Abstract
Salmonellosis is one of the most important food-borne diseases worldwide. For outbreak investigation and infection control, accurate and fast subtyping methods are essential. A triplex gene-scanning assay was developed and evaluated for serotype-specific subtyping of Salmonella enterica isolates based on specific single-nucleotide polymorphisms in fragments of fljB, gyrB, and ycfQ. Simultaneous gene scanning of fljB, gyrB, and ycfQ by high-resolution melting-curve analysis of 417 Salmonella isolates comprising 46 different serotypes allowed the unequivocal, simple, and fast identification of 37 serotypes. Identical melting-curve profiles were obtained in some cases from Salmonella enterica serotype Enteritidis and Salmonella enterica serotype Dublin, in all cases from Salmonella enterica serotype Ohio and Salmonella enterica serotype Rissen, from Salmonella enterica serotype Mbandaka and Salmonella enterica serotype Kentucky, and from Salmonella enterica serotype Bredeney, Salmonella enterica serotype Give, and Salmonella enterica serotype Schwarzengrund. To differentiate the most frequent Salmonella serotype, Enteritidis, from some S. Dublin isolates, an additional single PCR assay was developed for specific identification of S. Enteritidis. The closed-tube triplex high-resolution melting-curve assay developed, in combination with an S. Enteritidis-specific PCR, represents an improved protocol for accurate, cost-effective, simple, and fast subtyping of 39 Salmonella serotypes. These 39 serotypes represent more than 94% of all human and more than 85% of all nonhuman Salmonella isolates (including isolates from veterinary, food, and environmental samples) obtained in the years 2008 and 2009 in Austria.
INTRODUCTION
Salmonella enterica is one of the most important global food-borne pathogens. Every year, millions of human cases of salmonellosis are reported all over the world, resulting in thousands of deaths (28). In Austria, the incidence of salmonellosis was about 38 cases per 100,000 inhabitants in 2008 and 34 cases per 100,000 inhabitants in 2009 (9, 10). The gold standard for Salmonella subtyping is based on the scheme developed by Kauffmann, White, and Le Minor, using the serologic identification of O (somatic) and H (flagellar) antigens (4). Presently, more than 2,500 different Salmonella serotypes (or serovars) have been defined. Despite the usefulness of serotyping, testing with a complete set of antisera is time-consuming and requires a well-trained technician (5). Furthermore, serotyping does not reveal the genetic relatedness of different isolates (25). Thus, many molecular-based subtyping methods, i.e., pulsed-field gel electrophoresis (PFGE), amplified fragment length polymorphism (AFLP) (14), multilocus sequence subtyping (MLST), multilocus enzyme electrophoresis (MLEE) (25), multiple-locus variable-number tandem-repeat analysis (MLVA) (11), and microarray techniques (15), are routinely used for Salmonella subtyping. The advantage of sequence-based methods, like MLST or single-nucleotide polymorphism (SNP), over DNA band pattern-based methods, like PFGE or AFLP, is better and easier comparability of data (1, 6, 16).
For fast identification and subtyping of isolates, especially for outbreak situations and routine diagnostics, a PCR-based subtyping method is preferable in terms of cost, simplicity, turnaround time, and potential for standardization, since MLST and detection of multiple SNPs still represent time-consuming and cost-intensive approaches.
High-resolution melting (HRM) analysis represents a relatively new method using single-nucleotide polymorphisms for strain differentiation. The principle of HRM analysis is based on the generation of different melting-curve profiles due to sequence variations in double-stranded DNA. Single-nucleotide changes represent the smallest genetic changes and are divided into four classes distinguished by different melting-temperature (Tm) shifts (13, 27). SNP class 1 involves C/T and G/A, and SNP class 2 involves C/A and G/T, base exchanges that can easily be genotyped by HRM due to Tm differences of more than 0.5°C (13). In contrast, in SNP class 3, only C/G base exchange occurs, and SNP class 4 is described by A/T base exchange, producing very small Tm differences (<0.4°C for SNP class 3 and <0.2°C for SNP class 4) (21).
In general, HRM analysis is a simple, inexpensive, and rapid scanning method for known and unknown mutations (21, 22, 29) and can dramatically reduce the turnaround time for mutation screening and testing. Here, we describe the development and evaluation of a triplex HRM assay based on sequence variations in three genes (fljB, gyrB, and ycfQ), in combination with a Salmonella enterica serotype Enteritidis-specific real-time PCR assay, for the fast identification of Salmonella serotypes.
MATERIALS AND METHODS
Microorganisms.
A set of 33 Salmonella isolates representing the 14 most frequent Salmonella enterica serotypes found in 2008 in Austria (S. Enteritidis [n = 4], Salmonella enterica serotype Typhimurium biphasic variant [n = 3, including the reference strain, ATCC 14028], S. Typhimurium monophasic variant [n = 2], Salmonella enterica serotype Infantis [n = 2], Salmonella enterica serotype Saintpaul [n = 2], Salmonella enterica serotype Hadar [n = 2], Salmonella enterica serotype Agona [n = 2], Salmonella enterica serotype Newport [n = 2], Salmonella enterica serotype Thompson [n = 2], Salmonella enterica serotype Abony [n = 2], Salmonella enterica serotype Montevideo [n = 2], Salmonella enterica serotype Senftenberg [n = 2], Salmonella enterica serotype Dublin [n = 2], and Salmonella enterica serotype Tennessee [n = 2]) and one relatively rare serotype (Salmonella enterica serotype Indiana [n = 2]) was initially used to develop a triplex high-resolution melting-curve assay. To evaluate the performance of the assay, an additional collection of 385 Salmonella isolates comprising 46 serotypes was analyzed. Ten Salmonella enterica serotype Gallinarum and 10 Salmonella enterica serotype Choleraesuis (including nine S. Choleraesuis var. kunzendorf) isolates closely related to S. Enteritidis were analyzed by the specific S. Enteritidis edrI assay.
