Classification of Multiple DNA Dyes Based on Inhibition Effects on Real-Time Loop-Mediated Isothermal Amplification (LAMP): Prospect for Point of Care Setting

Than Linh Quyen(1), Tien Anh Ngo(2), Dang Duong Bang(2*), Mogens Madsen1 and Anders Wolff(1*).

• 1: Department of Biotechnology and Biomedicine, Technical University of Denmark (DTU-Bioengineering), Lyngby, Denmark
• 2: Laboratory of Applied Micro and Nanotechnology (LAMINATE), National Food Institute, Technical University of Denmark (DTU-Food), Lyngby, Denmark

LAMP has received great interest and is widely utilized in life sciences for nucleic acid analysis. To monitor a real-time LAMP assay, a fluorescence DNA dye is an indispensable component and therefore the selection of a suitable dye for real-time LAMP is a need. To aid this selection, we investigated the inhibition effects of twenty-three DNA dyes on real-time LAMP. Threshold time (Tt) values of each real-time LAMP were determined and used as an indicator of the inhibition effect. Based on the inhibition effects, the dyes were classified into four groups: (1) non-inhibition effect, (2) medium inhibition effect, (3) high inhibition effect, and (4) very high inhibition effect. The signal to noise ratio (SNR) and the limit of detection (LOD) of the dyes in groups 1, 2, and 3 were further investigated, and possible inhibition mechanisms of the DNA dyes on the real-time LAMP are suggested and discussed. Furthermore, a comparison of SYTO 9 in different LAMP reactions and different systems is presented. Of the 23 dyes tested, SYTO 9, SYTO 82, SYTO 16, SYTO 13, and Miami Yellow were the best dyes with no inhibitory effect, low LOD and high SNR in the real-time LAMP reactions. The present classification of the dyes will simplify the selection of fluorescence dye for real-time LAMP assays in point of care setting.

Over the last 18 years, LAMP has been used widely in the laboratory and in point of care (POC) settings to analyze nucleic acid as well as to detect pathogens (Notomi et al., 2000; Njiru, 2012; Seki et al., 2018). LAMP is faster than PCR (Espy et al., 2006; de Paz et al., 2014) and can be performed under simpler conditions, i.e., at a constant temperature in a range of 60–65°C without the thermal cycling required for PCR (Notomi et al., 2000; Sun et al., 2015). LAMP has advantages such as being rapid, sensitive, specific, simple to perform, less expensive (Njiru, 2012; Velders et al., 2018), and less affected by inhibitors (Stedtfeld et al., 2014; Kosti et al., 2015). LAMP has therefore been considered an attractive method for POC systems (de Paz et al., 2014; Sun et al., 2015). The LAMP reaction produces large amounts of amplified product (dsDNA) and the by-product (magnesium pyrophosphate) that allow visualization of results using a real-time PCR machine (Aoi et al., 2006; Chen and Ge, 2010) or a turbidimeter (Mori et al., 2004), and can even be detected by the naked eye when using appropriate DNA staining techniques (Amin Almasi, 2012; Xie et al., 2014).

Real-time LAMP has been employed since it is not only effective for DNA amplification under isothermal conditions but also simultaneously offers detection, which can be quantitative for e.g., the monitoring of specific gene expression (Mori et al., 2004; Cai et al., 2008; Chen and Ge, 2010; Tourlousse et al., 2012; Cao et al., 2015) or for detection of bacterial concentrations (Nixon et al., 2014). Measurement of turbidity has been used in real-time LAMP (Takano et al., 2019) and a commercial detection device is available (Eiken Chemical, Co., Ltd., Japan). The real-time LAMP turbidity is based on quantification of an increase of magnesium pyrophosphate precipitate as a by-product of the LAMP reaction (Mori et al., 2004). However, the turbidity-based detection method is 10 times less sensitive than that of fluorescence detection (Chen and Ge, 2010). Molecular beacon probes in the F1-B1 region and FRET-like fluorogenic probes in the loop region have been designed and tested, but the amplification efficiency of the assays was reported to be significantly reduced (Cai et al., 2008). The use of DNA dyes is therefore considered an excellent alternative detection method in real-time LAMP since it is simple and it can also enhance the sensitivity of the assay as compared to the turbidity measurement (Cao et al., 2015). In addition, it has been shown that DNA dyes such as SYTO 82 could be used for real-time LAMP using a low-cost charge-coupled device (CCD) based fluorescence imaging system that could be comparable to a commercial real-time PCR instrument (Ahmad et al., 2011). Sun et al. (2015) also reported the use of SYTO 62 for real-time LAMP in an integrated LOC system. Those reports showed the potential of using DNA dyes in POC devices.

