Evaluation of different target genes for the detection of Salmonella sp. by loop‐mediated isothermal amplification

The loop‐mediated isothermal amplification (LAMP) technique was used to investigate six salmonella‐specific sequences for their suitability to serve as targets for the pathogen identification. Sequences selected for designing LAMP primers were genes invA, bcfD, phoP, siiA, gene62181533 and a region within the ttrRSBCA locus. Primers including single nucleotide polymorphisms were configured as degenerate primers. Specificity of the designed primer sets was determined by means of 46 salmonella and 32 other food‐ and waterborne bacterial reference species and strains. Primers targeting the ttrRSBCA locus showed 100 % inclusivity of target and exclusivity of other test species and strains. Other primer sets revealed deficiencies, especially regarding Salmonella enterica subsp. II–IV and Salmonella bongori. Additionally, primers targeting the siiA gene failed to detect S. enterica subsp. enterica serotypes Newport and Stanley, whereas bcfD primers did not amplify DNA of S. enterica subsp. enterica serotype Schleissheim. TtrRSBCA primers, providing short detection times and constant melting temperatures of amplification products, achieved best overall performance.


Introduction
Salmonellosis represents one of the major foodborne zoonoses of great public concern worldwide. The World Health Organization reported that particularly non-typhoidal Salmonella enterica was involved in a total of 230 000 deaths caused by foodborne diarrhoeal disease agents in 2010 (World Health Organization 2015). As a contribution to food safety improvement, different loopmediated isothermal amplification (LAMP) assays for the detection of Salmonella sp. have been developed. In 2000, a Japanese research group introduced LAMP as a novel technique for identifying nucleic acids (Notomi et al. 2000). It soon became seen as a robust, sensitive and rapid alternative to the polymerase chain reaction (PCR) and was widely applied within the scope of diagnostic and scientific issues (Mori and Notomi 2009). The first LAMP assay for the detection of Salmonella sp. was developed in 2005. Primer design was based on host cell invasion gene invA (Hara-Kudo et al. 2005). Most of the subsequently established LAMP assays followed this example and focused on further validation of invA-based detection, various equipment for visualization of LAMP products and general method improvement by merging LAMP with additional technologies and providing multiplex approaches (Yang et al. 2018). One drawback of using invA as a target gene is that some serotypes including S. enterica subsp. enterica serotypes Kentucky, Senftenberg and Litchfield are known to occasionally lack this sequence and therefore cannot be detected using nucleic acid amplification techniques (Ginocchio et al. 1997;Turki et al. 2012). To supply optimal properties regarding inclusivity of Salmonella sp. and exclusivity of non-salmonella strains, several species-specific alternative target genes were tested in new LAMP assay development (see below). Since those corresponding sequences were not investigated in detail in subsequent studies, using them as targets for LAMP-based salmonella detection lacks verification regarding specificity. Therefore, in the present study, invA (Hara-Kudo et al. 2005) and four alternative target genes, including bcfD (Zhuang et al. 2014), phoP (Li et al. 2009), siiA (Zhao et al. 2017) and gene62181533 (Li et al. 2016), were tested for their suitability regarding LAMP detection. Additionally, the ttrRSBCA locus, which has so far only been published as a target in connection with a real-time PCR protocol (Malorny et al. 2004), was investigated. The aim of these analyses was to provide further information about target gene specificity and to point out suitable and more specific alternatives to the invA gene.

