INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The final identification of Salmonella enterica is based on biochemical tests followed by serotyping (22). However, a minor proportion of presumptive S. enterica isolates, defined as rough isolates (12), may lack the O-antigens or may lack both O- and H-antigens. In addition, reference laboratories occasionally receive strains suspected to be Salmonella from other laboratories for verification which do not result in unambiguous biochemical reactions or cannot be identified by subsequent serotyping procedures. At the Danish Veterinary Laboratory, nearly 10% of strains obtained for serotyping require verification to the species level. Most of these strains are obtained from a production environment, such as slaughterhouses, feed mills, food production units, and herds of livestock.

DNA testing methods, such as PCR, can circumvent the phenotypic variations seen in both biochemical patterns and the lack of detectable antigens (20). Several experimental Salmonella-specific PCR assays have been published (reviewed in reference 21). However, it can be difficult to implement PCR tests that can provide reproducible results within and among diagnostic laboratories, due to the well-known risk of contamination (carryover problem), the presence of PCR inhibitors, or variations in the performances of different thermal cyclers (17, 27).

The 5' nuclease (TaqMan) PCR assay is a closed-tube, fluorescence-based, online and end-point detection technique which can improve the reproducibility of results of conventional PCR testing (10). The method exploits the ability of DNA polymerase to cleave nucleotides from a double-fluorescence-labeled (by reporter and quencher dyes), specific oligonucleotide probe which is annealed to a target DNA strand (8). The emission of the reporter dye is normally quenched by virtue of its proximity to the quencher dye. However, when the reporter dye is cleaved from the quencher by the activity of the DNA polymerase, a charge-coupled-device camera detects its fluorescence.

The technology has been found valuable for detection of pathogenic bacteria, such as Listeria monocytogenes, Escherichia coli O157:H7, and Yersinia pestis (2, 9, 23). However, the only full report that is available on the application of 5' nuclease TaqMan PCR for the detection of Salmonella is based on the evaluation of a proprietary kit based on a concealed target sequence (13). Thus, the purpose of the present study was to develop an automated, simple, and ready-made PCR technique, based on the TaqMan technology, as a further tool in identification of problematic Salmonella isolates.

(This work was presented in part at the 99th General Meeting of the American Society for Microbiology, Chicago, Ill., May 30 to June 3, 1999.)

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains. In a "blind" experiment, a total of 110 Salmonella strains (Table 1), 120 non-Salmonella strains (Table 2), and 100 rough and problematic Salmonella-suspect strains (Table 3), grown on blood agar at 37°C overnight, were examined. The cultures were diagnostic isolates obtained from the collection at the Department of Microbiology, Danish Veterinary Laboratory. The collection is maintained in Luria Bertani broth containing 15% (vol/vol) glycerol and stored at 70°C.

View this table: [in this window] [in a new window]   TABLE 1.   The end-point, relative fluorescence (Rn), results of PCR tests for 110 Salmonella enterica strains from various serotypesa

View this table: [in this window] [in a new window]   TABLE 2.   The end-point, relative fluorescence (Rn), results of PCR test for 120 non-Salmonella isolatesa

View this table: [in this window] [in a new window]   TABLE 3.   The end-point, relative fluorescence (Rn), results of PCR tests for 100 rough and problematic Salmonella isolatesa

Oligonucleotides. The PCR oligonucleotide primers (Styinva-JHO-2-left and -right) and probe (Salmonella probe) were designed according to the published DNA sequence of the invasion (invA) gene (GenBank accession no. M90846 [6]) in order to amplify a chromosomal DNA sequence of 119 bp (Table 4). The primers were purchased from DNA Technology Ltd. (Århus, Denmark), which used conventional phosphoramidite chemistry for synthesis and reverse-phase fast-cartridge chemistry for purification (ScanPrimers). The Salmonella probe was labeled with 6-carboxyfluorescein (FAM) (the reporter dye) and 6-carboxytetramethylrhodamine (TAMRA) (DNA Technology Ltd.) (the quencher dye). A positive internal control probe was designed (see below) and labeled with tetrachloro-6-carboxyfluorescein, (TET) (the reporter dye) and TAMRA. A phosphate molecule was also attached to the 3' thymine residue to prevent extension of the bound probe during amplification. The chemical synthesis of the TaqMan-modified DNA oligonucleotides was performed using -cyanoethyl phosphoramidite chemistry (3) on an ABI DNA synthesizer (Applied Biosystems, Foster City, Calif.). A 500 A CPG linked with TAMRA was used as solid synthesis support. Because TAMRA is not a base-stable, phenoxyacetyl-protected deoxyribosyladenine, 4-isopropyl-phenoxyacetyl-protected deoxyguanosine and acetyl-protected deoxycytidine phosphoramidites were used to allow mild cleavage and deprotection with t-butylamine-methanol-water (1:1:2 vol/vol/vol) (19). The probe was purified in reverse-phase high-pressure liquid chromatography, where the fraction containing the 5' 6-FAM-3' TAMRA-modified DNA oligonucleotide with absorption maxima at 495 or 555 nm as detected with a diode array.

