Phage-host interactions analysis of newly characterized Oenococcus oeni bacteriophages: Implications for malolactic fermentation in wine

Phage-host interactions analysis of newly characterized Oenococcus oeni bacteriophages: Implications for malolactic fermentation in wine

International Journal of Food Microbiology 246 (2017) 12–19 Contents lists available at ScienceDirect International Journal of Food Microbiology jou...

1MB Sizes 0 Downloads 75 Views

International Journal of Food Microbiology 246 (2017) 12–19

Contents lists available at ScienceDirect

International Journal of Food Microbiology journal homepage: www.elsevier.com/locate/ijfoodmicro

Phage-host interactions analysis of newly characterized Oenococcus oeni bacteriophages: Implications for malolactic fermentation in wine Antonella Costantini a, Francesca Doria a, Juan-Carlos Saiz b, Emilia Garcia-Moruno a,⁎ a b

Consiglio per la ricerca in agricoltura e l'analisi dell'economia agraria, Centro di Ricerca per l'Enologia (CREA-ENO), Via P. Micca 35, 14100 Asti, Italy. Departamento de Biotecnología, Instituto Nacional de Investigación y Tecnología Agraria y Alimentaria (INIA), Ctra. Coruna Km 7.5, 28040 Madrid, Spain.

a r t i c l e

i n f o

Article history: Received 4 October 2016 Received in revised form 24 January 2017 Accepted 30 January 2017 Available online 31 January 2017 Keywords: Bacteriophage Oenococcus oeni Host-range Integrase gene sequences Malolactic fermentation Wine

a b s t r a c t Nowadays, only few phages infecting Oenococcus oeni, the principal lactic acid bacteria (LAB) species responsible for malolactic fermentation (MLF) in wine, have been characterized. In the present study, to better understanding the factors affecting the lytic activity of Oenococcus phages, fifteen O. oeni bacteriophages have been studied in detail, both with molecular and microbiological methods. No correlations were found between genome sizes, type of integrase genes, or morphology and the lytic activity of the 15 tested phages. Interestingly, though phage attack in a wine at the end of alcoholic fermentation seems not to be a problem, it can indeed represent a risk factor for MLF when the alcohol content is low, feature that may be a key point for choosing the appropriate time for malolactic starter inoculation. Additionally, it was observed that some phages genomes bear 2 or 3 types of integrase genes, which point to horizontal gene transfer between O. oeni bacteriophages. © 2017 Published by Elsevier B.V.

1. Introduction Malolactic fermentation (MLF), a biochemical process that usually occurs in wine after the alcoholic fermentation, is carried out by lactic acid bacteria (LAB), in particular by Oenococcus oeni that decarboxylates L-malic acid in L-lactic acid and carbon dioxide. MLF confers a deacidification and a greater microbiological stability to the wine. Moreover, because of MLF, many other metabolic reactions occur changing the organoleptic properties of wines, both in aroma complexity and in taste and color (Davis et al., 1985b). MLF can take place spontaneously by the action of LAB population present in wine, or by the use of commercial starters, which is becoming a common oenological practice in wineries to induce and achieve better control of MLF (Bauer and Dicks, 2004). Stuck and sluggish malolactic fermentations are however frequent, and inoculation with commercial starters do not often resolve this problem. Factors influencing the successful onset and completion of malolactic fermentation include ethanol content, pH, SO2 levels, temperature, nutritional requirements, and microbial incompatibility (Davis et al., 1985b). In addition, some authors have also pointed to the presence of phages of O. oeni in wine as a cause of difficulties for MLF. Thereby, Sozzi et al. (1976) first reported the presence of O. oeni

⁎ Corresponding author. E-mail address: [email protected] (E. Garcia-Moruno).

http://dx.doi.org/10.1016/j.ijfoodmicro.2017.01.020 0168-1605/© 2017 Published by Elsevier B.V.

bacteriophages in wine and associated their presence with abnormal MLF, which drove to propose the use of selected starter cultures composed by various O. oeni strains characterized by different phage sensitivity (Arendt et al., 1991; Davis et al., 1985a). However, data are still controversial about the practical implications of bacteriophages in MLF. For instances, it was proposed that the choice of a lysogenic strain used as a malolactic starter culture would be positive to prevent against the attack by phages present in the wine, since lysogenic strains are in general more resistant to phages attack (Poblet-Icart et al., 1998). However, the use of non-lysogenic strains has been also proposed (Torriani et al., 2011). Lysogeny appears to be frequent in O. oeni, 45–67.6% of O. oeni strains are lysogenic (Arendt et al., 1991; Doria et al., 2013; Poblet-Icart et al., 1998), as confirmed by pan-genome comparisons. In fact, several prophage sequences integrated in the O. oeni genome have been described (Borneman et al., 2012). In any case, until now, few studies about O. oeni bacteriophages have been reported, more probably because, under laboratory conditions, O. oeni has a very low growth rate, thus making O. oeni phages detection difficult. Recently, by using a new developed PCR method based on the amplification of the endolysin (lys) gene of the bacterial genome, we isolated a total of 15 O. oeni phages (Doria et al., 2013). Here, the phage–bacterium interactions have been addressed by analyzing the behaviors of different O. oeni strains in the presence of the 15 isolated phages, as well as the effect that pH and ethanol have on the phage

