The list of strains and their origin is shown in Table 2.

Tryptic soy broth (TSB; Sigma-Aldrich, Merck, St. Louis, MO, USA) was used as the culture medium for overnight propagation of Listeria, Bacillus, Staphylococcus, Pseudomonas and E. coli, while lactic acid bacteria were cultivated for 48 h in MRS broth (Biolab Ltd., Budapest, Hungary) under anaerobic conditions (Anaerocult A; Merck KGaA, Darmstadt, Germany) at 30 °C. An overnight propagation of Campylobacter jejuni was performed using a Columbia blood agar plate containing 5 % sterile defibrinated sheep blood (Sigma-Aldrich, Merck) at 37 °C in a microaerophilic environment (Anaerocult C; Merck KGaA).

DNA was isolated using MasterPure Complete DNA and RNA Purification Kit (Epicentre, Madison, WI, USA) according to the manufacturer’s instructions.

Overnight culture of L. monocytogenes CCM 4699 in TSB (Sigma-Aldrich, Merck) was inoculated into TSB or Fraser enrichment broth (Noack and Co. GmbH, Vienna, Austria) at 10⁴ cell/mL and incubated at 30 °C for 24 h by shaking (190 rpm). A volume of 1 mL culture containing 10⁸ cells was measured into an Eppendorf tube, cells were sedimented by centrifugation at 13 500×g for 4 min at 5 °C (Z216-MK microcentrifuge; HERMLE Benchmark, Sayreville, NJ, USA) and the supernatant was discarded. Cells were resuspended in 100 µL lysis buffers (HotSHOT+Tween and TZ) and treated as described in Table 1.

Cell debris was sedimented by centrifugation of 100 µL lysed cells (13 500×g, 4 min, 5 °C; Z216-MK microcentrifuge; HERMLE Benchmark) and the supernatant was transferred to a new Eppendorf tube. Nucleic acids (NA) were precipitated with equal volume of iso-propanol and sedimented by centrifugation (13 500×g, 4 min, 5 °C). The pellet was washed with 70 % ethanol and dried in SpeedVac Vacuum Concentrator (Thermo Fisher Scientific, Seattle, WA, USA). NA were dissolved in 50 µL TE (Tris-EDTA) buffer, 15 µL RNase solution (10 mg/mL) was added, and the mixture was incubated at 37 °C for 30 min. DNA precipitation was repeated, and the purified DNA was dissolved in 50 µL TE buffer. DNA concentration was measured with NanoDrop microvolume spectrophotometer (Thermo Fisher Scientific).

Cell number of the L. monocytogenes suspensions was determined by plating the cells in tryptic soy agar (TSA; Sigma-Aldrich, Merck). Genome copy number of the extracted DNA was calculated by the Staroscik Copy number calculator of Technology Networks (42):

where X is the number of copies per µL, γ(DNA) is the DNA concentration (ng/µL), N_A is Avogadro’s number (6.0221·10²³ mol^-1), C is the genome size of the organism expressed in number of base pairs (bp), and M_bp is the average mass of a base pair (650 Da).

According to the NCBI database, genome size of L. monocytogenes CCM 4699 (ATCC 19117) strain is 2 951 805 bp (Genome assembly ASM30702v1 (43)).

The total reaction volume was 25 µL containing the components as described by Tang et al. (44). The LAMP reaction consisted of a 60-minute amplification step at 65 °C and a 10-minute enzyme inactivation step at 80 °C.

LAMP products were separated in 1 % agarose gel (SeaKem LE Agarose; Lonza, Basel, Switzerland) by electrophoresis at 120 V for 45 min. DNA bands were visualised by ethidium bromide staining.

PrimerExplorer V5 software (45, 46) was used to design LAMP primer sets specific to the genes encoding the listeriolysin O (hlyA) and internalin A (inlA) of L. monocytogenes. The nucleotide sequences of the hlyA and inlA of L. monocytogenes strains N53-1 (accession number: HE999705; GenBank, NCBI) and NRRL B-33220 (accession number: DQ844405, GenBank, NCBI), respectively (43), provided the basis for the primer design.

For the detection of these genes, LAMP primer sets were designed, and the sensitivity and specificity of the LAMP inner primers (FIP and BIP) with an optimised PCR reaction were calculated. For the PCR detection of these genes, inner PCR primers (F2-B2) were designed based on the nucleotide sequences of the inner LAMP primers FIP and BIP.