All 438 Salmonella isolates were provided by the Austrian National Reference Centre for Salmonella and serotyped according to the White-Kauffmann-Le Minor scheme.
Bacteria were grown overnight at 37°C on BD BBL XLD agar (Becton, Dickinson and Company, Franklin Lakes, NJ). Genomic DNA (gDNA) was isolated with the Genelute Bacterial Genomic DNA kit (Sigma, St. Louis, MO) according to the instructions of the manufacturer. The concentration and quality of the purified gDNA were determined by UV spectrophotometry at 260 and 280 nm and agarose gel (1.5% [wt/vol]) electrophoresis.
Detection of serotype-specific SNPs.
Serotype-specific SNPs and fragments were identified through extensive sequence analysis of the published sequences of atpD, fljB, and gyrB, as well as through isolation, reamplification, and sequencing of AFLP fragments (23). Finally, loci within fljB, gyrB, and ycfQ that allowed the unequivocal differentiation of the initial set of isolates were chosen for amplification and subsequent HRM analysis. The Primer-BLAST online tool from the National Center for Biotechnology Information (NCBI) (http://www.ncbi.nlm.nih.gov) was used for specific HRM primer design.
High-resolution melting-curve PCR analysis.
HRM analysis was performed by specific coamplification of a 170-bp fragment of fljB using the primers fljB-forward (5′-GTGAAAGATACAGCAGTAACAACG-3′) and fljB-reverse (5′-ACAAAGTACTTGTTATTATCTGCG-3′), a 171-bp fragment of gyrB using the primers gyrB-forward (5′-AAACGCCGATCCACCCGA-3′) and gyrB-reverse (5′-TCATCGCCGCACGGAAG-3′), and a 241-bp fragment of ycfQ using the primers ycfQ-forward (5′-GCCTACTCTCTATGCGGAATTCAC-3′) and ycfQ-reverse (5′-GATATCGCGCGAGGAGGCG-3′). All the primers had the respective M13 sequence attached to the 5′ end of the gene-specific priming sequence for subsequent sequencing. PCR and HRM analysis were performed on the LightCycler LC480 (Roche Diagnostics, Penzberg, Germany). In a final volume of 10 μl, the HRM PCR mixture contained 2 ng of gDNA, 0.1 pmol of each gyrB primer, 0.075 pmol of each fljB primer, 0.075 pmol of each ycfQ primer, and 3 mM MgCl2 in the LightCycler 480 High Resolution Melting Master Mixture containing ResoLight dye (Roche Diagnostics). HRM PCR included an activation step at 95°C for 10 min, followed by 45 cycles of 95°C for 10 s, 60°C for 10 s, and 72°C for 10 s. Prior to HRM analysis, the amplification products were heated to 95°C for 1 min and then cooled to 40°C for 1 min. HRM analysis was performed from 60°C to 95°C, rising at 1°C/s, with 25 acquisitions per degree.
We used the LC 480 gene-scanning software version 1.5 with manual settings for sensitivity at 0.30 and for temperature shift at threshold 5, a premelt normalization range from 80.87 to 81.51, and a postmelt normalization range from 89.17 to 89.92 for HRM analysis. After normalizing and temperature shifting of the melting curves, difference plots were generated by selecting HRM profile 6, representing S. Thompson, as the baseline. Only amplification products reaching the plateau phase were analyzed. For optimal performance (i.e., correct assignment of melting-curve profiles to a known sequence) of HRM experiments, each run must contain well-characterized standards (i.e., strains with known sequences) because melting-curve profiles obtained from different runs cannot be compared directly.
Sequencing of HRM products.
For sequencing of HRM products, a singleplex reaction (20-μl final volume) was applied with the respective primer pair (10 pmol of each primer), 2 ng of gDNA, and the RedTaq Ready Mix (Sigma). The amplification reaction was performed in a Master Cycler Epgradient S (Eppendorf, Hamburg, Germany) programmed as follows: an initial step of 30 s at 95°C and 35 cycles of 30 s at 95°C, 30 s at 55°C, and 1 min at 72°C. After amplification, 5 μl of each PCR product was analyzed on a 1.5% (wt/vol) agarose gel. PCR products were purified using the QIAquick PCR purification kit (Qiagen, Inc., Chatsworth, CA). Sequence analysis was performed using a SequiTerm Excel II Cycle Sequencing Kit (Epicentre, Madison, WI) with fluorescence-labeled primers M13 universal (5′-TGTAAAACGACGGCCAGT-3′) and M13 reverse (5′-CAGGAAACAGCTATGACC-3′) (MWG-Biotech, Ebersberg, Germany) in a Licor 4300 automated DNA sequencer (Li-Cor Bioscience, Lincoln, NE) according to the manufacturer's instructions. All sequences obtained were assembled, edited, and compared to determine sequence variations.
Specific detection of S. enterica serovar Enteritidis.
In previous experiments, AFLP analysis resulted in the identification of an S. Enteritidis-specific fragment (23) identical to Salmonella enterica subsp. enterica serovar Enteritidis difference region I (AF370716) (2). For specific detection of S. Enteritidis, a 156-bp fragment of S. Enteritidis difference region I was amplified using the primers EntI-forward (5′-GACGAGCTCTTTACACTCCATCAGTT-3′) and EntI-reverse (5′-GAAAGTGTTTCCAGAACTCTTGTTGCAT-3′). PCR was performed on a LightCycler LC480 (Roche Diagnostics). The specificity of the assay was evaluated using 86 S. Enteritidis and 352 non-S. Enteritidis isolates (including the closely related S. Dublin, S. Gallinarum, and S. Choleraesuis) as described in “Microorganisms” above.
