Characterization of Listeria prophages in lysogenic isolates from foods and food processing environments

Prophages are commonly found in Listeria genomes, potentially enhancing survival or fitness of Listeria spp. Currently, there is still limited information on the distribution of prophages among Listeria isolates of different allelic types and from various sources. In this study, by using mitomycin C induction, prophages were found in 23/144 isolates (16.0%), including 13 L. monocytogenes and 10 Listeria spp. isolates, resulting in 28 and 11 induced phages, respectively. These prophage-carrying isolates (lysogens) were obtained from foods and food-related environments presenting 3 common allelic types (ATs) of L. monocytogenes (lineage I, II and IV), 4 ATs of L. innocua and 1 AT of L. welshimeri. The likelihood of prophage-carrying isolates of L. monocytogenes was 14.4 (95% CI: 4.9–35.4), and 18.5 (95% CI: 4.8–50.2) for Listeria spp. The 39 induced phages were classified into 3 lysis groups by the host range test against 9 major serotypes of L. monocytogenes and 5 species of Listeria. Most phages were host-specific with higher ability to lyse L. monocytogenes serotype 4 than other serotypes. The genome size of phages ranged from 35±2 kb to 50±2 kb and belonged to two common phage families, Myoviridae and Siphoviridae. Restriction analysis classified 19 selected phages into 16 restriction profiles, suggesting highly diverse prophages with at least 16 types. This may contribute to the variation in the genomes of Listeria. Information obtained here provides basic knowledge for further study to understand the overall role of prophages in Listeria, including roles in survival or fitness in foods and food processing environments.


Introduction
Listeria monocytogenes is an important foodborne pathogen that can cause listeriosis-a serious foodborne illness with mortality rate up to 30% [1]. The genus Listeria comprises 17 species. Several new species have been discovered in the past decade, for example, L. marthii from soil [2]; L. fleischmannii and L. weihenstephanensis in cheese and water, respectively [3,4]. Previous studies have reported that various types of foods and food processing environments can be contaminated with Listeria spp., including L. monocytogenes [5,6]. L. monocytogenes showed a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 hosts were subjected to serotype classification by multiplex PCR following the protocol of Doumith et al. [27] combined with the classification of sigB allelic types.

SigB allelic typing
Typing of a partial sigB gene has been used for allelic type classification and species confirmation of Listeria isolates in previous studies [5,28,29]. Of the 144 Listeria isolates used for prophage induction in this study, 60 have been assigned an allelic type (AT) previously [5,6]. SigB allelic typing was performed to 84 Listeria isolates in this study. The protocol of Nightingale et al. [28] was followed for PCR amplification of a partial sigB gene (780 bp) using the same primers. The purified PCR products were sequenced at Macrogen Inc. (Seoul, South Korea). Allelic type classification was performed using the SigB allelic type database of all species of Listeria (kindly provided by Prof. Martin Wiedmann and Dr. Renato H Orsi, Cornell University, Ithaca, New York). A phylogenetic tree was generated by MEGA5 program [30] using sequences of the allelic types found in this study with some closely related allelic types in the database mentioned above. The maximum likelihood method with gamma distribution was used for constructing the tree with 1,000 bootstrap replications [29,31].

Induction of Listeria prophage by mitomycin C and phage lysate preparation
The culture of an isolate was prepared by inoculating an isolated colony in 5 ml of Luria Bertani (LB) broth (Oxoid, UK) supplemented with 50 mM morpholinepropanesulfonic acid (MOPS), 1% (wt/vol) glucose, 10 mM CaCl 2 , and 10 mM MgCl 2 (LB-MOPS-Glu-salts) [26]. The culture was incubated at 30˚C (220 rpm) to reach an optical density (at 600 nm) of 0.4 to 0.5, and a 1 ml-aliquot was mixed with mitomycin C (Sigma-Aldrich, St Louis, USA) to a final concentration of 1 μg/ml (modified from [32]). The mixture was subsequently incubated for 7 h. Then, 200 μl of this mixture was later mixed with 100 μl of a given propagating host in a total volume of 2 ml of LB MOPS, followed by an incubation for 18 h at 30˚C (220 rpm). The double layer technique was applied with the filtered lysate of the overnight co-culture following the procedure described by Vongkamjan et al. [26]. Plaque formation was observed as appearance of induced phage. An isolated plaque representing a distinct plaque morphology type (A: translucent plaque, rather round shape, F � 1 mm; B: turbid at the edge, star-shape, F = 0.5-1 mm; C: turbid zone, F � 0.5 mm; D: clear zone, round, F = 1 mm; E: turbid zone, tiny, F � 0.2 mm) was selected for three-time-purification, followed by phage lysate preparation by the double layer method [26]. Titers of induced phages were determined by spotting 5μl of ten-fold serial dilutions of phage on the propagating host lawn. High titer of phage lysate was kept at 4˚C for further analysis. The likelihood of prophage-carrying isolates among L. monocytogenes and Listeria spp. was determined as odd ratio with 95% confidence interval

