Scholarly article on topic 'Detection and enumeration of Lactobacillus helveticus in dairy products'

Detection and enumeration of Lactobacillus helveticus in dairy products Academic research paper on "Animal and dairy science"

CC BY-NC-ND
0
0
Share paper
Academic journal
International Dairy Journal
OECD Field of science
Keywords
{}

Abstract of research paper on Animal and dairy science, author of scientific article — Aline Moser, Hélène Berthoud, Elisabeth Eugster, Leo Meile, Stefan Irmler

Abstract Lactobacillus helveticus, a lactic acid bacterium, is an important species in food fermentation, e.g., cheesemaking, and is considered beneficial to human health. We developed a quantitative real-time polymerase chain reaction (qPCR) method for the detection and quantification of L. helveticus in dairy products. The method uses a set of target-specific PCR primers and a fluorogenic probe and amplifies a part of the pheS gene that encodes the alpha subunit of the phenyalanine-tRNA synthetase. All 24 L. helveticus strains tested were qPCR positive; no signal was observed for 23 strains belonging to closely related species. The limit of detection was ten copies per reaction and the assay covered a linear dynamic range of eight logs. The method was used to detect and enumerate L. helveticus in milk and cheese during ripening; therefore it can be used to study the temporal and spatial distribution of L. helveticus during cheese manufacturing and ripening.

Academic research paper on topic "Detection and enumeration of Lactobacillus helveticus in dairy products"

Accepted Manuscript

Detection and enumeration of Lactobacillus helveticus in dairy products Aline Moser, Hélène Berthoud, Elisabeth Eugster, Leo Meile, Stefan Irmler

PII: S0958-6946(16)30362-4

DOI: 10.1016/j.idairyj.2016.12.007

Reference: INDA 4124

To appear in: International Dairy Journal

Received Date: 28 September 2016 Revised Date: 2 December 2016 Accepted Date: 4 December 2016

INTERNATIONAL DAIRY

JOURNAL

Please cite this article as: Moser, A., Berthoud, H., Eugster, E., Meile, L., Irmler, S., Detection and enumeration of Lactobacillus helveticus in dairy products, International Dairy Journal (2017), doi: 10.1016/j.idairyj.2016.12.007.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1 Detection and enumeration of Lactobacillus helveticus in dairy products

8 Aline Moser3,15, Hélène Berthouda, Elisabeth Eugstera, Leo Meileb, Stefan Irmlera*

9 10 11 12

14 a Agroscope, 3003 Bern, Switzerland

15 b Laboratory of Food Biotechnology, Institute of Food Science and Nutrition, ETH Zürich, 8092

16 Zurich, Switzerland

22 * Corresponding author. Tel.: +41 58 463 81 55

23 E-mail address: stefan.irmler@agroscope.admin.ch (S. Irmler)

26 ABSTRACT

28 Lactobacillus helveticus, a lactic acid bacterium, is an important species in food fermentation,

29 e.g., cheesemaking, and is considered beneficial to human health. We developed a quantitative

30 real-time polymerase chain reaction (qPCR) method for the detection and quantification of L.

31 helveticus in dairy products. The method uses a set of target-specific PCR primers and a

32 fluorogenic probe and amplifies a part of the pheS gene that encodes the alpha subunit of the

33 phenyalanine-tRNA synthetase. All 24 L. helveticus strains tested were qPCR positive; no

34 signal was observed for 23 strains belonging to closely related species. The limit of detection

35 was ten copies per reaction and the assay covered a linear dynamic range of eight logs. The

36 method was used to detect and enumerate L. helveticus in milk and cheese during ripening;

37 therefore it can be used to study the temporal and spatial distribution of L. helveticus during

38 cheese manufacturing and ripening.

41 1. Introduction

43 Lactobacillus helveticus is a member of the group of lactic acid bacteria (LAB) that

44 metabolises lactose and other carbohydrates predominantly into lactic acid (Hammes & Hertel,

45 2006). LAB are found in various habitats in which they encounter complex nutritional resources,

46 such as in soil, sewage, and fermented food, as well as on the body surface and in the intestinal

47 tract of humans and animals.

48 The species L. helveticus is commonly used as thermophilic starter in the manufacturing

49 of cheeses, such as Swiss-type and Italian hard cheeses (Slattery, O'Callaghan, Fitzgerald,

50 Beresford, & Ross, 2010). L. helveticus possesses cell-envelope bound proteases and

51 intracellular peptidases, the latter of which can be released into the cheese matrix upon

52 autolysis. Consequently, L. helveticus has also been used as adjunct in Cheddar cheeses to

53 accelerate protein degradation and enhance flavour development during cheese ripening

54 (Hannon et al., 2003; Kenny, FitzGerald, O'Cuinn, Beresford, & Jordan, 2006).

55 Because milk proteins are a source of peptides with various bioactivities and, can be, for

56 instance, released by the proteolytic activity of L. helveticus, extensive research has been

57 conducted on the use of L. helveticus strains as probiotics (see Taverniti & Guglielmetti, 2012

58 and Griffiths & Tellez, 2013, for reviews). Well-known examples include the development and

59 commercialisation of the antihypertensive milk drinks Evolus (Valio Ltd, Valio, Finland) and

60 Calpis (Calpis Food Industry Co. Ltd. Tokyo, Japan), which contain L. helveticus strains. Based

61 on this potential, the detection and enumeration of L. helveticus are essential for many

62 applications. A simple and rapid quantification method for L. helveticus enables, e.g., the

63 tracking of this species in milk production and fermentation.

