Spontaneously Induced Prophages in Lactobacillus gasseri Contribute to Horizontal Gene Transfer

Abstract

Lactobacillus gasseri is an endogenous species of the human gastrointestinal tract and vagina. With recent advances in microbial taxonomy, phylogenetics, and genomics, L. gasseri is recognized as an important commensal and is increasingly being used in probiotic formulations. L. gasseri strain ADH is lysogenic and harbors two inducible prophages. In this study, prophage ϕadh was found to spontaneously induce in broth cultures to populations of ∼10⁷ PFU/ml by stationary phase. The ϕadh prophage-cured ADH derivative NCK102 was found to harbor a new, second inducible phage, vB_Lga_jlb1 (jlb1). Phage jlb1 was sequenced and found to be highly similar to the closely related phage LgaI, which resides as two tandem prophages in the neotype strain L. gasseri ATCC 33323. The common occurrence of multiple prophages in L. gasseri genomes, their propensity for spontaneous induction, and the high degree of homology among phages within multiple species of Lactobacillus suggest that temperate bacteriophages likely contribute to horizontal gene transfer (HGT) in commensal lactobacilli. In this study, the host ranges of phages ϕadh and jlb1 were determined against 16 L. gasseri strains. The transduction range and the rate of spontaneous transduction were investigated in coculture experiments to ascertain the degree to which prophages can promote HGT among a variety of commensal and probiotic lactobacilli. Both ϕadh and jlb1 particles were confirmed to mediate plasmid transfer. As many as ∼10³ spontaneous transductants/ml were obtained. HGT by transducing phages of commensal lactobacilli may have a significant impact on the evolution of bacteria within the human microbiota.

INTRODUCTION

Lactic acid bacteria (LAB) are used extensively in foods because they are generally recognized as safe (GRAS) organisms that produce natural preservatives, such as lactic acid and other fermentation by-products (1). Species of LAB are commensal bacteria of the human gastrointestinal, oral, and vaginal tracts and are also utilized as starter cultures in dairy, vegetable, wine, and meat fermentations and as probiotics (2). Probiotics are “live microorganisms, which when administered in adequate amounts confer a health benefit on the host” (3). Potential mechanisms of action include competitive exclusion, such as that shown against diarrheagenic Escherichia coli (4), maintenance of intestinal epithelial barrier function (5), metabolic and antimicrobial effects, enhancement of a balanced microbiota, functional modulation of signal transduction (6), and immunomodulation of innate/adaptive host immunity (7).

Genomic sequences of several lactic acid bacteria have revealed multiple temperate phages and phage remnants residing within the genomes (8). The sequence of Lactobacillus gasseri ATCC 33323 (9), an autochthonous bacteria of the gastrointestinal tract (GIT) and a dominant Lactobacillus species of the mouth, GIT, and vaginal microbiota, revealed two adjacent tandem copies of prophage LgaI as well as other nonfunctional phage remnants (10). Potentially, phage transduction in L. gasseri could significantly mediate horizontal gene transfer (HGT) in mucosal populations.

Few transduction systems have been discussed in Lactobacillus species (11–13). The first high-frequency plasmid transduction system in lactobacilli was developed utilizing L. gasseri bacteriophage ϕadh (14). Transduction efficiency was increased by insertion of specific ϕadh genomic fragments into the transduced plasmid (14). In another study, the ϕadh cos fragment was introduced into a plasmid transduction system, enabling the transduction of HIV coreceptor antagonists (CC chemokines), which were then highly expressed in L. gasseri ADH (15). A high-frequency transduction system has been described in industrial strains of Lactobacillus delbrueckii subsp. lactis and subsp. bulgaricus using L. delbrueckii phage LL-H (12).

Although numerous bacteriophages have been well characterized in probiotic microbes, the spontaneous induction, lytic host range, and range of transduction of temperate phages of Lactobacillus have not been thoroughly investigated. Spontaneous induction has been studied in Lactococcus lactis (16, 17). In this study, we report that L. gasseri ADH harbors two distinct temperate phages residing within its genome: the previously characterized linear 43.8-kb cos-type Siphoviridae phage ϕadh (13, 14, 18, 19) and a newly discovered linear 38.3-kb Myoviridae pac-type phage, jlb1, closely related to the previously identified phage LgaI of L. gasseri ATCC 33323 (9, 20). The two temperate phages of L. gasseri ADH were characterized and functionally investigated for lysogenic properties and transduction capability.

MATERIALS AND METHODS

Bacteria, phage, and plasmids.

The bacteria, phages, and plasmids used in this study are listed in Table 1. L. gasseri ADH and its derivatives encode the uncharacterized native plasmid pTRK15. Frozen stock cultures were maintained at −20°C in MRS broth (Becton, Dickinson and Company, Sparks, MD) with 12.5% glycerol. Lactobacillus broth cultures were grown under aerobic conditions statically at 37°C in MRS (pH 6.5). Agar plates (1.5% agar), with or without soft overlays (0.5% agar), were incubated anaerobically at 37°C. In some assays of phage susceptibility, agar plates were incubated aerobically. The agar media used were MRS (pH 6.8) and Lactobacillus selection medium (LBS [pH 5.5]) (Becton, Dickinson and Company), supplemented when appropriate with 10 mM CaCl₂, 7 μg/ml chloramphenicol (Cm), 200 μg/ml gentamicin (Gen), 25 μg/ml rifamycin (Rif) 300 μg/ml spectinomycin (Spc), or 1 mg/ml streptomycin (Str). DiversiLab fingerprinting as shown in Table 1 was performed on a number of L. gasseri strains following the manufacturer's instructions (bioMérieux Clinical Diagnostics, Durham, NC).

