Detection and quantification of Bacillus cereus group in milk by droplet digital PCR

    Detection and quantification of Bacillus cereus group in milk by droplet digital PCR Davide Porcellato, Judith Narvhus, Siv Borghild Skeie PII: DOI: Reference:

S0167-7012(16)30099-9 doi: 10.1016/j.mimet.2016.05.012 MIMET 4896

To appear in:

Journal of Microbiological Methods

Received date: Revised date: Accepted date:

1 April 2016 3 May 2016 16 May 2016

Please cite this article as: Porcellato, Davide, Narvhus, Judith, Skeie, Siv Borghild, Detection and quantification of Bacillus cereus group in milk by droplet digital PCR, Journal of Microbiological Methods (2016), doi: 10.1016/j.mimet.2016.05.012

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.

ACCEPTED MANUSCRIPT Title Detection and quantification of Bacillus cereus group in milk by droplet digital

PT

PCR

SC RI

Authors

Davide Porcellatoa*, Judith Narvhusa, Siv Borghild Skeiea

Department of Chemistry, Biotechnology and Food Science, Norwegian

MA

a

NU

Affiliation

Corresponding author:

PT

*

ED

University of Life Science, P.O. Box 5003, N-1432 Ås, Norway

Davide Porcellato, Postbox 5003, 1432 Ås, Norway

CE

Tel.: +4764965143 - Fax: +4764965901

AC

E-mail address: [email protected]

1

ACCEPTED MANUSCRIPT Abstract

PT

Droplet digital PCR (ddPCR) is one of the newest and most promising methods for the detection and quantification of molecular targets by PCR. Here, we optimized

SC RI

and used a new ddPCR assay for the detection and quantification of the Bacillus cereus group in milk. We also compared the ddPCR to a standard qPCR assay. The new ddPCR assay showed a similar coefficient of determination and a better limit of

NU

detection compared to the qPCR assay during quantification of the target molecules in

MA

the samples. However, the ddPCR assay has a limitation during quantification of a high number of target molecules. This new assay was then tested for the

ED

quantification of the Bacillus cereus group in 90 milk samples obtained over three months from two different dairies and the milk was stored at different temperatures

PT

before sampling. The ddPCR assay showed good agreement with the qPCR assay for the quantification of the Bacillus cereus group in milk, and due to its lower detection

CE

limit more samples were detected as positive. The new ddPCR assay is a promising

AC

method for the quantification of target bacteria in low concentration in milk.

Keywords

Droplet digital PCR, quantitative PCR, Bacillus cereus group, milk

2

ACCEPTED MANUSCRIPT 1

Introduction The presence of spoilage bacteria, such as Bacillus spp., in dairy products

PT

results in economical losses for the industry during the storage and processing chain.

SC RI

Bacillus spp. are aerobic spore-forming bacteria and are found in a wide variety of potential contamination sources, such as soil, water, udder, equipment. Contamination with Bacillus spp., and in particular the Bacillus cereus group, in dairy products is

NU

both a safety and quality concern (Arnesen et al., 2008, De Jonghe et al., 2010). These bacteria are able to survive pasteurization and can be detected in several dairy

MA

products such as milk and cream before and after heat treatment, (Bartoszewicz et al., 2008, Granum, 2002). Fast methods for detection and quantification of the Bacillus

ED

cereus group in milk, and other dairy products, are essential tool for the dairy industry.

PT

Conventional methods to quantify the Bacillus cereus group include culturing

CE

techniques on selective media and identification of the isolates by biochemical and molecular methods. These techniques are time-consuming and laborious. Molecular-

AC

based quantification and identification of the Bacillus cereus group directly in the sample has been successfully explored by quantitative PCR (qPCR) in different food matrices (Dzieciol et al., 2013, Fernandez-No et al., 2011, Martinez-Blanch et al., 2009). The main advantage of this method is that results can be obtained in a few hours rather than days, as in the case of conventional methods. Recently, a new PCR-based technology has been introduced for the quantification and detection of molecular targets (Hindson et al., 2011). Droplet digital PCR is the third generation of PCR techniques, which allows the absolute quantification of molecular targets without the use of standard curves, due to the recent advent of compartmentalization. The PCR mix, similar to the standard qPCR,

3

ACCEPTED MANUSCRIPT is divided in a large number of partitions, each of which may contain zero or 1 (or more) copies of the target molecule. After PCR amplification, the partitions are

PT

counted as positive or negative. The absolute quantification of the original number of copies added to the PCR reaction is accomplished using binomial Poisson statistics.