64 We are interested in better understanding the microbiology of Gruyère PDO cheese and

65 its influence on cheese ripening and flavour development. Gruyère PDO is manufactured with

66 natural whey cultures (NWCs) (http://www.gruyere.com/en/specifications/,). Preliminary

67 microbiological investigations have shown that L. helveticus is one of the dominant species

68 present (data not shown). A species-specific and easy-to-use method of detecting and

enumerating L. helveticus is a valuable tool with which to study the temporal and spatial distributions of representatives of this species in the cheese manufacturing process, including those in the milk.

In this report, we present the development of a hydrolysis-probe-based real-time PCR assay targeting the phenylalanyl-tRNA synthase gene (pheS). This method allows both the detection and enumeration of L. helveticus in dairy products. The assay developed here was found to be specific for L. helveticus and was used to quantify this species in dairy products.

2. Material and methods

2.1. Bacterial strains, media and growth conditions

All strains used in this study are listed in Table 1. Strains were stored at -80 °C in sterile reconstituted skim milk powder (10%, w/v). Strains were reactivated and cultivated in 10 mL MRS broth (de Man, Rogosa, & Sharpe, 1960) at 37 °C.

2.2. DNA extraction

For the broth cultures, a bacterial pellet that had been harvested from a 1 mL culture via centrifugation [10,000 x g, 5 min, room temperature (RT)] was used for DNA extraction. In case of NWC samples, the pellet (10,000 x g, 5 min, RT) obtained from 10 mL NWC, to which 50 |jL 10% sodium dodecyl sulphate (SDS) was added, was used for DNA extraction.

With regard to cheese samples, 10 g grated cheese were added to 90 mL modified peptone water (10 g L-1 peptone from casein, 5 g L-1 sodium chloride, 20 g L-1 trisodium citrate dihydrate, pH 7.0). After incubation for 10 min at 40 °C, the sample was homogenised for 3 min in a stomacher (Masticator, IUL Instruments, Konigswinter, Germany). After an additional incubation at 40 °C for 10 min, 50 jL 10% (w/v) SDS were added to 10 mL of the homogenate

96 which was then centrifuged (4000 x g, 21 °C, 30 min). The pellet containing the bacteria was

97 used for DNA extraction.

98 To improve DNA extraction, the bacterial pellets were treated with the following pre-lysis

99 steps. First, the bacterial cells were incubated in 1 mL 50 mM sodium hydroxide for 15 min at

100 RT. Next, the cells were treated with lysozyme (2.5 mg mL-1 dissolved in 100 mM

101 Tris(hydroxymethyl)aminomethane, 10 mM ethylenediaminetetraacetic acid, 25% (w/v) sucrose,

102 pH 8.0) for 1 h at 37 °C. Finally, the bacteria were collected via centrifugation at 10,000 x g at

103 RT for 5 min. The genomic DNA (gDNA) was then extracted using the EZ1 DNA Tissue kit and

104 a BioRobot EZ1 workstation (Qiagen, Hombrechtikon, Switzerland). The final volume of the

105 eluted DNA was 100 |jL.

106 To determine the DNA from the whole and lysed cells in the cheese, a protocol

107 described by Gatti et al. (2008) was applied. Briefly, the bacterial pellet - called the whole-cell

108 fraction - was treated with the Turbo DNA-free Kit (ThermoFisher Scientific, Switzerland)

109 according to the manufacturer's instructions before DNA extraction. The lysed-cell fraction was

110 obtained by passing the supernatant of the cheese homogenate through a sterile filter with a

111 pore size of 0.22 ^m (Whatman, Dassel, Germany) and extracting the DNA from the filtrate.

112 DNA from the raw milk was extracted as described by Turgay, Schaeren, Wechsler, Bütikofer,

113 and Graber (2016).

115 2.3. Oligonucleotide primer and probe design

117 For various LAB, the nucleotide sequences of the pheS gene that encode the alpha

118 subunit of the phenylalanine-tRNA synthetase were aligned (supplementary Fig. S1) using CLC

119 Main Workbench version 7.5.1 (CLC bio, Aarhus, Denmark). Thereby, an 87-bp region from

120 nucleotide position 318 to 404 within the pheS gene was found to be a potential target site

121 because it was identical in all analysed L. helveticus strains but showed variability in other LAB

122 species. Primer3 software v.0.4.0 (Untergasser et al., 2012) was used to design the primers and

the hydrolysis probe for this gene region. The primers and hydrolysis probe were synthesised by Microsynth (Balgach, Switzerland).

2.4. Real-time PCR assay

Quantitative real-time PCR (qPCR) assays were carried out on a Rotor-Gene 6000 (Corbett Life Science, Sydney, Australia) in a final volume of 12 jL containing 2 jL DNA extract, 300 nM of LbhelvF1 (5'-AGGTTCAAAGCATCCAATCAATATT-3'), 300 nM of LbhelvR1 (5'-TCGGGACCTTGCACTACTTTATAAC-3'), 100 nM of hydrolysis probe (5'-(FAM) ATACCGATGAAGTAGCTTTCCAA ATCATCCA(BHQ-1 )-3'), and 6 jL of 2x qPCR MasterMix No ROX (Eurogentec, Seraing, Belgium). All reactions were performed under the following conditions: 50 °C for 2 min and 95 °C for 10 min followed by 40 cycles of 95 °C for 15 s and 60 °C for 60 s. The fluorescence of the reporter dye (FAM) was measured during amplification at 510 nm. Data were analysed using Rotor-Gene 1.7 software with a threshold of 0.03 and the 'Dynamic Tube Normalization' option. The standard curve was generated by plotting the threshold cycle (Cq) values as a function of the concentration of recombinant pheS standard gene copies jL-1. All standard and sample reactions were run in triplicate.