The reaction mixture contained 2 ng of gDNA, 0.25 pmol of each primer, and 3 mM MgCl2 in the LightCycler 480 SYBR green I Master Mix (Roche Diagnostics) with PCR grade water adjusted to a final volume of 10 μl. The reaction conditions included an initial denaturation step at 95°C for 10 min, followed by 35 cycles of 95°C for 10 s, 60°C for 10 s, and 72°C for 40 s.
Multilocus sequence subtyping.
MLST analysis was performed on 84 isolates, representing the serotypes S. Enteritidis (n = 21), S. Dublin (n = 13), S. Montevideo (n = 19), S. Saintpaul (n = 13), S. Newport (n = 13), and Salmonella enterica serotype Paratyphi B var. d-tartrate+ (n = 5), by determining the sequences of seven housekeeping genes (aroC, dnaN, hemD, hisD, purE, sucA, and thrA), as described previously (8). Amplification was performed in a 20-μl reaction mixture using RedTaq Ready Mix (Sigma) and 10 pmol of each primer. The amplification conditions were as follows: an initial step of 30 s at 95°C and 45 cycles of 30 s at 95°C, 30 s at 53°C, and 1 min at 72°C. After amplification, 5 μl of each PCR product was analyzed on a 1.5% (wt/vol) agarose gel. Prior to sequencing, the amplification products were purified with Exo Sap-It (GE Healthcare, Buckinghamshire, United Kingdom). Two microliters of purified amplification product was used for subsequent sequencing with the BigDyeTerminator v3.1 sequencing kit (Applied Biosystems, Carlsbad, CA). The products were analyzed on an ABI Genetic Analyzer 3500Dx (Applied Biosystems). The sequences obtained were compared to the reference sequences available at the Salmonella MLST database (http://mlst.ucc.ie/mlst/dbs/Senterica) for allele identification. Novel alleles and sequence types (STs) were submitted for allele, ST, and clonal complex (CC) designations. CLUSTAL W was used for sequence comparison and determination of the genetic relatedness of the different MLSTs (http://www.clustal.org/) (12).
Nucleotide sequence accession numbers.
The sequences of the eight different fljB sequence types were deposited in GenBank under accession numbers JQ514786 and JQ629418 to JQ629424, the 21 different gyrB sequence types under GenBank accession numbers JQ514787 and JQ629443 to JQ629462, and the 19 different ycfQ sequence types under GenBank accession numbers JQ595558 and JQ629425 to JQ629442.
RESULTS
High-resolution melting-curve PCR analysis.
We performed parallel gene scanning of three amplification products of gyrB, fljB, and ycQf on the initial set of 33 Salmonella isolates, comprising 15 Austrian S. enterica serotypes (the 14 most common serotypes and one rare serotype, accounting for 85% of all tested Salmonella isolates in 2008 in Austria). They yielded 16 different melting-curve profiles (HRM-CP 1 to 16), as shown in Fig. 1. Most S. enterica serotypes had one specific melting-curve profile, except S. Newport, which had two specific HRM profiles (HRM-CP 3 and HRM-CP 4) (Fig. 1 and Table 1). The triplex HRM assay was blindly tested on an arbitrary collection of 385 Salmonella isolates. Two hundred sixty-eight (69.6%) of these 385 additional isolates could be assigned to the initial 16 HRM-CPs. The remaining 117 isolates yielded 39 new melting-curve profiles derived from 36 serotypes (curves not shown). Classical serotyping revealed that the isolate collection (all 418 isolates) contained 46 different serotypes that yielded 55 distinct melting-curve profiles (Table 1).
Fig 1.
Difference plot of melting-curve profiles of normalized and temperature (Temp)-shifted amplification products of the genes fljB, gyrB, and ycfQ from the 15 different Salmonella serotypes evaluated in the initial set of isolates showing HRM-CPs 1 to 16.
Table 1.
Sequence types and triplex HRM profiles of fljB, gyrB, and ycfQa
| Serotype (no. of strains) | Curve profile/sequence type |
||||
|---|---|---|---|---|---|
| HRM-CP | HRM-ST | fljB-ST | gyrB-ST | ycfQ-ST | |
| S. Abony (21) | 9 | 9 | 5 | 4 | |
| S. Agona (13) | 13 | 13 | 7 | 5 | |
| S. Amsterdam (1) | 34 | 35 | 21 | 3 | |
| S. Blockley (4) | 19 | 19 | 3 | 10 | 7 |
| S. Coeln (3) | 38 | 39 | 1 | 12 | 2 |
| S. Corvallis (4) | 49 | 51 | 1 | 5 | |
| S. Derby (1) | 44 | 45 | 16 | 18 | |
| S. Derby (2) | 43 | 44 | 16 | 17 | |
| S. Hadar (14) | 8 | 8 | 1 | 3 | |
| S. Havana (1) | 53 | 55 | 16 | 12 | |
| S. Heidelberg(7) | 31 | 31 | 1 | 1 | 4 |
| S. Indiana (21) | 7 | 7 | 4 | 4 | 5 |
| S. Infantis (13) | 5 | 5 | 2 | 1 | 3 |
| S. Javiana (1) | 47 | 49 | 7 | 16 | 9 |
| S. Kedougou (2) | 36 | 37 | 3 | 3 | 2 |
