divergent genomic lineages were observed, with some ecotypes representing evolutionary intermediates between L. lactis and L. garvieae (8). The presence of insertion sequences (ISs) was also highly variable among L. garvieae strains (10). Analysis of these ISs within individual genomes revealed limited intragenomic sequence diversity, suggesting that many ISs in L. garvieae are evolutionarily young and may have been only recently acquired. Of note, the similarity of most L. garvieae ISs to L. lactis ISs suggests possible genetic exchange between these two species.

Little is known about the presence of other mobile genetic elements such as phages in L. garvieae. Phages are usually found in the same niches as their bacterial hosts and can have two main life cycles, lytic or lysogenic. During the lytic cycle, the phage redirects the host metabolism to produce new virions, resulting in host cell lysis and the release of numerous virulent phages that can infect other targeted cells. During the lysogenic cycle, the phage genome integrates into the host genome, remains in the host as a prophage, and replicates along with the host genome, until the lytic cycle is induced by stress factors. Phages play key roles in bacterial evolution, influencing the dynamics of bacterial populations and their adaptability through gene exchange (11, 12).

The emergence of multidrug-resistant bacteria has sparked renewed interest in virulent phages for potential use in therapy or in sanitation strategies in the agricultural, food processing, and fishing industries as well as in humans and animals (13, 14). However, phages can also destroy bacteria that play key roles in production or fermentation processes, particularly in the food and biotechnology industries (15, 16).

Very few lytic phages of L. garvieae have been identified to date (17–19). One L. garvieae phage belonging to the Siphoviridae family (double-stranded DNA [dsDNA] genome, long noncontractile tail) was investigated for controlling L. garvieae infections in fish (19). To our knowledge, only one L. garvieae-virulent phage has been investigated at the genomic level (20). L. garvieae phage WP-2 belongs to the Podoviridae family (dsDNA genome, short noncontractile tail) of the order Caudovirales (12). It was classified as a member of the Ahjdlikevirus genus in the Picovirinae subfamily because it has a small genome (18,899 bp, 24 open reading frames [ORFs]) and a protein-primed DNA polymerase (20).

Here, we report the molecular characterization of the novel L. garvieae phage GE1. This virulent phage was isolated from a Canadian compost soil in 2014 and belongs to the Siphoviridae family. Comparative genome analyses suggest that this new siphophage is related to virulent phages infecting L. lactis.

MATERIALS AND METHODS

Bacterial host and phage isolation.

A total of 73 L. garvieae strains previously isolated from different ecological niches (fish and dairy farms, vegetables, meat, and cereals) (21) and 58 industrial L. lactis strains were used to determine the host range of the phage. Lactococcal strains were grown statically at 30°C in M17 broth (Oxoid) supplemented with 0.5% glucose (GM17) or lactose (LM17). We tested Canadian samples of raw milk, wastewater, soil compost, and aquaculture water to isolate phages infecting L. garvieae. The soil compost from which we isolated phage GE1 was made from green waste (domestic compost) obtained from the Compo-Recycle Center at Chertsey (Québec, Canada), and the samples used for phage isolation were taken from the top.

All samples were kept at 4°C until used, and the same phage isolation protocol was used (22). Five milliliters of each environmental sample (for the solid compost soil, 5 g was resuspended in 5 ml of sterile water) was added to 5 ml of double-strength GM17 supplemented with 10 mM CaCl₂ and inoculated with 100 μl of an overnight culture of an L. garvieae strain. Nineteen representative strains of previously studied L. garvieae (8) were used for lytic phage isolation. The mixtures were incubated overnight at 30°C. The cultures were centrifuged at 7,000 rpm for 10 min using an IEC clinical centrifuge (USA), and the supernatants were filtered (0.45-μm cellulose acetate filter). Five milliliters of the filtered supernatant was added to double-strength GM17 inoculated as described above. This second amplification procedure was repeated three more times. The presence of phages in the supernatants was tested by spot test, depositing 5 μl of the last amplification on a GM17 plate with 10 mM CaCl₂ containing a single L. garvieae strain. Clear plaques were picked, purified three times, and amplified as described previously (23). To obtain concentrated phage preparations, 1 liter of lysate was mixed with polyethylene glycol 8000 (10%) and separated on a discontinuous CsCl gradient followed by a continuous CsCl gradient, as previously reported (24).