2.5. Construction of the plasmid standard

A part of the pheS gene, including the target sequence for the qPCR assay, was amplified with the primers StdLhF (5'-CGTGATGTTGCCCCAGAAAA-3') and StdLhR (5'-GGTGTGAGTGAGTAGCATCG-3') from the gDNA of L. helveticus FAM1450. The amplification was performed at a final volume of 50 jL, containing 500 nM of each primer, 2.5 U AmpliTaq Gold® Polymerase (Roche, Bale, Switzerland), 0.2 mM dNTPs, and 5 jL 10x PCR buffer (Roche). The reactions were run on a thermocycler at 95 °C for 10 min followed by 35 cycles of 95 °C for 20 s, 60 °C for 30 c, and 72 °C for 30 s, as well as a final extension for 5 min at 72 °C. The amplicon was purified using the QIAquick PCR Purification kit (Qiagen) and then inserted

151 into the pGEM®-T Easy vector (Promega, Madison WI, USA). Escherichia coli MAX Efficiency®

152 DH5a™ competent cells (Invitrogen, California, USA) were transformed with the recombinant

153 plasmids. Clones carrying inserts in the plasmid were selected via blue/white screening on

154 plates containing X-gal. Plasmid DNA (pGEM/pheS) was isolated from 1 mL of a bacterial

155 overnight culture with the QIAprep Spin Miniprep kit (Qiagen) according to the manufacturer's

156 protocol. The orientation and DNA sequence of the insert were determined via Sanger

157 sequencing.

159 2.6. qPCR specificity, sensitivity, and efficiency

161 First, the specificity of the primer design was evaluated using the NCBI Primer-BLAST

162 tool. Furthermore, the specificity of the primer/probe was assayed experimentally with gDNA

163 isolated from various bacterial strains (Table 1).

164 The sensitivity of the qPCR assay was evaluated by preparing a serial tenfold dilution of

165 the plasmid pGEM/pheS, which had been linearised via restriction with the PstI restriction

166 enzyme, in 10 mM Tris(hydroxymethyl)aminomethane (pH 8.0). The DNA concentration was

167 determined using the Qubit dsDNA BR Assay kit (LubioScience, Lucerne, Switzerland), and the

168 molar concentration of the plasmid was calculated using the OligCalc oligonucleotide properties

169 calculator (Kibbe, 2007). Eleven independent dilution series were prepared by different

170 operators to control for inter-operator variations. Three dilution series per operator were

171 prepared by Operators A and B, and five dilution series were prepared by Operator C. The

172 dilutions were tested by qPCR and the calculated DNA copy numbers were used to assess the

173 limit of detection (LOD) and the linearity of the qPCR assay. The LOD was defined as the

174 concentration at which 95% of runs gave a positive result. The standard curve was generated

175 by plotting the Cq values of all dilutions as a function of the concentration and calculating the

176 linear regression in R (version 3.1.3, R Core Team, 2015). The qPCR efficiency was calculated

177 from the standard curves with an equation PCR efficiency of (10-1/slope) - 1.

The efficiency of the sum of DNA extraction plus qPCR was determined by comparing the calculated copy numbers with the colony forming units (cfu). For this, a culture of L. helveticus FAM1450 consisting of 9.93 x 108 cfu mL-1 was serially diluted tenfold in raw milk. Two dilution series were prepared and DNA from each dilution sample was extracted as described above before subjecting it to qPCR. The concentration of cfu in the MRS broth was determined via plate counting using modified MRS agar plates that contained lactose (20 g L-1) instead of glucose. The plates were incubated at 37 °C for 48 h under aerobic conditions. To test for qPCR inhibitors, tenfold serial dilution series of DNA extracts from raw milk, cheese, and NWCs were measured in triplicate via qPCR. The Cq values were plotted as a function of the dilution steps, and the linear regression was calculated using R v3.1.3 (R Core Team, 2015).

2.7. Enumeration of L. helveticus in dairy products

The applicability of the qPCR assay was evaluated by analysing various dairy samples (Table 2). The estimated number of copies per jL reaction (CReaction) for each reaction tube was calculated using the following formula:

Cq-b CReaction = 10 m

where b is the intercept of the regression line of the standard curve and m is the slope of the regression line of the standard curve. To account for the dilution steps and volume transformation during sample preparation, the CReaction values were multiplied by 10,000, 100, 10 and 25 for the lysed-cell fraction, cheese, NWC and raw milk samples, respectively, to obtain the gene copy number g-1 or mL-1, respectively (CSample). Only samples for which all experimental replicates gave a Cq value above the LOD were regarded as positive.

2.8. Cheesemaking

205 Model raclette-type cheeses were produced from 300 L of pasteurised milk, using the

206 mesophilic starters CHOOZIT MD 88 and CHOOZIT MA 14 (Danisco, France), without and with

207 a mixture of L. helveticus (FAM13019, FAM23236, and FAM23237) as an adjunct culture. Milk

208 was pre-ripened at 28-32 °C for 40 min. Coagulation occurred for 30 min at 32 °C in the

209 presence of rennet. Cutting and stirring were performed at 38 °C for 30 min. The whey-curd

210 mixture was filled in moulds and pressed at 30 °C for 0.75 h. After brining (11-13 °C, 16 h), the

211 cheeses were smear-ripened at 10-11 °C and 90-96% relative humidity. Samples were taken

212 after 24 h, 80 days, and 120 days of ripening.

