Detection of Salmonella spp. Using a Generic and Differential FRET-PCR

To facilitate the detection of Salmonella and to be able to rapidly and conveniently determine the species/subspecies present, we developed and tested a generic and differential FRET-PCR targeting their tetrathionate reductase response regulator gene. The differential pan-Salmonella FRET-PCR we developed successfully detected seven plasmids that contained partial sequences of S. bongori and the six S. enterica subspecies. The detection limit varied from ∼5 copies of target gene/per PCR reaction for S. enterica enterica to ∼200 for S. bongori. Melting curve analysis demonstrated a T m of ∼68°C for S. enterica enterica, ∼62.5°C for S. enterica houtenae and S. enterica diarizonae, ∼57°C for S. enterica indica, and ∼54°C for S. bongori, S. enterica salamae and S. enterica arizonae. The differential pan-Salmonella FRET-PCR also detected and determined the subspecies of 4 reference strains and 47 Salmonella isolated from clinically ill birds or pigs. Finally, we found it could directly detect and differentiate Salmonella in feline (5/50 positive; 10%; one S. enterica salamae and 4 S. enterica enterica) and canine feces (15/114 positive; 13.2%; all S. enterica enterica). The differential pan-Salmonella FRET-PCR failed to react with 96 non-Salmonella bacterial strains. Our experiments show the differential pan-Salmonella FRET-PCR we developed is a rapid, sensitive and specific method to detect and differentiate Salmonella.


Introduction
Salmonella are important zoonotic pathogens which affect the health of people and animals worldwide. Advances in methods to detect Salmonella and a better understanding of their genomics have resulted in an improved classification system for the organisms. Currently, the genus Salmonella consists of only 2 species, S. bongori and S. enterica, with the latter containing 6 subspecies: enterica (I), salamae (II), arizonae (IIIa), diarizonae (IIIb), houtenae (IV), and indica (VI) [1,2]. Using the White-Kauffmann-Le Minor Scheme based on somatic, flagellar and capsular antigens, around 2600 serovars are recognized of which over 99.5% belong to the S. enterica species and more than half are S. enterica subspecies enterica [1]. The S. enterica enterica serotypes are primarily associated with warmblooded animals while S. bongori and the remaining five S. enterica subspecies are mainly found in cold-blooded animals and the environment [3]. While most human and animal infections are associated with S. enterica enterica, the growing popularity of reptiles as pets has led to an increasing number of infections with S. bongori and S. enterica subspecies other than S. enterica enterica [2,4,5,6]. A retrospective analysis of over 75,000 isolates collected between  showed that S. enterica subspecies salamae, arizonae, diarizonae and houtenae caused invasive disease in people significantly more frequently than did S. enterica enterica [2].
The diagnosis of salmonellosis using conventional culture and serotyping is time consuming, laborious and requires large numbers of reagents. Effective nucleic acid-based methods could greatly simplify the procedure and a number of DNA-based techniques have been described to target specific Salmonella serovars [7][8][9][10][11][12]. Most commonly, they specifically identify S. enterica or only S. enterica subsp. enterica [13][14][15][16][17][18][19]. Although a multiplex PCR assay has been described to differentiate species and subspecies of Salmonella [20], the procedure requires organisms isolated in culture, is time consuming, and may be associated with decreased sensitivity and specificity due to complex interactions among five sets of primers. The FRET-PCR assay largely eliminates these problems and we therefore attempted to use this technique to design and test a differential pan-Salmonella FRET-PCR to sensitively and specifically detect and differentiate Salmonella species and subspecies. The ability to respire tetrathionate is a characteristic of Salmonella and tetrathionate broth is used as a standard enrichment medium for Salmonella species. The tetrathionate reductase locus (ttr), including the tetrathionate reductase response regulator gene (ttrR), is located within the Salmonella pathogenicity island 2 and is highly conserved with minimal genetic polymorphism among Salmonella species/subspecies. The inclusivity and exclusivity of the PCR technology targeting ttrR gene was found to be 100% while applying on 110 Salmonella strains and 87 non-Salmonella strains [16,18]. This makes the ttrR an ideal target for molecular probes. The results of our experiments using the ttrR to develop a FRET-PCR to differentiate Salmonella are described in this report.

Fecal samples
The study was reviewed and approved by the Institutional Animal Care and Use Committee of the Ross University School of Veterinary Medicine (RUSVM). Between June and November of 2011, convenience free-catch fecal samples were obtained from dogs (n = 114) owned by consenting RUSVM veterinary students, and from stray cats (n = 50) involved in the RUSVM Small Animal Spay/Neuter Outreach Program. The animals from which feces were collected were clinically healthy and had no signs of diarrhea. Within 2 hours of collection into clean plastic fecal pots, around 1 g of the feces was transferred with a sterile spatula into a 2.0 ml sterile Eppendorf tube and frozen at 280uC until thawed at room temperature for DNA extraction.