Internal control DNA. An artificially created chimerical DNA was used as an internal positive control in every reaction mixture, except for the nontemplate controls. The control DNA consisted of a fragment (116 bp) of a coding region of the viral hemorrhagic septicemia virus from rainbow trout (25) (accession no. X66134), flanked by the target for the Salmonella-specific PCR primers. This product was created by a two-step PCR as follows. The first step comprised amplification of DNA from viral hemorrhagic septicemia virus using chimeric PCR primers flanked with the Salmonella-specific primers (Flank-1-left and Flank-2-right; Table 4) by 35 cycles at the following PCR conditions. A 2-µl sample was inoculated into 48 µl of prepared reaction mixture containing 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl₂, 100 µM concentrations of the four deoxynucleotides, 65 ng of each of the oligonucleotide primers, and 0.5 U of Taq polymerase (Applied Biosystems). The samples were subjected to an initial step of denaturation at 94°C for 3 min followed by 35 cycles of denaturation at 94°C for 1 min, annealing at 55°C for 1 min, and extension at 72°C for 1 min in a thermocycler. To ensure complete strand extension, the reaction mixture was kept at 72°C for 10 min after the final cycle. In the second step, the amplicon of step 1 was diluted 1:1,000 in double-distilled water and used as a template in a second amplification using the Salmonella-specific primers (Styinva-JHO-2; Table 4) and the aforementioned amplification condition. The final amplicon was purified using a spin column (QIAquick nucleotide removal kit, catalog no. 28104; QIAGEN, Hilden, Germany), as recommended by the supplier. The eluate, which contained 25 µg of DNA/ml as measured by optical density (at 260 nm) using a spectrophotometer (GeneQuant RNA/DNA Calculator; Pharmacia, Uppsala, Sweden), was diluted 1:10,000 in distilled water before use. The final dilution of the internal control target was established empirically to reduce the competition with target DNA (1). The control DNA (125 pg/reaction) was used as a positive amplification control in all sample wells.

Salmonella 5' nuclease PCR. One colony from each agar plate incubated overnight was transferred, directly and without any treatment, to 50 µl of predisposed PCR mixture in a 96-well microwell plate (MicroAmp, catalog no. 403012; Applied Biosystems). The pre-PCR mixture contained a 900 nM concentration of each primer, 100 nM Salmonella probe, 100 nM internal control probe, 0.05 U of rTth (1) DNA polymerase with Buffer Packs (N808-0098; Applied Biosystems)/µl, 5 µl of 10× chelating buffer (N808-0098; Applied Biosystems), 2.5 mM MgCl₂, 200 µM deoxynucleoside triphosphate blend (N808-0260; Applied Biosystems), 125 pg of internal positive control DNA, 8% (vol/vol) glycerol (molecular biology grade, catalog no. 44448-2V; BDH, Kebolab, Denmark), and double-distilled water to 50 µl. In each microwell plate, two wells were used as nontemplate controls (NTC), which contained all the reagents except for the internal control template and sample.

Durability and reproducibility. The aim was to investigate the possibility of using microwell plates containing ready-prepared Salmonella PCR mixtures containing all the reagents except for the sample. This would alleviate the need for daily preparation of reagents. The microwell plates were added to the PCR mixture as described before, stored at 20°C in the dark, and tested by addition of samples at various time intervals of up to 3 months (Table 5).

View this table: [in this window] [in a new window]   TABLE 5.   The end-point (Rn) fluorescence results for Salmonella enterica isolates tested in premixed and predispensed Salmonella PCR reagents in microwell platesa

Data analysis. The cutoff values were set above the highest TET and FAM end-point fluorescence signals of the non-Salmonella strains (Table 2). Due to the clear difference in the Rn values between the non-Salmonella strains and the Salmonella strains, there was no need for statistical analysis (H. Stryhn, personal communication).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Testing a total of 110 Salmonella strains, covering many serogroups (Table 1) and 120 non-Salmonella isolates (Table 2) assessed the accuracy of the PCR assay developed. The test correctly identified all the Salmonella strains by resulting in Rn end-point values of 1.2 to 10.8 for FAM (the Salmonella probe) and 1.0 to 3.3 for TET (the control probe). Testing of 120 closely related, non-Salmonella isolates resulted in FAM values of 0.1 to 0.5 (Table 2). All but three samples (Campylobacter jejuni) resulted in high TET values (0.7 to 1.7). Three strains of C. jejuni completely inhibited the PCR, resulting in no fluorescence signal for either FAM or TET. The inhibitory effect was not reversed despite the use of purified DNA in various concentrations (data not shown). The end-point results shown in Table 2 are for plate colonies without any treatment.