214 3. Results

216 3.1. Specificity of the pheS targeted primers

218 In silico analysis using the Primer-BLAST tool against the RefSeq database revealed

219 that the primer aligned with 100% identity to the pheS gene of L. helveticus only (data not

220 shown). To experimentally assess the specificity of the primer, qPCR experiments were

221 performed with gDNA extracted from 24 L. helveticus strains and 23 strains belonging to other

222 LAB species. The target region was amplified in all L. helveticus strains. No signal was detected

223 in the 23 strains from the other LAB species tested (Table 1).

225 3.2. Copy number of pheS in L. helveticus genomes

227 Twenty-two L. helveticus genomes available in the GenBank database (for accession

228 numbers, see Fig. S1) were searched for the pheS gene. All genomes were found to contain a

229 single copy of the gene.

231 3.3. Sensitivity and linearity of the pheS-specific qPCR

The sensitivity and efficiency of the qPCR assay were evaluated with a plasmid containing a 490-bp region of the pheS gene and by preparing a dilution series of L. helveticus FAM1450 in raw milk.

The plasmid was restricted to obtain a linearised form. Eleven dilution series of the linear plasmid showed a linear relationship from 10 to 108 copies jL-1 (Fig. 1). The equation of the linear regression was Cq = -3.46x + 37.59 with a correlation coefficient (R2) of 0.994. The efficiency was calculated as 95%. Below 10 copies, no signal was obtained in all triplicate measurements. Therefore, the limit of detection was set to 10 copies jL-1.

L. helveticus FAM1450 was also serially diluted in milk. The relationship between the logarithm of bacterial cells per mL milk and the DNA extracted from the milk samples was linear from 9.9 to 9.9 x 108 cfu mL-1 (Fig. 2). The equation of the linear regression was Cq = -3.408x + 38.357, with a correlation coefficient (R2) of 0.999. Triplicate measurements of lower dilutions did not produce a signal.

3.4. Applicability of the novel qPCR assay

First, the applicability of the qPCR assay was examined with model raclette-type cheeses that were produced under defined conditions. The cheeses were produced with and without an L. helveticus adjunct culture. L. helveticus was detected in all cheeses to which it had been added during cheesemaking, whereas no L. helveticus was detected in the control cheese samples. The population level of L. helveticus, represented by the copy number, was 2.52 (± 0.59) x 108 copies g-1 cheese on average after 24 h of ripening dropping to 6.03 (± 0.51) x 106 copies g-1 cheese after 80 days and 2.32 ± 1.1 x 106 after 120 days of ripening (Fig. 3). We applied a sterilising filtration to the cheese extracts, as described by Gatti et al. (2008), to quantify the number of lysed cells. After 24 h of ripening, no signal of the presence of L. helveticus was detected in the lysed-cell fraction. After ripening periods of 80 d and 120 d, the lysed-cell fraction contained more than 107 pheS copies g-1 cheese (Fig. 3).

260 The qPCR assay was also tested with NWCs, cheese, and milk samples derived from

261 three Gruyère PDO cheese factories (Table 2). All of the NWCs tested showed an amplification

262 signal for the pheS gene within a range of 7.54 x 106 copies mL-1 and 2.43 x 108 copies mL-1.

263 All three Gruyère PDO cheeses tested positively for the presence of L. helveticus and contained

264 a mean of 1.8 x 108 and 1.8 x 107 pheS copies g-1 after 24 h and 6 months, respectively.

265 We also analysed retail cheeses (Table 2). All cheeses that we knew to be produced

266 with natural whey cultures tested positively for the presence of L. helveticus. Remarkably, a

267 cheese sample from the USA that we found to exhibit high leucyl aminopeptidase activity (133

268 IU kg-1), which is an indicator for the presence of L. helveticus (unpublished data), also tested

269 positively for L. helveticus.

270 We detected L. helveticus in two cheesemaking milk samples collected from Gruyère

271 PDO cheese factories (Table 2). The counts for the milk samples were low, but Sanger

272 sequencing revealed that the sequence of both amplicons aligned with 100% identity to the

273 pheS gene of L. helveticus DPC4571 (data not shown).

275 3.5. Absence of inhibitory compounds in the DNA extractions used for qPCR

277 DNA extracts from a milk, a NWC and a cheese were tested for PCR inhibitors by

278 examining tenfold serial dilutions using qPCR. All samples showed a linear correlation between

279 the Cq value and the dilution step, with correlation coefficients (R2) of 0.998, 0.995, and 0.999

280 for the NWC, cheese, and milk, respectively (supplementary data Fig. S2).

282 4. Discussion

284 DNA-based methods, such PCR and qPCR, are recognised as valuable alternatives to

285 culture-dependent methods for the enumeration of bacteria (Postollec, Falentin, Pavan,

286 Combrisson, & Sohier, 2011). The former methods are sensitive, precise, and rapid and can be

287 used to specifically detect and quantify bacteria with a variety of origins and in mixed

populations. Some PCR-based methods for the identification of L. helveticus have been reported in the past. These techniques amplify parts of the characteristic genes encoding Slayer proteins, peptidases, or peptidoglycan hydrolases, or the 16S-23S rRNA spacer region (Fortina, Ricci, Mora, Parini, & Manachini, 2001; Jebava, Chuat, Lortal, & Valence, 2014; Tilsala-Timisjarvi & Alatossava, 1997; Ventura, Callegari, & Morelli, 2000).