SC RI

This allows ddPCR to obtain an absolute quantification without the use of standard curves. Droplet digital PCR has previously been applied over a wide range of fields to detect and quantify molecular targets such as food pathogens (Bian et al., 2015,

NU

Klancnik et al., 2015), GMO (Koppel et al., 2015, Morisset et al., 2013), HIV

MA

(Kiselinova et al., 2014), and soil bacteria (Kim et al., 2014). In this study, we optimized and applied a ddPCR assay for the detection and

ED

quantification of the Bacillus cereus group in milk. We used a previously developed qPCR assay, which targets the gyrB gene, and we compared both qPCR and ddPCR

PT

assays on a set of 90 milk samples which were obtained from different dairies, and had different storage times and conditions. The advantages and limitations of the

AC

CE

ddPCR assay are also discussed.

4

ACCEPTED MANUSCRIPT 2 2.1

Materials and Methods Bacteria samples and culturing

PT

A pure culture of Bacillus cereus ATCC 14579, obtained from the laboratory

SC RI

collection was used to validate and optimize the method. Further, eight isolates of Bacillus (2 strains of B. cereus, and one strain each of B. mycoides, B. weihenstephanensis, B. smithinii, B licheniformis, B. amyloliquefaciens and B.

NU

pumilus) previously isolated from milk and stored in our collection were used to assess the inclusivity and exclusivity testing of the assay used in this study as

MA

previously described (Dzieciol, Fricker, Wagner, Hein and Ehling-Schulz, 2013). All the cultures were kept in brain-heart infusion (BHI, Oxoid Ltd., Basingstoke, United

ED

Kingdom) broth with added 15% (v/v) glycerol and stored at -80 °C. All the isolates were previously identified by 16S rRNA sequencing as described by Porcellato et al.

Milk samples and artificially contaminated milk samples

CE

2.2

PT

(2012).

AC

The milk samples (n = 90) used in this experiment were collected during three different months (from July to September) from two different dairies and included: raw milk from silo tanks (n = 18), pasteurized milk from the bulk tank before packaging (n = 18), pasteurized milk in cartons sampled immediately after packaging (n = 18) and pasteurized milk in cartons kept for 13 days at 4 °C (n = 18) and 8 °C (n = 18). The milk samples which were not stored were collected on the day of pasteurization and kept for 24 hours at 4 °C before analysis. On the day of the analysis, 10 mL of milk were collected for total DNA extraction. Each category of milk (raw milk, pasteurized milk, pasteurized milk in carton) collected over the three months, and from the two dairies, was independently analyzed in triplicate.

5

ACCEPTED MANUSCRIPT Artificially contaminated milk was used to validate the DNA extraction performance by both qPCR and ddPCR assay. Ten mL of UHT milk were spiked with Bacillus

PT

cereus ATCC 14579 cells from an initial concentration of log 0 CFU mL-1 to log 4.54

2.3

SC RI

CFU mL-1 and used for DNA extraction.

DNA extraction from bacteria and milk samples

DNA extraction from the pure cultures was performed from 1 mL of 18 h

NU

culture in BHI at 30 °C. DNA was extracted using the UltraClean Microbial DNA isolation kit (Mobio laboratories Inc, Carlsbad, CA, USA) according to the

MA

manufacturer´s instructions. Ten mL of milk samples or spiked milk samples were used for DNA extraction. The milk samples were centrifuged for 10 min at 8000 x g

ED

and the pellet was resuspended and washed twice with 1 mL of 2% sodium citrate water (w/v). DNA was extracted using the same method as for the pure isolates, with

PT

minor changes. After resuspension of the pellet in 300 μl of bead solution (Mobio), it

CE

was heat-treated at 70 °C for 10 min. The bead-beating time was increased to 15 min and elution of DNA was performed twice using the same eluate. DNA was stored at -

AC

20 °C until used. These changes were performed to optimize the DNA extraction from vegetative cells of Bacillus spp.

2.4

Real-time PCR Real-time PCR (qPCR) was performed in a LightCycler 480 II system (Roche,

Mannheim, Germany). The final PCR volume was 20 μl containing 1X of LightCycler® 480 Probes Master (Roche), 0.3 μM of each forward and reverse primer and

0.2

μM

of

the

probe.

