A Proof of Principle for the Detection of Viable Brucella spp. in Raw Milk by qPCR Targeting Bacteriophages

Brucellosis is still a global health issue, and surveillance and control of this zoonotic disease in livestock remains a challenge. Human outbreaks are mainly linked to the consumption of unpasteurized dairy products. The detection of human pathogenic Brucella species in food of animal origin is time-consuming and laborious. Bacteriophages are broadly applied to the typing of Brucella isolates from pure culture. Since phages intracellularly replicate to very high numbers, they can also be used as specific indicator organisms of their host bacteria. We developed a novel real-time PCR (qPCR) assay targeting the highly conserved helicase sequence harbored in all currently known Brucella-specific lytic phages. Quality and performance tests determined a limit of detection of <1 genomic copy/µL. In raw milk artificially contaminated with Brucella microti, Izv phages were reliably detected after 39 h of incubation, indicating the presence of viable bacteria. The qPCR assay showed high stability in the milk matrix and significantly shortened the time to diagnosis when compared to traditional culture-based techniques. Hence, our molecular assay is a reliable and sensitive method to analyze phage titers, may help to reduce the hands-on time needed for the screening of potentially contaminated food, and reveals infection risks without bacterial isolation.


Introduction
Brucellosis is a zoonotic disease usually transmitted from animals to humans by direct contact with infected livestock or through the consumption of derived products such as raw meat, milk, and cheese [1][2][3][4][5]. Various species of the genus Brucella have been described but human infections are mainly caused by B. melitensis, B. abortus and B. suis [6][7][8]. Brucellae can be isolated from products of animal origin using a selective culture medium such as Farrell's medium. Classical microbiological methods (oxidase, catalase, agglutination with monospecific sera, CO 2 requirement, H 2 S production, urease activity, growth inhibition by thionin and basic fuchsin, phage-typing) and PCR assays are subsequently used for identification and sub-differentiation of species [9,10]. However, isolation from dairy products using culture methods is time-consuming (taking at least three days up to three weeks) and laborious due to the endogenous milk flora which may overgrow the pathogens. Serological tests like the milk ring test suffer from low specificity due to cross-reactivity with various other bacteria [10,11]. PCR assays are faster than culture and are more specific and sensitive than serological tests [12,13]. Due to DNA amplification, the detection limit can be as low as a single copy of the respective target sequence. The quality of PCR results, however, depends on DNA isolation 2 of 11 methods and PCR inhibitors in the matrix, which may reduce sensitivity and can lead to false-negative results [5,13,14]. Furthermore, these methods cannot discriminate between live and dead bacteria which may hamper the evaluation of a positive test result in case of DNA persistence after successful therapy or through contaminations [15,16]. Therefore, a combination of serological and PCR-based methods is highly recommended for epidemiological investigations [13,17,18].
In diagnostic laboratories, bacteriophages are routinely used for subtyping of Brucella species and biovars because they show host specificity despite high genomic similarity within the genus [19][20][21][22][23][24]. However, phage concentrations cannot be easily determined due to inconsistent lysis efficiency of the various Brucella phages and their differing ability to produce clear plaques. In our study, we therefore developed a qPCR assay for the reliable detection and quantification of all known lytic Brucella phages.
In the last decades, phages have been increasingly considered as tools for detection and control of pathogenic bacteria in food production [25][26][27][28][29][30]. The application of phages is of advantage because they specifically prove the presence of viable bacteria needed for their own replication. Hence, we used Brucella phages in combination with the newly established qPCR assay to detect viable brucellae in raw milk.

Phage Propagation
Brucella phages Iz v , Bk2 v and Wb v were propagated on Brucella (B.) microti 4910 T using available phage supernatant stocks as previously described [24]. Brucella microti was chosen to allow for test application under biosafety level 2 conditions. In brief, B. microti was grown in Brucella-broth to McFarland 1-2 in ventilated T250 cell culture flasks (VWR International GmbH, Darmstadt, Germany). Phage stocks were added to a multiplicity of infection (MOI) of 1 to 0.01 and incubated aerobically at 37 • C on a rotational shaker for 24-48 h. Lysates were centrifuged for 30 min at 8,500× g to remove bacterial debris and the supernatant was sterile filtered (0.2 µm). Afterwards, samples were treated with DNaseI/RNaseA (20 µg/mL wt/vol; Roche, Mannheim, Germany) for 30 min at 37 • C. Phage titers were determined by spot tests on B. microti soft-overlay agar (10 µL phage lysate of 10-fold dilutions) and plaque tests (100 µL phage lysate of 10-fold dilutions) as previously described [24]. All assays were run in duplicates of two independent experiments.