Desfosses-Focault et al. (2012) describe a qPCR that targets the tuf gene of L. helveticus, and the authors determined a limit of detection with 105 cfu per g cheese. Multiplex real-time PCR systems have also been described that amplify parts of the pheS, hsp60, prtH, or 16S rRNA genes of L. helveticus (Bottari, Agrimonti, Gatti, Neviani, & Marmiroli, 2013; Cremonesi et al., 2011; Herbel et al., 2013; Lu, Kong, Yang, & Kong, 2015). The purpose of these assays is the simultaneous detection and enumeration of L. helveticus in LAB mixtures. The detection limits described ranged from 104 cfu per mL (Herbel et al., 2013) to 183 copies (Bottari et al., 2013).

Remarkably, none of the mentioned qPCR systems was tested with gDNA from Lactobacillus gallinarum, which is the closest relative of L. helveticus. To our knowledge, only two PCR methods that can be used for the qualitative detection of L. helveticus have used L. gallinarum for specificity testing (Jebava et al., 2014; Ventura et al., 2000). Although L. gallinarum is typically found in the intestinal tracts of poultry, it has also been detected in cheese (Van Hoorde, Verstraete, Vandamme, & Huys, 2008). Therefore, we assume that differentiating between L. gallinarum and L. helveticus is an important requirement for a qPCR detection method in the dairy field.

In the present study, a simple and fast qPCR assay for the detection and quantification of Lactobacillus helveticus in food was developed. Specific primers and a hydrolysis probe were designed to target a region of the phenylalanyl-tRNA synthase (pheS) gene, which has previously been proven to be a suitable target gene for the specific identification of LAB species (Naser et al., 2007). We tested our qPCR assay with 48 strains belonging to 17 LAB species (Table 1). We found that the assay was highly specific to L. helveticus because neither the

closely related L. gallinarum, L. kefiranofaciens, L. crispatus, and L. acidophilus nor any of the other LAB species were detected with the assay.

We used a plasmid standard for absolute quantification. Because previous studies have shown that the use of circular plasmid standards in qPCR leads to a significant overestimation of the expected concentration (Hou, Zhang, Miranda, & Lin, 2010; Lin, Chen, & Pan, 2011), we used linearised plasmids in our qPCR assays. The qPCR showed a linear quantification over a range of 8 logs, with a limit of detection of ten copies per reaction (Fig. 1). Given the assumption that the pheS gene is a single copy gene, ten bacterial cells can be detected. The method described herein is therefore more sensitive than the methods mentioned above.

In comparison with colonies identified via plate counting, we determined approximately one additional log of copies of the pheS gene. We assume that this difference is mainly explained by cell chains that are not separated during plating. In fact, upon microscopic inspection, we observed that the L. helveticus strains used in this study form cell chains (data not shown). A further explanation could be the presence of dead bacterial cells. Discrepancies between culture-dependent and culture-independent methods were also reported by other researchers (Postollec et al., 2011). Viability dyes, such as ethidium monoazide or propidiume monoazide, which covalently modify the DNA of bacteria with compromised membranes and consequently cannot be amplified further by PCR, could be used to estimate the number of dead cells (Elizaquivel, Aznar, & Sanchez, 2014).

To evaluate the applicability of this method, we assayed NWCs, cheese, and milk (Table 2). The NWCs were collected from three cheese factories producing Gruyère PDO. We found that all NWCs contained L. helveticus ranging from 106 to 108 cells (copies) mL-1, indicating that L. helveticus is one of the predominant species present in these NWCs. L. helveticus was also found to be one of the predominant species present in the natural starter cultures used for the production of Grana Padano and Parmigiano Reggiano (Gatti, Bottari, Lazzi, Neviani, & Mucchetti, 2014). It is noteworthy that the production parameters for Gruyère PDO differ significantly from those for Italian cooked, extra-hard cheeses and that the microbial composition of these NWCs has not yet been studied in detail. Further studies will reveal

343 whether the microbial composition of the NWCs used for Gruyère PDO production are similar to

344 or different from that of Italian natural whey starters.

345 The detection of L. helveticus at a low population density in raw milk implies that this

346 ecosystem is also a habitat for L. helveticus strains. Coppola et al. (2006) also reported the

347 presence of this species in raw milk that was used for the manufacturing of a pasta-filata

348 cheese.

349 With regard to cheese, we were able to detect and enumerate L. helveticus in Gruyère

350 PDO cheese and various other cheese types. The appearance of intracellular high-molecular-

351 weight compounds, such as DNA, RNA, and bacterial enzymes, can be viewed as indicators of

352 bacterial autolysis, which is associated with enhanced protein degradation during cheese

353 ripening (Lortal & Chapot-Chartier, 2005). DNA, as long as it is not completely degraded, has

354 the advantage of being a species-specific and sensitive indicator of autolysis in a mixed

355 population. By applying a filter technique that removes whole cells, we observed, in a model-

356 type cheese manufactured with controlled production parameters, that L. helveticus reached

357 more than 108 copies g-1 cheese after 24 h. No free DNA of this species was detected at this

358 time-point. However, more than 107 free pheS gene copies g-1 were detected after 80 and 120

359 days of ripening, indicating the leakage of DNA out of the cells into the cheese matrix. This free

360 DNA was obviously not rapidly consumed by other bacteria. This is in agreement with Gatti et

361 al. (2008) who also detected significant amounts of free DNA from L. helveticus in raw milk

362 cheese that had ripened for 6 months. Furthermore, Treimo, Vegarud, Langsrud, and Rudi

363 (2006) amplified regions of the 16S rDNA gene of lactococci and propionibacteria present in the

364 supernatant of a liquid cheese model after five weeks of cultivation. Therefore, it can be

365 assumed that bacterial free DNA, even if considerably fragmented, is stable enough to be

366 amplified by PCR during the first months of cheese ripening. Compared with the more laborious

367 immunoassays, which also detected species-specific intracellular constituents, a PCR-based

368 method, combined with filter techniques, seems to be an easy-to-use alternative to investigate

369 bacterial autolysis and its impact on a cheese ecosystem harbouring a mixed population.