Results and Discussion
In the present study, six LAMP primer sets were designed, targeting a region within the ttrRSBCA (ttr) locus and genes invA, bcfD, phoP, siiA and gene62181533 (g62). All sequences have been reported to be exclusively present in Salmonella sp. in the literature. A total of 78 bacterial strains were used to determine specificity and to evaluate overall performance of the different primer sets. LAMP primers were configured as degenerate primers to ensure inclusivity of salmonella strains in case of single nucleotide polymorphisms within the target genes. Inclusivity was 97Á8% for invA, 93Á5% for bcfD, 97Á8% for phoP, 87Á0% for siiA, 95Á7% for g62 and 100% for ttr primers. Except for g62 primers that occasionally showed unspecific reactions after 40 min, all primer sets revealed exclusivity of 100% towards all tested non-salmonella strains (Table 1). InvA (x = 10Á28 min) and ttr primers (x = 10Á60 min) provided the shortest detection times (Fig 1; for supporting information also see Fig S1). Melting temperatures showed the greatest range (R = 1Á8°C) when invA primers were used (Fig 2).
Except for ttr primers, there were deficiencies mainly concerning inclusivity of S. enterica subsp. II-IV and S. bongori for all tested primer sets. Corresponding target genes were not intensively applied and studied for constant occurrence in these rather rarely appearing salmonella species and subspecies. LAMP-associated investigations of siiA, bcfD, g62 and the widely used invA sequences focused on S. enterica subsp. I and IIIa (Hara-Kudo et al. 2005;Zhuang et al. 2014;Li et al. 2016;Zhao et al. 2017). In the present study, DNA of S. enterica subsp. IV could not be amplified by invA primers, while this subspecies was reliably identified during a multicentre validation study for an invA-based salmonella PCR (Malorny et al. 2003). Primers targeting genes phoP, bcfD, siiA and g62 also failed to detect that strain. In a previous LAMP study, phoP was shown to be included in S. enterica subsp. IV (Li et al. 2009), whereas corresponding data for genes bcfD, siiA and g62 are not available. In contrast to the findings of Li et al. (2016), primers for g62 were not able to amplify DNA of S. enterica subsp. IIIa. In previous studies, appearance of gene siiA was proven in S. enterica subsp. II and IIIa, whereas bcfD was additionally shown to be inclusive in S. bongori (Zhuang et al. 2014;Zhao et al. 2017). These results do not comply with the findings in our study. The siiA sequence could not be amplified in one strain of S. enterica subsp. II and S. enterica subsp. IIIa-IIIb. Furthermore, S. enterica subsp. enterica serotypes Newport and Stanley gave negative results. Although salmonella strains of serotype Newport were successfully tested within an siiA-based PCR study, there are no data available for S. enterica subsp. IIIb and S. Stanley strains (Ben Hassena et al. 2015). In contrast to the findings of Zhuang et al. (2014), no amplification occurred when bcfD primers were used for detection of S. bongori. In addition, S. enterica subsp. enterica serotype Schleissheim could not be identified and was not included in previous studies either.
Since invA, phoP, g62, siiA and bcfD primers showed positive results for most of the tested salmonella strains, this implies that the corresponding sequences may not be genetically stable in all Salmonella sp. or were presumably absent in unidentifiable strains. Besides the possibility of completely lacking sequences, point mutations  not specified in detail (Kong et al. 2013). It can be assumed that the expression of these genes is not necessary for the survival of certain Salmonella sp. lineages. Thus, their sequences are likely to be rather susceptible to genetic variation or to be even missing in the salmonella genome. In contrast, the ttr locus is required for tetrathionate respiration in anaerobic environments and is likely to be significant within the life cycle of Salmonella sp. (Hensel et al. 1999). Malorny et al. (2004) concluded that the ttr genes should be genetically stable in all salmonella strains and emphasized the advantages of ttr-based salmonella detection. The assumption that a correlation exists between the constant presence of genes and their importance for survival is sustained by the fact that, except for ttr primers, weaknesses in primer specificity mainly occurred in S. enterica subsp. II-IV and S. bongori. These strains are primarily associated with coldblooded animals and known as opportunistic pathogens with low human virulence. However, they should not be neglected since cold-blooded animals play an increasing role as pet animals or as a food source for human consumption in various countries. Although non-enterica subspecies are less virulent, they can induce extraintestinal infections causing a broad spectrum of serious diseases (Lamas et al. 2018).
It was shown that all primer set-specific LAMP products are characterized by different melting temperatures ( Fig. 2; for supporting information also see Fig S2). Except for invA-based amplification products, no wideranging measurements were apparent. Since melting temperatures depend on sequence content and should not differ more than AE1°C in specific amplification products, divergent measurements indicate unspecific amplification of non-target species. Wide-ranging measurements as shown in invA-based LAMP products can complicate the evaluation of results. When using different sample processing techniques, additional impacts of matrices could lead to increased variability in measured melting temperatures.
In the present study, the ttr locus was identified as most suitable target for reliable detection of all salmonella species and subspecies by LAMP. Ttr primers showed the best overall performance, providing short detection times and constant melting temperatures. Deficiencies in other primer sets involved poor inclusivity of cold-blooded animal-associated strains. This reveals the importance of choosing essential target genes for the reliable nucleic-acid-based detection of Salmonella sp. However, further validation of the established LAMP assays is necessary to evaluate their suitability for salmonella detection in food matrices. Validation should include artificial contamination experiments and investigation of naturally contaminated products (Feldsine et al. 2002).
Close attention should be given to predestined foodstuffs and also to exotic meat sources as suggested by the results of this study.