The cutoff Rn values for the FAM and TET signals to be positive were thus more than 0.6, assigning a "gray-zone" between 0.5 and 0.6 for retesting of possible suspect results. Using this cutoff level, a total of 100 rough Salmonella isolates, previously identified as Salmonella in another PCR and by biochemical tests (12) were all positive in the present 5' nuclease PCR test (Table 3). The FAM values for the rough isolates ranged from 1.0 to 4.1, and the TET values ranged from 1.2 to 3.4.

The reproducibility testing was aimed at investigating the possibility of preparing microwell plates in advance in order to avoid batch variation, reduce the risk of contamination, and facilitate the routine application of the method. The test results on five positive S. enterica strains showed positive Rn values for both the FAM and TET signals after up to 3 months of storage at 20°C (Table 5).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

One of the important potential applications of diagnostic PCR is identification testing. However, the availability of methods and reagents for the identification of Salmonella has been historically limited to biochemical and serological approaches. New tests based on novel reagents tend to supplement existing tests instead of replacing them (26). The emerging exception is nucleic acid technology, which is replacing biochemical and agglutination tests (26).

In order to increase the sensitivity, specificity, and speed of detection of Salmonella, several different genetics-based methods have been developed (4, 7, 15, 24). However, due to the lack of common genes for toxins or other virulence factors, the approach for the design of specific DNA probes has been to select randomly cloned chromosomal fragments. Several conventional PCR tests for detection of Salmonella have been published based on this technique (18, 21). Furthermore, rRNA-directed oligonucleotide probes have been used successfully in a single-phase hybridization assay to detect a large number of serovars of Salmonella, except those belonging to subspecies V (24).

The major limitation to PCR is contamination of specimens with either postamplification products from previous analyses, the so-called carryover problem, or contamination of negative specimens with positive specimens prepared at the same workstation (5). Several extensive guidelines have been developed to prevent false-positive results (14). Application of a premixed, closed-tube PCR can substantially minimize the risk of carryover contamination. This in turn can facilitate the implementation of diagnostic PCR testing in accredited laboratories. In particular, the test developed can simplify the identification procedure by testing presumptive Salmonella colonies directly from indicative agar plates.

Occurrence of false-negative results, mainly due to the presence of DNA polymerase inhibitors (1, 16) or poor quality of target DNA, can be controlled by construction of an internal control sequence, amplified by the same set of primers as the target sequence. The Salmonella PCR used in the present study included such a positive control in the very same tube, reflecting the amplification condition of the samples. The concentration of the internal control template was designed to be suboptimal, so that it reduces the competition with the target template. Thus, the TET Rn values for the non-Salmonella cultures were lower than those for the Salmonella cultures. Due to the known overlap in the emission spectra of the FAM (_em = 518) and the TET (_em = 538), part of the TET signal measured in the Salmonella-positive samples was actually due to the emission of a FAM signal. Another important detail of the test presented was that the results are expressed as end-point values and not as threshold cycle (C_T) values. The C_T values are usually used in quantitative assays. In addition, interpretation of end-point data was simpler than that of the C_T values. With regard to the stability of the premixed PCR mixture, there was no remarkable decrease in the signal values after 3 months. The possibility of storage for a longer period is currently being investigated. However, the shelf life for TaqMan probes is usually 6 months.

Despite many efforts, three C. jejuni strains completely inhibited the PCR although C. jejuni does not grow on Salmonella-selective agar plates and therefore does not create a verification problem. We do not have any explanation for this phenomenon.

The variation in the end-point values presented can be partially due to the amount of colony material, resulting in various concentrations of the amplicon, although the size of the amplicon always corresponded to the theoretical size of the target sequence, as detected in gel electrophoresis (data not shown). In addition, two parameters were found to be important to the correct performance of the Salmonella test developed. First, the use of rTth as DNA polymerase was crucial, since preliminary experiments with other enzymes, e.g., AmpliTaq Gold, resulted in false-positive or false-negative results (J. Hoorfar, R. Knutsson, and P. Rådström, Abstr. 99th Gen. Meet. Am. Soc. Microbiol. 1999, abstr. P-99, p. 530, 1999). Second, in the software setup prior to data analysis, the reference dye for interwell calibration was assigned to be TAMA and not ROX.

Although the automated on-line PCR is rather complicated and requires costly equipment, an increasing number of reference laboratories are converting the traditional gel-based detection PCR to online, fluorescence-based detection in order to facilitate application of the increasing number of quantitative PCR kits.

In conclusion, the identities of presumptive Salmonella isolates can be conveniently confirmed using the TaqMan PCR test described here. The test is currently implemented in our accredited routine Salmonella laboratory and can be easily implemented in other accredited laboratories with limited experience in molecular biology. However, further studies are warranted to assess the application of Salmonella PCR to complex biological samples, since inhibitory substances inherent in various samples can interfere with the amplification (16, 27).