TABLE 1.

Bacteria, phages, and plasmids used in this study

Bacterial straina	Relevant characteristicsb	Similarity groupc	Origin or reference
L. gasseri ADH (NCK099) and derivatives
NCK099	ϕadh+ jlb1+(pTRK15)	<98% similarity	ATCC 19992, VPI6033
NCK100	ϕadh+ jlb1+(pTRK15)	ND	13, 14, 33, 34
NCK102	ϕadh− jlb1+(pTRK15)	ND	13
NCK102a	ϕadh− jlb1+(pTRK15, pTRK170)	ND	This study
NCK110	ϕadh+ jlb1+(pTRK15) Rifr Genr	ND	34
NCK374	ϕadh+ jlb1+ Spcr Strr(pTRK15, pTRK170) (Cmr)	ND	14
Other L. gasseri
NCK1338 (WD19)		Group 1	35
NCK1340 (AM1)		<98% similarity	35
NCK1341 (JK12)		ND	35
NCK1342 (JG141)		<98% similarity	35
NCK1343 (SD10)		<98% similarity	35
NCK1344 (FR2)		Group 1	35
NCK1345 (FR4)		Group 2	35
NCK1346 (ML1)		<98% similarity	35
NCK1347 (ML3)		Group 2	35
NCK1348 (RF14)		Group 1	35
NCK1349 (RF81)		Group 1	35
NCK1557 (S-B)		ND	36
NCK2140 (UFVCC 1083)		<98% similarity	37
NCK2141 (UFVCC 1091)		<98% similarity	37
NCK334 (ATCC 33323)		<98% similarity	14
VPI 6033 (ATCC 19992)		ND	14
VPI 11089		ND	14
VPI 11759		ND	14
VPI 12601		ND	14
L. acidophilus
MS01		ND	14
NCK90		ND	14

^aNCK, Culture Collection of the Department of Food, Bioprocessing, and Nutrition Sciences, North Carolina State University, Raleigh.

^bϕadh⁺, ϕadh lysogen; jlb1⁺, jlb1 lysogen; Rif^r, rifampin resistance (25 μg/ml); Gen^r, gentamicin resistance (200 μg/ml); Spc^r, spectinomycin resistance (300 μg/ml); Str^r, streptomycin resistance (1 mg/ml); Cm^r, chloramphenicol resistance (10 μg/ml).

^cGroups have 98% or higher similarity among the strains by DNA fingerprinting. ND, not determined.

Induction of prophages using MC.

L. gasseri NCK100 or L. gasseri NCK102 (ϕadh-cured derivative of NCK100) cells were propagated in 10 ml MRS broth to an optical density at 590 nm (OD₅₉₀) of 0.35 to 0.50. The cells were collected by centrifugation and resuspended in 10 ml of fresh MRS broth, and 100 μl was inoculated into 10 ml of fresh MRS broth. Once the culture reached an OD₅₉₀ of 0.1, mitomycin C (MC) Sigma, St. Louis, MO) was added at concentrations between 0.1 and 0.5 μg/ml, and aerobic incubation continued for 18 h at 37°C in the dark. After 18 h, the culture and phage lysates were centrifuged, and the cell-free supernatants were filtered using a 0.45-μm-pore membrane and stored at 4°C.

Sequencing and analysis.

Genomic DNA from overnight L. gasseri ADH (NCK099) was extracted using a PowerMicrobial Midi DNA isolation kit (MoBio Laboratories, Inc., Carlsbad, CA) according to the manufacturer's protocol and sequenced in the Genome Sciences Laboratory at North Carolina State University (http://research.ncsu.edu/gsl). A library was prepared according to the manufacturer's instructions and sequenced with the Genome Sequencer FLX (Roche Applied Science, Indianapolis, IN) using 454 sequencing with the GS FLX Titanium chemistry. Coverage for the ADH genome was approximately 50-fold. Contigs containing phage sequences were annotated using GAMOLA (21). Additional sequence analysis was performed with BLASTN and BLASTP (http://blast.ncbi.nlm.nih.gov/Blast.cgi) and Mauve genome alignment software (22). Optical mapping was performed by OpGen, Inc., Gaithersburg, MD.

Phage susceptibility screening.

The susceptibility of 16 L. gasseri strains (Table 2) to phages ϕadh and jlb1 was initially screened by spotting filtered induced phage lysates onto cell lawns on MRS and LBS plates, at pHs of 6.8 and 5.5, respectively. For each L. gasseri strain, 200 μl of cells at an OD₅₉₀ of 0.5 was mixed with 3 ml of soft overlay agar and poured over base agar plates supplemented with 10 mM CaCl₂. Once the seeded top overlay solidified, phage suspensions (5 μl of phage dilutions at 10⁰, 10⁻¹, and 10⁻²) were applied directly to the overlay. Duplicate plates were incubated both aerobically and anaerobically at 37°C for 24 h. Double-layer plaque assays were performed on strains that exhibited zones of clearing in the spot assays, as described previously (13).