Microbiological assays.

The host range of phage GE1 was assessed by spotting 5 μl of a 10⁻² dilution of a phage lysate on top agar containing one lactococcal strain. One-step growth assays were performed in triplicate with a multiplicity of infection (MOI) of 0.05 and the host strain L. garvieae INS1 (21). The burst size was calculated by dividing the average phage titer after the exponential phase by average titer before the infected cells began to release virions (25).

Electron microscopy.

Ten microliters of uranyl acetate (2%) was deposited on a Formvar carbon-coated grid (200 mesh; Pelco International). Cesium chloride-purified phages (10 μl) were mixed with the stain by pipetting up and down. After 2 min, the grid was deposited on blotting paper and dried for at least 5 min. Phages were observed at 80 kV using a JEOL1230 transmission electron microscope at the Plateforme d'Imagerie Moléculaire et Microscopie of the Université Laval. Capsid size and tail length were determined by measuring at least 10 phage specimens.

Phage DNA extraction, sequencing, and genome analysis.

Phage DNA was extracted from 500 ml of lysate using the Qiagen Lambda Maxi DNA purification kit with modifications as described previously (26). To determine the genome sequence of phage GE1, a sequencing library was first prepared with the Nextera XT DNA sample preparation kit (Illumina) according to the manufacturer's instructions. The library was sequenced using a MiSeq reagent kit (v2) on a MiSeq system (Illumina). De novo assembly was performed with the Ray assembler (27). Genome extremities were amplified using converging primers, and the PCR product was sequenced with an ABI 3730xl sequencer at the sequencing platform of the Centre Hospitalier de l'Université Laval. The cos site was also confirmed by direct sequencing of the phage DNA using forward (5′-GCAAGGAGGTAATCAGATGCA-3′) and reverse (5′-GAACGCATTCTGTGAGCTTG-3′) primers.

Open reading frame (ORF) prediction was carried out using ORF Finder (http://www.ncbi.nlm.nih.gov/gorf/gorf.html) and RAST Server (28). An ORF was considered valid only if its starting codon was AUG, UUG, or GUG and it was at least 30 amino acids (aa) in length. Ribosomal binding sites (RBSs) were also identified. Functions and domains were attributed by comparison of the translated products using BLASTp (29). PSI-BLAST and InterProScan (http://www.ebi.ac.uk/) were used to search for more distant homologous proteins and conserved domains, respectively, when significant similarity was found by BLAST searches. Theoretical molecular masses (MM) and isoelectric points (pI) of the phage proteins were obtained using ProtParam (http://web.expasy.org/protparam/). tRNAs were identified using the tRNAscan-SE server (30) and confirmed using ARAGORN (31). Virulence Factor Databases (32) together with DBETH (33) were used to identify virulence factors.

Structural protein profile.

Purified GE1 and c2 phages (∼10¹¹ PFU/ml) were dialyzed (0.02 M Tris-HCl, pH 7.4, 0.1 M NaCl, 0.1 M MgSO₄), mixed with 4× loading buffer (0.25 M Tris-HCl, pH 6.8, 40% [vol/vol] glycerol, 8% [wt/vol] SDS, 20% [vol/vol] β-mercaptoethanol, and 0.1% [wt/vol] bromophenol blue), and boiled for 5 min. Proteins were separated on an Any kD Mini-Protean TGX precast protein gel (Bio-Rad) and stained with Coomassie blue.

Proteomic tree design.