5. Conclusions

The qPCR assay developed herein for the detection of L. helveticus proved to be highly specific; we were able to discriminate between L. helveticus and its closest relative, L. gallinarum. Its high specificity and sensitivity make this qPCR assay a suitable tool for the detection and quantification of L. helveticus in dairy products. The assay can be used for the rapid identification of colonies on agar plates; for the study of the spatial and temporal distributions of L. helveticus during cheese ripening; for the identification of autolytic strains indicated by the presence of "free" DNA in medium or food; and for the detection of habitats of L. helveticus in the process of milk production and transformation. Overall, the qPCR assay presented here is a useful tool for various applications in research and industry.

Acknowledgements

We thank Claudia Wenzel for providing the model raclette-type cheese samples. References

Bottari, B., Agrimonti, C., Gatti, M., Neviani, E., & Marmiroli, N. (2013). Development of a multiplex real time PCR to detect thermophilic lactic acid bacteria in natural whey starters. International Journal of Food Microbiology, 160, 290-297. Coppola, S., Fusco, V., Andolfi, R., Aponte, M., Blaiotta, G., Ercolini, D. et al. (2006). Evaluation of microbial diversity during the manufacture of Fior di Latte di Agerola, a traditional raw milk pasta-filata cheese of the Naples area, Journal of Dairy Research, 73, 264-272. Cremonesi, P., Vanoni, L., Morandi, S., Silvetti, T., Castiglioni, B., & Brasca, M. (2011).

Development of a pentaplex PCR assay for the simultaneous detection of Streptococcus thermophilus, Lactobacillus delbrueckii subsp. bulgaricus, L. delbrueckii subsp. lactis, L.

398 helveticus, L. fermentum in whey starter for Grana Padano cheese. International Journal

399 of Food Microbiology, 146, 207-211.

400 de Man, J. C., Rogosa, M., & Sharpe, M. E. (1960). A medium for the cultivation of lactobacilli.

401 Journal of Applied Bacteriology, 23, 130-135.

402 Elizaquivel, P., Aznar, R., & Sanchez, G. (2014). Recent developments in the use of viability

403 dyes and quantitative PCR in the food microbiology field. Journal of Applied

404 Microbiology, 116, 1-13.

405 Fortina, M. G., Ricci, G., Mora, D., Parini, C., & Manachini, P. L. (2001). Specific identification of

406 Lactobacillus helveticus by PCR with pepC, pepN and htrA targeted primers. FEMS

407 Microbiology Letters, 198, 85-89.

408 Furet, J.-P., Quenee, P., & Tailliez, P. (2004). Molecular quantification of lactic acid bacteria in

409 fermented milk products using real-time quantitative PCR. International Journal of Food

410 Microbiology, 97, 197-207.

411 Gatti, M., Bottari, B., Lazzi, C., Neviani, E., & Mucchetti, G. (2014). Microbial evolution in raw-

412 milk, long-ripened cheeses produced using undefined natural whey starters. Journal of

413 Dairy Science, 97, 1-19.

414 Gatti, M., De Dea Lindner, J., De Lorentiis, A., Bottari, B., Santarelli, M., Bernini, V., et al.

415 (2008). Dynamics of whole and lysed bacterial cells during Parmigiano-Reggiano cheese

416 production and ripening. Applied and Environmental Microbiology, 74, 6161-6167.

417 Griffiths, M. W, & Tellez, A. M. (2013). Lactobacillus helveticus: The proteolytic system.

418 Frontiers in Microbiology, 4, Article no. 30.

419 Hammes, W. P., & Hertel, C. (2006). The genera Lactobacillus and Carnobacterium.

420 Prokaryotes, 4, 320-403.

421 Hannon, J. A., Wilkinson, M. G., Delahunty, C. M., Wallace, J. M., Morrissey, P. A., & Beresford,

422 T. P. (2003). Use of autolytic starter systems to accelerate the ripening of Cheddar

423 cheese. International Dairy Journal, 13, 313-323.

Herbel, S. R., Lauzat, B., von Nickisch-Rosenegk, M., Kuhn, M., Murugaiyan, J., Wieler, L. H., et al. (2013). Species-specific quantification of probiotic lactobacilli in yoghurt by quantitative real-time PCR. Journal of Applied Microbiology, 115, 1402-1410.

Hou, Y., Zhang, H., Miranda, L., & Lin, S. (2010). Serious overestimation in quantitative PCR by circular (supercoiled) plasmid standard: Microalgal pcna as the model gene. PloS One, 5, e9545.

Jebava, I., Chuat, V., Lortal, S., & Valence, F. (2014). Peptidoglycan hydrolases as species-specific markers to differentiate Lactobacillus helveticus from Lactobacillus gallinarum and other closely related homofermentative lactobacilli. Current Microbiology, 68, 551557.