E. coli NCAIM B01909 was cultivated in TSB at 30 °C for 24 h with shaking (190 rpm), and a suspension of 10⁸ cell/mL was prepared, which was added at a 1:10 ratio to the TSB and Fraser broth cultures of L. monocytogenes CCM 4699 (10⁸ cell/mL). Decimal dilution series were prepared, lysed by TZ buffer at 65 and 100 °C, respectively, and the LOD was determined.

The effect of the alkaline lysis buffer HotSHOT+Tween and the Triton X-100 and sodium azide-based TZ buffer on the LAMP reaction were determined at their original concentrations (35, 37) as well as at 3× and 10× dilutions. A volume of 1 µL (50 ng) of purified template DNA, extracted from L. monocytogenes CCM 4699, was added to 10 µL of lysis buffers (Table 1) and used as a template in the LAMP mixtures targeting the hlyA gene (44). Separation of the reaction products by gel electrophoresis is shown in Fig. 1, which indicates that the HotSHOT 5×+Tween lysis buffer inhibited the LAMP reaction at the original concentration, but not at the diluted concentrations. In contrast, the HotSHOT 2×+Tween and TZ buffers had no inhibitory effect at all. The same results were obtained using turbidity for endpoint detection. We can conclude that using up to 10 µL of the template in HotSHOT 2×+Tween and TZ lysis buffers per LAMP reaction does not inhibit the amplification.

Next, we investigated whether lysing L. monocytogenes CCM 4699 cells with HotSHOT+Tween and TZ lysis buffers provides amplifiable DNA templates for the LAMP reaction. For the amplificability test, 1 µL of DNA solution (50 ng) extracted from the cell lysates or 1 µL of cell lysates was used as templates for the LAMP reaction. As shown in Table 3, positive results were obtained for every cell lysate and DNA extract using turbidity or gel electrophoresis for endpoint detection. Treating the cells with lysozyme prior to buffer administration did not significantly affect the results. We investigated whether reducing the temperature of the cell lysis in TZ buffer from 100 to 65 °C would result in adequate cell disintegration and DNA extraction. The results in Table 3 show that heat treatment at 65 °C for 15 min is sufficient to yield amplifiable template DNA, which is advantageous for microfluidic application.

The efficiency of DNA extraction using the HotSHOT+Tween and TZ lysis buffers was determined by measuring the amount of DNA extracted from the cell lysates. The genome copy number in the extracted DNA was calculated based on the known genome size of L. monocytogenes CCM 4699. The ratio of the genome copy number per mL of the extracted DNA to the cell number per mL of the initial cell suspension indicated the efficiency of the DNA extraction. The best result was obtained with the TZ lysis buffer, extracting up to 15 % of the genomic DNA. Much smaller amounts of DNA were extracted from the cells using the HotSHOT+Tween buffer, with only around 5 % of the genomic DNA present in the cell lysates. It should be mentioned that the efficiency of DNA extraction in the cell lysates exceeds the calculated values, as significant losses occur during DNA purification and concentration, particularly in small volumes (V≤100 µL).

The main conclusion is that all three lysis buffers result in amplifiable DNA template, with TZ lysis buffer being the most efficient in terms of the amount of extracted DNA. The amplification and sensing reactions of the LAMP assay can be performed in the same tube as the cell lysis when using HotSHOT 2×+Tween and TZ lysis buffers.

For the detection of L. monocytogenes in food, environmental and human samples, conventional and multiplex PCR are the most frequently used molecular techniques, typically targeting virulence genes (3). Most LAMP assays are also based on the detection of L. monocytogenes virulence genes. However, LAMP requires a much more extensive primer design than PCR, as it uses four or six different primers for amplification, which target six different regions of the gene.

In our previous work, we used a LAMP assay for the detection of L. monocytogenes strains based on the amplification of hlyA gene (44). However, in some cases we obtained false positive or negative results, even when the respective negative or positive controls worked properly. Consequently, in the development of a novel LAMP assay, we targeted different regions of the hlyA gene using LAMP primers and also included the inlA virulence gene. To detect these genes, we designed LAMP primer sets and determined the sensitivity and specificity of the LAMP inner primers, FIP and BIP, with an optimised PCR reaction. For this reaction, we designed inner PCR primer pairs F2-B2, based on the nucleotide sequences of the inner LAMP primers. Sensitivity and specificity of the PCR primers were tested against 12 L. monocytogenes and 15 other Listeria strains, respectively, which were reliably identified by phenotypic and molecular methods (Table 2). Results summarised in Table 4 indicate that the HlyA-new and InlA PCR primer pairs generated specific PCR products (in the range of 100–200 bp) for all L. monocytogenes strains, while the PCR reaction was negative for one strain with the HlyA-old primer pair. The HlyA-new and InlA PCR reactions did not yield amplicons of the expected size for any of the 15 non-monocytogenes Listeria strains; however, non-specific PCR products could be visualised by gel electrophoresis in all three PCR reactions in rare cases. PCR products of various sizes and low intensity appeared primarily in L. innocua and L. ivanovii strains, although their visibility was different in the three replicates in several cases. Based on the results, the calculated sensitivity and specificity of both the HlyA-new and InlA PCR reactions were 100 %, while these were only 92 and 87 % for the HlyA-old, respectively. Considering that the InlA PCR reaction generated the fewest non-specific products, we selected the InlA LAMP assay for further experiments.