| S. Kedougou (4) | 23 | 23 | 3 | 1 | 3 |
| S. Kottbus (6) | 37 | 38 | 20 | 5 | |
| S. Livingstone (1) | 41 | 42 | 8 | 4 | |
| S. Livingstone (2) | 48 | 50 | 1 | 2 | |
| S. Manhattan (5) | 22 | 22 | 3 | 3 | 4 |
| S. Mississippi (2) | 35 | 36 | 5 | 7 | 2 |
| S. Montevideo (10) | 16 | 16 | 8 | 9 | |
| S. Montevideo (9) | 17 | 17 | 9 | 10 | |
| S. Muenchen (3) | 21 | 21 | 1 | 13 | 7 |
| S. Napoli (1) | 32 | 32 | 8 | 13 | |
| S. Newport (1) | 28 | 28 | 1 | 13 | 11 |
| S. Newport (2) | 18 | 18 | 1 | 2 | 4 |
| S. Newport (2) | 27 | 27 | 1 | 3 | 11 |
| S. Newport (4) | 3 | 3 | 1 | 1 | 3 |
| S. Newport (4) | 4 | 4 | 1 | 3 | 3 |
| S. Orion (1) | 45 | 46 | 2 | 19 | 14 |
| S. Paratyphi B var. d-tartrate+ (1) | 55 | 57 | 1 | 14 | 5 |
| S. Paratyphi B var. d-tartrate+ (2) | 25 | 25 | 1 | 8 | 4 |
| S. Paratyphi B var. d-tartrate+ (2) | 54 | 56 | 13 | 19 | |
| S. Poona (1) | 51 | 53 | 8 | 16 | 9 |
| S. Saintpaul (1) | 26 | 26 | 1 | 3 | 2 |
| S. Saintpaul (13) | 2 | 2 | 1 | 2 | 2 |
| S. Sandiego (1) | 24 | 24 | 14 | 9 | |
| S. Senftenberg (10) | 12 | 12 | 7 | 6 | |
| S. Senftenberg (3) | 30 | 30 | 3 | 12 | |
| S. Stanley (3) | 50 | 52 | 1 | 14 | 4 |
| S. Tennessee (21) | 10 | 10 | 6 | 5 | |
| S. Thompson (1) | 29 | 29 | 3 | 17 | 4 |
| S. Thompson (12) | 6 | 6 | 3 | 1 | 4 |
| S. Typhimurium biphasic variant (25) | 1 | 1 | 1 | 1 | 1 |
| S. Typhimurium monophasic variant (21) | 11 | 11 | 1 | 1 | |
| S. Virchow (3) | 20 | 20 | 1 | 11 | 4 |
| S. Worthington (1) | 52 | 54 | 13 | 12 | |
| S. IIIb 61:k:1,5,7 (3) | 39 | 40 | 6 | 18 | 15 |
| S. IIIb 59:z10:z53 (1) | 40 | 41 | 18 | 16 | |
| S. Dublin (6) | 15 | 15 | 2 | 8 | |
| S. Dublin (7) | 14 | 14 | 2 | 7 | |
| S. Enteritidis (86) | 14 | 14 | 2 | 7 | |
| S. Kentucky (3) | 42 | 43 | 13 | 2 | |
| S. Mbandaka (10) | 42 | 43 | 13 | 2 | |
| S. Ohio (1) | 46 | 47 | 15 | 4 | |
| S. Rissen (4) | 46 | 48 | 15 | 5 | |
| S. Give (3) | 33 | 33 | 4 | 16 | 9 |
| S. Schwarzengrund (3) | 33 | 33 | 4 | 16 | 9 |
| S. Bredeney (5) | 33 | 34 | 4 | 8 | 9 |
Serotypes characterized by single, unique melting-curve profiles are shown in Table 1. In addition, subtyping within a serotype was possible for S. Saintpaul, S. Thompson, S. Senftenberg, S. Montevideo, Salmonella enterica serotype Kedougou, Salmonella enterica serotype Derby, and Salmonella enterica serotype Livingstone, each characterized by two distinct melting-curve profiles; S. Paratyphi B var. d-tartrate+ yielded three unique melting-curve profiles, and S. Newport showed five characteristic melting-curve profiles (Table 1). All 86 investigated S. Enteritidis isolates and 7 out of 13 S. Dublin isolates shared HRM-CP 14; all 10 Salmonella enterica serotype Mbandaka and all three Salmonella enterica serotype Kentucky isolates shared HRM-CP 42; the single Salmonella enterica serotype Ohio isolate and all four Salmonella enterica serotype Rissen isolates shared HRM-CP 46; all five Salmonella enterica serotype Bredeney, all three Salmonella enterica serotype Schwarzengrund, and all three Salmonella enterica serotype Give isolates shared HRM-CP 33 (Table 1).
Sequencing of the three amplification products of fljB, gyrB, and ycfQ detected eight different fljB sequence types, 21 different gyrB sequence types, and 19 different ycfQ sequence types (Tables 2 to 4).
Table 2.
Mutations detected among the 8 different fljB STsa
| fljB ST | Nucleotide at position: |
|||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 566 | 568 | 576 | 577 | 599 | 600 | 603 | 606 | 610 | 611 | 616 | 618 | 621 | 624 | 628 | 630 | 633 | 635 | 636 | 646 | 647 | 651 | 652 | 654 | 655 | 658 | 659 | 666 | 669 | 670 | 673 | 675 | |
| 1 | A | G | C | A | T | A | G | T | G | A | G | A | T | T | G | G | T | C | G | A | A | T | A | G | G | T | C | C | T | G | G | G |
| 2 | C | G | A | C | A | C | ||||||||||||||||||||||||||
| 3 | G | C | C | A | C | G | C | C | C | T | T | C | G | G | A | A | ||||||||||||||||
| 4 | G | A | C | A | G | A | A | |||||||||||||||||||||||||
| 5 | G | C | C | C | T | A | C | C | G | G | C | A | A | |||||||||||||||||||
| 6 | T | G | A | G | C | G | C | A | G | C | ||||||||||||||||||||||
| 7 | G | C | C | A | C | G | C | C | A | C | T | T | C | G | G | A | A | |||||||||||||||
| 8 | A | G | C | C | A | C | A | C | T | A | C | T | C | A | A | |||||||||||||||||
In the GenBank entry of S. Typhimurium LT2 (AE006468), which shows sequence type 1, the amplified fragments correspond to positions 541 to 710 of fljB. Numbering starts with the A of the start codon.
Table 4.