Artificially inoculated dog feces
Two hundred ml of the ATCC strain S. Typhimurium 14028 with a concentration of 10 6 CFU/ml was thoroughly mixed for 1 minute with 1 gram of Salmonella culture-negative canine feces suspended in 0.8 ml of sterile PBS. An aliquot of 100 ml was streaked onto MacConkey agar (Remel, Lenexa, Kansas, 66215, USA) and cultured for 18-hour culture at 37uC. As described below, the DNA was extracted from a further aliquot of 200 ml and tested with the differential pan-Salmonella FRET-PCR.

Extraction of DNA from cat and dog feces
The QIAamp DNA Stool Mini Kit (Qiagen Inc., Valencia, CA, USA) was used to extract fecal DNA from cats and dogs following the manufacturer's instructions. Briefly, the fecal sample (100 mg) was lysed for 2 h at 55uC with buffer ATL and proteinase K before the reaction was stopped with buffer AL at 70uC for 10 min. After centrifugation, anhydrous ethanol was added to the supernatant and the mixture passed through the QIAamp kit column. Following two washes with buffers AW1 and AW2, DNA was eluted in 50 mL of elution buffer and stored at 280uC until thawed at room temperature for testing.
Thermal cycling. Differential pan-Salmonella FRET-PCR was performed in a LightCyclerH 2.0 real-time PCR platform using a thermal protocol and PCR conditions as described [22]. Each reaction was performed in a 20 mL final volume containing 10 mL of extracted DNA. Thermal cycling consisted of 18 highstringency step-down cycles followed by 30 relaxed-stringency fluorescence acquisition cycles. The 18 high-stringency step-down thermal cycles were 660 sec @ 95uC, 12 sec @ 64uC, 8 sec @ 72uC; 960 sec @ 95uC, 12 sec @ 62uC, 8 sec @ 72uC; 360 sec @ 95uC, 12 sec @ 60uC, 8 sec @ 72uC. The relaxed-stringency fluorescence acquisition cycling consisted of 3060 sec @ 95uC, followed by fluorescence acquisition of 12 sec @ 56uC, and 10 sec @ 72uC. Following the completion of differential pan-Salmonella FRET-PCR, the melting curve analysis for probes annealing to the PCR products was determined by monitoring the fluorescence from 45uC to 80uC, and the first derivatives of F4/F1 were evaluated to determine the probe melting temperature (T m ). This setting is only possible on LightCycler 1.5 and 2.0 models which can accommodate very rapid temperature ramps because of their capillary tube reaction system. PCR products were verified using electrophoresis through 4% MetaPhor agarose gels and purified for automated DNA sequencing with a QIAquick PCR Purification Kit according to the manufacturer's instructions (Qiagen, Valencia, CA, USA). Both strands of DNA of the PCR products were sequenced at the Genomic Sequencing Laboratory (Davis Sequencing, Davis, USA) using the forward and downstream primers (Figure 1).

Development of the differential pan-Salmonella FRET-PCR
Our differential pan-Salmonella FRET-PCR detected each of the 7 plasmids containing partial ttrR sequences of S. bongori, the 6 S. enterica subspecies ( Table 2). It also detected all 47 Salmonella isolates made from clinically ill birds and pigs in China and the four reference strains; the T m for these isolates varied from 66.5uC to 68.5uC (Table 1). No reaction products were obtained when the differential pan-Salmonella FRET-PCR was performed on 96 strains of non-Salmonella bacteria.
Using the Salmonella ttrR-containing plasmids as quantitative standards, we determined that the detection limit of the FRET-PCR varied from ,5 copies of the S. enterica enterica ttrR gene per PCR reaction to ,200 copies per PCR reaction for S. bongori ( Table 2). Melting curve analysis enabled a high level of differentiation of Salmonella with strains clustering in four groups based on their melting points: i) S. enterica subsp. enterica had a T m of ,68uC; ii) ,62.5uC for S. enterica subsp. houtenae and S. enterica subsp. diarizonae; iii) ,57uC for S. enterica subsp. indica; iv) ,54uC for S. bongori, S. enterica subsp. salamae, and S. enterica subsp. arizonae (Figure 2). High-resolution melting analysis on samples containing mixtures of plasmids revealed multiple distinct melting peaks that enabled us to differentiate the Salmonella species/subspecies involved (data not shown).

Artificially inoculated dog feces
Following incubation of the fecal sample typical black Salmonella colonies were observed on the inoculated MacConkey agar and confirmed to be S. Typhimurium by standard biochemical testing. Differential pan-Salmonella FRET-PCR of the fecal sample was positive with a T m of 68uC which was consistent with S. Typhimurium. Direct detection of fecal Salmonella DNA by the differential pan-Salmonella FRET-PCR Of the 164 fecal samples studied, 10% of the feline specimens (5/50) and 13.2% of the canine specimens (15/114) were positive for Salmonella by the differential pan-Salmonella FRET-PCR. Of the positive specimens, 95% (19/20) were S. enterica enterica based on their T m of ,68uC and this was confirmed by gel electrophoresis and sequencing ( Figure-S1; Figure-S2). One fecal DNA from a cat had a T m of 54uC which was consistent with S. enterica salamae and this was also confirmed by sequencing of the reaction products.