It is worth noting that the higher primer annealing temperature in the LAMP than in the PCR reaction (65 vs 52 °C) can significantly reduce non-specific primer binding and the generation of false positive LAMP products.

Biosensors for LAMP with turbidity-based real-time monitoring and endpoint detection are very simple and widespread. The drawback of applying these systems in microfluidic devices is that visual assessment is subjective and uncertain, and nephelometric microsensors are not yet commercially available. Colorimetric endpoint detection seems more convenient and cheaper in LAMP, especially in microfluidic sensors.

We tested the robustness of the hydroxynaphthol blue (HNB) colorimetric detection method (22) and found that although the colour change was easily detected, the reaction needed to be optimised and standardised for the conditions of each LAMP reaction. The HNB concentration and LAMP reaction time were particularly important factors (data not shown). We also investigated the applicability of another metal indicator, eriochrome black T (EBT), for colorimetric endpoint detection (23) and compared it with the turbidity visualisation, using the InlA LAMP assay, and including L. monocytogenes and non-monocytogenes Listeria strains in the tests. Results in Fig. 2 show that L. monocytogenes strains were positive with both the EBT (Fig. 2a) and turbidity (Fig. 2b) reactions. For non-monocytogenes Listeria strains, the EBT reaction was clearly negative, while turbidity was only rarely weakly positive in one of the three parallel samples (2/3 and 3/1). Therefore, we applied EBT endpoint detection in the subsequent LAMP experiments.

Performance characteristics of the InlA LAMP assay were evaluated using ten strains of each of L. monocytogenes, non-monocytogenes Listeria and non-Listeria (Table 2). Sensitivity, specificity, and LOD were calculated as the most critical performance characteristics.

Fig. 3 shows the results obtained with L. monocytogenes (Fig. 3a), non-monocytogenes Listeria (Fig. 3b), and non-Listeria bacterial strains (Fig. 3c), presenting typical series of EBT reactions.

For sensitivity (SE), the ratio of the number of L. monocytogenes strains resulting positive in the LAMP test to the number of true positives was determined, while in the case of specificity (SP), the ratio of the number of non-monocytogenes Listeria and non-Listeria bacterial strains that gave negative LAMP results to the true negatives was calculated.

The sensitivity of the InlA LAMP assay was 100 %, as all 10 L. monocytogenes strains gave positive LAMP results in repeated experiments. This finding is consistent with the results obtained in the InlA PCR reaction. The specificity was 92 % for the non-monocytogenes Listeria strains, as 8 % of the LAMP tests were false positives; it was 100 % for the non-Listeria bacteria, and 96 % when all non-L. monocytogenes strains were considered collectively. In some cases, one of the three parallel reactions was positive, likely due to a slight decrease in Mg²⁺ concentration in the reaction mixture. This is because non-specific amplification can occur in the presence of a substantial amount of DNA released from the background microbiota (47, 48). As shown by Francois et al. (19), weak amplification of template DNA can also occur in the absence of the target gene. Furthermore, the negative control sample can be significantly amplified if sample preparation conditions (temperature and time) and the amplification period are not optimal. Notably, the false positive results did not originate from the same Listeria species that produced weak, non-specific PCR products in the InlA PCR reaction.

LOD refers to the smallest amount of template DNA and the corresponding calculated genome copy number, as well as the smallest amount of living cell count that can be detected by the applied LAMP assay (19). The lowest detectable amount of DNA by LAMP (LOD) was determined using decimal dilutions of DNA extracted from L. monocytogenes CCM 4699, with the last positive result at the 10^-5 dilution. Based on this result, the calculated amount of genomic DNA that can be reproducibly detected in the LAMP reaction (V=25 µL) is 500 fg, which corresponds to 157 genome copy numbers. The calculated LOD per unit reaction volume is 20 fg/µL, or 6.3 copy/µL. This is close to the LOD value of LAMP obtained by Nathaniel et al. (29) for the lmo0753 gene of L. monocytogenes and by Francois et al. (19) for Salmonella; however, it is ten times lower than that obtained by Busch et al. (28) for LAMP developed for the mpl gene of L. monocytogenes. Oh et al. (23) reported a similar LOD using LAMP with EBT sensing for the detection of E. coli O157:H7 in a centrifugal microfluidic device.