Phage DNA Isolation and Quantification
For DNA isolation, Iz v phage was enriched using polyethylene glycol (PEG) precipitation. PEG6000 (Fluka Biochemica, St. Gallen, Switzerland) was added to the phage lysate to a final concentration of 10% and NaCl to a final concentration of 1 M. The suspension was incubated for 3 h at 4 • C and then centrifuged for 20 min at 8,500× g and 4 • C. DNA was isolated from the resulting pellet with the PureLink™ Genomic DNA Mini Kit (Invitrogen, Fisher Scientific GmbH, Schwerte, Germany) according to manufacturer's instructions. DNA concentration was measured using the Invitrogen Qubit™ fluorometer (Fisher Scientific GmbH, Schwerte, Germany). Genomic copy numbers were calculated with an open access online tool (http://cels.uri.edu/gsc/cndna.html, Genomics and Sequencing Center, University of Rhode Island) assuming a template length of 41,446 bp for the Iz v phage (GenBank accession no. KY056619.1). The phage DNA was serially diluted in PureLink™ Elution Buffer (Invitrogen) with final concentrations of 1 × 10 6 to 1 × 10 0 genomic copies/PCR reaction.

Brucella Phage Helicase (BPHeli) qPCR Assay Design and Running Conditions
Primers were designed with Primer3 (http://bioinfo.ut.ee/primer3/) [31,32] using the helicase as target (accession no. KY056619.1, Figure 1) because this gene showed 100% sequence identity among all available genomes of lytic Brucella phages [24,33,34] and was assumed to be present in all other known non-reference phages. Primers were analyzed for standard parameters like melting temperature (Tm) and guaninecytosine (GC)-content as well as probability of dimer creation using the Multiple Primer Analyzer (Thermo Fisher Scientific, Waltham, MA, USA). Furthermore, specificity and coverage were checked against the whole NCBI database using BLAST (https://blast.ncbi.nlm.nih.gov/Blast.cgi). Our primers turned out to be 100% specific, with 100% coverage exclusively for the lytic Brucella phages. No unspecific binding sites in other bacterial or viral genomes were found. The total reaction mixture of 25 μL contained 12.5 μL of QuantiFast SYBR Green Mastermix (Qiagen, Hilden, Germany), 1 μL of each primer (BPHeli(+1226): 5′-ctggacactatggaaggcgg-3′, BPHeli(+1399): 5′-gccgttcattggggtctaca-3′, 10 μM), 2 μL of the sample and 8.5 μL PCR water. The qPCR was carried out with a CFX96 Touch Realtime PCR machine (BioRad, Hercules, CA, USA) choosing the following settings: 5 min of initial denaturation at 95 °C, 40 cycles of 10 s at 95 °C and 30 s at 60 °C (fluorescent measurement at the end of each cycle), with a final melting curve analysis.

Performance Indicators of the BPHeli qPCR Assay
Analytical sensitivity of the BPHeli qPCR assay was determined using Izv phage DNA purified and serially diluted as described above. Quantification cycle (Cq) values were plotted against the respective number of genome equivalents, and standard deviations and log-linear regression were calculated using IBM SPSS Statistics v. 21. In addition, phage lysates were analyzed to test the applicability of the assay to quantify phage concentrations in propagation experiments. Standards were included in each run. Specificity was tested both in silico using BLAST against the NCBI database and in vitro using lysates of the 19 Brucella phages available, namely Izv, Wbv, Bk2v, F25, Tbv, R/Cv, F25u, F1m, P, 371/XXIX, F1u, 12m, 3, MP51, 6, 45/III, L19, Fiv, and F1 [24]. Genomic DNA of B. microti 4910 T , B. abortus 544, and B. melitensis 16M were used as negative controls to exclude cross-reactivity and inhibitory effects in the analysis of phage lysates.
Performance of the BPHeli assay was also tested in raw cow's milk to exclude inhibitory effects of the food matrix. To this end, we prepared log10 dilutions of Izv phage DNA in raw cow's milk (1:10, 1:5, and 1:2 in DNA elution buffer). Furthermore, we assessed the ability of the BPHeli qPCR to directly detect Izv, Bk2v, and Wbv phage lysates in milk by comparing the molecularly detected concentrations in PFU/mL with the concentrations determined by spot and plaque tests.  Primers were analyzed for standard parameters like melting temperature (Tm) and guanine-cytosine (GC)-content as well as probability of dimer creation using the Multiple Primer Analyzer (Thermo Fisher Scientific, Waltham, MA, USA). Furthermore, specificity and coverage were checked against the whole NCBI database using BLAST (https://blast.ncbi.nlm.nih.gov/Blast.cgi). Our primers turned out to be 100% specific, with 100% coverage exclusively for the lytic Brucella phages. No unspecific binding sites in other bacterial or viral genomes were found. The total reaction mixture of 25 µL contained 12.5 µL of QuantiFast SYBR Green Mastermix (Qiagen, Hilden, Germany), 1 µL of each primer (BPHeli(+1226): 5 -ctggacactatggaaggcgg-3 , BPHeli(+1399): 5 -gccgttcattggggtctaca-3 , 10 µM), 2 µL of the sample and 8.5 µL PCR water. The qPCR was carried out with a CFX96 Touch Real-time PCR machine (BioRad, Hercules, CA, USA) choosing the following settings: 5 min of initial denaturation at 95 • C, 40 cycles of 10 s at 95 • C and 30 s at 60 • C (fluorescent measurement at the end of each cycle), with a final melting curve analysis.