The

GCCCTGGTATGTATATTGGATCTAC-3´ GGTCATAATAACTTCTACAGCAGGA-3´)

primer

pair

and and

(forward

reverse probe

(FAM

5´5´5´-

6

ACCEPTED MANUSCRIPT CCATTTTTTCTTGTATACCAACT-3 MGB) used in this study targeted the gyrB gene of the Bacillus cereus group (Dzieciol, Fricker, Wagner, Hein and Ehling-

PT

Schulz, 2013). PCR mix made with Bacillus cereus ATCC 14579 used for the construction of the standard curve contained 1 μl of DNA while PCR mixes made

SC RI

with milk samples contained 5 μL of the extracted DNA. PCR amplification was performed with initial denaturation at 95 °C for 5 min followed by 45 cycles of denaturation at 95 °C for 10 sec and annealing and elongation at 60 °C for 1 min. The

NU

quantification cycle (Cq) was calculated with the second derivative method using the

MA

Lightcycler 480 software ver. 1.5 (Roche). Samples with Cq higher than 38 were considered negative. In each run, standard curve samples and negative controls were

2.5

ED

added in triplicate.

Standard curve construction and qPCR data analysis

PT

A standard curve for qPCR was constructed from genomic DNA of Bacillus

CE

cereus ATCC 14579. The DNA concentration was calculated using the Qubit dsDNA HS assay kit (ThermoFischer Scientific, Waltham, MA, USA). The number of

AC

Bacillus cereus ATCC 14579 genome copies were calculated with the following formula: (amount * 6.022x1023) / (length * 1x109 * 650), where amount was the concentration of DNA used to construct the standard curve and length was the length of the Bacillus cereus ATCC 14579 genome (5427080 bp). The standard curve was constructed by plotting the Cq value from the qPCR against the number of genome copies obtained from a serial dilution of Bacillus cereus ATCC 14579 DNA.

2.6

Digital droplet PCR reaction and data analysis Optimization of the ddPCR assay was done with respect to the annealing

temperature (thermal gradient from 55 to 65 °C) and the probe concentration (from

7

ACCEPTED MANUSCRIPT 0.05 μM to 3 μM). Droplet digital PCR (ddPCR) reaction were performed using 1x ddPCR Supermix for probe (Bio-Rad, Pleasanton, CA), 0.4 μM of each forward and

PT

reverse primer and 0.1 μM of the probe (after optimization) and 1 or 5 μl of DNA for standard curve and milk samples, respectively. The final volume of the reaction was

SC RI

22 μl. Twenty microliter of the ddPCR reaction were used to generate the droplet mix in an 8-well cartridge using the QX100 droplet generator (Bio-Rad). The emulsion (40 μl) was then transferred to a 96-well plate and amplified with the following

NU

condition: denaturation for 10 min at 95 °C, followed by 40 cycles of denaturation at

MA

94 °C for 30 sec and annealing and elongation at 60 °C for 1 min (after optimization of the annealing temperature). The PCR products were denatured at 98 °C for 10 min and kept at 4 °C until the droplets were read. Ramp rate used in the droplet PCR was

ED

2 °C sec-1. The 96-well plate was then transferred to the QX100 droplet reader (Bio-

PT

Rad) and data acquisition and analysis was performed using QuantaSoft software ver 1.7 (Bio-Rad). The fluorescence amplitude threshold, used for the discrimination of

CE

the positive and negative droplets in QuantaSoft software was set between 2000 and

AC

2200. This value was chosen after optimization of the probe concentration and annealing temperature in the ddPCR assay. The concentration values were calculated by QuantaSoft software (in copies μL-1) and multiplied by 22 (the initial PCR volume) to obtain the absolute number of copies added to the PCR reaction. If a sample contained all positive droplets and quantification was not possible, the same sample was diluted 1:100 and reanalyzed. The number of copies was then multiplied by the dilution factor to calculate the original number of copies in the sample. The same Bacillus cereus ATCC 14579 DNA which was used for preparation of the standard curves in the qPCR assay, was also run in triplicate with the ddPCR assay to quantify the exact number of genome copies. The concentration values obtained from the

8

ACCEPTED MANUSCRIPT ddPCR were plotted against the theoretical number of genomes or cells to obtain the standard curve. Further, ten replicates of the 3 lowest dilutions of the DNA used for

Results and Discussion

SC RI

3

PT

the standard curve were run with the ddPCR to identify the limit of detection.