Traditional detection of L. monocytogenes in food, environmental and human samples by LAMP includes culturing the samples in a selective medium, followed by DNA isolation from the typical colonies for detection. When the number of L. monocytogenes cells is low, a one- or two-step enrichment process is initiated using appropriate culture media, applying a quantity of food samples corresponding to the limit value. These samples are then cultured to assess the presence or absence of L. monocytogenes. This can be achieved by further cultivation in a selective culture medium and confirmation of the typical colonies or identification by culture-based or culture-independent (e.g. genomic) techniques (11, 12, 15).

It is necessary to check whether enrichment media inhibit the reaction at the applied concentrations. Since Fraser selective broth is one of the most frequently used and well-established media for selective enrichment of L. monocytogenes in food (13), its inhibitory effect on LAMP was investigated. According to the results, adding the maximum template volume (V=4 µL) to the reaction mixture did not inhibit the LAMP reaction. This indicates that the enriched culture can be used directly for cell lysis, and then as a template for the LAMP reaction. Since enrichment occurs in two consecutive steps, the inhibitory effect of the food matrix on the LAMP reaction should not be expected, as the first broth (half-Fraser or other) containing the food sample is diluted 100-fold in the second (Fraser) broth.

Since the TZ buffer proved to be the most efficient of the tested lysis buffers, it was selected for the disintegration of L. monocytogenes cells. The applicability of the resulting cell lysates for direct detection of L. monocytogenes using the InlA LAMP assay was then tested. Cultures of L. monocytogenes CCM 4699 grown in either TSB or Fraser enrichment medium were lysed in TZ buffer. The efficiency of cell lysis at 100 and 65 °C for 15 min was compared by the determination of LOD values. Results are summarised in Table 5.

Adding diluted suspensions of TSB cultures lysed by TZ buffer at 100 °C as templates in the InlA LAMP reaction resulted in an LOD similar to that obtained with purified DNA. The LOD was in the range of 100 cells per reaction, indicating that L. monocytogenes could be reliably detected in suspensions of ≥10⁴ cell/mL. Cell lysis with TZ buffer at 65 °C increased the LOD by one order of magnitude. In Fraser cultures, the LOD was one order of magnitude higher than in TSB at both lysis temperatures, which could be attributed to the presence of several inhibitory compounds in the Fraser broth that make the cells tolerant to the lysing effect of the TZ buffer.

The number of L. monocytogenes cells at the end of the enrichment period depends on the initial number of cells in the sample and their ability to grow in the pre-enrichment (half-Fraser) and enrichment (Fraser) broths (49). Therefore, it is expected that L. monocytogenes will need to be detected in positive samples with different cell numbers. To test the effect of cell numbers on the LOD, decimal dilution series of TSB and Fraser cultures (10⁸ cell/mL) were prepared in the corresponding culture medium. The dilution series members were then lysed individually with TZ buffer at 65 and 100 °C, and the lysed samples were used as templates for the InlA LAMP assay. Lysing the dilution series members individually resulted in a one-order-of-magnitude decrease in the LOD most of the time for both the TSB and Fraser enrichment media at 65 and 100 °C (Table 5). It can be concluded that the efficiency with which TZ buffer lyses cells increases at concentrations below 10⁸ cell/mL at both 100 and 65 °C, and that L. monocytogenes can be reliably detected in samples even at low cell number (≤10⁴ cell/mL) after enrichment.

During the selective enrichment of L. monocytogenes, non-Listeria cells present in the food sample may also grow at a low rate. Therefore, we tested whether adding E. coli cells at a ratio of 1:10 to L. monocytogenes TSB and Fraser broth cultures affects the LOD values. Results in Table 5 show that the presence of E. coli cells during cell lysis has practically no influence on the LOD values, confirming high specificity and robustness of the developed InlA LAMP assay.

As the continuation of the present work, validation of the developed InlA LAMP assay according to the ISO 16140-2:2016 (50) is the next step using different food categories artificially inoculated and naturally contaminated with L. monocytogenes.