The predicted proteome of L. garvieae phage GE1 was compared with other lactococcal phage proteomes, including the L. garvieae phage WP-2 (20) and members of the 10 L. lactis phage groups (34). The genomes were downloaded from NCBI, and the proteins were extracted using the published annotations. The terminase was used as the starting point to generate a long phage protein concatemer for each phage (35 phages in total). The order of the proteins was identical to the gene order presented in the genome. For the two L. lactis phage genomes (KSY1 and P087) in which a terminase was not found, the protein concatemers were constructed according to the gene order as deposited in GenBank. MEGA5 (35) software and a BLOSUM matrix were used for multiple alignments. An unrooted phylogenetic tree was constructed using the neighbor-joining algorithm (36). The phylogeny was tested with 100 bootstrap replicates.

Nucleotide sequence accession number.

The annotated genomic sequence of phage GE1 was deposited in GenBank under accession number KT339177.

RESULTS

Characterization of phage GE1.

L. garvieae phage GE1, a virulent phage isolated from a compost soil sample, belongs to the Siphoviridae family (Fig. 1). It has a prolate capsid of 56 nm ± 3 nm by 38 nm ± 3 nm and a noncontractile tail of 123 nm ± 12 nm in length and 11 nm ± 2 nm in width. The phage morphology is highly reminiscent of phages c2 and Q54, members of two distinct L. lactis phage groups that bear their names (34). However, the tail length of phage GE1 is slightly greater (34).

FIG 1.

Morphology and structural proteome of phage GE1. (A) Electron micrograph of phage GE1. (B) Structural protein profiles of phages GE1 and c2. Numbers on the left indicate molecular mass of the markers (Bio-Rad) (lane M).

The burst size of phage GE1 was calculated to be 139 ± 13 virions per infected cell, and its latency period was 31 ± 4 min, similar to that of other lactococcal phages (36). The host range of phage GE1 was assessed using a collection of 73 L. garvieae strains and 58 industrial L. lactis strains. Phage GE1 was highly specific as it infected only two L. garvieae strains, its host INS1 and strain INS2, which was isolated from a similar ecological niche (vegetables) (21). Phage GE1 did not infect any of the 58 L. lactis strains tested.

Genome of phage GE1 and its cos site.

Phage GE1 has a 24,847-bp linear double-stranded DNA genome with a GC content of 37.8%, similar to the GC content of its host genome (8) but slightly higher than that of L. lactis phage genomes (35%) (36). Analysis of the phage GE1 genome did not identify any tRNA or recognizable virulence factors. The genome has cohesive ends consisting of single-stranded 9-bp 3′ overhangs (Fig. 2A). Analysis of the DNA region surrounding the cos site revealed characteristics similar to L. lactis phages belonging to the c2 group (37). This region is characterized by inverted repeat sequences as well as 22 direct repeats. Overall, this region is AT-rich (69%) compared to the rest of the GE1 genome (62%). Three G boxes and no C clusters were found near the cos site (Fig. 2A).

FIG 2.

Analysis of GE1 cos region. (A) Analysis of the cos site and the flanking regions. Direct repeats (bold arrows), inverted repeats (bold, dashed arrows), A/T-rich regions, and G-rich segments are indicated. Details of the cohesive termini (rectangle) are also shown together with the site of cleavage (vertical arrows). Shading indicates nucleotides that belong to the consensus sequence found in all lactococcal phages for which the cos site has been determined. (B) Multiple alignments of λ-like R consensus sequences from Q54, GE1, and c2. Conserved nucleotides are shaded.

The inverted repeat found in the cos site of phage GE1 (GCAA) lacks a conserved T in one of the consensus sequences [TCAN(N_4–7)NACT], typically found within a 15-bp segment spanning the cos site of many L. lactis phages (38). The absence of this conserved base in the consensus cos site is reminiscent of the cos site of lactococcal phage Q54, which has a CCAA inverted repeat.