Droplet digital PCR has a potential of being an efficient method for the detection and quantification of spoilage bacteria in food. Here, we investigate the

NU

potential of ddPCR for the detection and quantification of bacteria in the Bacillus cereus group in milk samples. We adapted a qPCR method, based on a hydrolysis

in the Bacillus cereus group.

MA

probe, previously used in a multiplex qPCR assay for the detection of the gyrB gene

ED

To validate the amplification of the target gene and the specificity of the assay, qPCR was performed on a set of different Bacillus species previously isolated and

PT

characterized at species level by 16S rRNA. Successful amplifications of the target

CE

DNA from isolates included in the Bacillus cereus group (B. cereus, B. mycoides and B. weihenstephanensis) were detected, while no amplification after 40 cycles was

AC

detected for the Bacillus isolates not included in the Bacillus cereus group (B. smithinii, B. licheniformis, B. amyloliquefaciens and B. pumilus). This was in accordance with previous results obtained by using the same primers/probe set (Dzieciol, Fricker, Wagner, Hein and Ehling-Schulz, 2013). After optimization of the ddPCR assay, DNA from the same Bacillus species (with a final DNA concentration from 0.1 to 0.3 ng reaction-1) was analyzed by ddPCR and no positive droplets were detected for the isolates not included in the Bacillus cereus group. Two steps were used to optimize the ddPCR assay. First, ddPCR results from four different concentrations of the probe were assessed. The amount of DNA used in the experiment was log 4 of B. cereus ATCC 14579 genome copies. This

9

ACCEPTED MANUSCRIPT concentration of DNA was chosen to obtain an identical number of both positive and negative droplets. Except for the lowest concentration (0.05 μM), the other 3

PT

concentrations showed good separation between the positive and negative droplets (Figure 1A). The concentration with the highest amplitude difference between the

SC RI

positive and the negative droplets was the highest concentration of the probe. However, the concentration of 0.1 μM was chosen for further analyses, as the cost per samples was lower. The second optimization step was performed by using a thermal

NU

gradient PCR from 55 °C to 65 °C to check for the annealing temperature which

MA

would best separate the positive droplets from the negative. No differences were seen in the negative droplets amplitude (~2000) between all the temperatures. A decrease

ED

in the amplitude of the positive droplets was detected as the annealing temperature increased (Figure 1B). However, at a lower annealing temperature (< 59 °C) the

PT

amplitude distribution of the positive droplets was high (Figure 1B). From 59 °C to 61.4 °C the amplitude distribution of the droplets was minimal (Figure 1B) and

CE

therefore the annealing temperature used for further analysis was 60 °C.

AC

Evaluation of the qPCR and ddPCR sensitivity, or limit of detection (LOD) was performed using serial dilution of DNA from Bacillus cereus ATCC 14579. The LOD is the minimum target copy number that can be precisely measured in an assay (Bustin et al., 2009). The number of copies from the ddPCR assay and the Cq values from the qPCR assay were plotted against the theoretical number of genomes obtained after quantification of the initial DNA concentration. Both methods showed good linearity within the range of quantification and obtained a coefficient of determination (R2) of 0.999 and 0.995 for ddPCR and qPCR, respectively (Figure 2). Yang et al. (2014) suggested constructing a precise standard curve for the qPCR assay after precise quantification of the molecular target by using ddPCR. Here, we

10

ACCEPTED MANUSCRIPT constructed a new standard curve after precise quantification of the number of copies by ddPCR. This new standard curve showed good linearity and a coefficient of

PT

determination (R2) of 0.988 (data not shown). This standard curve was used for the quantification of the target molecules in milk samples. To identify the lowest LOD in

SC RI

our ddPCR assay, 10 replicates were run with the 3 lowest concentrations of the DNA samples used to construct the standard curve. The LOD was determined as the lowest concentration amplified in all the 10 replicates. The LOD for the ddPCR assay was

NU

8.4 copies ± 2.2 (25.9%, Figure 2). The negative control in the ddPCR assay showed

MA

no positive amplification (Figure 1C). The LOD in the qPCR assay was also calculated. The dilution with 100% amplification in all the 10 replicates was 49 (log

ED

1.69) target copies (Figure 2B). The LOD in ddPCR for the quantification of the low target molecules in the samples is in agreement with previous findings, which showed

PT

similar or lower sensitivity of this method compared to qPCR (Henrich et al., 2012, Kim, Jeong and Cho, 2014, Morisset, Stebih, Milavec, Gruden and Zel, 2013).