Discussion
By systematically aligning the ttrR sequences of representative Salmonella species, subspecies and serovars, we were able to identify a region suitable for targeting in a differential pan-Salmonella FRET-PCR. The primers we designed proved to be highly conserved and enabled us to amplify all the Salmonella species and subspecies we tested, including four reference strains and 47 isolates made from clinically ill birds and pigs. We were unable, however, to amplify non-Salmonella bacteria which shows our test has a high specificity. The probes we designed were also conserved but had a low level of polymorphism with the different Salmonella species and subspecies which resulted in different melting points for the organisms. Based on their melting points we were able to rapidly (under 4 hours from DNA extraction to melting curve analysis) and clearly detect and differentiate S. enterica enterica which is the most important subspecies, and in addition the less common and important pathogen S. enterica indica. Although our test did not enable us to differentiate between S. enterica subsp. houtenae and S. enterica subsp. diarizonae, and also between S. bongori, S. enterica subsp. salamae, and S. enterica subsp. arizonae, these organisms are only isolated comparatively rarely in clinical samples [3]. In some areas, however, they are being recognized more frequently [2] and once the possible subspecies have been found in our FRET-PCR, differentiation between them would require sequencing of the reaction products. If biochemical tests are to be used to make the identification, knowledge of the potential subspecies gained from the FRET-PCR would enable more specific testing to be performed.
While a multiplex PCR has been described which identifies all the Salmonella species and subspecies, the test requires cultures of organisms and uses 5 primers [20], the products of which need to be resolved on gels by electrophoresis. Our differential pan-Salmonella FRET-PCR requires only a single set of primers and one set of probes, and the products are revealed as the differential pan-Salmonella FRET-PCR assay is performed. This greatly simplifies the procedure and assures the sensitivity and specificity of Salmonella detection, and the rapidity of the test means it will be very useful in situations where screening of large numbers of  samples is required in a timely and convenient fashion. Compared to the classical gel-based PCR, our differential pan-Salmonella FRET-PCR might be more expensive but has the advantages of reducing labor costs, eliminating the need for electrophoresis equipment, gels and expensive biohazard disposal, and, most importantly, in minimizing the chances of carry-over contamination.
Numerous PCR assays have been described for the detection of Salmonella DNA in a variety of samples, food in particular [7][8][9][10][13][14][15][16][17][18][19][20]. As with routine cultures, however, these PCR methods described to date for Salmonella detection usually require a preenrichment step which delays testing by 6 to 24 hours [19,23], and make the tests unsuitable for rapid screening of large numbers of samples. Our differential pan-Salmonella FRET-PCR was effective in directly detecting Salmonella DNA in experimentally inoculated dog feces when we tested a sample containing approximately 10 5 organisms, at least some of which were viable and could be cultured. This experiment indicates the differential pan-Salmonella FRET-PCR is sensitive and capable of differentiating its target in a complex matrix containing a large variety of microbial flora and host genome. We were also able to use the differential pan-Salmonella FRET-PCR to detect Salmonella in dog and cat feces from apparently healthy animals. Although we did not culture the feces from the dogs and cats and were therefore unable to precisely determine the sensitivity and specificity of the test when performed directly on fecal samples, our findings are comparable with other studies using PCR which found Salmonella in 6% and 2% of dogs [24] and cats [25] with normal feces, respectively. Further studies are underway in our laboratory to determine the sensitivity of the PCR in detecting DNA of Salmonella directly in tissue, fecal and environmental samples.
In conclusion, our study established a novel differential pan-Salmonella FRET-PCR with high sensitivity and specificity to quantitatively detect all Salmonella species and subspecies. Melting curve analysis following amplification and real-time detection conclusively and conveniently differentiated Salmonella species and subspecies and enabled us to diagnose, for the first time, the presence of S. enterica salamae in the feces of a clinically healthy cat. Figure S1 Representative melting curves of the pan-Salmonella FRET-PCR of canine fecal samples. DNA extracted from canine fecal samples was used for the pan-Salmonella FRET-PCR described in this study. The positive samples showed an identical T m of 68uC while the melting curves of negative samples remained flat. (TIF) Figure S2 Gel electrophoresis (4.0% agarose) analysis of the pan-Salmonella FRET-PCR's amplified products. Lane 1: Trans2K Plus DNA Marker (Beijing Transgen Biotech Co., Ltd.); Lanes 2-3: positive fecal samples from cats; Lane 4: positive fecal sample from dog; Lanes 5-11: 7 plasmids containing part of ttrR gene in the following order: S. enterica enterica, S. enterica houtenae, S. enterica diarizonae, S. enterica indica, S. enterica salamae, S. enterica Arizonae and S. bongori; Lanes 12-13: negative fecal samples from cat and dog; Lanes 14-17: DNAs extracted from Escherichia coli, Campylobacter spp., Pseudomonas aeruginosa, and Enterococcus faecalis, respectively. (TIF)