Sequence differences detected among the 19 identified ycfQ STsa
| ycfQ ST | Nucleotide at position: |
||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 180 | 186 | 190 | 210 | 243 | 258 | 264 | 279 | 284 | 285 | 288 | 292 | 300 | 315 | 321 | 330 | 333 | 342 | 351 | 369 | 378 | 385 | 386 | |
| 1 | C | C | A | C | A | C | T | A | C | G | G | G | C | C | C | T | A | G | C | C | G | T | C |
| 2 | G | ||||||||||||||||||||||
| 3 | T | G | A | G | G | C | |||||||||||||||||
| 4 | G | G | G | C | |||||||||||||||||||
| 5 | G | G | G | ||||||||||||||||||||
| 6 | T | G | G | G | A | T | C | ||||||||||||||||
| 7 | T | G | G | G | C | ||||||||||||||||||
| 8 | T | G | G | G | A | C | |||||||||||||||||
| 9 | G | G | G | G | C | ||||||||||||||||||
| 10 | G | C | |||||||||||||||||||||
| 11 | G | G | G | T | |||||||||||||||||||
| 12 | G | G | G | G | |||||||||||||||||||
| 13 | T | G | A | G | G | G | C | ||||||||||||||||
| 14 | T | G | A | G | G | G | C | ||||||||||||||||
| 15 | G | G | C | G | G | C | T | T | T | C | T | ||||||||||||
| 16 | G | G | C | G | G | C | T | T | T | C | |||||||||||||
| 17 | G | G | G | A | C | T | |||||||||||||||||
| 18 | G | C | T | ||||||||||||||||||||
| 19 | G | G | |||||||||||||||||||||
In the GenBank entry of S. Typhimurium LT2 (AE006468), which shows sequence type 1, the amplified fragments correspond to positions 159 to 399 of ycfQ. Numbering starts with the A of the start codon.
Table 3.
SNPs detected between the 21 different gyrB STsa
| gyrB ST | Nucleotide at position: |
|||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 726 | 753 | 756 | 762 | 768 | 780 | 786 | 795 | 807 | 828 | 831 | 837 | 840 | 846 | |
| 1 | T | C | G | A | G | T | C | C | T | C | C | T | T | T |
| 2 | C | C | T | T | ||||||||||
| 3 | C | A | C | G | ||||||||||
| 4 | C | T | T | C | G | |||||||||
| 5 | C | |||||||||||||
| 6 | C | T | G | |||||||||||
| 7 | C | T | A | C | C | G | ||||||||
| 8 | C | C | G | |||||||||||
| 9 | T | C | G | |||||||||||
| 10 | T | T | T | C | T | T | G | |||||||
| 11 | C | T | G | C | G | |||||||||
| 12 | C | T | T | |||||||||||
| 13 | C | C | C | G | ||||||||||
| 14 | C | T | G | |||||||||||
| 15 | C | C | G | |||||||||||
| 16 | C | G | ||||||||||||
| 17 | C | A | A | C | G | |||||||||
| 18 | C | C | G | A | C | T | T | C | G | |||||
| 19 | C | A | C | G | ||||||||||
| 20 | T | C | G | |||||||||||
| 21 | G | |||||||||||||
In the GenBank entry of S. Typhimurium LT2 (AE006468) which shows sequence type 1, the amplified fragments correspond to positions 698 to 868 of gyrB. Numbering starts with the A of the start codon.
Amplification of the fljB fragment was observed in 22 serotypes (Table 1).
Amplification of the gyrB and the ycfQ fragment was possible with all Salmonella isolates. Based on the combination of the sequences of the amplified fragments, 57 different HRM sequence types (HRM-STs) could be defined, resulting in HRM-CP 1 to 55 (Table 1).
For accurate differentiation of S. Enteritidis from S. Dublin isolates showing HRM-CP 14, a PCR assay specific for the S. Enteritidis edrI region was developed. The specific edrI fragment could be amplified from all 86 S. Enteritidis isolates tested, but not from the 352 non-S. Enteritidis isolates (Fig. 2).
Fig 2.
Specific detection of S. Enteritidis by amplification of an edrI gene fragment (red). No edrI fragment was detectable in isolates of all other investigated serotypes (blue).
Multilocus sequence typing analysis.
To assess the HRM results obtained for S. Enteritidis and S. Dublin, as well as serotypes displaying multiple HRM profiles, a set of arbitrarily chosen isolates was analyzed by multilocus sequence typing. MLST of 21 S. Enteritidis isolates and all seven S. Dublin isolates showing HRM-CP 14 revealed that all 21 S. Enteritidis isolates belonged to MLST 11 (ST-11) and the seven S. Dublin isolates to ST-1494. ST-11 and ST-1494 could be distinguished due to sequence differences in hisD (13 SNPs), purE (2 SNPs), sucA (8 SNPs), and thrA (4 SNPs) (Table 5). HRM-CP 15, characteristic of the remaining six S. Dublin isolates, belonged to ST-10 (5 isolates) and ST-1487 (1 isolate). Neighbor-joining placed ST-10, ST-1487, and ST-1494 into one cluster (Fig. 3). Differences between these MLSTs were only detected in the sequences of aroC and hisD. ST-11 had the highest sequence similarity to ST-1494 (Table 5).
Table 5.