CE

Milk samples were artificially contaminated with serial dilutions of the Bacillus

AC

cereus cells in low concentration (from log 0.5 to log 4.5 cell mL-1) to validate the new ddPCR assay in milk. This low concentration of inoculum was chosen due to the maximum number of copies that can be amplified by the ddPCR assay (log 5 copies reaction-1). The artificially contaminated milk samples were subjected to the same DNA extraction method used for the milk samples, and qPCR and ddPCR results were plotted against the CFU mL-1 count. The standard curve obtained from the qPCR assay showed a PCR efficiency of 94.6%. The LOD in the qPCR assay was higher (log 1.5 cell mL-1) compared to the ddPCR assay (log 1.01 cell mL-1, results of three replicates, data not shown). Real-time PCR has previously shown to be more affected by the presence of PCR inhibitors compared to ddPCR and this might explain the

11

ACCEPTED MANUSCRIPT higher detection limit in qPCR compared to ddPCR in milk samples (Racki et al., 2014, Yang, Paparini, Monis and Ryan, 2014). This lower inhibition in the ddPCR

PT

might be due to the PCR reaction partitioning, which reduces the exposure to PCR inhibitors in the droplets (Dingle et al., 2013, Hindson et al., 2011).

SC RI

A total of 90 samples of raw milk, pasteurized milk and packaged milk from cartons kept for 1 day at 4 °C and after 13 days at 4 and 8 °C were analyzed and compared to the standard qPCR assay. The milk samples were collected from two

NU

different dairies and kept at 4 and 8 °C for 13 days as previous results from our

MA

laboratory showed growth of Bacillus spp. by plate count (> log 4 CFU mL-1 after 13 days) at 8 °C and no increase of Bacillus spp. count at the lowest temperature. The limits of detection used to consider 1) a sample as positive and 2) to quantify the

ED

Bacillus cereus group were 8.2 copies reaction-1 and 49 copies reaction-1 for ddPCR

PT

and qPCR, respectively. A total of 53 (58%) and 18 (20%) samples were positive for the Bacillus cereus group as determined by ddPCR and qPCR, respectively. The

CE

coefficient of determination (R2) between the copies of the target quantified by both

AC

methods was 0.948 (Figure 3). The samples, which were positive in the ddPCR but negative in the qPCR, were shown to have a lower concentration of the target molecules by ddPCR (< log 3.2 copies). This indicates how ddPCR might be a more suitable method to quantify Bacillus cereus in milk samples containing a low amount of the target. Bacillus spp. was detected at 5 of 6 sampling time for raw milk, and it was detected in all the samples of pasteurized milk before and after packaging, with concentration from log 0.4 to 3.41 genome equivalents mL-1 (Table 1). Milk samples kept at 4 °C for 13 days did not show an increase in Bacillus spp. count. However, after storage at 8 °C for 13 days the concentration of Bacillus spp. was higher than log 4.2 genome equivalents mL-1 and in some samples the concentration reached log 6.35

12

ACCEPTED MANUSCRIPT genome equivalents mL-1 (Table 1). In some of the samples kept at 8 °C, the concentrations were over the maximum detection limit of the ddPCR (log 5 copies

PT

reaction-1). These samples were reanalyzed after a 2-log dilution. However, in the qPCR assay, this was not a problem, as qPCR was able to quantify target molecules in

SC RI

concentrations higher than log 7 copies reaction-1.

In this study, advantages and limitations of the new ddPCR method were identified. First, standard curves were not needed during ddPCR assays. The ddPCR

NU

methodology quantified the absolute number of copies added to the reaction by

MA

partitioning the PCR volume and by using the Poisson distribution estimation (Hindson et al., 2011, Pinheiro et al., 2012). This might reduce the bias introduced in

ED

the qPCR assays as calculation of the target concentration had to be made to construct the standard curve. The use of ddPCR to quantify the number of copies in the

PT

standard curve used in the qPCR might reduce this bias (Yang, Paparini, Monis and Ryan, 2014). The ddPCR was also showed to perform better, compared to qPCR,

CE

during the quantification of a low numbers of target molecules and this would be an