Primer design
Numerous sequence data used for primer design were obtained from the DNA sequence database GenBank of the National Center for Biotechnology Information (NCBI) (https://www.ncbi.nlm.nih.gov/). For creating the six gene-specific LAMP primer sets, the same GenBank accession numbers were used as in the aforementioned target gene-related publications. Therefore, the primer design for the invA gene was based on S. enterica subsp. enterica serotype Typhimurium (GenBank accession no. M90846), for the bcfD gene on S. enterica subsp. enterica serotype Newport (GenBank accession no. JX026810), for the phoP gene on S. enterica subsp. enterica serotype Paratyphi A ATCC 9150 (GenBank accession no. NC006511), for the siiA gene on S. enterica subsp. enterica serotype Typhimurium str. LT2 (GenBank accession no. NC_003197), for the g62 gene on S. enterica subsp. enterica serotype Choleraesuis str. SC-B67 (GenBank accession no. NC006905) and for the ttr locus on S. enterica subsp. enterica serotype Typhimurium str. LT2 (GenBank accession no. AF282268), whereby the ttr target region was limited to 701 bp including parts of the ttrC and ttrA gene. All primer sequences were configured by the software LAMP Designer (Premier Biosoft, San Francisco, CA) and subsequently submitted to basic local alignment search tool (BLAST) analysis (NCBI) against 53 salmonella whole-genome sequences obtained from GenBank (NCBI). Polymorphisms within the primer sequences were marked in accordance with the nucleotide code of the International Union of Pure and Applied Chemistry (IUPAC) and considered by ordering appropriate degenerate primers via Eurofins Genomics GmbH (Ebersberg, Germany). Primer sequences are shown in Table 2.

Bacterial strains and DNA extraction
A total of 46 salmonella and 32 non-salmonella strains were used to determine the specificity of the LAMP primers (Table 3). Salmonella strains were selected to represent all known S. enterica subspecies including various epidemiologically important serotypes, and S. bongori. Non-salmonella strains were chosen because of the close relation to salmonella or because they are found in the same environment and grow under the same conditions. Generally, isolates were cultured aerobically on

LAMP assay
LAMP reactions were carried out using the real-time fluorometer Genie ® II (OptiGene Ltd, Horsham, UK). Each reaction mixture with a total volume of 25 µl consisted of 15 µl GspSSD isothermal master mix ISO-001 (Opti-Gene), 2Á5 µl primer mix containing 5 pmol F3 and B3 primer, 20 pmol FIP and BIP primer and 10 pmol LoopF and LoopB primer, 2Á5 µl nuclease-free water (Qiagen) and 5 µl DNA template (5 pg). Runs took place at 65°C and lasted 60 min. Melting temperatures of the LAMP products were determined using a ramp rate of 0Á05°C/s within temperatures ranging from 98 to 80°C. In each run, one reaction mixture with 5 pg DNA of S. enterica subsp. enterica serotype Typhimurium DSM 19587 and one reaction mixture containing 5 µl nuclease-free water instead of template DNA served as positive and negative control, respectively.

Data analyses
Raw data analysis was conducted using the software Genie Explorer (OptiGene) and processed using Microsoft Excel 2016 (Microsoft Corporation, Redmond, WA). Positive predictive values were calculated as (number of true positives)/(number of true positives + number of false positives) and negative predictive values were calculated as (number of true negatives)/ (number of true negatives + number of false negatives). The accuracy was calculated as (number of true positives + number of true negatives)/(number of isolates) × 100.

Supporting Information
Additional Supporting Information may be found in the online version of this article: Figure S1. Amplification curves after LAMP reaction with 5 pg DNA of Salmonella Typhimurium DSM 19587 obtained by primer sets targeting different salmonellaspecific genes. Figure S2. Melting curves after LAMP reaction with 5 pg DNA of Salmonella Typhimurium DSM 19587 obtained by primer sets targeting different salmonellaspecific genes.