TABLE 2.

Host ranges of ϕadh and phage jlb1

L. gasseri strain	Host rangea
	ϕadh		jlb1
	Spot test	PFU/ml	Spot test	Plaque formation
NCK100 ϕadh+ jlb1+	−	−	−	−
NCK102 ϕadh− jlb1+	+++	2.4 × 108	−	−
NCK1338	−	−	−	−
NCK1340	−	−	−	−
NCK1341	−	−	+	−
NCK1342	−	−	+	−
NCK1343	−	−	+	−
NCK1344	−	−	+	Indistinct
NCK1345	−	−	++	−
NCK1346	−	−	−	−
NCK1347	−	−	++	−
NCK1348	−	−	++	−
NCK1349	+	4.8 × 103	+++	Indistinct
NCK1557	−	−	−	−
AJ02	−	−	−	−
AJ10	−	−	−	−

^a+, zone of lysis or inhibition in the bacterial cell lawn; −, no apparent lysis or inhibition.

Induction of transducing particles.

NCK374 (L. gasseri ADH containing pTRK170) and the ϕadh-cured derivative NCK102 (containing pTRK170) (Table 1) were propagated, and prophages were induced as described above with the following modifications: the concentrations of MC were 0.3 μg/ml for NCK374 and 0.5 μg/ml for NCK102 (containing pTRK170). Eighteen hours after MC induction, cultures were centrifuged and then filtered using a 0.45-μm-pore membrane, and the supernatants were stored at 4°C.

Transduction assays.

Transduction of pTRK170 to L. gasseri recipient cells was carried out as described previously by Raya and Klaenhammer (14). Experimental controls were recipient cells without phage, phage without cells, and cells plus phage with DNase (Sigma) (70 μg/ml), to rule out transformation.

Plasmid purification.

Extraction of pTRK170 from overnight-grown cells was performed using the Qiagen QIAprep spin miniprep kit (Valencia, CA) with the following modifications: 1 ml of L. gasseri cells was pelleted for plasmid extraction; 3 μl of 30 U/ml mutanolysin (Sigma) and 3 mg lysozyme (Sigma) in 250 μl of buffer P1 were added to the pelleted cells, and the tubes were vortexed and incubated at 37°C for 15 min before continuing the manufacturer's protocol.

PCR verification of pTRK 170.

An 861-bp fragment containing the chloramphenicol gene of pTRK170 was amplified with flanking primers cmF (5′-GCGGCCGCAACTAAAGCACCCATTAGTTC-3′) and cmR (5′-ACTAGTAGTACAGTCGGCATTATC-3′) using the Choice Taq Blue DNA polymerase PCR kit (Denville Scientific, Inc.) with standard protocols.

Phage purification.

A 2-liter portion of phage lysate was treated with 1 μg of DNase I (Sigma) and 1 μg of RNase A (Sigma) per ml for 30 min at 37°C. Cell debris was removed by centrifugation (7,000 × g for 10 min at 4°C). Phage were concentrated with 10% polyethylene glycol 8000 (Fisher Scientific) and 0.5 M NaCl as described by Yamamoto et al. (23). Phage particles were purified through CsCl discontinuous density gradients (1.3, 1.5, and 1.7 g/cm³) by ultracentrifugation for 3.5 h at 35,000 rpm in a Beckman SW40 rotor (24). Phage bands were extracted and dialyzed twice (6,000- to 8,000-Da membrane; Spectrum Laboratories, Inc.) against 2 liters of phage buffer (20 mM NaCl, 10 mM MgCl₂, 20 mM Tris hydrochloride [pH 8.0]) at 4°C.

Electron microscopy of phage.

Negative staining of bacteriophage was performed using 400-mesh Ladd Formvar-carbon grids. Grids were treated with 10 μl bacitracin (50 mg/ml) for 1 min. Bacitracin was removed by blotting with filter paper and rinsing with one drop of distilled deionized H₂O. Ten microliters of purified phage sample was applied to the grid for 1 min, and the excess sample was removed using a filter paper. The grid was stained with 10 μl 2% aqueous uranyl acetate (pH 4.5). Photographs were taken with JEOL JEM 1200EX transmission electron microscope at a magnification of 55,000× at 80 kV.

Monitoring spontaneous ϕadh production.

Overnight cultures (10 ml) of NCK100 were centrifuged, and the pellets were resuspended in 10 ml of fresh MRS and then inoculated into 800 ml of MRS in a pH-controlled fermentor at an OD₅₉₀ of ∼0.01. Fermentations were performed under anaerobic gas sparging at 37°C with controlled pHs of 5.5, 6.5, 6.75, and 7.0. Uncontrolled pH fermentations were performed in batch cultures of 40 ml MRS incubated anaerobically at 37°C. Fermentation samples were taken every hour, and the, pH, OD₅₉₀, and PFU/ml were determined.

Spontaneous phage-mediated transduction.