A. Costantini et al. / International Journal of Food Microbiology 246 (2017) 12–19

13

Table 1 Host range test (HRT) conducted in MRSTJ pH 4.5 with 16 strain of O. oeni and 15 phage lysates. Bacterial characteristic

Strain

foe1

foe2

foe3

foe17

foe20

foe21

foe22

foe25

foe26

foe29

foe31

foe33

foe34

foe35

foe37

Prophage free

OE6 OE10 OE12 OE14 OE28 OE39 OE9 OE11 OE32 OE30 OE36 OE3 OE21 OE26 OE42 OE41

± ± + ± − − + ± − − − + ± − − −

− ± − ± − − ± ± − − ± − − − − −

− ± − ± − − + ± − − − − − − − −

± ± − ± − − ± − − − − − − − − −

± ± − ± − − ± ± − − − ± − − + +

− − − ± − − − ± − − − − − − − −

+ ± − − − − ± − − − − − − − + −

+ + + + + − + + − − − + ± − + +

− + ± ± ± − ± − − − − − − − − −

− ± − ± − − ± ± − − − − − − − −

+ + − − − − + ± − − − − − − − −

+ + − + + − + + + + − + − − + +

− ± − ± − − − ± − − − − − − − −

± ± + ± ± ± − ± − − − − − − − −

− − − − − − − − − − − + − − ± +

Non-inducible

Inducible

“+” = lysis. “±” = slowing of bacterial growth. “−” = absence of lysis.

room temperature, and bacterial cell pellets were resuspended in 5 ml of sterile 0.1 M MgSO4. Cell suspensions were irradiated with a Sankyo Denki G30T8 UV germicidal lamp (Kanagawa, Japan) for 25 s at 52 cm distance, as described by Raya and Hébert (2009) and optimized by Doria et al. (2013) and then, transferred into sterile tubes containing 5 ml of double-strength MRSTJ medium to obtain a working MRSTJ final concentration. Cell-lysis was monitored every 4 h for 24 to 32 h, or until a decrease in optical density or differences between irradiated and non-irradiated cell suspension were observed.

lytic activity. In addition, molecular characterization of the phages has been conducted by analyzing their genome size and their protein profile, and by sequencing their integrase genes. 2. Materials and methods 2.1. Bacterial strains and growth conditions O. oeni strains were isolated from Piedmont typical wines (Doria et al., 2013). Bacteria identification at species level was conducted using a species-specific PCR for O. oeni (Zapparoli et al., 1998), and confirmed by sequencing the 16S gene with universal primers 68f-1387r (Marchesi et al., 1998) and comparison with O. oeni sequences available at the Gen Bank database. Strain typing was performed by multiplex RAPD-PCR analysis (Reguant and Bordons, 2003). O. oeni strains were routinely grown in MRS (VWR, Milan, Italy) supplemented with 20% tomato juice, MRSTJ (BD, Milan, Italy), with the pH adjusted to 4.5 with 37% fuming hydrochloric acid (VWR), at 25 °C under aerobic conditions until OD600nm reached 0.1–0.3 (early exponential growth).

2.3. Preparation and storage of phage lysates. UV treated O. oeni strains cultures were centrifuged (3000 × g, 12 min, 4 °C), and supernatants containing the bacteriophages were filtered through a 0.22 μm sterile cellulose acetate membrane syringe filter (VWR International, Milan, Italy), and then preserved at − 80 °C with 10% glycerol. Phages tested in this study were foe1, foe2, foe3, foe17, foe20, foe21, foe22, foe25, foe26, foe29, foe31, foe33, foe34, foe35, foe37 previously isolated by Doria et al. (2013).

2.2. Phage induction by UV 2.4. Lysis plaques count (DL, double layer plates) quantification of phages present in the bacterial lysates

UV induction of prophages was carried out as reported (Raya and Hébert, 2009) with the modifications described by Doria et al. (2013). Briefly, O. oeni strains were inoculated in 10 ml MRSTJ medium and incubated at 25 °C. During early exponential phase (OD600nm ~ 0.1–0.3 measured with DU730 spectrophotometer Beckman Coulter, Brea, CA) 5 ml bacterial cell cultures were centrifuged at 6000 × g 10 min at

Double layer plates were prepared according to the method reported by Adams (1959) adapted for O. oeni as follows. Petri plates containing two medium layers with different concentrations of agar were prepared. Bottom agar composed of 10 ml TJ agar (1.5% agar), pH 5.5,