Host range determination of the induced phages
Phage host range determination was performed by spotting 5 μl of diluted phage representing 100×RTD (routine test dilution) [26], approximately 10 6 -10 7 PFU/ml, on the 31 Listeria hosts mentioned above. Each spot on the lawn was examined and recorded for lysing (+) or no lysing (-) after overnight incubation at 30˚C [26]. The experiment was carried out in triplicate. Then, a clustering analysis based on lysis ability of the induced phages against the tested hosts was performed using R program.

Estimation of phage genome size
Pulsed-Field Gel Electrophoresis (PFGE) analysis was used to estimate genome size of the phages as described previously [25,26,33]. Briefly, high-titer lysate of a given phage (10 7 -10 9 PFU/ml) was used to prepare a plug with 1.3% low-melting-point agarose. PFGE analysis was performed on a CHEF-DR III system (Bio-Rad, Hercules, CA, USA) in 0.5×TBE buffer (1 M Tris, 0.5M EDTA and boric acid, pH 8.0) for 22 h with a 0.5 s to 5.0 s switch time, 6 V/cm, and an included angel of 120˚ [26]. Genome estimation was performed using Uvitec UVI-1D software (Uvitec Limited Co., Cambridge, UK) with the tool for molecular weight estimation.

Restriction enzyme analysis
A total of 19 phages representing each genome size group, but from different lysis groups and sources were selected for restriction enzyme analysis. DNA of the induced phages was extracted by phenol/chloroform as described previously [26]. Restriction analysis was performed using Fast digest HindIII (Thermo Fisher Scientific, MA, USA) and EcoRI (Vivantis, Selangor Darul Ehsan, Malaysia), following the manufacturers' instructions. Two restriction profiles were considered different when at least one distinguishing band was present [32].

Transmission electron microscopy (TEM)
Examination of the phage morphology and family was performed with four induced phages selected to represent a given genome size group. A 3 μl-drop of freshly prepared lysate of a given phage (

Classification of allelic types based on Listeria partial sigB sequences
Overall, a total of 144 L. monocytogenes and Listeria spp. isolates were subjected to partial sigB sequencing analysis to classify them into allelic types (Table 3 and Fig 1). Classification of these isolates showed 12 allelic types, which represented lineages I, II or IV of L. monocytogenes, or other species, including L. innocua and L. welshimeri. In the 90 L. monocytogenes isolates used in this study, five allelic types were observed. Of which, AT 58, AT 60 and AT 74 were commonly found in the majority of isolates. Allelic types 58 and 60 represented L. monocytogenes lineage I while AT 74 represented L. monocytogenes lineage IV. In the 54 Listeria spp. isolates, AT 111 and AT 30 were common allelic types found in this study; these isolates could be classified as L. welshimeri and L. innocua, respectively.

Distribution of prophages in Listeria isolates obtained from foods and food-related environments
Mitomycin C could induce the prophages in 23/144 (16.0%) of L. monocytogenes and Listeria spp. isolates (Table 3). These prophage-carrying isolates belonged to 8 out of 12 allelic types found among the tested isolates (Table 3 and Fig 1). Based on the presence of distinct types of plaque morphology based on the size, shape and the turbidity of each examined plaque, 39 inducible phages were obtained from 23 lysogenic

Lysis ability of the induced phages against 31 Listeria hosts
The host range of 39 induced phages was evaluated on 19 Listeria reference strains and 12 Listeria isolates from the collection of those isolates used in this study. These hosts were selected to represent 9 major L. monocytogenes serotypes and 5 distinct Listeria species. Clustering analysis based on the lysis similarity classified these 39 phages into 28 lysis profiles, presenting 3 major lysis groups (A, B and C) (Fig 2). Each lysis group included 10-17 induced phages. Group B contained those phages (n = 12) that were host-specific with the ability to lyse only 1-5 host strains (<16%) of L. monocytogenes serotype 4 and L. marthii. Phages (n = 17) in group A showed similar host range as those in group B (as host-specific phages), however, they could also lyse Mack (1/2a) and L. welshimeri. In comparison, phages (n = 10) in group C had a broader host range than those in groups A and B. These phages could lyse 10-18 hosts (32-58%) of L. monocytogenes serotype 4 and other Listeria species used as hosts, except L. seeligeri.