Ten-milliliter overnight cultures of donor strain NCK374, an Spc^r Str^r derivative of L. gasseri ADH containing pTRK170 (Cm^r), and the recipient strain, NCK110, a Rif^r Gen^r derivative of L. gasseri ADH, were centrifuged and then resuspended in 10 ml of fresh MRS, and both were used to equally coinoculate 800 ml of MRS to an OD₅₉₀ of ∼0.01. Fermentations were carried out at pH (5.5) under anaerobic conditions at 37°C. Fermentation samples were collected every 2 h. The OD₅₉₀ of the coculture, PFU/ml, total CFU/ml (MRS plates), and CFU/ml spontaneous transductants (Cm^r, Gen^r, and Rif^r) were determined for each sample. Transductants were confirmed by PCR of the Cm^r gene of pTRK170, and PCR products were visualized on agarose gels.

Nucleotide sequence accession number.

The GenBank accession number for the sequence of phage jlb1 is KF767351.

RESULTS

Prophage induction of NCK100.

Figure 1A shows the lysis of L. gasseri ADH (NCK100 cells following the addition of 0, 0.1 μg, 0.2 μg, and 0.5 μg of MC per ml to a log-phase culture). Consistent cell lysis was induced with 0.2 μg/ml MC added to early log-phase cultures at OD₅₉₀ 0.1. Within 4 to 5 h after the addition of MC to the cultures, the increase in cell density stopped, and a continuous decrease in OD₅₉₀ was observed until nearly complete lysis of the culture after 22 h. The PFU/ml of ϕadh was determined on the phage-sensitive indicator strain NCK102 (ϕadh-cured derivative), using LBS agar plates supplemented with CaCl₂ (pH 5.5) (13). Titers of ϕadh ranged from ∼10⁸ to 10¹⁰ PFU/ml. The addition of 0.2 μg/ml MC consistently caused complete lysis of the culture and was used for all subsequent prophage inductions of the ϕadh lysogen NCK100.

FIG 1.

Mitomycin C (μg/ml) induction of bacteriophage from the ϕadh lysogen NCK100 (A) and from the ϕadh-cured derivative of L. gasseri NCK102 (B). The data shown are representative of several replicate experiments.

Host range of ϕadh.

Table 2 shows 16 L. gasseri strains, previously isolated from human fecal or endoscopy samples, which were evaluated as potential hosts for ϕadh. Initial susceptibility screening was performed by spotting ϕadh dilutions onto L. gasseri lawns on MRS and LBS agars incubated both aerobically and anaerobically, to investigate conditions for individual strain susceptibility. As expected, strain NCK102 (ϕadh-cured derivative) showed complete cell lysis and clearing where the ϕadh lysates had been applied. No other strains exhibited clearing, except NCK1349, which showed incomplete clearing (Table 2). Plaque assays were performed using each of the 16 L. gasseri strains on MRS and LBS agars under both aerobic and anaerobic conditions. Optimal conditions for plaque formation on NCK102 were on LBS (pH 5.5) overlay and plates, incubated aerobically, as described previously (13). In contrast, NCK1349 indicator cells showed visible plaque formation on MRS overlay (pH 5.5) and base agars (pH 6.8) under anaerobic conditions. The highest PFU/ml for ϕadh on NCK1349 was 4.8 × 10³ PFU/ml (Table 2), with an efficiency of plaquing (EOP) of 2 × 10⁻⁵ relative to NCK102 (EOP, 1.0). The host range of ϕadh on the 16 L. gasseri strains under the assay conditions was narrow, with only NCK102 and NCK1349 showing susceptibility.

Prophage induction of NCK102.

In this study, another prophage appeared to be induced with MC from early-log-phase cultures of NCK102 (Fig. 1B). Within 4 h after the addition of 0.5 μg/ml MC, the increase in cell density of NCK102 ceased, and a continuous decrease in OD₅₉₀ was observed, until nearly complete lysis of the culture after 22 h. These results indicated the presence of a second inducible prophage in NCK102, designated jlb1. Efforts to isolate a prophage-cured derivative of NCK102 via previous strategies (13) were unsuccessful.

Morphology of phage jlb1.

When examined by transmission electron microscopy, jlb1 samples induced from NCK102 showed intact Myoviridae phage particles, empty capsids, broken tails, and sheaths, while ϕadh phage induced from NCK100 are clearly Siphoviridae phage (Fig. 2A, ϕadh, and B, jlb1). The dimensions of phage jlb1 match the dimensions of LgaI (20) and identical phage ϕgaY, which each have an isometric capsid with a diameter of approximately 63 nm and a contractible tail of about 180 nm in length with a diameter of about 20 nm (Fig. 2B). Tail fibers are visible, attached to the distal end of the tail sheath. The sheath of phage jlb1 does not extend up to the capsid, resulting in a phage “neck” clearly present between the capsid and the tail sheath. In the morphological characterization of LgaI, phages with both contracted and noncontracted sheaths were reported (20). In this study, all observed jlb1 phages appeared to have a contracted tail sheath with an average length of 62 nm between the sheath and the phage capsid. The overall length of ϕadh has been described as 460 nm, consisting of a long ∼400-nm tail, no sheath, and few tail fibers and with a capsid diameter of 62 nm (13). The transmission electron microscopy (TEM) photographs show the extreme morphological differences between the Siphoviridae phage ϕadh and the Myoviridae phage jlb1 that reside within the genome of L. gasseri ADH (Fig. 2).

FIG 2.