Kenny, O., FitzGerald, R. J., O'Cuinn, G., Beresford, T., & Jordan, K. (2006). Autolysis of selected Lactobacillus helveticus adjunct strains during Cheddar cheese ripening. International Dairy Journal, 16, 797-804.

Kibbe, W. A. (2007). OligoCalc: an online oligonucleotide properties calculator. Nucleic Acids Research, 35, W43-46.

Lin, C.-H., Chen, Y.-C., & Pan, T.-M. (2011). Quantification bias caused by plasmid DNA conformation in quantitative real-time PCR assay. PLoS One, 6, e29101.

Lortal, S., & Chaot-Chartier, M.-P. (2005). Role, mechanisms and control of lactic acid bacteria lysis in cheese. International Dairy Journal, 15, 857-871.

Lu, W., Kong, W., Yang, P., & Kong, J. (2015). A one-step PCR-based method for specific

identification of 10 common lactic acid bacteria and Bifidobacterium in fermented milk. International Dairy Journal, 41, 7-12.

Naser, S. M., Dawyndt, P., Hoste, B., Gevers, D., Vandemeulebroecke, K., Cleenwerck, I., et al. (2007). Identification of lactobacilli by pheS and rpoA gene sequence analyses. International Journal of Systematic and Evolutionary Microbiology, 57, 2777-2789.

Postollec F., Falentin, H., Pavan, S., Combrisson, J., & Sohier D. (2011) Recent advances in

quantitative PCR (qPCR) applications in food microbiology. Food Microbiology, 28, 848861.

452 R Core Team (2015). R: A language and environment for statistical computing. R Foundation

453 for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/

454 Slattery, L., O'Callaghan, J., Fitzgerald, G., Beresford, T., & Ross, R. (2010). Invited review:

455 Lactobacillus helveticus - a thermophilic dairy starter related to gut bacteria. Journal of

456 Dairy Science, 93, 4435-4454.

457 Taverniti, V., & Guglielmetti, S. (2012). Health-promoting properties of Lactobacillus helveticus.

458 Frontiers in Microbiology, 3, Article no. 392.

459 Tilsala-Timisjarvi, A., & Alatossava, T. (1997). Development of oligonucleotide primers from the

460 16S-23S rRNA intergenic sequences for identifying different dairy and probiotic lactic

461 acid bacteria by PCR. International Journal of Food Microbiology, 35, 49-56.

462 Treimo, J., Vegarud, G., Langsrud, T., & Rudi, K. (2006). Use of DNA quantification to measure

463 growth and autolysis of Lactococcus and Propionibacterium spp. in mixed populations.

464 Applied and Environmental Microbiology, 72, 6174-6182.

465 Turgay, M., Schaeren, W., Wechsler, D., Butikofer, U., & Graber, H. U. (2016). Fast detection

466 and quantification of four dairy propionic acid bacteria in milk samples using real-time

467 quantitative polymerase chain reaction. International Dairy Journal, 61, 37-43.

468 Untergasser, A., Cutcutache, I., Koressaar, T., Ye, J., Faircloth, B. C., Remm, M., et al. (2012).

469 Primer3 - new capabilities and interfaces. Nucleic Acids Research, 40, e115.

470 Valence, F., Deutsch, S.-M., Richoux, R., Gagnaire, V., & Lortal, S. (2000). Autolysis and

471 related proteolysis in Swiss cheese for two Lactobacillus helveticus strains. Journal of

472 Dairy Research, 67, 261-271.

473 Van Hoorde, K., Verstraete, T., Vandamme, P., & Huys, G. (2008). Diversity of lactic acid

474 bacteria in two Flemish artisan raw milk Gouda-type cheeses. Food Microbiology, 25,

475 929-935.

476 Ventura, M., Callegari, M. L., & Morelli, L. (2000). S-layer gene as a molecular marker for

477 identification of Lactobacillus helveticus. FEMS Microbiology Letters, 189, 275-279.

Figure legends

Fig. 1. Plasmid standard curve obtained by plotting the threshold cycle (Cq) values against the logarithm of the calculated gene copy numbers from serial tenfold dilutions of plasmid pGEM/pheS. Error bars indicate the standard deviations (n = 11).

Fig. 2. Comparison of qPCR against plate counting of L. helveticus FAM1450 in milk. The solid line represents the regression curve determined experimentally; the dashed line illustrates the theoretical expected curve. The DNA extractions were performed twice. Error bars represent standard deviations of these two assays.

Fig. 3. Enumeration of L. helveticus in cheese; whole cells and lysed cells corresponding to pheS copy numbers were enumerated by qPCR in model-type cheeses. Samples were taken after 24 h, 80 days, and 120 days of ripening. Error bars represent the standard deviation of two cheese samples.