The cos region of GE1 also contains five λ-like R consensus sequences, which are usually involved in terminase recognition and binding, as well as in packaging termination (39). The sequences were compared with those previously reported for L. lactis phages c2, bIL67 (c2-like), and Q54. As reported for phage Q54 (37), phage GE1's five λ-R sequences are distributed unevenly but across a smaller genomic region (500 bp compared to 900 bp of phage Q54). This contrasts with coliphage λ and lactococcal phage c2, whose λ-R sequences are regularly spaced. Alignment analysis of the R sites from GE1 (Fig. 2B) found bases in common involved in the packaging process, similar to those observed for c2 and Q54.

Analysis of phage GE1 genes.

A total of 48 ORFs longer than 30 codons were predicted from the genome sequence (Table 1). The sizes of the gene products range from 44 (ORF17 and ORF26) to 839 (ORF37) amino acids, with an average predicted protein size of 153 amino acids. Coding regions represent approximately 88% of the genome, and the longest noncoding region is 657 bp, occurring between orf27 and orf28. As for c2-like lactococcal phages, the GE1 genome is organized in two main gene clusters, grouping the presumed early- and late-expressed genes (40, 41). The late-expressed genes code for proteins involved in packaging (orf40), cell lysis (orf29 and orf38), and phage morphology (orf32, orf34, and orf37). No lysogeny module was found in the genome, confirming the lytic nature of phage GE1.

TABLE 1.