Negatively stained TEM images of the temperate phages that reside in L. gasseri NCK100, taken at a magnification of 55,000×. (A) Siphoviridae phage ϕadh. (B) Myoviridae phage jlb1.

Host range of phage jlb1.

With the clear demonstration of intact and distinct phage particles induced from the ϕadh-cured NCK102, the potential host range of phage jlb1was determined against the 16 L. gasseri strains. The similarity of 14 strains were determined by DNA fingerprinting (Table 1), and they were classified into three groups based on whether they had >98% similarity (group 1 [4 strains] and group 2 [2 strains]) or had <98% similarity (group 3 [8 strains]). L. gasseri strains NCK1341, NCK1342, NCK1343, NCK1344, NCK1345, NCK1347, NCK1348, and NCK1349 showed various degrees of susceptibility to spot-plated NCK102 lysates (Table 2). The pH of the phage jlb1 lysates (pH range of 5.2 to 6.1) was adjusted to 6.5, to verify that zones of inhibition were not due to acid inhibition. Plaque assays using each of the 16 L. gasseri strains with phage jlb1 resulted in no definitive plaque formation. However, based on spot inhibition tests, the potential host range of phage jlb1 appeared much broader than that of ϕadh (potentially 8 hosts for jlb1 compared to 2 for ϕadh). Interestingly, zones of inhibition, likely due to lysis from without, occurred on some strains representing all three groups: groups 1 and group 2 and the group with <98% similarity. The lawns of NCK1344 and NCK1349 (both group 1 strains) had indistinct nonenumerable plaques on aerobically incubated MRS (pH 5.5) overlay and base agar (pH 6.8). In the absence of a sensitive indicator strain, the concentration of infective jlb1 particles in NCK102 lysates could not be determined.

Sequence analysis of ϕadh and phage jlb1.

A draft genome sequence of L. gasseri ADH (NCK099) was obtained (Roche 454 sequencing), and two contigs were identified that contained bacteriophage-related genes. One of the contigs contained the complete sequence of ϕadh and the other the complete sequence of phage jlb1 (38,269 bp). Using optical mapping, the jlb1 contig was localized to its position in the L. gasseri ADH genome and joined to upstream and downstream contigs with PCR and sequencing. The repeat sequence TGGTCTCCTAAACCGTAGAGGTGAGTTCGAATCTCACCGGGGTCA was identified as the potential att sequence. It occurs within an Arg tRNA upstream of the integrase gene, jlb1_001, and downstream of jlb1_054. The same sequence was identified as the core att region in L. gasseri JCM 1131 phage ϕgaY (25). Fifty-four ORFs were identified covering 90% of the sequence of jlb1 (see Table S1 in the supplemental material). A schematic map of the positions of the open reading frames (ORFs) and identified genetic features is shown in Fig. 3A. The sequence of jlb1 was found to have a G-C content of 37.4% and to be a mosaic of regions highly homologous to several other L. gasseri and L. johnsonii temperate phages. Importantly, jlb1 was closely related but distinct from LgaI (9) and represents a new temperate phage discovered from L. gasseri.

FIG 3.

(A) Map of the open reading frames and known features of phage jlb1. (B) Alignment of genomes of phages KC5a, jlb1, LgaI, Lj771, and ϕadh (from top to bottom). Similarly colored blocks are regions of local DNA colinearity that are homologous and internally free from genomic rearrangement. Small rectangles below each genome are open reading frames. Colored graphs within the blocks show the similarity profiles of the genome sequences. Areas that are completely white were not aligned and probably contain sequence elements specific to a particular genome.

DNA-level alignment with the most homologous Lactobacillus gasseri phages is shown in Fig. 3B. A high level of homology exists with phages Lj771 (26), KC5a (S. Pavlova and L. Tao, unpublished data), and LgaI (20). Only a very small region around ORFs 24 and 25 was unique to jlb1. Phage ϕadh has very limited DNA homologies with any of these phages.

The amino acid sequence of each jlb1 open reading frame was compared with the amino acid sequences of the related phages using directed BLASTP searches. Homology with ϕadh sequence occurred only with ORFs 15, 16, 18, and 54, encoding three hypothetical proteins and lysin, respectively.

In general, the first 12 ORFs of jlb1, which control phage integration and retention, are highly homologous to phage KC5a and LgaI ORFs. ORFs 15 to 27, which are largely of undefined function, but most likely involved in phage replication, share more homology with Lj771 ORFs. Most of the packaging and structural genes from jlb1 ORF 31 to the end of the genome are highly related across all four genomes. Phage jlb1 proteins which are highly conserved among all of the 4 phages are represented by ORFs 19 (Holliday junction protein), 28 (portal protein), 34 to 36 (undefined function), and 38 to 48 (structural proteins, including the tail proteins and baseplate, as well as undefined late genes), and 51 (undefined function).

There are only 3 of 54 jlb1 proteins that have less than 50% amino acid identity with an ORF of at least one of the closely related phages. These ORFs are carried in regions of jlb1 which lack DNA homology with the four closely related phages. The unique proteins are encoded by ORFs 16, 24, and 25. ORFs 16 and 24 have no defined functions. ORF 25 codes for a putative ParB nuclease with 53% amino acid identity to a prophage methylase from Acidaminococcus intestini CAG:325.