Performance Indicators of the BPHeli qPCR Assay
Analytical sensitivity of the BPHeli qPCR assay was determined using Iz v phage DNA purified and serially diluted as described above. Quantification cycle (Cq) values were plotted against the respective number of genome equivalents, and standard deviations and log-linear regression were calculated using IBM SPSS Statistics v. 21. In addition, phage lysates were analyzed to test the applicability of the assay to quantify phage concentrations in propagation experiments. Standards were included in each run. Specificity was tested both in silico using BLAST against the NCBI database and in vitro using lysates of the 19 Brucella phages available, namely Iz v , Wb v , Bk2 v , F25, Tb v , R/C v , F25u, F1m, P, 371/XXIX, F1u, 12m, 3, MP51, 6, 45/III, L19, Fi v , and F1 [24]. Genomic DNA of B. microti 4910 T , B. abortus 544, and B. melitensis 16M were used as negative controls to exclude cross-reactivity and inhibitory effects in the analysis of phage lysates.
Performance of the BPHeli assay was also tested in raw cow's milk to exclude inhibitory effects of the food matrix. To this end, we prepared log10 dilutions of Iz v phage DNA in raw cow's milk (1:10, 1:5, and 1:2 in DNA elution buffer). Furthermore, we assessed the ability of the BPHeli qPCR to directly detect Iz v , Bk2 v , and Wb v phage lysates in milk by comparing the molecularly detected concentrations in PFU/mL with the concentrations determined by spot and plaque tests.

Stability of the BPHeli qPCR Assay in Cow's Milk
Iz v phage lysates were diluted in 50 mL Brucella-broth (BrucB), selective Brucella-broth (BrucSB), raw cow's milk 1:10 in Brucella-broth (MBrucB) and raw cow's milk 1:10 in selective Brucella-broth (MBrucSB), to a final concentration of 10 5 PFU/mL. Higher concentrations of the Iz v phage were used to facilitate the monitoring of changes during the experiments. The mixtures were incubated aerobically in ventilated T250 cell culture flasks at 37 • C for 39 h and subsamples for qPCR were taken after 0, 13, 15.5, 17.5, 22, and 39 h.

Phage Propagation and Growth of B. Microti in Raw Cow's Milk
Iz v phage lysates were diluted in each of the four broth mixtures as described above. Brucella microti 4910 T was added to final concentrations of 10 1 and 10 2 cfu/mL. A B. microti control was prepared for all mixtures (10 2 cfu/mL). The flasks were incubated at 37 • C for 39 h and subsamples for qPCR were taken after 0, 13, 15.5, 17.5, 22, and 39 h. Changes in phage concentrations were measured using our newly developed BPHeli qPCR assay, while B. microti concentrations were determined by the genus-specific IS711 qPCR, as published by Matero and colleagues with slight modifications [35,36]. Briefly, we used the QuantiFast Pathogen Kit (Qiagen, Hilden, Germany) according to manufacturer's instructions. The reaction mixture contained 0.5 µLprobe, 1 µL of each primer using 10 µM working solutions and 5 µl DNA template in a final volume of 25 µL. The PCR was carried out in a CFX96 Touch Real-Time PCR Detection System (BioRad) at 95 • C for 5 min, 45 cycles of 15 s at 95 • C and 30 s at 60 • C, and a final step at 60 • C for 1 min.