Estimated genome size of the induced phages
Genome estimation of the induced phages by PFGE classified them into four groups (Table 4), including group 1 (35±2 kb, 14 phages), group 2 (40±2 kb, 16 phages), group 3 (45±2 kb, 4 phages), and group 4 (50±2 kb, 5 phages). Most phages were in groups 1 and 2. Group 1 contained phages induced from lysogens of L. monocytogenes lineage I and the only lysogen of L. monocytogenes lineage IV, while group 2 contained phages induced from lysogens of all L. monocytogenes lineages and Listeria species found in this study (except L. monocytogenes lineage IV). In comparison, groups 3 and 4 included phages induced from lysogens of L. monocytogenes lineage I and L. innocua. Interestingly, many Listeria lysogens harbored different induced phages with different genome size representing different prophage types in their genomes.

Restriction analysis of the induced phages
Restriction enzyme analysis using HindIII was performed with 19 representative phages, resulting in 16 restriction profiles (H 1 to H 16 ) (Fig 3). In each genome size group, 2-7 different restriction profiles were obtained. Seven different profiles were obtained in phage genome size group 2 (40±2 kb). The profiles H 3 (group 1) and H 6 (group 2) were both observed in two phages from the lysogens of L. monocytogenes lineage I obtained from the same source. For group 2, profile H 4 was found in two phages (LP009 and LP029) from the lysogens of L. welshimeri obtained from different sources. Two and four distinct restriction profiles were observed in phages of genome groups 3 and 4, respectively. Phages with the same HindIII restriction profiles (H 3 , H 4 , or H 6 ) were further analyzed by enzyme EcoRI. Results showed that two phages with identical HindIII profiles had identical EcoRI profiles (E 1 , E 2 or E 3 ). In summary, by using two restriction enzymes, these 19 phages could be classified into 16 restriction profiles, suggesting 16 different prophage types. Interestingly, up to seven prophage types were obtained from the lysogens of L. monocytogenes lineage I (AT 58). In addition, three and four prophage types were obtained from the lysogens of L. welshimeri (AT 111) and L. innocua (AT 11, AT 22, AT 30, AT 31), respectively.

TEM analysis of the induced phages
Morphology of four phages selected to represent the four genome size groups was examined by TEM (Fig 4). Morphology observation showed that three phages belonged to the Myoviridae family with an isometric head (diameter 56 to 63 nm) with long contractile tail (167 to 228 nm) with a sheath. These phages were from the lysogens of L. monocytogenes lineage I or L. innocua with the genome sizes 35±2 kb, 45±2 kb and 50±2 kb. Another phage was classified to the Siphoviridae family with a hexagonal head (diameter of 64 nm) and longer non-contractile tail (239 nm). This phage was from the lysogen of L. welshimeri and had genome size 40±2 kb.