Phage LgaI has the greatest number of homologous proteins with jlb1: there are 43 homologous proteins with over 50% identity and 39 with over 90% homology. The first seven ORFs of jlb1 are highly homologous to LgaI ORFs. These include an ORF coding for the phage integrase and ORFs 3 and 4, which potentially encode a “toxin-antitoxin” system. Alternatively, the ORF 3 and 4 genes may be involved in regulating lytic gene expression, as they occur just upstream of the cI gene (27). Another pair of potential “toxin-antitoxin” genes occurs in jlb1 ORFs 12 and 13. In this case, ORF 12 has an XRE DNA binding protein domain and codes for a predicted antitoxin. ORF 13 has an ANT antirepressor domain and its product is the predicted toxin component. The two ORFs overlap slightly. Phage jlb1 ORFs 18 (undefined function), 19 (coding for a putative Holliday junction protein), and 23 (undefined function) and the majority of the structural genes or hypothetical genes in the second half of the genome have homologues in LgaI. However, the DNA packaging regions of jlb1 and LgaI differ significantly in genes that encode the endonuclease (ORF 24), small subunit terminase (ORF 26), and large subunit terminase (ORF 27).

Phage jlb1 shares significant homologies with KC5a, having 40 homologous ORFs with over 50% amino acid identity and 28 with 90% or more identity. Phage jlb1 ORFs 2 to 11 share 97 to 100% amino acid identity with homologues in KC5a. These include putative proteins associated with lysogeny and replication: ORF5 (100% identity), encoding the putative cI repressor, and ORF11 (100% identity), encoding a putative replication protein. In addition, the majority of the late genes are homologous, with the exception of the lysin gene: ORFs 43, 44, 45, 46, 47, and 51, have over 90% identity, ORFs 48, 49, and 52 (holin) have 75 to 79% identity, and ORFs 50, 53 (lysin), and 54, have no significant homology.

Phage jlb1 shares at least 50% homology with 39 proteins of Lj771 and greater than 90% homology with 22 proteins. Again, as for phage LgaI, the majority of the proteins are structural or hypothetical and are found in the second half of the jlb1 genome. Phage jlb1 ORFs 15, 17, 18 (all of unknown function), and 19 (putative Holliday junction protein) are also close homologues of Lj771 proteins.

Spontaneous induction of ϕadh.

Plaque assays against NCK102 (ϕadh cured) using filter-sterilized supernatants from NCK100 (ϕadh lysogen) cultures, sampled at different growth phases (early log, mid-log, and stationary), showed that ϕadh was being spontaneously induced during normal growth from its lysogenic host at concentrations of ∼10⁶ to 10⁷ PFU/ml at the stationary phase. To determine if pH conditions have an impact on the rate of spontaneous production of ϕadh, NCK100 was grown under pH-controlled and pH-uncontrolled conditions, and phage ϕadh production was monitored (Fig. 4).

FIG 4.

(A) Growth of the ϕadh and jlb1 lysogen NCK100 in MRS broth under controlled-pH and batch culture conditions. (B) Spontaneous induction of ϕadh in broth medium maintained at various pHs over time. The data shown are representative of several replicate experiments.

Changes in pH conditions had a significant impact on the growth rate of NCK100 (Fig. 4A). In an uncontrolled pH batch fermentation, NCK100 encountered pH changes ranging from pH 6.7 at 0 h, pH 6.2 at 4 h, and pH 5.0 at 8 h to pH 4.5 at 12 h. During this fermentation, the cells quickly began to clump and pellet by 9 h. Under controlled pH levels, NCK100 grew fastest at pH 6.5 (OD₅₉₀ of 1.3 by 6 h), with a long lag phase, but began to grow quickly after it had acclimated to the low pH. The growth of NCK100 at pH 7.0 was stunted. It was concluded that NCK100 grows well at pH 6.5 and reasonably well at pH 5.5 but was not well adapted for growth at a neutral pH of 7.0.

NCK100 cells were harvested from broth cultures and resuspended in fresh medium prior to inoculation of the fermentor, in order to reduce any background of phage that might be introduced by the initial inoculum. Nevertheless, the initial ϕadh concentrations ranged from 6.0 × 10¹ to 1.3 × 10³ PFU/ml at inoculation. The concentrations of ϕadh in an uncontrolled pH batch fermentation continuously increased as the NCK100 population increased (Fig. 4B). At a controlled fermentation of pH 5.5, ϕadh reached a concentration of 6.7 × 10⁷ PFU/ml, and phage production closely mirrored that of the uncontrolled batch culture (Fig. 4B). During the pH 6.5 fermentation, a steady concentration of ϕadh was observed, with phage concentrations reaching 1.2 × 10³ to 3.6 × 10³ PFU/ml from 4 to 9 h. At 10 h and 11 h, the phage concentration significantly increased to as high as 1.9 × 10⁷ PFU/ml; however, the phage concentration by 12 h plunged to 3.7 × 10³ PFU/ml. Although less pronounced, this phenomenon of a ϕadh production spike at 10 h followed by a reduction in concentration by 12 h was also observed at pH 6.75. These results clearly indicated that ϕadh was spontaneously produced during normal growth of NCK100. Overall, low-pH conditions either at a controlled pH of 5.5 or in batch culture, stimulated ϕadh induction. The cause for the spike in ϕadh production at 10 h and 11 h under higher-pH conditions is unknown.