Features of the ORFs of L. garvieae phage GE1^b

ORF	Stranda	Position		Size (aa)	MM (kDa)	pI	Putative RBS and start codonc	Predicted functiond	Best-match BLASTp result	No. of aa shared with best match/total no. of aa in best match (% identity)	E value	Sizee (aa)	Accession no.
		Start	Stop
1	−	529	104	141	16.9	6.8	AGGAGActagaATG	—
2	−	678	526	50	6.1	9.8	aAGGAGtgcaagctATG	—
3	−	854	675	59	7.2	9.3	AAGAAAGgcaatgtagtgaaagaATG	—
4	−	1039	842	65	7.7	9.5	AAGAAAGAggtagaaaATG	—
5	−	1295	1029	88	10.1	9.8	AAGAAAGaaaacatctaaaATG	—
6	−	1650	1477	57	6.8	8.7	gAAGAGGtatagaaATG	—
7	−	1949	1653	98	11.1	4.5	AGAGGAttgagATG	—
8	−	2256	1951	101	11.5	4.5	GAGgtGAAaagagcATG	—
9	−	2678	2253	141	16.2	10.1	AGGAGtttatagaccaGTG	HNH endonuclease	ORF31 Lactococcus phage 712	66/137 (48)	3e−32	141	YP_764291.1
10	−	2947	2675	90	10.4	7.9	GAAAGGGGAtatatATG	—
11	−	3588	3073	171	20.0	9.4	AGGAGAactttaATG	—	ORF41 Lactococcus phage 191	54/162 (33)	5e−11	177	AFE86775.1
12	−	3770	3588	60	7.1	4.7	GAAAGGccattATG	tRNA synthetase	ORF6 Lactococcus phage c2	24/52 (46)	0.15	54	NP_043532.1
13	−	4179	3754	141	16.4	9.8	ctgacttgccattttaacATG	HNH endonuclease	ORF48 Lactococcus phage P475	74/142 (52)	3e−36	145	AGI11121.1
14	−	4510	4181	109	11.9	4.7	AAGAGAatataaattATG	ssDNA-binding protein	ORF40 Lactococcus phage 7	49/117 (42)	2e−18	119	YP_008318242.1
15	−	5031	4531	166	18.3	7.8	AGGAAAttgaattcgcATG	ERF recombinase	Rec, S. pseudopneumoniae	58/123 (47)	1e−29	207	WP_000163473.1
16	−	5231	5028	67	7.8	5.2	AGGAGGacaagaagaaATG	—
17	−	5402	5268	44	5.4	6.7	atacAGAGGtggatagATG	—
18	−	5637	5389	82	10.0	9.5	AtgAGGGAtttaagttagATG	—	ORF10 Lactococcus phage c2	31/77 (40)	2e−05	83	NP_043536.1
19	−	5920	5621	99	11.9	4.9	CGTAGAGGActttggcaagtATG	—
20	−	6176	5913	87	10.3	4.6	AGGAGAgtaatATG	—	gp25 Lactococcus phage KSY1	27/60 (45)	3e−07	99	YP_001469023.1
21	−	6531	6352	59	7.1	7.9	AGGttGAGAtattagaATG	—
22	−	7478	6528	316	36.1	5.4	AAGGAGtAaatacaGTG	DNA polymerase	ORF42 Lactococcus phage jm3	166/317 (52)	6e−106	315	YP_008318192.1
23	−	7737	7465	90	10.5	8.9	GGAAAGGctttggattATG	—
24	−	8148	7921	75	9.0	8.9	AGGAGAaaacaaaATG	—
25	−	8355	8218	45	5.4	9.4	AAAGGAcggtgttattATG	—
26	−	8476	8342	44	5.1	9.7	GAGAGGtttagaaaATG	—
27	−	8651	8466	61	7.3	4.7	AGGAGActattatATG	—
28	+	9309	9461	50	5.8	4.5	AGGAGAaaggtttATG	—
29	+	9463	9951	162	18.2	5.2	tactcattacctatttaatATG	Holin	ORF24 Lactococcus phage c2	92/159 (58)	1e−59	161	NP_043550.1
30	+	10496	11167	223	24.9	9.0	AAAGGAAAtaattaaTTG	—
31	+	11177	12109	310	34.7	5.2	AAAtAGAGGtaacaaATG	—
32	+	12096	13667	523	57.6	4.9	GAAGAGtgaggtagaagATG	MCP	MCP Lactococcus phage CB17	130/535 (24)	9e−22	489	AAZ95019.1
33	+	13684	14484	266	28.9	5.2	AAAGAGgtaacttATG	—
34	+	14486	15100	204	22.4	4.7	AAGAAAGGtataatttaatATG	Major tail protein	ORF23 Lactococcus phage Q54	65/200 (33)	1e−21	212	YP_762592.1
35	+	15151	15342	63	6.7	6.0	AGGGGGAAggTTG	Ig-like domain
36	+	15366	15617	83	9.6	5.6	AGGGGActtaatATG	—
37	+	15798	18317	839	87.8	5.5	AGGAAAttaatTTG	Tail protein	Pl10 Lactococcus phage 5447	123/471 (26)	3e−22	622	AAT73602.1
38	+	18331	18606	91	10.2	9.3	AGGGAAaataATG	Holin	Gl17 Lactococcus phage 923	54/85 (65)	1e−27	96	AAT81366.1
39	+	18684	18968	94	10.9	8.3	AGcttgtGGGGGtAtcATG	HNH endonuclease	Gp27 Lactococcus phage Q54	44/101 (44)	4e−16	103	YP_762596.1
40	+	18968	20605	545	61.3	6.6	AtGGGGtGGtaATG	Terminase	ORF28 Lactococcus phage Q54	191/507 (38)	2e−96	514	YP_762597.1
41	+	20605	20901	98	11.4	9.1	AGGGAAattaatctaATG	—	ORF35 Lactococcus phage c2	20/65 (31)	0.005	99	NP_043561.1
42	+	20901	22100	399	46.6	5.5	AAGGAGtttatgtaATG	—
43	+	22100	22759	219	24.9	5.8	GAAAGGttacgggtataatttaATG	—
44	+	22759	23526	255	27.5	4.8	GAAGGGttatgttggtggattaATG	—
45	−	23708	23568	46	4.9	9.5	AGGAGAacatcATG	—
46	−	24237	23797	146	16.4	10.2	GAAAGGtggttatatgtaATG	—
47	−	24410	24237	57	6.6	4.0	AAGcAGGttatataaaagATG	—
48	−	24576	24394	60	6.8	4.7	AGGAAAtatattATG	—

^aOrientation of the gene in the genome.