Results
Our study pursued two major objectives: (i) to develop a fast and reliable assay for the detection of Brucella phages and for monitoring phage titers in propagation experiments, and (ii) to apply phage detection as an indirect evidence of the presence of viable bacterial pathogens in food using Brucella in raw cow's milk as a model.

Phage Propagation and DNA Isolation
We selected three Brucella phages with a broad host range for our experiments. After propagation of Iz v , Bk2 v and Wb v on B. microti reference strain 4910 T , titers of 4 × 10 7 , 2.8 × 10 6 and 6 × 10 6 PFU/mL were determined in the spot assay, and titers of 2.1 × 10 7 , 7.1 × 10 5 , and 1.8 × 10 6 PFU/mL in the plaque assay, respectively. Brucella microti was used to facilitate lab work under biosafety level 2. Since Brucella phages are highly conserved, we only isolated DNA from the purified Iz v phage (resulting in 30 ng/µL).

Brucella Phage Helicase (BPHeli) qPCR Assay Performance
Serial 1:10 DNA dilutions from 10 6 to 10 0 Iz v phage genome copies were used to assess the analytical sensitivity of our newly developed BPHeli qPCR assay. Data were obtained from 17 independent runs. Cq (cycle of quantification)-values were plotted against the respective numbers of genome equivalents, and standard deviations and log-linear regression were calculated as performance indicators using IBM SPSS Statistics v. 21 (IBM Corp., Armonk, NY, USA) ( Figure 2, Table 1). The limit of detection was determined to be lower than 1 genomic copy/µL. There was a log-linear correlation between Cq values and the number of phage genome equivalents. The standard deviations of the mean Cq values ranged between 0.17 and 0.19 with slightly higher values at lower bacteriophage concentrations (<10 2 Iz v phage genome equivalents). The detection of 10 0 genome equivalents failed in three runs. Positive fluorescence signals were routinely checked by melting curve analysis.
For specificity testing, lysates of 19 different Brucella phages as well as Brucella spp. genomic DNA were tested. Using the BPHeli qPCR all Brucella phages could be detected. Although genomic sequences were only available for a few Brucella phages, our findings showed that the chosen target region is highly conserved among the phages. Calculation of the phage DNA concentration in the lysates of Iz v , Bk2 v and Wb v revealed 6 × 10 9 , 1.1 × 10 10 , and 1.2 × 10 8 PFU/mL, respectively. The control samples were tested negative in the qPCR assay. Unspecific amplification signals only occurred in phage-free samples and could be easily identified by melting curve analysis (expected melting point at 85.5-86 • C, false-positive samples at 76.5 • C; Figure 3). Primer-dimer formation (self-dimers, cross-dimers) was excluded in silico as well as by melting curve analysis of negative control samples.  The limit of detection was determined to be lower than 1 genomic copy/μL. There was a loglinear correlation between Cq values and the number of phage genome equivalents. The standard deviations of the mean Cq values ranged between 0.17 and 0.19 with slightly higher values at lower bacteriophage concentrations (<10 2 Izv phage genome equivalents). The detection of 10 0 genome equivalents failed in three runs. Positive fluorescence signals were routinely checked by melting curve analysis.
For specificity testing, lysates of 19 different Brucella phages as well as Brucella spp. genomic DNA were tested. Using the BPHeli qPCR all Brucella phages could be detected. Although genomic sequences were only available for a few Brucella phages, our findings showed that the chosen target region is highly conserved among the phages. Calculation of the phage DNA concentration in the lysates of Izv, Bk2v and Wbv revealed 6 × 10 9 , 1.1 × 10 10 , and 1.2 × 10 8 PFU/mL, respectively. The control samples were tested negative in the qPCR assay. Unspecific amplification signals only occurred in phage-free samples and could be easily identified by melting curve analysis (expected melting point at 85.5-86 °C, false-positive samples at 76.5 °C; Figure 3). Primer-dimer formation (self-dimers, crossdimers) was excluded in silico as well as by melting curve analysis of negative control samples.  To assess a possible matrix effect on the performance of the BPHeli qPCR assay, Izv phage DNA was diluted in raw cow's milk. Phage DNA was detected in all milk samples independent of the concentrations introduced. However, there was a fourfold increase of Cq-values as well as a reduction in the relative fluorescence units (RFU) of the amplification curves with an increasing proportion of milk in the test sample.