Table 2 Host range test in modified MRSTJ and in Chardonnay wine. Evaluation of the effect of ethanol and pH on the ability of bacteriophage to lyse host cells. Bacterial characteristics

Phage-free Non inducible phages Inducible phages

MRSTJ

OE6 OE28 OE30 OE32 OE3 OE42

MRSTJ type A

MRSTJ type B

MRSTJ type C

Chardonnay

foe25

foe33

foe25

foe33

foe25

foe33

foe25

foe33

foe25

foe33

+ + − − + +

+ + + + + +

+ − − − ± −

± ± − − ± −

+ − − − − −

+ − − − − −

− − − − − −

− − − − − −

− − − − − −

− − − − − −

MRSTJ modified type A: pH = 3.5 without ethanol. MRSTJ modified type B: 10% ethanol, pH = 4.5. MRSTJ modified type C: 10% ethanol, pH = 3.5. “+” = lysis. “±” = slowing of bacterial growth. “−” = absence of lysis.

14

A. Costantini et al. / International Journal of Food Microbiology 246 (2017) 12–19

2.5.2. Sensitivity tests in modified MRSTJ A, B, C, and Chardonnay wine HRT was carried out also in three kind of modified MRSTJ-media to simulate alcohol and pH wine conditions, as well as in Chardonnay wine. Modified MRSTJ media (A, B, C) were prepared using MRSTJ diluted ten times with distilled water as follows: MRSTJ modified type A: pH = 3.5 without ethanol; MRSTJ modified type B: 10% ethanol, pH = 4.5; MRSTJ modified type C: 10% ethanol, pH = 3.5. Chardonnay wine (characterized by a content of 3 g/l malic acid, 11% ethanol, and a pH of 3.34) was previously treated at 90 °C for 10 min to eliminate phages that could eventually be present in wine. Six O. oeni strains were assayed by HRT: OE3, OE42 (with an inducible phage), OE30, OE32 (with a non-inducible phage) and OE6, OE28 (free-phage strains), which were tested with 2 phage lysates (foe25, foe33) selected based on their virulence during the preliminary HRT in MRSTJ pH 4.5. Fig. 1. Electron micrographs of purified bacteriophages foe3, foe33, foe20 and foe25.

50 μl 1 M CaCl2 and top, or soft, agar made by 4 ml TJ agar (0.7% agar), pH 5.5, 200 μl of glycerol, 50 μl 1 M CaCl2 (Arendt et al., 1991; Tenreiro et al., 1993). Bacterial cultures (100 μl) at the beginning of the exponential phase (OD600nm ~ 0.1–0.3) were transferred into a sterile tube and, then, 100 μl of phage suspension and 50 μl of 1 M CaCl2 were added. Cultures were incubated at 37 °C under gentle stirring for 15 min, mixed with 200 μl of glycerol and 4 ml of melted soft agar, spread on bottom agar, and incubated at 25 °C for 6 days to determine the plaque-forming units/ml, PFU/ml (Cavin et al., 1991; Santos et al., 2009). 2.5. Host range test (HRT) To establish the host range of the isolated phages, sensitivity tests were conducted by comparing bacterial cultures growth trends with or without added phages. To this purpose, 9.75 ml of broth were inoculated with 100 μl of bacterial culture in exponential growth and 100 μl of phage lysate (multiplicity of infection, MOI, =10), incubated at 37 °C for 30 min under gentle stirring, and then at 25 °C for 15 days. Phage lysate was omitted in the control sample. All samples were supplemented with 0.5% of CaCl2 1 M, and culture growth was monitored by measuring the OD600nm every three days. 2.5.1. Sensitivity tests in MRSTJ HRT in MRSTJ pH 4.5 was performed with 16 strains of O. oeni and 15 phage lysates. O. oeni strains belong to the three previously identified groups by Doria et al. (2013): 5 strains with inducible phages (OE3, OE21, OE26, OE42, OE41), 5 with non-inducible phages (OE9, OE11, OE30, OE32, OE36), and 6 free-phage bacterial strains (OE6, OE10, OE12, OE14, OE28, OE39).

2.6. Electron microscopy The morphology of viral particles was observed by transmission electron microscopy at the Centro di Ricerche Biotecnologiche (University of Piacenza, Italy). Phage sample suspensions were negatively stained with 2% uranyl acetate and examined in a JEOL JEE 1200 EXII (JEOL, London, United Kingdom) transmission electron microscope at an accelerating voltage of 80 kV. 2.7. Molecular characterization 2.7.1. Genome size analysis Phages obtained after lysis induction by UV were purified as described (Doria et al., 2013), and the final suspensions mixed 1:1 with 2% low melt agarose (Sigma, Milan, Italy) and poured into molding blocks. DNA was extracted according to Lingohr et al. (2009). Plugs were loaded onto a 1% agarose gel and subjected to electrophoresis in 0.5 × Tris-borate-EDTA (TBE) buffer for 20.5 h at 14 °C using a CHEF Mapper pulsed field gel electrophoresis apparatus (Bio-Rad, Milan, Italy) at 6 V/cm, 120°, ramping: 0.11–92 s linear slope. Gels were visualized under UV illumination Geldoc (Biorad) after ethidium bromide staining. 2.7.2. Integrase gene characterization Both bacteriophage and prophage DNA were extracted according to Doria et al. (2013), and presence of integrase A, B, C and D genes were assayed by PCR using primer sets IntA f/r, IntB f/r, IntC f/r, IntD f/r, respectively (Jaomanjaka et al., 2013). However, modifications on the primers sets for some integrase gene have to be applied for proper amplification due to what seems to be a mistake in the original description (see below). Amplicons were visualized under UV on ethidium bromide stained 2.5% agarose gel in TBE buffer, purified, and sequenced (BMR