Transduction of pTRK170 into lactobacilli by ϕadh.

The ability of ϕadh to transduce plasmid pTRK170 (14), containing two BglII fragments of ϕadh (13, 14), was evaluated against 16 L. gasseri strains. Plasmid pTRK170 was successfully transduced via ϕadh into 12 strains: NCK100, NCK102, NCK1338, NCK1340, NCK1342, NCK1343, NCK1346, NCK1347, NCK1349, NCK1557, AJ02, and AJ10. Plasmid DNA extraction and PCR amplification of an 861-bp region of the Cm^r gene of pTRK170 verified that transductants from each of the 12 L. gasseri strains did contain the plasmid (data not shown), while recipient strains showed no amplicon. Transduction assays performed using DNase-treated lysates did not impact the transduction frequencies, ruling out transformation events.

The frequencies of plasmid transduction differed greatly among the L. gasseri recipient strains, with surprisingly NCK100 (ϕadh lysogen) having the highest transduction frequency of 1.6 × 10⁻³ transductants/PFU. (Approximately 1.6 per 1,000 PFU are transducing phage containing replicating pTRK170.) The results of this study combined with those of Raya et al. (14) showed successful transduction of pTRK170 via ϕadh into 17 L. gasseri strains and 2 L. acidophilus strains (Table 3), despite the relatively narrow range of susceptible lytic hosts for ϕadh. These results indicate that bacterial susceptibility to phage lysis was not in and of itself an indicator of “transducibility” or the ability of a recipient strain to take up and express horizontally transferred genes via phage. In our hands, efforts to transduce plasmid pGK12 with the ϕadh cos site fragment (15) from induced lysates of L. gasseri ADH to known transduction recipients defined herein and previously by Raya et al. (14) were unsuccessful after numerous attempts (data not shown).

TABLE 3.

Transduction of plasmid pTRK170 in lactobacilli^a

Recipient strain	No. of Cmr transductants
	ϕadh (per PFU)	jlb1 (per ml)b
L. gasseri
NCK100	1.6 × 10−3
NCK102	4.4 × 10−5
NCK1338	1.3 × 10−6
NCK1340	1.6 × 10−6
NCK1341	<1.5 × 10−9
NCK1342	2.8 × 10−5	3.4 × 101
NCK1343	1.9 × 10−7
NCK1344	<1.5 × 10−9
NCK1345	<1.5 × 10−9
NCK1346	5.0 × 10−5
NCK1347	2.1 × 10−5	5.0 × 101
NCK1348	<1.5 × 10−9
NCK1349	5.2 × 10−5	5.0 × 103
NCK1557	2.2 × 10−8
AJ02	3.7 × 10−6
AJ10	1.4 × 10−6
NCK334 (ATCC 33323)	4.7 × 10−7
VPI 6033 (ATCC 19992)	3.8 × 10−4
VPI 11089	4.6 × 10−7
VPI 11759	3.6 × 10−8
VPI 12601	8.6 × 10−8
L. acidophilus
MS01	1.4 × 10−8
NCK90	1.3 × 10−7

^aResults for strains not transducible by ϕadh are denoted in boldface.

^bPFU of jlb1 could not be determined.

Transduction of pTRK170 into lactobacilli by phage jlb1.

The ability of jlb1 to transduce genes to recipient L. gasseri strains was also investigated by inducing phage jlb1 from the ϕadh-cured derivative harboring jlb1 and pTRK170. Plasmid pTRK170 contains two ϕadh fragments (fragment 1, partial phage portal protein; fragment 2, conserved hypothetical protein 5) that are not homologous with jlb1. Nevertheless, low numbers of transductants were observed in recipients NCK1342, NCK1347, and NCK1349, with 3.5 × 10¹, 5.0 × 10¹, and 5.0 × 10³ transductants per ml, respectively. These results indicated that jlb1 is a transducing phage capable of transferring genes to other L. gasseri recipient strains, even though these strains were not sensitive indicators of jlb1.

Spontaneous phage-mediated transduction in coculture.

Donor strain NCK374 (ϕadh⁺ jlb1⁺ Spc^r Str^r Cm^r ADH derivative containing pTRK170) and recipient NCK110 (ϕadh⁺ jlb1⁺ Rif^r Gen^r ADH derivative) were grown in a batch coculture fermentation in MRS (pH 5.5) to determine if spontaneous phage-mediated transduction occurs via ϕadh and/or jlb1 (Fig. 5). Fermentation samples were collected every 2 h to monitor the total CFU/ml, PFU/ml of ϕadh, OD₅₉₀, and transductants (Cm^r Rif^r Gen^r). The spontaneous production of ϕadh increased quickly as the L. gasseri ADH population increased. The concentration of ϕadh spontaneously induced from NCK110 and NCK374 reached its peak at 12 h at 1.1 × 10⁶ PFU/ml. This compares to 6.7 × 10⁷ PFU/ml of free ϕadh observed for the parent L. gasseri NCK100 when grown in pure culture under the same pH-controlled conditions.

FIG 5.