Several DNA dyes such as SYBR Green I (Patel et al., 2013), SYTO 9 (Chen and Ge, 2010; Patel et al., 2013), SYTO 82 (Ahmad et al., 2011), YOPRO-1 (Hara-Kudo et al., 2005) and Pico Green (Tian et al., 2012) have been utilized in real-time LAMP. However, none of these studies reported any further investigation on inhibitory effects of the dyes on the real-time LAMP assay. We previously investigated the inhibitory effects of six dyes (SYTO 24, SYTO 62, SYTO 82, SYBR Green I, Eva Green and SYBR Safe) at 5 μM concentration on the real-time LAMP assay (Sun et al., 2015). Seyrig et al. (2015) and Oscorbin et al. (2016) demonstrated the inhibitory effects of fourteen dyes belonging to green dyes (SYBR Green I, Pico Green, Eva Green, Calcein, SYTO 9, SYTO 13, and SYBR Gold) and orange dyes (SYTO 80, SYTO 81, SYTO 82, SYTO 83, SYTO 84, SYTO 85, and SYTOX Orange) on real-time LAMP (Seyrig et al., 2015; Oscorbin et al., 2016). However, these studies were performed on a narrow range of two groups of DNA dyes. Moreover, previous publications on the use of different dyes in real-time LAMP have not included detailed data of fluorescence signal strength or investigation of LAMP reaction efficiencies – a critical aspect of a real-time reaction. Furthermore, it is difficult to predict the behavior of different dyes in real-time LAMP due to the lack of detailed structural information of most of the dyes.

In this study, we thoroughly investigated the inhibitory effects and classified 23 DNA dyes that span a wide range of four groups of dyes with different optical properties in a real-time LAMP assay for detection of Salmonella Enteritidis. The dyes were used at various concentrations. All the dyes were further evaluated for their fluorescence intensity at the end of the LAMP assay at different dye concentrations. Also, the detection limit of the real-time LAMP assay was investigated and compared when using different dyes from different groups. Furthermore, the inhibition effect of SYTO 9 using different LAMP reactions and different systems has been studied. The classification of the dyes into four groups based on inhibition effect does not only suggest which dyes have the greatest potential for the development of real-time LAMP that can be integrated into a POC device, but also provides essential information for selecting DNA dyes for other applications in real-time LAMP detection.

DNA Preparation

Salmonella Enteritidis (S. Enteritidis) strain CCUG-32352 (originating from the University of Gothenburg Culture Collection) was obtained from the culture collection of National Food Institute, Technical University of Denmark (DTU-Food). S. Enteritidis genomic DNA was isolated using DNeasy Blood and Tissue kit (Qiagen, Germany) according to the supplier instructions. The DNA concentration was determined by Nanodrop 1000 (Thermo Scientific, United States) and the DNA preparation was stored at minus 20°C before use.

Primers and Real-Time LAMP Conditions

We designed a Salmonella LAMP primer set (Supplementary Table S1) based on the hilA gene sequence alignment of S. enteritica (NCBI GenBank accession no. CP010.280.1) using the PrimerExplorer V4 (Eiken Chemical Co. Ltd., Tokyo, Japan). A Campylobacter primer set targeting the Cj0414 gene of Campylobacter jejuni was also selected for this study (Supplementary Table S1; Yamazaki et al., 2009).

The LAMP assay was carried out in 10 μl master mixture containing 0.2 μM of F3; 0.2 μM of B3; 1.4 μM of FIP; 1.4 μM of BIP; 0.8 μM of LF; 0.8 μM of LB; 1.4 mM dNTP mix (DNA Technology, Aarhus, Denmark), 0.5 M Betaine (Sigma-Aldrich, Denmark), 4 U of Bst 2.0 DNA polymerase (New England BioLabs), 1× isothermal amplification buffer (comprising 20 mM Tris–HCl, 10 mM (NH4)2SO4, 50 mM KCl, 2 mM MgSO4, and 0.1% Tween® 20, pH 8.8), various concentrations ranging from 0.5 μM to 10 μM of each dye that included SYTO 9, SYTO 13, SYTO 16, SYTO 24, SYTO 60, SYTO 62, SYTO 64, SYTO 82, SYBR Green I, SYBR Gold, YOPRO1, TOTO1, TOTO3, BOBO3, POPO3, and TOPRO3 (Invitrogen, United States); Eva Green (Biotium, United States); Boxto (TATA Biocentre, Sweden); Miami Green, Miami Yellow, and Miami Orange (Kerafast, United States), Pico 488 (Lumiprobe, Germany) and Nuclear Green DCS1 (Abcam, United Kingdom), sterilized water and DNA template.

All the LAMP assays were conducted in a DNA engine thermocycler with a Chromo4 real-time detector (Bio-Rad Laboratories Inc., Hercules, CA, United States) using thin-walled 100 μl white PCR tube strips (Abgene, Surrey, United ingdom), in an Mx3005P (Stratagene, AH diagnostics, Denmark), and in a Piko real-time PCR system (Thermo Fisher Scientific, Finland). The reactions were performed at 65°C for 60 min and the reactions were then terminated by heating to 90°C for 10 min. The fluorescent signal was recorded every minute of amplification. Excitation and emission wavelengths of all the dyes and the fluorescence recording channels used for the dyes in this study are listed in Supplementary Table S2. The excitation and emission of each channel are shown in Supplementary Table S3.