AC

advantage in samples containing a low number of the target bacteria. Pasteurized milk normally contains low bacterial counts, but the presence of the Bacillus cereus group might significantly influence the quality and safety during processing and cold storage (Bava et al., 2011, De Jonghe, Coorevits, De Block, Van Coillie, Grijspeerdt, Herman, De Vos and Heyndrickx, 2010, Ivy et al., 2012, Rasolofo et al., 2010). The first limitation of the ddPCR encountered in this study was quantification of the target molecules when in high abundance. Samples with a high abundance of the target (>105) cannot be quantified. This was the case of milk stored for 13 days at 8 °C. These samples were diluted and run again on the ddPCR, thus increasing cost and time needed for the analysis. Therefore, foreknowledge of the amount of target

13

ACCEPTED MANUSCRIPT molecules is needed. This limitation was not encountered in the qPCR assay, which has a better range of quantification. Lastly, compared to standard qPCR, the time and

PT

cost were almost doubled in ddPCR. In conclusion, we investigated a new ddPCR assay for the quantification of the

SC RI

Bacillus cereus group. We adapted a previous developed qPCR assay and with few optimization steps we showed the potential applicability of ddPCR to target spoilage bacteria in a food sample. The ddPCR assay showed a lower detection limit compared

NU

to qPCR and this might be an advantage in samples with low copies of the target

MA

molecules, such as milk. Droplet digital PCR is a promising technology, which has so far been little utilized for food samples. To our knowledge, this is the first study

ED

reporting the use of ddPCR to target spoilage bacteria in milk. Molecular quantification of specific targets in food samples can be a useful tool in the industry to

PT

evaluate quality and safety of the products. Acknowledgements

CE

The authors would like to acknowledge Marte Monshaugen, Ahmed Abdelghani

AC

and Tine SA for their assistance during sampling and Trude Gjæver Reinholdtsen for the collection of isolates. The authors also acknowledge Department of Chemistry, Biotechnology and Food Science, TINE SA, the Norwegian Foundation for Research Levy on Agricultural Products (FFL) and the Norwegian Agricultural Agreement Research Fund (JA) for financial support.

14

ACCEPTED MANUSCRIPT References Arnesen, L.P.S., Fagerlund, A., Granum, P.E., 2008. From soil to gut: Bacillus

PT

cereus and its food poisoning toxins. Fems Microbiol Rev. 32, 579-606. Bartoszewicz, M., Hansen, B.M., Swiecicka, I., 2008. The members of the

SC RI

Bacillus cereus group are commonly present contaminants of fresh and heat-treated milk. Food Microbiol. 25, 588-596.

Bava, L., Zucali, M., Sandrucci, A., Brasca, M., Vanoni, L., Zanini, L.,

NU

Tamburini, A., 2011. Effect of cleaning procedure and hygienic condition of milking

MA

equipment on bacterial count of bulk tank milk. J Dairy Res. 78, 211-219. Bian, X.J., Jing, F.X., Li, G., Fan, X.Y., Jia, C.P., Zhou, H.B., Jin, Q.H., Zhao,

ED

J.L., 2015. A microfluidic droplet digital PCR for simultaneous detection of pathogenic Escherichia coli O157 and Listeria monocytogenes. Biosens Bioelectron.

PT

74, 770-777.

Bustin, S.A., Benes, V., Garson, J.A., Hellemans, J., Huggett, J., Kubista, M.,

CE

Mueller, R., Nolan, T., Pfaffl, M.W., Shipley, G.L., Vandesompele, J., Wittwer, C.T.,

AC

2009. The MIQE Guidelines: Minimum Information for Publication of Quantitative Real-Time PCR Experiments. Clin Chem. 55, 611-622. De Jonghe, V., Coorevits, A., De Block, J., Van Coillie, E., Grijspeerdt, K., Herman, L., De Vos, P., Heyndrickx, M., 2010. Toxinogenic and spoilage potential of aerobic spore-formers isolated from raw milk. Int J Food Microbiol. 136, 318-325. Dingle, T.C., Sedlak, R.H., Cook, L., Jerome, K.R., 2013. Tolerance of DropletDigital PCR vs Real-Time Quantitative PCR to Inhibitory Substances. Clin Chem. 59, 1670-1672.

15

ACCEPTED MANUSCRIPT Dzieciol, M., Fricker, M., Wagner, M., Hein, I., Ehling-Schulz, M., 2013. A novel diagnostic real-time PCR assay for quantification and differentiation of emetic

PT

and non-emetic Bacillus cereus. Food Control. 32, 176-185. Fernandez-No, I.C., Guarddon, M., Bohme, K., Cepeda, A., Calo-Mata, P.,

SC RI

Barros-Velazquez, J., 2011. Detection and quantification of spoilage and pathogenic Bacillus cereus, Bacillus subtilis and Bacillus licheniformis by real-time PCR. Food Microbiol. 28, 605-610.