Detected MLSTs compared to the observed HRM profilesa
| Serotype (no. of isolates) | Allele |
MLST | HRM-ST | HRM profile | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| aroC | dnaN | hemD | hisD | purE | sucA | thrA | ||||
| S. Enteritidis (21) | 5 | 2 | 3 | 7 | 6 | 6 | 11 | 11 | 14 | 14 |
| S. Dublin (7) | 5 | 2 | 3 | 491 | 5 | 5 | 10 | 1494 | 14 | 14 |
| S. Dublin (5) | 5 | 2 | 3 | 6 | 5 | 5 | 10 | 10 | 15 | 15 |
| S. Dublin (1) | 410 | 2 | 3 | 6 | 5 | 5 | 10 | 1487 | 15 | 15 |
| S. Newport (4) | 2 | 2 | 15 | 14 | 15 | 20 | 12 | 31 | 3 | 3 |
| S. Newport (3) | 10 | 7 | 21 | 14 | 15 | 12 | 12 | 45 | 4 | 4 |
| S. Newport (1) | 10 | 61 | 18 | 67 | 15 | 12 | 12 | 1496 | 4 | 4 |
| S. Newport (2) | 5 | 14 | 6 | 12 | 5 | 14 | 58 | 166 | 18 | 18 |
| S. Newport (1) | 10 | 7 | 21 | 12 | 15 | 12 | 12 | 46 | 27 | 27 |
| S. Newport (1) | 16 | 2 | 61 | 43 | 36 | 39 | 42 | 223 | 27 | 27 |
| S. Newport (1) | 10 | 387 | 21 | 12 | 15 | 12 | 12 | 1497 | 28 | 28 |
| S. Montevideo (5) | 43 | 31 | 16 | 42 | 35 | 13 | 111 | 1488 | 16 | 16 |
| S. Montevideo (2) | 43 | 31 | 16 | 13 | 34 | 13 | 4 | 1489 | 16 | 16 |
| S. Montevideo (1) | 411 | 41 | 18 | 13 | 12 | 13 | 4 | 1492 | 16 | 16 |
| S. Montevideo (1) | 11 | 41 | 55 | 42 | 34 | 58 | 4 | 138 | 16 | 16 |
| S. Montevideo (1) | 201 | 41 | 16 | 42 | 35 | 13 | 111 | 1493 | 16 | 16 |
| S. Montevideo (8) | 412 | 204 | 8 | 247 | 216 | 6 | 194 | 1490 | 17 | 17 |
| S. Montevideo (1) | 201 | 204 | 8 | 247 | 216 | 6 | 194 | 1491 | 17 | 17 |
| S. Saintpaul (10) | 5 | 14 | 18 | 9 | 6 | 12 | 17 | 27 | 2 | 2 |
| S. Saintpaul (2) | 201 | 14 | 18 | 9 | 6 | 12 | 17 | 680 | 2 | 2 |
| S. Saintpaul (1) | 5 | 14 | 21 | 9 | 6 | 12 | 17 | 49 | 26 | 26 |
| S. Paratyphi B var. d-tartrate+ (1) | 46 | 44 | 46 | 46 | 38 | 18 | 34 | 88 | 25 | 25 |
| S. Paratyphi B var. d-tartrate+ (1) | 2 | 14 | 24 | 14 | 2 | 19 | 140 | 1560 | 25 | 25 |
| S. Paratyphi B var. d-tartrate+ (2) | 20 | 4 | 23 | 14 | 16 | 19 | 140 | 423 | 56 | 54 |
| S. Paratyphi B var. d-tartrate+ (1) | 15 | 15 | 19 | 17 | 5 | 19 | 18 | 28 | 57 | 55 |
The numbers in boldface represent newly detected alleles or MLSTs.
Fig 3.
Phylogenetic tree of 84 Salmonella isolates based on alignment of the seven MLST genes (aroC, dnaN, hemD, hisD, purE, sucA, and thrA) sequenced. Identified MLSTs were compared to the respective HRM-CPs.
The five HRM-CPs (HRM-CP 3, HRM-CP 4, HRM-CP 18, HRM-CP 27, and HRM-CP 28) specific for 13 S. Newport isolates consisted of seven MLSTs. HRM-CP 3, HRM-CP 18, and HRM-CP 28 could be assigned to ST-31, ST-166, and ST-1497. HRM-CP 4 comprised ST-45 (3 isolates) and ST-1496 (1 isolate), and HRM-CP 27 comprised ST-46 (1 isolate) and ST-223 (1 isolate). Neighbor-joining analysis revealed one cluster for the MLSTs: ST-166, ST-31, ST-45, ST-1496, ST-46, and ST1497, with the highest sequence homology between ST-46 and ST-1497 (Fig. 3 and Table 5). The 19 S. Montevideo isolates investigated belonged to two HRM-CPs and to seven MLSTs. HRM-CP 16 comprised ST-1488 (5 isolates), ST-1489 (2 isolates), ST-1492 (1 isolate), ST-138 (1 isolate), and ST-1493 (1 isolate). HRM-CP 17 comprised ST-1490 (8 isolates) and ST-1491 (1 isolate). All S. Montevideo MLSTs could be assigned to a serotype-specific cluster (Fig. 3 and Table 5).
Thirteen S. Saintpaul isolates, characterized by HRM-CP 2 and HRM-CP 26, consisted of three MLSTs. HRM-CP 2 comprised ST-27 (10 isolates) and ST-680 (2 isolates), and HRM-CP 26 was assignable to ST-49. Comparative sequence analysis resulted in a serotype-specific cluster comprising ST-27, ST-680, and ST-49 (Fig. 3 and Table 5).
The three S. Paratyphi B var. d-tartrate+-specific HRM-CPs were assignable to four MLSTs. HRM-CP 54 and HRM-CP 55 could be assigned to the MLSTs ST-423 and ST-28. HRM-CP 25 was assignable to ST-88 (1 isolate) and ST-1560 (1 isolate). S. Paratyphi B var. d-tartrate+ ST-423, ST-1560, and ST-28 formed a cluster with S. Newport ST-223 (HRM-CP 27) that was adjacent to the S. Newport cluster comprising all other Newport STs. Finally, the novel alleles (aroC10, aroC411, aroC412, dnaN387, and hisD491) identified and new sequence types, ST-1487, ST-1488, ST-1489, ST-1490, ST-1491, ST-1492, ST-1493, ST-1494, ST-1496, ST-1497, and ST-1560, were submitted to the MLST database.