Coculture experiment using L. gasseri NCK374 Cm^r (containing pTRK170) and transduction recipient L. gasseri NCK110 (Rif^r and Gen^r at 25 μg/ml and 200 μg/ml, respectively). Every 2 h, samples monitored the PFU/ml of ϕadh, total cell count, and the number of spontaneous transductants/ml (L. gasseri Rif^r Gen^r Cm^r). The data shown are representative of several replicate experiments.

Spontaneous transductants at 2.3 × 10³ spontaneous transductants/ml were observed at 12 h (Fig. 5), with a frequency of 6.4 × 10⁻⁶ transductants (Cm^r Rif^r Gen^r CFU/Spc^r Str^r CFU) observed. Although spontaneous transduction was shown to occur at a relatively low frequency, these results demonstrated that ϕadh, jlb1, or possibly both can spontaneously transduce pTRK170 to recipient L. gasseri strains in coculture. Therefore, temperate phages residing within commensal lactobacilli are likely to be induced spontaneously and transduce genes to other closely related species occupying the same ecological niche.

DISCUSSION

In this study, we demonstrated that L. gasseri ADH (NCK100) contains a second inducible temperate phage, jlb1, that shares a high degree of sequence homology with prophage LgaI found in L. gasseri ATCC 33323 (8, 20). L. gasseri ATCC 33323 and L. gasseri JCM1131 are different designations for the same strain (25); therefore, their resident phages, ϕgaY (28) and LgaI (29), are presumably identical. The second prophage in NCK100 (now designated phage jlb1) was previously described as a defective (noninducible) LgaI phage by NheI restriction digest patterns and Southern blotting using phage LgaI DNA as a probe against NCK102 genomic DNA (20). In contrast, we have shown that jlb1 can be induced from early-log-phase cells by using a higher concentration of mitomycin C than that required for induction of ϕadh. The annotated sequence of jlb1 showed that the majority of structural genes, like many other L. gasseri and L. johnsonii phages, are genetically identical to LgaI, while the two phages have distinct differences in genes associated with lysogeny and capsid DNA packaging.

Mitomycin C-induced lysates of LgaI from L. gasseri ATCC 33323 contain intact phage particles as well as empty capsids, broken tails, and contracted sheaths (20). Purified LgaI intact phages clearly belong to the Myoviridae phages. Electron micrographs of jlb1 consistently show a single phage morphology that consists of a phage sheath that does not touch the phage capsid but remains at the distal end of the phage particle near the tail fibers. Differences in purification techniques can impact the morphology of phages; however, sequence analysis showed a significant sequence inversion in the open reading frame of the minor tail protein of jlb1 that may alter or truncate the minor phage tail proteins.

The genome of L. gasseri ADH harbors two inducible intact phages that have significant differences in genomic sequence and organization, morphological structure, and host range. The cos-type Siphoviridae phage ϕadh has many genes that are unique and shares only limited sequence homology to jlb1. Distinct differences were observed in the host susceptibility ranges of jlb1 and ϕadh against the 16 L. gasseri strains exposed to the phage lysates. While ϕadh was able to completely lyse the ϕadh-cured strain (NCK102) and, to a lesser extent, NCK1349, jlb1 was shown to produce zones of spot lysis on eight L. gasseri strains. The ability of jlb1 to impact the growth of these strains does not indicate that the phage successfully infected and lysed these bacteria; however, it does suggest that the phage was at least able to successfully attach to the bacterial cell walls, potentially causing lysis from without.

The ϕadh-based transduction system using pTRK170 was found to effectively transduce pTRK170 into 12 of the 16 L. gasseri strains used in this study. Combined with previous work using ϕadh to transduce pTRK170 (14), 17 L. gasseri strains and 2 L. acidophilus strains have been transduced using ϕadh. Using pTRK170 (despite no sequence homology between jlb1and the plasmid), jlb1 was also shown to be a transducing phage that could transfer pTRK170 into 3 L. gasseri strains. It must be noted that because NCK100 harbors both ϕadh and jlb1, phage jlb1 could also be contributing to the transduction range and frequency of induced phage that are spontaneously released from L. gasseri ADH (NCK100).

The ability of silent temperate phage to be spontaneously induced into the lytic cycle has previously been observed in Salmonella and E. coli (30) and was recently observed in Lactobacillus paracasei (31). This study showed that ϕadh is spontaneously induced during normal growth conditions in NCK100. The spontaneous production of ϕadh was significantly impacted by the pH of the growth medium. Lower-pH conditions were shown to consistently produce large concentrations of phage.

This study demonstrated that spontaneous production of ϕadh can produce transducing particles able to transfer genes. Despite the broad transduction host range of the ϕadh-pTRK170 transduction system within strains of L. gasseri species, the phage has not transferred the plasmid outside this species (14). Horizontal gene transfer is an important mechanism of evolutionary change, but recent analysis shows that prokaryotes are more likely to bias the transfer of genetic material to closely related species than distantly related species (32).

In order to gain better understanding of the human microbiome and the evolutionary changes occurring in the gastrointestinal tract (GIT), it is important to study temperate phages capable of high rates of spontaneous induction and able to acquire bacterial genes and transduce them to related strains. Future in vitro studies should be performed to determine the population dynamics of phages within the mammalian GIT over time and the frequency at which transduction events occur. It is clear that for the species L. gasseri, a commensal of the mouth, GIT, and vagina, temperate phages are common and likely serve as key vehicles for horizontal gene transfer among commensal lactobacilli.