BPHeli qPCR Assay Stability in Cow's Milk
The Izv phage lysate was added to four different broth mixtures to a final concentration of 10 5 PFU/mL based on previous qPCR results. Results from the BPHeli qPCR assay indicated reduced amounts of phage target genomes in the broth (up to 3 log10 values) which was most likely a matrix effect because samples from the mixtures were directly used for PCR without preceding DNA extraction (Figure 4). To assess a possible matrix effect on the performance of the BPHeli qPCR assay, Iz v phage DNA was diluted in raw cow's milk. Phage DNA was detected in all milk samples independent of the concentrations introduced. However, there was a fourfold increase of Cq-values as well as a reduction in the relative fluorescence units (RFU) of the amplification curves with an increasing proportion of milk in the test sample.

BPHeli qPCR Assay Stability in Cow's Milk
The Iz v phage lysate was added to four different broth mixtures to a final concentration of 10 5 PFU/mL based on previous qPCR results. Results from the BPHeli qPCR assay indicated reduced amounts of phage target genomes in the broth (up to 3 log10 values) which was most likely a matrix effect because samples from the mixtures were directly used for PCR without preceding DNA extraction (Figure 4).

BPHeli qPCR Assay Stability in Cow's Milk
The Izv phage lysate was added to four different broth mixtures to a final concentration of 10 5 PFU/mL based on previous qPCR results. Results from the BPHeli qPCR assay indicated reduced amounts of phage target genomes in the broth (up to 3 log10 values) which was most likely a matrix effect because samples from the mixtures were directly used for PCR without preceding DNA extraction (Figure 4).  The matrix effect could not be exclusively explained by the added raw milk because there was also a decrease in the phage concentration (approx. 1 log10) when BrucB (Brucella broth) and BrucSB (selective Brucella broth) were compared. The measured Iz v phage concentration remained stable over 39 h except for MBrucB (raw cow's milk in Brucella-broth) (Figure 4). After 13 h, we observed a decline of the phage concentration in MBrucB of about 1 log10 value.

Growth of B. microti and Iz v Phage Propagation in Raw Cow's Milk
Multiplication of B. microti 4910 T and Iz v phage in raw cow's milk was determined using qPCR assays for the detection of Brucella and phage DNA, respectively. The experiments were carried out in two different Brucella media with and without adding raw cow's milk (1:10 dilution in the respective medium). Iz v was used because it has the widest lysis spectrum among Brucella spp. and, therefore, the greatest application potential. assays for the detection of Brucella and phage DNA, respectively. The experiments were carried out in two different Brucella media with and without adding raw cow's milk (1:10 dilution in the respective medium). Izv was used because it has the widest lysis spectrum among Brucella spp. and, therefore, the greatest application potential.
Brucella microti growth and propagation of Izv phage were measured at six points in time over a period of 39 h ( Figure 5).