NU

Granum, P.E., 2002. Bacillus cereus and food poisoning. Applications and

MA

Systematics of Bacillus and Relatives Blackwell Science Ltd, Oxford, UK. Henrich, T.J., Gallien, S., Li, J.Z., Pereyra, F., Kuritzkes, D.R., 2012. Low-level

ED

detection and quantitation of cellular HIV-1 DNA and 2-LTR circles using droplet digital PCR. J Virol Methods. 186, 68-72.

PT

Hindson, B.J., Ness, K.D., Masquelier, D.A., Belgrader, P., Heredia, N.J., Makarewicz, A.J., Bright, I.J., Lucero, M.Y., Hiddessen, A.L., Legler, T.C., Kitano,

CE

T.K., Hodel, M.R., Petersen, J.F., Wyatt, P.W., Steenblock, E.R., Shah, P.H., Bousse,

AC

L.J., Troup, C.B., Mellen, J.C., Wittmann, D.K., Erndt, N.G., Cauley, T.H., Koehler, R.T., So, A.P., Dube, S., Rose, K.A., Montesclaros, L., Wang, S.L., Stumbo, D.P., Hodges, S.P., Romine, S., Milanovich, F.P., White, H.E., Regan, J.F., KarlinNeumann, G.A., Hindson, C.M., Saxonov, S., Colston, B.W., 2011. High-Throughput Droplet Digital PCR System for Absolute Quantitation of DNA Copy Number. Anal Chem. 83, 8604-8610. Ivy, R.A., Ranieri, M.L., Martin, N.H., den Bakker, H.C., Xavier, B.M., Wiedmann, M., Boor, K.J., 2012. Identification and Characterization of Psychrotolerant Sporeformers Associated with Fluid Milk Production and Processing. Appl Environ Microb. 78, 1853-1864.

16

ACCEPTED MANUSCRIPT Kim, T.G., Jeong, S.Y., Cho, K.S., 2014. Comparison of droplet digital PCR and quantitative real-time PCR for examining population dynamics of bacteria in soil.

PT

Appl Microbiol Biot. 98, 6105-6113. Kiselinova, M., Pasternak, A.O., De Spiegelaere, W., Vogelaers, D., Berkhout,

SC RI

B., Vandekerckhove, L., 2014. Comparison of Droplet Digital PCR and Seminested Real-Time PCR for Quantification of Cell-Associated HIV-1 RNA. Plos One. 9. Klancnik, A., Toplak, N., Kovac, M., Marquis, H., Jersek, B., 2015.

NU

Quantification of Listeria monocytogenes cells with digital PCR and their biofilm

MA

cells with real-time PCR. J Microbiol Meth. 118, 37-41. Koppel, R., Bucher, T., Frei, A., Waiblinger, H.U., 2015. Droplet digital PCR

ED

versus multiplex real-time PCR method for the detection and quantification of DNA from the four transgenic soy traits MON87769, MON87708, MON87705 and FG72,

PT

and lectin. Eur Food Res Technol. 241, 521-527. Martinez-Blanch, J.F., Sanchez, G., Garay, E., Aznar, R., 2009. Development of

CE

a real-time PCR assay for detection and quantification of enterotoxigenic members of

AC

Bacillus cereus group in food samples. Int J Food Microbiol. 135, 15-21. Morisset, D., Stebih, D., Milavec, M., Gruden, K., Zel, J., 2013. Quantitative Analysis of Food and Feed Samples with Droplet Digital PCR. Plos One. 8. Pinheiro, L.B., Coleman, V.A., Hindson, C.M., Herrmann, J., Hindson, B.J., Bhat, S., Emslie, K.R., 2012. Evaluation of a Droplet Digital Polymerase Chain Reaction Format for DNA Copy Number Quantification. Anal Chem. 84, 1003-1011. Porcellato, D., Ostlie, H.M., Liland, K.H., Rudi, K., Isaksson, T., Skeie, S.B., 2012. Strain-level characterization of nonstarter lactic acid bacteria in Norvegia cheese by high-resolution melt analysis. J Dairy Sci. 95, 4804-4812.