Table 1

Strain Reference Isolation source qPCR result

Lactobacillus helveticus

FAM1450 ACC NA +

FAM1476 ACC NA +

FAM21493 ACC MSS +

FAM22081 ACC NA +

FAM21339 ACC MSS +

FAM22076 ACC NWC +

FAM21456 ACC MSS +

FAM8104 ACC Raw milk cheese (Tilsit) +

FAM13019 ACC NWC +

FAM8627 ACC NA +

FAM22155 ACC NWC +

FAM2888 ACC NA +

FAM20575 ACC NWC +

FAM22330 ACC MSS +

FAM1213 ACC NA +

FAM22074 ACC NWC +

FAM1172 ACC NA +

FAM8105 ACC Raw milk cheese (Tilsit) +

FAM22079 ACC NWC +

FAM22472 ACC NA +

FAM13019 ACC NWC +

B02 Chr. Hansen Holding A/S, Denmark Starter culture +

LH32 Chr. Hansen Holding A/S, Denmark Commercial strain +

DSM 20075T DSMZ Emmental cheese +

Lactobacillus gallinarum DSM 10532T

DSMZ ^Chicken crop -

LMG 14751 BCCM Chicken faeces -

LMG 14754 BCCM k Chicken faeces -

LMG 14755 BCCM Chicken faeces -

LMG 18181 BCCM Chicken intestine -

LMG 22870 BCCM Laying hen vagina -

Lactobacillus kefiranofaciens subsp. kefiranofaciens

DSM 5016T DSMZ Kefir grains -

Lactobacillus kefiranofaciens subsp. kefirgranum

DSM 10550T DSMZ Kefir grains -

Lactobacillus crispatus

DSM 20584T DSMZ Eye -

Lactobacillus acidophilus

DSM 20079T DSMZ Human -

Lactobacillus johnsonii

DSM 10533T DSMZ Human blood -

Lactobacillus delbrueckii subsp. bulgaricus

DSM 20081T DSMZ Yoghourt -

Lactobacillus delbrueckii subsp. lactis

DSM 20072T DSMZ Emmental cheese -

Lactobacillus fermentum

DSM 20052T DSMZ Fermented beets -

Lactobacillus casei

ATCC 334 ATCC, Virginia, USA Emmental cheese -

FAM18121 ACC Gruyère PDO cheese -

Lactobacillus casei subsp. tolerans

DSM 20012 DSMZ Pasteurised milk -

DSM 20258T DSMZ Pasteurised milk -

Lactobacillus paracasei subsp. paracasei

DSM 5622T DSMZ Milk products -

Lactobacillus rhamnosus

CCUG 34291 CCUG Human faeces -

Enterococcus faecium

FAM8492 ACC MSS -

Streptococcus thermophilus

DSM 20617T DSMZ Pasteurised milk -

Lactobacillus amylovorus DSMZ Cattle waste corn

DSM 20531T fermentation -

4 a Abbreviations are: ACC, Agroscope Culture Collection (Switzerland); DSMZ, German Collection of Microorganisms

5 and Cell Cultures; BCCM, Belgian Co-Ordinated Collections of Micro-Organisms, CCUG, Culture Collection University

6 of Göteborg (Sweden); NWC, natural whey culture; MSS, mixed strain starter; NA, information about isolation source

7 not available. T Indicates type strain; + and - indicate presence (+) or absence (-) of a qPCR signal.

Table 2

Sample log pheS copies g-1 or mL-1

Long-ripened hard cheese (produced with NWC), Switzerland 7.16 ± 0.04

Tilsit (produced with NWC), Switzerland 8.72 ± 0.42

Goat cheese, Switzerland 6.32 ± 0.03

Cheese with high leucyl aminopeptidase activity (133 IU kg-1), USA 6.98 ± 0.11

Mozzarella di Bufala Campana PDO, Italy 8.55 ± 0.01

Grana Padano PDO, Italy 7.34 ± 0.03

Parmigiano Reggiano PDO, Italy 6.02 ± 0.04

Provolone, Italy 8.75 ± 0.07

Tête de Moine PDO, Switzerland No L . helveticus detected

Tilsit (no NWC used for the production), Switzerland No L . helveticus detected

Emmental PDO, Switzerland No L . helveticus detected

Gruyère PDO (ripened for 24 h), cheese factory 1, Switzerland 8.29 ± 0.01

Gruyère PDO (ripened for 6 months), cheese factory 1, Switzerland 6.92 ± 0.08

NWC (incubated at 38 °C for 20 h), cheese factory 1, Switzerland 6.82 ± 0.29

Gruyère PDO (ripened for 24 h), cheese factory 2, Switzerland 8.49 ± 0.01

Gruyère PDO (ripened for 6 months), cheese factory 2, Switzerland 7.11 ± 0.4

NWC (incubated at 32 °C for 20 h), cheese factory 2, Switzerland 7.23 ± 0.28

NWC (incubated at 38 °C for 20 h), cheese factory 2, Switzerland 7.48 ± 0.05

Gruyère PDO (ripened for 24 h), cheese factory 3, Switzerland 7.58 ± 0.06

Gruyère PDO (ripened for 6 months), cheese factory 3, Switzerland 7.06 ± 0.21

NWC (incubated at 32 °C for 20 h), cheese factory 3, Switzerland 7.46 ± 0.03

NWC (incubated at 38 °C for 20 h), cheese factory 3, Switzerland 8.38 ± 0.04

NWC (incubated at 38 °C for 10 h), cheese factory 3, Switzerland 7.41 ± 0.04

Raw cheesemaking milk (6th January 2015), cheese factory, Switzerland No L . helveticus detected

Raw cheesemaking milk (13th January 2015), cheese factory, Switzerland No L . helveticus detected

Raw cheesemaking milk (14th April 2015), cheese factory, Switzerland 2.36 ± 0.19

Raw cheesemaking milk (21st April 2015), cheese factory, Switzerland 3.59 ± 0.07

12 a Mean values for triplicate measurements (± standard deviation) are shown. PDO, Protected Designation of Origin;

13 NWC, natural whey culture.

Figure 1

Figure 2

10« _ 10s _

1Q7 _. 106 _ 105 10" -to3 102 . 101 . 10° .

whole-cells lysed-cells

whole-cells lysed-cells

whole-cells lysed-cells

Figure 3