Fig. 2. Genome size analysis made on the 15 purified bacteriophages by pulse-field gel electrophoresis (PFGE). A: DNA size standards 8–48 kb (Biorad), foe1, foe2, foe3, foe17, foe29, foe20, foe21, foe22, foe25, foe26, DNA size Standards 5 kb Ladder (Biorad), foe31, foe35, DNA size standards 8–48 kb (Biorad), foe37, foe33, foe 34.

A. Costantini et al. / International Journal of Food Microbiology 246 (2017) 12–19

Genomics, Padua University, Italy). As control, to verify the absence of contaminating bacterial DNA in the phage sample, PCR amplification of 16S rDNA from phage lysates was tested using primers 63f/1387r

15

(Marchesi et al., 1998). Amplified integrase sequences (GenBank accession numbers KX384824-KX384845 and KX910692-KX910697) were aligned and compared (Muscle, www.ebi.ac.uk).

Fig. 3. Alignment of the sequences obtained with primers IntBf/IntBr and phage 10MC integrase gene (int B). The sequences of phages foe2, foe3, foe26, and foe37 are identical to foe1.

16

A. Costantini et al. / International Journal of Food Microbiology 246 (2017) 12–19

Fig. 3 (continued).

2.7.3. SDS PAGE Phage lysates were analyzed by 4–12% SDS-PAGE as described (Boulanger, 2009) in a Miniprotean equipment (Bio-Rad) at 200 V for 30 min. After electrophoresis, gels were stained with Coomassie R-250 blue.

3. Results and discussion 3.1. Host range test (HRT) in MRSTJ (pH = 4.5) We have previously detected 25 lysogenic O. oeni strains by amplification of the endolysin gene that were then classified into three groups (Doria et al., 2013). Here, for HRT, 16 O. oeni strains (6 prophage-free strains, 5 with non-inducible phages, and 5 with inducible phages) were tested three times in MRSTJ (pH = 4.5) with 15 phage lysates

(Table 1). As a consequence, 3 different bacterial behaviors were observed: i) strains sensitive to the phage in which bacterial growth was almost or completely stopped; ii) strains in which the presence of the bacteriophage slowed bacterial growth; and iii) strains with a normal growth not affected by the presence of viral particles. Our results showed that the more susceptible strains to phages attacks were the prophage-free strains, then the ones with non-inducible phage, and finally those with inducible phages, which were by far the most resistant (Table 1). These data are in accordance with previous reports (Poblet-Icart et al., 1998) confirming a higher resistance to lysis in lysogenic strains with inducible phages. On the other hand, to our knowledge, this is the first time that resistance to the lysis of non-inducible O. oeni strains is reported. Remarkably, OE26 strain, which has an inducible phage (foe26) integrated in its genome, showed up to be resistant to all 15 phages tested. In contrast, not a single phage affects all 16 bacterial strains tested. In fact, a different lytic ability among the

Fig. 4. SDS-PAGE phage lysates. Marker (BSA 66 kDa, Lysozyme 14 kDa),1–9: foe3, foe20, foe21, foe22, foe25, foe29, foe33, foe34, foe37.