DISCUSSION
Conventional Salmonella serotyping according to the White-Kauffmann-Le Minor scheme (4, 19) is a time-consuming and demanding method and requires vertebrate animals for serum production (5). Therefore, several molecular-based subtyping methods have been developed for Salmonella subtyping (5, 7, 14). PFGE and MLST are still expensive and time-consuming methods and are therefore of limited value for routine subtyping (14, 24). Most PCR-based subtyping procedures allow the detection of only a single serotype (17, 20), and even in multiplex-PCR-based approaches, the number of identifiable serotypes is low and requires a multistep protocol (18). Advances in whole-genome sequencing will further facilitate the identification of suitable markers and therefore improve molecular subtyping procedures (3). Here, we describe an improved protocol for accurate, cost-effective, simple, and fast subtyping of Salmonella serotypes in a single closed-tube assay format by parallel high-resolution melting-curve analysis based on serotype-specific SNPs within distinct loci of fljB, gyrB (26), and ycfQ (23). Evaluation of the triplex HRM assay we developed, using an arbitrary collection of 418 Salmonella isolates comprising a total of 46 different serotypes, generated 55 distinct melting-curve profiles. Subsequent sequencing revealed that these 55 melting curves consisted of 57 different HRM sequence types (sequences of fljB, gyrB, and ycfQ amplicons). All 86 S. Enteritidis isolates investigated and 7 out of 13 S. Dublin isolates yielded a unique melting-curve profile (HRM-CP 14) due to identical sequences of the three amplification products. Six S. Dublin isolates produced HRM profile 15 and could therefore be differentiated from the S. Enteritidis isolates. At nucleotide position 288 of ycfQ, an SNP (G-to-A transition) specific to these six S. Dublin (HRM-CP 15) isolates was detected. This SNP represents the only difference between the six isolates of S. Dublin and all tested S. Enteritidis isolates that could be detected in the three amplification products mentioned above. The other seven S. Dublin isolates investigated lacked this SNP and were therefore indistinguishable from the S. Enteritidis isolates. For unambiguous identification of the most frequent serotype, S. Enteritidis, a PCR assay specific for S. Enteritidis difference region I was developed (2, 23). Evaluation of this S. Enteritidis-specific assay with 438 Salmonella isolates comprising 48 serotypes revealed 100% specificity and 100% sensitivity.
The triplex HRM assay we developed failed to differentiate between S. Mbandaka and S. Kentucky (both HRM-CP 42), between S. Ohio and S. Rissen (both HRM-CP 46), and between S. Give, S. Schwarzengrund, and S. Bredeney (all HRM-CP 33). Sequencing of the three HRM amplification products revealed no sequence variations between the serotypes S. Mbandaka and S. Kentucky (both HRM-ST 43) or between S. Schwarzengrund and S. Give (both HRM-ST 33).
However, a single SNP class 3 G-to-C base exchange within the gyrB amplicon differentiated S. Bredeney (HRM-ST 34) from S. Schwarzengrund and S. Give (both HRM-ST 33). Furthermore, a single G-to-C base exchange within the ycfQ amplicon differentiated S. Ohio (HRM-ST 47) from S. Rissen (HRM-ST 48). In both cases, the G-to-C base exchange was not detectable using high-resolution melting analysis. A Tm difference of less than 0.4°C for 412 bp is insufficient for accurate detection of SNP class 3 mutations (13, 21).
All other HRM profiles detected had specific HRM-STs. S. Newport had five, S. Paratyphi B var. d-tartrate+ had three, and S. Saintpaul, S. Thompson, S. Senftenberg, S. Montevideo, S. Kedougou, S. Derby, and S. Livingstone each had two different but unique melting-curve profiles. Thus, this triplex assay even allows the subtyping of S. Dublin, S. Newport, S. Paratyphi B var. d-tartrate+, S. Saintpaul, S. Montevideo, S. Thompson, S. Senftenberg, S. Kedougou, S. Derby, and S. Livingstone.
In-depth molecular analysis of S. Enteritidis, S. Dublin, S. Montevideo, S. Saintpaul, S. Newport, and S. Paratyphi B var. d-tartrate+ using multilocus sequence typing showed a good correlation between HRM profiles and MLSTs. However, MLST discriminated better within serotypes than HRM profiling.
In conclusion, this triplex HRM assay, in combination with the amplification of the S. Enteritidis difference I region, allows serotype identification of the 15 most frequent human, as well as nonhuman (including isolates from veterinary, food, and environmental samples), Salmonella serotypes found in the years 2008 and 2009 in Austria: S. Enteritidis, S. Typhimurium biphasic variant, S. Newport, S. Infantis, S. Thompson, S. Hadar, S. Abony, S. Tennessee, S. Typhimurium monophasic variant, S. Senftenberg, S. Agona, S. Dublin, S. Montevideo, Salmonella enterica serotype Virchow, and S. Paratyphi B var. d-tartrate+ (9, 10). Within a pool of 46 different Salmonella serotypes, 39 serotypes could be clearly characterized by this triplex HRM assay (Table 1) when applied in combination with the amplification of a region specific for S. Enteritidis. Thus, HRM analysis has the potential to complement classical serotyping of Salmonella isolates due to its discriminatory power and simplicity.