^bAbbreviations: MM, molecular mass; RBS, ribosomal binding site; ssDNA, single-stranded DNA; MCP, major capsid protein; Rec, recombinase.

^cUppercase letters represent hypothetical RBS sequences (also underlined); bold letters represent starting codons.

^d—, no significant matches.

^eTotal size of the aligned proteins.

Seven ORFs (1, 11, 14, 20, 24, 27, and 28) possess a typical RBS domain (AGGAGA). Of these, orf27 and orf28 are located at the ends of the 657-bp intergenic region between early- and late-expressed genes. Moreover, orf27 (early gene) is also preceded by the consensus −10 and −35 promoter sequences (TTGACA-17 bp-TATAAT), whereas upstream of orf28 (late gene) only a −10 consensus sequence was found. These characteristics suggest that the region between orf27 and orf28 could act as an origin of replication, as already reported for phage c2 (42).

Proteome of phage GE1.

The functions of phage GE1 deduced proteins were determined by comparison (BLASTp) with the GenBank database. When low similarity was found between GE1 proteins and other phage proteins, putative functions were reinforced by identifying conserved domains and/or position-specific iterative comparisons. Only ORFs with a significant protein hit in GenBank are presented in Table 1. Predicted functions were attributed to 14 ORFs (out of 48 ORFs; 29%), and only 17 ORFs (35%) matched proteins in the GenBank database, indicating that new viral genes/proteins were identified by characterizing this novel phage. Details are given below for some ORFs.

Homing endonucleases and recombination genes.

Three ORFs (ORF9, ORF13, and ORF39) were predicted to be homing endonucleases. These ORFs had similarity with deduced proteins found in Lactococcus phages 712 and P745 (43), both 936/Sk1-like phages, as well as with L. lactis phage Q54 (37). Since homing endonucleases promote genetic exchange and horizontal transfer (44), this part of the phage GE1 genome may have arisen through the acquisition of mobile DNA, as already proposed for phage 712 (45). ORF15 and ORF22 had similarity to proteins with putative recombination functions, found in Streptococcus pseudopneumoniae and 936/Sk1-like phage jm3 (43).

Replication.

ORF12 and ORF14 have homology with a glutamyl-tRNA synthase protein found in Lactococcus phage c2 and with a single-stranded binding protein of lactococcal 936/Sk1-like phages, such as phi7 and P113G (43). Glutamate tRNA synthase catalyzes the attachment of the amino acid glutamine to its cognate tRNA molecule in a highly specific two-step reaction (46). Single-stranded binding proteins are required during DNA replication, repair, and recombination (40, 47, 48).

Morphogenesis and lysis.

Phage GE1 structural proteins (ORF32/major capsid protein, ORF34/major tail protein, and ORF37/minor tail protein) and proteins involved in packaging (ORF40/terminase) and lysis (ORF29/holin and ORF38/holin) showed similarities to proteins found in other phages possessing the same morphological features, particularly c2, Q54, CB17, 5447, and 923 (34, 49). ORF35 possesses an immunoglobulin-like domain but did not match with other proteins in databases. Immunoglobulin-like proteins may be involved in host recognition (50). The structural protein profile of phage GE1 was determined by SDS-PAGE (Fig. 1B) and found to be highly related to that of phage c2 (37).

Comparison of prolate-headed phages.

As demonstrated above, phage GE1 has similarities to L. lactis phages c2 and Q54, both prolate-headed phages. The genome organization and proteome of phage GE1 were specifically compared to those of L. lactis phages c2 and Q54. Figure 3 clearly confirms the relatedness of phage GE1 to c2 and Q54 phages. GE1 possesses 13 ORFs that align with those of phage c2 (amino acid identity of 25 to 58%). They are principally involved in replication (six ORFs) and morphogenesis (seven ORFs), and five of these seven proteins were also common to phage Q54 (amino acid identity of 27 to 44%). However, the genome organization of phage GE1 is closer to that of phage c2 than Q54, as illustrated by gene orientation (Fig. 3).

FIG 3.