A. Costantini et al. / International Journal of Food Microbiology 246 (2017) 12–19

different viruses was observed, being foe25 and foe33 the most virulent, as they have lytic effects on most of the bacterial strains tested (Table 1). 3.2. HRT in modified MRSTJ (A, B, and C media) and in Chardonnay wine To evaluate the effect of alcohol and low pH (3.5) on the lysing capacity of the phages, HRT were conducted in modified MRSTJ. HRT were applied in 6 representative bacterial strains of each group (2 with inducible phage, 2 with non-inducible phages, and 2 phage-free) in three types of modified MRSTJ (A, B, and C) and in Chardonnay wine, with the two most virulent phages, foe25 and foe33 (Table 2). OE30 and OE32 (two bacteria with a non-inducible phage) were chosen because of their resistance to foe25 and sensitivity to foe33 in MRSTJ pH 4.5 (Table 1). These two bacteria remain resistant to foe25 in modified pH and alcohol conditions, but turned out resistant to foe33 in presence of ethanol (MRSTJ B and C) and in media without ethanol at pH = 3.5 (MRSTJ A). In addition, it was observed that, in MRSTJ type A (pH = 3.5 without ethanol), phage foe25 had a lytic activity with the prophage-free strain OE6 and caused a slower growth in OE3, while phage foe33 caused a slower growth in OE6, OE28 and OE3 strain cultures. In contrast, in MRSTJ type B (10% ethanol, pH = 4.5), lytic activity of the 2 phages was observed only in the prophage-free strain OE6. Finally, in MRSTJ type C (10% ethanol, pH = 3.5, close to wine pH), all bacteria assayed resulted resistant to the two phages tested, results that were confirmed when Chardonnay wine was used as culture medium (Table 2). In resume, although a pH = 3.5 characteristic of wines confers the bacteria a greater resistance to phage lysis, the effect of ethanol seems to be even more important. In fact, a combination of low pH and presence of ethanol conferred resistance to lysis in all the tested strains. These results are unlikely due to difficulties in bacterial growth, since O. oeni has a comparable growth in MRSTJ C and MRSTJ pH 4.5 media. These findings may have a clear practical implication in the utilization of commercial bacterial starters for the MLF. Currently, there is a tendency to proceed with an early inoculation (at the initial phase of the alcoholic fermentation) to gradually adapt the cells to alcohol. If so, our results point that selection of malolactic starters should be based on the resistance of O. oeni strains to phages attack. At this point, it should be noted that published data are still controversial about the practical implications of bacteriophages in MLF, their role during vinification, and their influence on the dynamic of fermenting bacteria. For instances, the notion that pH can influence phage attack has been demonstrated for phage infection in Lactobacillus, which optimum pH in laboratory medium was 5.5–6 (Watanabe and Takesue, 1972), and it has also been reported that phages cannot interfere with malolactic fermentation in wine having pH lower than 3.23 (Henick-Kling et al., 1986). In this regard, our results clearly indicated that pH, and certainly ethanol, affects the lytic activity of the phages tested. This effect could be due to the modification of the bacterial cell surface induced by stress conditions, such as ethanol and acid pH. In fact, under stress conditions, O. oeni produces large amounts of stress proteins and changes its membrane composition, largely during ethanol shock (Da Silveira et al., 2003; Grandvalet et al., 2008; Guzzo et al., 1994, 2000). This has been further demonstrated by us by means of both gene expression and proteomic (Costantini et al., 2015). Cell membrane has been later confirmed as the main target of ethanol damage (Olguín et al., 2015), and it has been reported that cells grown in ethanol and acid conditions present changes in their membrane lipid composition that lead to a significant increase in palmitic and dihydrosterulic acids, increasing membrane rigidity (Grandvalet et al., 2008). Therefore, all these changes can alter phage-host cell surface interactions, protecting bacteria against phage attack. However, the isolation of phages from wines having problems with MLF suggested that phages can potentially induce failures in the fermentation (Arendt and Hammes, 1992). Accordingly, our results point that these cell membrane modifications in presence of stressing conditions (low pH and