Footnotes
Published ahead of print 17 February 2012
REFERENCES
- 1. Achtman M. 2008. Evolution, population structure, and phylogeography of genetically monomorphic bacterial pathogens. Annu. Rev. Microbiol. 62:53–70 [DOI] [PubMed] [Google Scholar]
- 2. Agron PG, et al. 2001. Identification by subtractive hybridization of sequences specific for Salmonella enterica serovar Enteritidis. Appl. Environ. Microbiol. 67:4984–4991 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Baker S, Hanage WP, Holt KE. 2010. Navigating the future of bacterial molecular epidemiology. Curr. Opin. Microbiol. 13:640–645 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Brenner FW, Villar RG, Angulo FJ, Tauxe R, Swaminathan B. 2000. Salmonella nomenclature. J. Clin. Microbiol. 38:2465–2467 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Herrera-León S, et al. 2004. Multiplex PCR for distinguishing the most common phase-1 flagellar antigens of Salmonella spp. J. Clin. Microbiol. 42:2581–2586 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Holt KE, et al. 2010. High-throughput bacterial SNP typing identifies distinct clusters of Salmonella Typhi causing typhoid in Nepalese children. BMC Infect. Dis. 10:144. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Hu H, Lan R, Reeves PR. 2002. Fluorescent amplified fragment length polymorphism analysis of Salmonella enterica serovar Typhimurium reveals phage-type-specific markers and potential for microarray typing. J. Clin. Microbiol. 40:3406–3415 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Kidgell C, et al. 2002. Salmonella typhi, the causative agent of typhoid fever, is approximately 50,000 years old. Infect. Genet. Evol. 2:39–45 [DOI] [PubMed] [Google Scholar]
- 9. Kornschober C, Orendi U. 2009. Nationale Referenzzentrale für Salmonellen Jahresbericht 2008. http://www.bmg.gv.at/cms/home/attachments/4/6/2/CH1186/CMS1237532383420/jb_salmonellen_2008.pdf
- 10. Kornschober C, Orendi U. 2010. Nationale Referenzzentrale für Salmonellen Jahresbericht 2009. Newsletter public health. http://www.bmg.gv.at/cms/home/attachments/0/6/6/CH1187/CMS1268584751914/jb_salmonellen_2009.pdf
- 11. Kruy SL, van Cuyck H, Koeck JL. 2011. Multilocus variable number tandem repeat analysis for Salmonella enterica subspecies. Eur. J. Clin. Microbiol. Infect. Dis. 30:465–473 [DOI] [PubMed] [Google Scholar]
- 12. Larkin MA, et al. 2007. Clustal W and Clustal X version 2.0. Bioinformatics 23:2947–2948 [DOI] [PubMed] [Google Scholar]
- 13. Liew M, et al. 2004. Genotyping of single-nucleotide polymorphisms by high-resolution melting of small amplicons. Clin. Chem. 50:1156–1164 [DOI] [PubMed] [Google Scholar]
- 14. Lindstedt B-A, Heir E, Vardund T, Kapperud G. 2000. Fluorescent amplified-fragment length polymorphism genotyping of Salmonella enterica subsp. enterica serovars and comparison with pulsed-field gel electrophoresis typing. J. Clin. Microbiol. 38:1623–1627 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Litrup E, et al. 2010. DNA microarray analysis of Salmonella serotype Typhimurium strains causing different symptoms of disease. BMC Microbiol. 10:96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Maiden MC, et al. 1998. Multilocus sequence typing: a portable approach to the identification of clones within populations of pathogenic microorganisms. Proc. Natl. Acad. Sci. U. S. A. 95:3140–3145 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. McCarthy N, et al. 2009. Sensitive and rapid molecular detection assays for Salmonella enterica serovars Typhimurium and Heidelberg. J. Food Prot. 72:2350–2357 [DOI] [PubMed] [Google Scholar]
- 18. Muñoz N, Diaz-Osorio M, Moreno J, Sánchez-Jiménez M, Cardona-Castro N. 2010. Development and evaluation of a multiplex real-time polymerase chain reaction procedure to clinically type prevalent Salmonella enterica serovars. J. Mol. Diagn. 12:220–225 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Nair S, Lin TK, Pang T, Altwegg M. 2002. Characterization of Salmonella serovars by PCR-single-strand conformation polymorphism analysis. J. Clin. Microbiol. 40:2346–2351 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Perera K, Murray A. 2008. Development of a PCR assay for the identification of Salmonella enterica serovar Brandenburg. J. Med. Microbiol. 57:1223–1227 [DOI] [PubMed] [Google Scholar]
- 21. Pietzka AT, et al. 2009. Rapid identification of multidrug-resistant Mycobacterium tuberculosis isolates by rpoB gene scanning using high-resolution melting curve PCR analysis. J. Antimicrob. Chemother. 63:1121–1127 [DOI] [PubMed] [Google Scholar]
- 22. Pietzka AT, Stöger A, Huhulescu S, Allerberger F, Ruppitsch W. 2011. Gene scanning of an internalin B gene fragment using high-resolution melting curve analysis as a tool for rapid typing of Listeria monocytogenes. J. Mol. Diagn. 13:57–63 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Pietzka A, et al. 2008. Amplified fragment length polymorphism of diverse Salmonella enterica serovars for serotype differentiation and identification of serotype specific genetic markers. Infection 36(Suppl. 1):72–73 [Google Scholar]
- 24. Schouls LM, et al. 2009. Multiple-locus variable number tandem repeat analysis of Staphylococcus aureus: comparison with pulsed-field gel electrophoresis and spa-typing. PLoS One 4:e5082. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Sukhnanand S, et al. 2005. DNA sequence-based subtyping and evolutionary analysis of selected Salmonella enterica serotypes. J. Clin. Microbiol. 43:3688–3698 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Tankouo-Sandjong B, et al. 2007. MLST-v, multilocus sequence typing based on virulence genes, for molecular typing of Salmonella enterica subsp. enterica serovars. J. Microbiol. Methods 69:23–36 [DOI] [PubMed] [Google Scholar]
- 27. Venter JC, et al. 2001. The sequence of the human genome. Science 291:1304–1351 [DOI] [PubMed] [Google Scholar]
- 28. WHO 2005. Drug-resistant Salmonella. Fact sheet 139. WHO, Geneva, Switzerland [Google Scholar]
- 29. Wittwer CT, Reed GH, Gundry CN, Vandersteen JG, Pryor RJ. 2003. High-resolution genotyping by amplicon melting analysis using LCGreen. Clin. Chem. 49:853–860 [DOI] [PubMed] [Google Scholar]