Genomic comparison of GE1, Q54, and c2 phages. ORFs from GE1 showing similarity with other phages are linked by gray shading, and the percent amino acid identity indicated is representative of the aligned region.

Phylogenetic studies of L. garvieae phages.

Finally, a phylogenetic analysis was performed using L. garvieae phages (GE1 and WP-2) as well as members of the 10 currently recognized groups of L. lactis phages to confirm the relationships described above. Figure 4 shows a proteomic phylogenetic tree constructed using MEGA5 software and the neighbor-joining method. This global comparison revealed that the two L. garvieae phages are distributed among L. lactis phages, indicating that they probably evolved from a common ancestor. The genomic comparison also demonstrated that phage GE1 originates at the same point as other prolate-headed phages, diverging, together with phage Q54, in a separate branch.

FIG 4.

Phylogenetic tree of different lactococcal phage proteomes determined using the neighbor-joining algorithm. The bar shows the difference between the phages in amino acids.

DISCUSSION

To date, L. garvieae phages have been poorly characterized. GE1 is the first L. garvieae siphophage to be characterized at the genomic level. Phage GE1 shares some characteristics with other lactococcal phages, specifically those infecting the food-grade bacterial species L. lactis. Phage GE1 has a prolate capsid (morphotype B2), a narrow host spectrum, and a burst size and latency period similar to those of other L. lactis phages.

Significant research attention has focused on phages infecting L. lactis in recent decades because these phages pose a significant risk to milk fermentation processes (51). The main interest in studying these phages is to understand their diversity and evolution to enable development of efficient phage control tools (52). L. lactis phages are currently grouped into 10 taxonomic groups based on morphology and genome analysis (17), and of these, phages 936/Sk1, c2, and P335 (all Siphoviridae phages) are the most frequently encountered in the dairy industry (53). Over 80 L. lactis phage genome sequences are now publicly available.

L. garvieae strains are found in raw milk and in a range of aquatic and terrestrial environments (21). Some L. garvieae strains may also be major components of the autochthonous microbial populations of certain artisanal cheeses (5). In fact, it is believed that the activity of L. garvieae strains may contribute to the final sensory characteristics of some dairy products (54). The adaptation of L. garvieae to the milk environment was likely enabled by the acquisition of certain plasmids (55). Not surprisingly, L. garvieae phages have similarities to L. lactis phages, also found in dairy environments.

Comparative genome and proteome analyses clearly demonstrated that phage GE1 is related to L. lactis phage c2 and Q54 groups (Fig. 3 and 4). Its genome size of 24,847 bp is also intermediate between those of c2 (22,172 bp) and Q54 (26,537 bp). The GE1 genome has cohesive extremities containing a cos site [GCAA(N₉)AACT], differing from the c2 [TCAA(N₄)AACT] and general lactococcal phage [TCAN(N₄)NACT] consensus cos sites (38) but similar to the phage Q54 cos site [CCAA(N₁₀)AACT]. The regions flanking the cos site contain several features common to other lactococcal phages such as the presence of direct and inverted repeats and an A/T-rich region.

A global analysis of the genomes of phages GE1, c2, and Q54 shows that c2 and GE1 are similarly organized, with two large modules in opposite orientations starting at the probable origin of replication. In phage GE1, an extra module (4 genes) was detected at the end of the late gene cluster, as in phage Q54. The predicted proteome of phage GE1, however, was closer to that of phage c2, with 13 ORFs sharing similarity. In the phage GE1 early-expressed module, five genes (ORFs 9, 11, 13, 14, and 22) code for proteins homologous to those found in L. lactis 936/Sk1-like phages. These latter phages, which also belong to the Siphoviridae family but have a small isometric capsid, are highly abundant in dairy environments (56), suggesting possible recombination events. Members of the rarely isolated Q54 group were proposed to be the result of past recombination events between 936/Sk1- and c2-like phages (37). Another striking feature of phage GE1 is the low level of similarity of its proteome with proteins in existing databases, confirming its novelty.