17

ethanol) could make the Oenococcus cells resistant to phage attack; however, other factors (sensitivity of phages to ethanol, pH, etc.) may also play a role on the observed resistance. However, during winemaking process, when ethanol is absent or at low level, phages can lyse bacterial cells and interfere with the malolactic fermentation; this could be particularly important in the case of inoculation of MLF starters in the must, or in the early stages of the alcoholic fermentation. 3.3. Morphology Fig. 1 shows electron microscopy images of the most (foe25 and foe33) and less (foe3 and foe20) harmful phages for the tested bacteria. They all present an isometric head and a flexible tail, but although differing in head diameters and tail lengths: foe 25 (55 nm and 200 nm), foe 33 (60 nm and 260 nm), foe3 (60 nm and 160 nm), and foe20 (47 nm and 220 nm), this variability does not seem to be related with virulence. The morphology of the isolated phages can be ascribed to the class B of Bradley (1967) as previously showed for others O. oeni bacteriophages (Arendt and Hammes, 1992; Davis et al., 1985a, b). 3.4. Molecular characterization As displayed in Fig. 2, pulsed field gel electrophoresis (PFGE) showed that most of the analyzed bacteriophages have a genome size of about 38 kb, except foe 29, foe33 and foe34 that have it of about 46 kb. Currently, only 12 full genomes of phages infecting Leuconostoc sp. are available at the databases, with genomes sizes of 25.7 to 38.7 kb (Kot et al., 2014), while the sizes of the only 3 available Oenococcus phage complete sequences are 46.1, 43, and 46.2 Kb for phi9805 (accession KF147927), phiS11 (accession KF183314) and phiS13 (accession KF183315), respectively. Our data indicated that foe29, foe33 and foe34 have a genome size similar to that of phi9805 and phiS13, while the others analyzed phages have a size of about 38 kb, more similar to that of Leuconostoc sp. phages. However, in accordance to what has been described for Salmonella phages (Switt et al., 2013), phage virulence was not related to genome size, as phage with different lysis profiles have similar genome sizes. Assessment of oenococcal phage diversity has been limited by the lack of a proper molecular characterization methodology. Recently, using databank available integrase genes, a classification for oenococcal prophages has been proposed (Jaomanjaka et al., 2013) that resulted in four groups. Integrase A group includes previously reported fog44 and fog30 bacteriophages; integrase B Leuconostoc oenos bacteriophage 10MC; integrase C bacteriophage fogPSU-1; and integrase D Oenococcus phages phiS11, phi9805, and phiS13. These authors reported that most (54%) of O. oeni prophages belong to the intA group and 27% possess intD gene, being rare the prophages of the intB and intC groups. Nonetheless, four out of the 22 lysogenic strains analyzed, resulted poly-lysogenic (Jaomanjaka et al., 2013). Therefore, to further characterize the isolated O. oeni prophages, the presence of int sequences was determined using specific primers described by Jaomanjaka et al. (2013) both, in the isolated O. oeni prophages and the phage genomes, where absence of Oenococcus DNA was verified by 16S rDNA amplification. PCR amplification of integrase A gene with primers IntAf/IntAr amplified all isolated bacteriophage, being the sequences 100% homologous within them and with that of previously reported fog44 (Ac. No. AF047001.2) and fog30 integrases (AJ629110.1). Surprisingly, using IntBf/IntBr, 850 bp amplicons were obtained for phage foe1, foe2, foe3, foe26, foe29, and foe37, greater than the 443 bp reported by Jaomanjaka et al., 2013. Sequence analysis of the amplicons (Fig. 3) showed, with the exception of foe26 which differed in 10 nucleotides, a 100% homology within them and a 97% similarity with Leuconostoc oenos bacteriophage 10MC (Ac. No. U77495) putative integrase gene (named as integrase B by Jaomanjaka et al., 2013). On the other hand, no amplification was achieved with primers IntCf/IntCr, as IntCr primer

18

A. Costantini et al. / International Journal of Food Microbiology 246 (2017) 12–19

does not match the so-called integrase C gene (see below). Finally, PCR with intDf/intDr primers only amplified two phages (foe21 and foe37), which integrase D sequences were 100% homologous between them and with that of Oenococcus phages phiS11, phi9805, and phiS13 integrase gene (Ac. No KF183314.1, KF147927.1 and KF183315.1 respectively). Unexpected amplifications sizes to that previously reported (Jaomanjaka et al., 2013) and sequenced similarities observed using IntBf/IntBr moved us to carry out a Blast analysis of this set of primers, as well as of the IntCf/IntCr set. Alignment analyses showed that IntBr matches with PSU-1 (int C) integrase gene instead of 10MC (int B), while IntCr matches with that of 10MC (int B) instead of PSU-1. In fact, amplification with intBf/intCr resulted in a band of about 460 bp for phages foe1, 2, 3, 26, 29, 37 similar to that previously reported for integrase B (Jaomanjaka et al., 2013). Sequence analyses of these amplicons showed a 100% homology within them, with the exception of foe26 which differed in 10 nucleotides, and a 97% similarity with phage 10MC (Fig. 3). Even more, amplifications with both primer sets IntBf/IntBr (850 bp) and IntBf/IntCr (460 bp size) was achieved in the same phages. Alignments of primer IntBr with the integrases of phages 10MC (Ac. No. U77495) and PSU (Ac. No. AJ629109) showed only two nucleotides mismatches between them, which likely explains the amplification of a 850 bp fragment of integrase B. PCR using set intCf/intBr gave amplicons of about 430 bp with identical sequences for phages foe21, 25, 26, 35, 37, and with a 98% of similarity with PSU-1 (int C) integrase gene. Based on these results, it seems clear that primers were erroneously referred in the original report (Jaomanjaka et al., 2013). Summarizing, all isolated bacteriophages analyzed possess integrase A; six (foe1, foe2, foe3, foe26, foe29, foe37) integrase B; five (foe21, foe25, foe26, foe35, foe37) integrase C; and two (foe21 and foe37) integrase D. Hence, the genome of some phages includes more than one integrase gene without necessarily having a poly-lysogenic state. These findings, to our knowledge, have never been reported for O. oeni bacteriophages, although GenBank search found the presence of two integrase genes (position 15250–16698 and 41435–42555) in the partial genome sequence of LLC-1 Leuconostoc phage (accession number KJ608189). At this point, it should be noted that bacteriophage genomes exhibit genetic mosaicism, appearing as mixtures of genetic modules that encode apparently exchangeable functional elements such as the virion coat, the virion tail, or the genome integration proteins (Rokyta et al., 2006). So that, horizontal gene transfer is thought to play an important role in the evolution of bacteriophage genomes. In fact, Siphoviridae (as O. oeni phages), Podoviridae, and Myoviridae families of tailed dsDNA phages are characterized by relatively large genomes (b20 to hundreds Kb) and often a lysogenic lifecycle, which can likely facilitate horizontal gene transfer either by minimizing constraints on gene acquisition or loss, or by increasing the opportunities for phage recombination, respectively (Rokyta et al., 2006). Finally, SDS-PAGE analysis of the phages protein profile showed the presence of two main bands with a molecular weight of around 36 and 45 kDa, and two less intense bands of around 15 and 64 kDa, respectively (Fig. 4). This pattern is similar to that of Leuconostoc oeni phage ɸ1002 (Huang et al., 1996), where 3 main proteins of 43, 30 and 14 kDa can be distinguished. 4. Conclusion In conclusion, molecular and physiological characterization of the 15 Oenococcus bacteriophage analyzed showed that they present different genome sizes and, remarkably, that phages can contain more than one type of integrase gene. In addition, HRT analysis demonstrates a quite great variability in the phage-bacteria interactions, as two phages resulted more virulent than the others, and one O. oeni strain, which has an inducible phage integrated in its genome, turns out to be resistant