Distribution of prophages among L. monocytogenes, L. innocua, and L. welshimeri obtained from foods and food-related environments
A total of 39 induced phages were obtained from the 23/144 isolates of L. monocytogenes and Listeria spp. tested (16%). Prophages were detected in most allelic types found in this study (8/ 12) of L. monocytogenes lineage I, II and IV, L. innocua, and L. welshimeri. No isolates of L. monocytogenes in the tested collection belonged to lineage III, therefore the results lack information on the prophage distribution from isolates of this lineage.
Prophages were found to be absent or rare in five specific allelic types representing L. monocytogenes lineages I, IV and L. innocua. There is still limited information to explain why those allelic types are lack of presence of prophage. However, we speculate that the isolates of these lineages are particularly resistant to uptake of extraneous DNA as suggested previously [35]. Another possible reason could be the lack of prophage insertion sites for these lineages as opposed to L. monocytogenes lineage I and II [17,18]. Interestingly, finding here may link to the host range data; L. monocytogenes serotype 1/2 and 3 (mostly belonged to lineages I, II) are resistant to phages as those hosts could be lysogens containing prophage sequences that are homologous to the induced phages [36]. Another possibility could be appropriated propagating host in the induction as mentioned in previous studies [14,37]. This hypothesis was supported when different sets of prophages were obtained using different propagating hosts, even if those hosts represented serotype 4a (FSL F2-695 and FSL J1-208).  (Fig 3A). EcoRI-restriction profiles of induced phages that showed similar restriction profiles by HindIII (Fig 3B). a "M" is a 1-kb molecular maker. Induced phages obtained from a single lysogen are marked with the same symbol next to the phage ID. Phages within a box had the same restriction pattern. b Refer to Table 1  In addition, our study showed the likelihood of having prophages in L. monocytogenes isolates as 14.4 (95% CI: 4.9-35.4) and 18.5 (95% CI: 4.8-50.2) for the isolates of Listeria spp. (including L. innocua and L. welshimeri). It seems reasonable as L. innocua and L. welshimeri were reported that derived from L. monocytogenes through early evolutionary events of gene acquisition by phage transduction [38]. Therefore, isolates of these Listeria species may have more insertion site for prophages to incorporate into the host chromosomes. In addition, in this study, different prophage types were found among Listeria lysogens or even within a single lysogen. Similarly, genome analysis has previously revealed multiple prophages in Listeria genomes, especially in L. monocytogenes and L. innocua [17][18][19] Induced Listeria phages appear to be host-specific with higher ability to lyse L. monocytogenes serotype 4 than other serotypes Induced Listeria phages in this study showed different lysis profiles with 74% of these represented host-specific phages. A previous study revealed that lysogenic Listeria phages have narrower host range than isolated Listeria phages [39]. The majority of Listeria phages isolated from a turkey processing plant showed broad host ranges [40]. In related bacteria, phages obtained from lysogenic strains of Streptococcus iniae were reported to have narrow lytic spectrum [41]. This can be explained by the unique characteristic of induced phages as they have the ability to incorporate their genome in the host chromosome instead of lysing the host.
In this study, most induced phages could lyse the hosts of L. monocytogenes serotype 4. Similarly, Listeria phages from silage or turkey processing environments were highly susceptible to L. monocytogenes serotype 4 strains [26,40]. Moreover, the induced phages could not lyse L. monocytogenes serotype 1/2 and serotype 3 hosts. This is of interest since serotype 1/2a is linked to the increasing cases of listeriosis in the last decade [42,43]. Therefore, we speculated that prophages may facilitate the survival of Listeria hosts. Another potential support is that the differences in phage susceptibility between serotypes of L. monocytogenes can be explained by the different in structure of cell wall teichoic acids (WTA). This is because L. monocytogenes serotype 4 contains WTA with terminal glucose and galactose residues, which is important for phage adsorption [44,45] and further facilitation for phages to lyse the host.
Most Listeria species used as hosts for the host range determination were sensitive to the induced phages, except L. seeligeri. This may be because the L. seeligeri strain is a lysogen. We also observed that the induced phages were likely to be resistant to 10/31 lysogenic Listeria hosts. This supports the hypothesis that lysogens could have phage resistance in certain bacterial hosts, because the bacteria may harbor specific (pro)phage sequences that could increase their survival without being affected by phage with homologous sequence as reported in Oenococcus oeni [36]. However, sequencing analysis of the induced phage and its host is still needed to elucidate the phage resistance characteristics.
Induced phages show highly similar genome size as previously reported temperate Listeria phages, but rather high genetic diversity and belong to two common phage families The induced phages in this study showed genome size ranged from 35±2 kb to 50±2 kb and were classified into four groups. Similarly, genomes of previously reported temperate Listeria phages also had sizes from 35 kb to 48 kb [46][47][48]. Temperate phages typically have smaller genomes, that may contain only the necessary sequences for their replication and encapsulation [49,50]. Only the basic, important genes coding for six main functional modules were present in the genomes of temperate phages [47,51]. However, genomes of lytic phage contained a number of additional gene coding sequences with no function [34].
Restriction analysis of 19 selected phages using enzymes HindIII and EcoRI resulted in 16 restriction profiles, suggesting considerable diversity of prophages in the genome of Listeria lysogens. The diversity of prophages may contribute to the variation of host genomes by prophage incorporation as mentioned in a previous study [18]. Another study has also shown that the differences in a 42-kb-prophage sequence could differentiate four examined L. monocytogenes strains [52]. In this current study, seven and three prophage types were found in a single allelic type (AT 58 and AT 111, respectively). This suggests the usefulness of prophage to the classification of Listeria isolates of the same allelic types.
Two common families Myoviridae and Siphoviridae were observed among the induced phages in this study. LP014 belonged to Siphoviridae family with long, non-contractile tails, which is the most common among phages (60%) [53]. The morphological characteristics revealed that our Siphoviridae phage was similar to the Listeria phages LP-032-2 and LP-032-3 with a head diameter of 53-55 nm and a tail length of 160-297 nm reported previously [34].
In summary, this is the first study that investigated the distribution of phages induced from Listeria isolates of different allelic types of distinct L. monocytogenes lineages or Listeria species. Characterization of induced phages allows us to better understand their diversity and lysis ability against Listeria hosts representing different L. monocytogenes serotypes and distinct Listeria species. Diversity of prophage may have contributed to the genetic diversity of Listeria spp. isolated from foods and food-related environments in Thailand. Recombination and mosaicism caused by prophages in Listeria genomes may occur and gene transfer may be affected and could later drive host survival and fitness in foods or food-associated environments.