Discussion
We developed a real-time PCR assay for the detection of 19 different Brucella phages. A previously published PCR targeting the DNA polymerase gene of phage S708 was only tested for four out of them, namely Tb, Bk, Fz, and Wb [37]. Although our newly established BPHeli qPCR performed well, the phage concentrations (PFU/mL) determined by spot test, plaque test and the molecular assay varied, also depending on the phages under study. This observation could be partially explained by the amplification of phage DNA originating from non-functional or damaged phage particles using the qPCR assay [38]. Furthermore, some of the Brucella phages did not produce clear plaques, which makes the assessment of exact phage titers of a lysate technically demanding. Phage particles may also stick together resulting in different sizes of spots and plaques leading to higher variations in PFU counts [38,39]. In contrast, our qPCR assay revealed less variability than the biological tests and provided results within two hours, whereas spot and plaque tests took up to two days of incubation.
To evaluate our newly developed phage-based qPCR assay for the detection of bacteria in complex food matrices, we initially conducted stability and performance tests and then analyzed raw cow's milk contaminated with Brucella. Our study showed that qPCR results, and thus indirectly phages, remained stable over 39 h of incubation, which makes Brucella phages a useful tool for the indirect detection of viable brucellae. The BPHeli qPCR assay was able to detect Brucella phages in raw milk samples without any preceding DNA isolation. By omitting the DNA pre-processing step, we could significantly reduce hands-on-time and speed up time-to-result. However, assay sensitivity was lower in raw cow's milk than in culture medium, which might have been a result of PCR inhibitors in the food matrix, but DNA isolation does not necessarily improve the analytical sensitivity of qPCR assays [14].
Indirect detection of bacteria by phages is based on the addition of phages or phage cocktails to a target sample and subsequent analysis of phage propagation [40,41]. In our study, we inoculated raw cow's milk with B. microti, added Brucella phages and used a qPCR assay to evaluate the moment of the phage burst in the sample. Thereby we tested whether Brucella phages can serve as a proof of viable Brucella cells in raw milk. Enrichment culture and the isolation of Brucella spp. from raw milk products is laborious and time consuming taking up to several weeks. Because of the natural milk microbiome selective media are needed and periodic sub-culturing has to be performed [9]. Usually serological and molecular methods are combined to prove the presence of Brucella in raw milk products [17]. The application of phages can accelerate diagnosis compared to traditional culture methods used for bacterial isolation from raw milk products. As phages only propagate in viable bacterial cells, their detection also provides more meaningful results on the contamination of dairy products with Brucella spp. Using IS711 and BPHeli qPCR assays simultaneously allowed for tracking the growth of B. microti and the propagation of the Iz v phage in raw cow's milk, respectively. As anticipated, B. microti was able to grow slowly in a milk-broth mixture including selective medium (MBrucSB) whereas bacteria did not grow using non-selective medium (MBrucB). Under the latter conditions, the natural milk microbiome inhibited or prevented successful bacterial growth. After 22 h of incubation, the number of B. microti cells in MBrucSB measured by IS711 qPCR has risen by 1 to 2 log10 values. The detection of IS711 is very sensitive since this genetic element occurs in multiple copies in the Brucella genome. However, the number of IS711 elements in B. microti 4910 T is higher (n = 13, Genbank CP001578.1 + CP001579.1) than in B. melitensis 16M (n = 7, Genbank CP007763.1 + CP007762.1) or B. abortus S19 (n = 4, Genbank CP000887.1 CP000888.1). This might lead to an earlier detection of the growth of B. microti compared to other, Brucella spp. Hence, the reliability of the IS711 qPCR assay for an early detection of Brucella spp. multiplication in raw milk must be questioned. In contrast, at 39 h of incubation the reproduction of Iz v phages was more clearly visible with concentrations changing from 10 2 to 10 8 PFU/mL. In a previous study, this phage burst appeared after 48-72 h depending on the B. abortus concentration in the sample [37]. The observed discrepancy might be explained by diverging sampling schedules (leading to a later detection) or by the use of different Brucella spp. in the experiments. Classical Brucella spp. are slow-growing bacteria while B. microti is known to grow fast, within 1-2 days on standard media [42,43].
Sergueev and colleagues used the S708 phage for their diagnostic experiments in clinical samples whereas we decided to use the Iz v phage because its host range comprises all known Brucella spp. relevant for human infections [24]. S708 is lytic for B. suis, B. abortus, and B. neotomae [34], but lacks activity against B. melitensis. However, B. melitensis is the most common cause of human infections worldwide and is highly related to outbreaks due to the consumption of contaminated dairy products from goats and sheep [44,45]. The detection of Brucella in raw milk products using phages should, therefore, comprise all major zoonotic Brucella spp.
In our experiments, we further observed that even after the phage burst bacterial cells were still viable and culturable implying that Brucella might not be fully eradicated from a food sample using phages or gained phage resistance [46]. To overcome resistance, phage cocktails can be applied [47]. Moreover, phage propagation might depend on a certain number of viable bacterial cells as the phage burst only occurred after 17.5 h and 22 h of incubation in BrucB and BrucSB, respectively. As already shown, the starting concentration of Brucella phages also influences the timing of the phage burst [37]. Brucella phages display a varying host range, even though they are genetically closely related [24]. Except for the genomic information, there is still a huge lack of knowledge on gene translation and regulation, protein structures, phage-host interactions, and the general propagation cycle of these phages. For the Escherichia coli T7 phage more comprehensive information is already available, making it a useful tool for pathogen detection and phage display experiments [48,49]. Hence, there is an urgent need for further molecular and biological investigations to better understand Brucella phages and to evaluate their application in the context of food safety.

Conclusions
We developed a highly sensitive and robust qPCR assay for the detection of Brucella phages both in lysates and in raw milk and products thereof. In this way, our assay indirectly proves the presence of viable Brucella spp. after adding Brucella phages to a contaminated food sample, which is a much faster approach than traditional culture methods. Rapid molecular diagnostics, which are capable of detecting live bacteria, allow for better risk assessment and may significantly improve food safety standards in the future.

Conflicts of Interest:
The authors declare no conflict of interest.