to all 15 phages tested. An interesting aspect that must be also underlined is that, in presence of ethanol and low pH, the phages were unable to attack the cells. This fact leads to a general consideration: though phage attack in a wine at the end of alcoholic fermentation seem not to be a problem, considering the overall vinification process, phages can represent a risk factor when the alcohol content is low, since, under this condition, they can attack O. oeni cells driving to subsequent troubles in malolactic fermentation. This fact is a very important feature when choosing the appropriate time for malolactic starter inoculation that may have consequences for malolactic fermentation success.

Acknowledgments This work was supported by grant D.G.R.47 from the Regione Piemonte (Italy). References Adams, M.H., 1959. Bacteriophages. Interscience Publishers, New York. Arendt, E.K., Hammes, W.P., 1992. Isolation and characterization of Leuconostoc oenos phages from German wines. Appl. Microbiol. Biotechnol. 37, 643–646. Arendt, E.K., Lonvaud-Funel, A., Hammes, W.P., 1991. Lysogeny in Leuconostoc oenos. J. Gen. Microbiol. 137, 2135–2139. Bauer, R., Dicks, L.M.T., 2004. Control of malolactic fermentation in wine. A review. S. Afr. J. Enol. Vitic. 25, 74–88. Borneman, A.R., McCarthy, J.M., Chambers, P.J., Bartowsky, E.J., 2012. Comparative analysis of the Oenococcus oeni pan genome reveals genetic diversity in industrially-relevant pathways. BMC Genomics 13, 373. Boulanger, P., 2009. Purification of bacteriophages and SDS-PAGE analysis of phage structural proteins from ghost particles. In: Clokie, M.R.J., Kropinski, A.M. (Eds.), Bacteriophages: Methods and Protocols, Molecular and Applied Aspects 502. Humana Press, New York, NY. Bradley, D.E., 1967. Ultrastructure of bacteriophage and bacteriocins. Bacteriol. Rev. 31, 230–314. Cavin, J.F., Drici, F.Z., Prevost, H., Divies, C., 1991. Prophage curing in Leuconostoc oenos by mitomycin C induction. Am. J. Enol. Vitic. 42, 163–166. Costantini, A., Rantsiou, K., Majumder, A., Jacobsen, S., Pessione, E., Svensson, B., GarciaMoruno, E., Cocolin, L., 2015. Complementing DIGE proteomics and DNA subarray analyses to shed light on Oenococcus oeni adaptation to ethanol in wine-simulated conditions. J.Proteomics 18, 114–127. Da Silveira, M.G., Golovina, E.A., Hoekstra, F.A., Rombouts, F.M., Abee, T., 2003. Membrane fluidity adjustments in ethanol-stressed Oenococcus oeni cells. Appl. Environ. Microbiol. 69, 5826–5832. Davis, C.R., Silveira, N.F.A., Fleet, G.H., 1985a. Occurrence and properties bacteriophages of Leuconostoc oenos in Australian wines. Appl. Environ. Microbiol. 50, 872–876. Davis, C.R., Wibowo, D., Eschenbruch, R., Lee, T.H., Fleet, G.H., 1985b. Practical implications of malolactic fermentation: a review. Am. J. Enol. Vitic. 36, 290–301. Doria, F., Napoli, C., Costantini, A., Berta, G., Saiz, J.C., Garcia-Moruno, E., 2013. Development of a new method for the detection and identification of O. oeni bacteriophages based on endolysin gene sequence and RAPD. Appl. Environ. Microbiol. 79, 4799–4805. Grandvalet, C., Assad-Garcia, J.S., Chu-Ky, S., Tollot, M., Guzzo, J., Gresti, J., TourdotMaréchal, R., 2008. Changes in membrane lipid composition in ethanol and acidadapted Oenococcus oeni cells: characterization of the cfa gene by heterologous complementation. Microbiology 154, 2611–2619. Guzzo, J., Cavin, J., Divies, C., 1994. Induction of stress proteins in Leuconostoc oenos to perform direct inoculation of wine. Biotechnol. Lett. 16, 1189–1194. Guzzo, J., Jobin, M., Delmas, F., Fortier, L., Garmyn, D., Tourdot-Maréchal, R., Lee, B., Diviès, C., 2000. Regulation of stress response in Oenococcus oeni as a function of environmental changes and growth phase. Int. J. Food Microbiol. 55, 27–31. Henick-Kling, T., Lee, T.H., Nicholas, D.J.D., 1986. Characterisation of the lytic activity of bacteriophages of Leuconostoc oenos isolated from wine. J. Appl. Bacteriol. 61, 525–534. Huang, C.M., Asmundson, R.V., Kelly, W.J., 1996. Characterization of a temperate phage isolated from Leuconostoc oenos strain 1002. Appl. Microbiol. Biotechnol. 45, 472–476. Jaomanjaka, F., Ballestra, P., Dols-lafargue, M., Le Marrec, C., 2013. Expanding the diversity of oenococcal bacteriophages: insights into a novel group based on the integrase sequence. Int. J. Food Microbiol. 166, 331–340. Kot, W., Neve, H., Heller, K.J., Vogensen, F.K., 2014. Bacteriophages of Leuconostoc, Oenococcus, and Weissella. Front. Microbiol. 5, 186. Lingohr, E., Frost, S., Johnson, R.P., 2009. Determination of bacteriophage genome size by pulsed-field gel electrophoresis. Methods Mol. Biol. 502, 19–25. Marchesi, J.R., Sato, T., Weightman, A.J., Martin, T.A., Fry, J.C., Hiom, S.J., Wade, W.G., 1998. Design and evaluation of useful bacterial-specific PCR primers that amplify genes coding for bacterial 16S rRNA genes. Appl. Environ. Microbiol. 64, 795–799. Olguín, N., Champomier-Verges, M., Anglade, P., Baraige, F., Cordero-Otero, R., Bordons, A., Zagorec, M., Reguant, C., 2015. Transcriptomic and proteomic analysis of Oenococcus oeni PSU-1 response to ethanol shock. Food Microbiol. 51, 87–95.