17

ACCEPTED MANUSCRIPT Racki, N., Dreo, T., Gutierrez-Aguirre, I., Blejec, A., Ravnikar, M., 2014. Reverse transcriptase droplet digital PCR shows high resilience to PCR inhibitors

PT

from plant, soil and water samples. Plant Methods. 10. Rasolofo, E.A., St-Gelais, D., LaPointe, G., Roy, D., 2010. Molecular analysis

SC RI

of bacterial population structure and dynamics during cold storage of untreated and treated milk. Int J Food Microbiol. 138, 108-118.

Yang, R.C., Paparini, A., Monis, P., Ryan, U., 2014. Comparison of next-

NU

generation droplet digital PCR (ddPCR) with quantitative PCR (qPCR) for

MA

enumeration of Cryptosporidium oocysts in faecal samples. Int J Parasitol. 44, 1105-

AC

CE

PT

ED

1113.

18

ACCEPTED MANUSCRIPT Table 1. Average concentration of Bacillus spp. genome equivalent copies quantified by digital droplet PCR of raw milk, pasteurized milk from the process tank

PT

before packaging, pasteurized milk in cartons analyzed 1 day after packaging and pasteurized milk in cartons kept for 13 days at 4 °C and 8 °C. The table shows the

MA

Raw milk Pasteurized milk before packaging Packaged milk (1 day at 4 °C) Packaged milk (13 days at 4 °C) Packaged milk (13 days at 8 °C)

July Stdeva 1.59 0.58 3.41 0.16 1.25 0.23 0.40 0.69 6.35 0.32

NU

Dairy 1

SC RI

mean and standard deviation (Stdev) of three independent replicates.

Dairy 2

September Mean Stdeva 0.00 0.00 3.02 0.28 0.60 0.49 1.73 0.27 5.99 0.75

August Mean Stdeva 0.86 0.75 0.72 0.63 0.73 0.45 1.18 0.20 4.78 1.56

September Mean Stdeva 1.23 0.16 1.08 0.09 0.65 0.57 0.00 0.00 5.05 0.45

AC

CE

PT

ED

Raw milk Pasteurized milk before packaging Packaged milk (1 day at 4 °C) Packaged milk (13 days at 4 °C) Packaged milk (13 days at 8 °C)

July Mean Stdeva 1.09 0.95 0.70 0.39 0.64 0.30 0.00 0.00 5.98 0.24

August Mean Stdeva 0.73 0.35 0.51 0.28 1.71 1.05 0.00 0.00 4.24 1.67

19

ACCEPTED MANUSCRIPT Figure captions Figure 1. Droplet digital PCR plots obtained during assay optimization. A:

PT

droplet amplitude distribution obtained from 4 different probe concentrations (0.05, 0.1, 0.2, 0.3 μM). B) Droplet amplitude distribution obtained from thermal gradient

SC RI

PCR from 55 to 65 °C. C) Example of droplet distribution obtained from the serial dilution of Bacillus cereus ATCC 14579 genomic DNA. Duplicates were used for all the dilutions and for the negative control. Theoretical concentration of the target

NU

copies is present on the top of each column. Lines in the plots are the cutoff value to

MA

determine if the droplets are positive or negative.

Figure 2. Standard curves obtained from serial dilution of Bacillus cereus

ED

ATCC 14579 genomic DNA plotted against A1) ddPCR copy concentration and B) quantification cycle (Cq) values. Both plots show the fitted linear regression line,

PT

equation and coefficient of determination (R2). A2) enlarged plot obtained from 10 replicates of the 3 last dilutions used to calculate the limit of detection in droplet

CE

digital PCR assay (*: limit of detection).

AC

Figure 3. Plot showing the amount of target copies quantified by ddPCR and qPCR for the milk samples, which were positive for the presence of the Bacillus cereus group.

20

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

MA

NU

SC RI

PT

Figure 1

21

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

MA

NU

SC RI

PT

Figure 2

22

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

MA

NU

SC RI

PT

Figure 3

23

ACCEPTED MANUSCRIPT Highlights 

Droplet digital PCR (ddPCR) assay for the identification of Bacillus in milk is

PT

presented ddPCR assay showed better detection limit compared to real-time PCR



ddPCR is a promising method for the quantification of target bacteria in low

SC RI



concentration

CE

PT

ED

MA

NU

Advantages and limitations of droplet digital PCR are discussed.

AC



24