Gel Electrophoretic Analysis

After completion of real-time LAMP reactions, the LAMP amplified products were confirmed by gel electrophoresis. Five μl of each LAMP amplified products were loaded on 2% agarose gel containing 1× of SYBR® Safe DNA Gel Stain (Invitrogen, Life Technologies, United States). The gel electrophoresis was carried out at 25 volts for 30 min and was visualized under UV light using a BioSpectrum® AC imaging system (AH diagnostics, Denmark).

Sensitivity Experiments

The sensitivity of real-time LAMP was performed at an optimal concentration for each of the dyes SYTO 9, SYTO 13, SYTO 16, SYTO 64, Boxto, Miami Yellow, TOPRO 3, SYTO 60, SYTO 62, Eva Green, SYBR Green I and Nuclear Green DCS1. Ten-fold dilutions of S. Enteritidis chromosomal DNA ranging from 2000 to 0.02 pg were prepared. One microliter of each DNA concentration was used as a template for LAMP reactions in the sensitivity test. It has been estimated that 2000 pg of S. Enteritidis genomic DNA is corresponding to 3.86 × 105 copies of S. Enteritidis genome (Chin et al., 2016). The sensitivity experiments were repeated four times. The lowest DNA concentration that gave a positive result in three experiments was considered as the limit of detection (LOD) of the assay.

Data Analysis

Similar to the threshold cycle (Ct) in a real-time PCR (Bustin et al., 2009), the threshold time value (Tt) has been used to determine the real-time LAMP efficiency (Kubota et al., 2011). The Tt was defined as the time required for the fluorescence to reach a threshold value (Ft). Ft was determined using equation 1 (eq. 1), which was based on the starting fluorescence value (Fi) plus the average fluorescence value (ΔF ̄Δ⁢F ̄) and three times of the standard deviation of the fluorescence values (3sΔF) observed in triplicate negative reactions without DNA template (Kubota et al., 2011).

Ft=Fi+ΔF ̄+3sΔF(1)Ft⁢=⁢Fi+Δ⁢F ̄+3⁢sΔ⁢F (1)

Signal-to-noise ratio was determined and used to analyze the fluorescence intensity of the dyes in real time. The SNR was calculated by the fluorescence signal (arbitrary unit) (X) minus the average baseline signal in the first five points (μ) and divided by the standard deviation of the baseline signal at the first five points (σ) as in equation 2 (eq. 2) (Seyrig et al., 2015). The fluorescence signal of a real-time LAMP reaction was analyzed at an optimal concentration of each dye.

SNR=(X−μ)/σ(2)S⁢N⁢R⁢=⁢(X-μ)/σ (2)

To evaluate the speed of the real-time LAMP reactions for rapid detection, we determined the doubling time (DT) for double-stranded DNA assuming exponential amplification (Keremane et al., 2015; Kubota and Jenkins, 2015). The DT was calculated by multiplying the negative log (2) with the slope of threshold time versus log DNA concentration using equation 3 (eq. 3).

DT=−log(2)×slope (3)

Inhibition Effect of DNA Dyes on Real-Time LAMP

The inhibition effect of the dyes on real-time LAMP reaction was investigated using different concentrations of each dye in the real-time LAMP assay. The efficiency of a real-time LAMP was evaluated based on the Tt value, which was used as an indicator of the inhibitory effects because an increase of the inhibitory effects would raise the Tt value. To define the inhibitory effect, the Tt values were plotted against the dye concentrations. The slope of the plot based linear relationship was used for further indicating the degree of the inhibition. An optimal concentration for each dye was defined as the concentration that resulted in the shortest Tt value. In addition, besides the determination of the Tt values, agarose gel electrophoresis was performed to confirm the LAMP amplification in case the dye used did not give a fluorescence signal in the real-time LAMP reaction.

In the real-time LAMP reactions using the 23 dyes, fluorescent signals were observed for 20 of these dyes including SYTO 9, SYTO 13, SYTO 16, SYTO 64, SYTO 82, Boxto, Miami Green, Miami Yellow, Miami Orange, YOPRO 1, SYTO 62, TOPRO 3, SYTO 60, EvaGreen, POPO 3, DCS1, SYBR Green I, BOBO 3, Pico 488, and TOTO 3. The Tt values of these 20 dyes were calculated and plotted against the dye concentrations at an initial DNA template concentration of 2 ng of DNA S. Enteritidis per reaction (Figure 1 and Supplementary Figures S1–S20). In contrast, no fluorescent signals were observed with the remaining 3 dyes (TOTO 1, SYTO 24, and SYBR Gold) in the real-time LAMP reaction (Supplementary Figures S21–S23). According to the results of the real-time LAMP reaction and slopes of linear relationship of all these dyes, the inhibitory effect of these dyes on the real-time LAMP reaction was classified into four different groups: (1) non-inhibition effect, (2) medium inhibition effect, (3) high inhibition effect, and (4) very high inhibition effect (Figure 1 and Table 1).

Read the details here: https://www.frontiersin.org/articles/10.3389/fmicb.2019.02234/full