A. Costantini et al. / International Journal of Food Microbiology 246 (2017) 12–19 Poblet-Icart, M., Bordons, A., Lonvaud-Funel, A., 1998. Lysogeny of Oenococcus oeni (syn. Leuconostoc oenos) and study of their induced bacteriophages. Curr. Microbiol. 36, 365–369. Raya, R.R., Hébert, E.M., 2009. Isolation of phage via induction of lysogens. In: Clokie, M.R.J., Kropinski, A.M. (Eds.), Bacteriophages Methods and Protocol, Vol. 1: Isolation, Characterization and Interactions. Humana Press, pp. 23–32. Reguant, C., Bordons, A., 2003. Typification of Oenococcus oeni strains by multiplex RAPDPCR and study of population dynamics during malolactic fermentation. J. Appl. Microbiol. 95, 344–353. Rokyta, D.R., Burch, C.L., Caudle, S.B., Wichman, H.A., 2006. Horizontal gene transfer and the evolution of microvirid coliphage genomes. J. Bacteriol. 188, 1134–1142. Santos, S.B., Carvalho, C.M., Sillankorva, S., Nicolau, A., Ferreira, E.C., Azaredo, J., 2009. The use of antibiotics to improve phage detection and enumeration by double-layer agar technique. BMC Microbiol. 9, 148. Sozzi, T., Maret, R., Poulin, J.M., 1976. Mise en evidence de bacteriophage dans le vin. Separatum experiential. 32 pp. 17–23.

19

Switt, A.I., den Bakker, H.C., Vongkamjanm, K., Hoelzer, K., Warnick, L.D., Cummings, K.J., Wiedmann, M., 2013. Salmonella bacteriophage diversity reflects host diversity on dairy farms. Food Microbiol. 36, 275–285. Tenreiro, R., Santos, R., Brito, L., Paveia, H., Vieira, G., Santos, M.A., 1993. Bacteriophage induced by mitomycin C treatment of Leuconostoc oenos strains from Portuguese wines. Lett. Appl. Microbiol. 16, 207–209. Torriani, S., Felis, G.E., Fracchetti, F., 2011. Selection criteria and tools for malolactic starters development: an update-a review. Ann. Microbiol. 61, 33–39. Watanabe, K., Takesue, S., 1972. The requirement for calcium in infection with Lactobacillus phages. J. Gen. Virol. 17, 19–30. Zapparoli, G., Torriani, S., Pesente, P., Dellaglio, F., 1998. Design and evaluation of malolactic enzyme gene targeted primers for rapid identification and detection of Oenococcus oeni in wine. Lett. Appl. Microbiol. 27, 243–246.