Evaluation of reverse transcription-loop-mediated isothermal amplification for rapid detection of SARS-CoV-2

The main strategy for response and control of COVID-19 demands the use of rapid, accurate diagnostic tests aimed at the first point of health care. During the emergency, an increase in asymptomatic and symptomatic cases results in a great demand for molecular tests, which is promoting the development and application of rapid diagnostic technologies. In this study, we describe the development and evaluation of RT-LAMP to detect SARS-CoV-2 based on three genes (ORF1ab, M and N genes) in monoplex and triplex format. RT-LAMP assays were compared with the gold standard method RT-qPCR. The triplex format (RdRp, M and N genes) allowed obtaining comparable results with de RT-qPCR (RdRp and E genes), presented a sensitivity of 98.9% and a specificity of 97.9%, opening the opportunity to apply this method to detect SARS-CoV-2 at primary health-care centers.


RT-qPCR.
RT-LAMP in RNA samples from patients with clinically suspected COVID-19. Five RT-LAMP assays were performed with each sample from the patients considered in this study, using four sets of oligonucleotides that amplify a fragment of nsp3 (non-structural protein 3-ORF1a), RdRp (RNA dependent RNA polymerase-ORF1b), M, N genes, and in the triplex format (RdRp, M and N mix). RT-LAMP assays were developed with the WarmStart Colorimetric LAMP 2X Master Mix -M1800 (DNA and RNA) (New England Biolabs, USA) 15 , using a mixture of RT-LAMP oligonucleotides at a final concentration of 1.6 µM FIP and (BIP); 0.2 µMof F3 and B3 primers; 0.4 μM of LF and LB. A reaction mixes with a total volume of 25 μL was prepared: containing 12.5 μl of WarmStart Colorimetric LAMP 2X Master Mix, 2.5 μl of LAMP primer mix, 5 μl of RNase-free molecular grade water (Sigma) and 5 μl of RNA. Positive and negative controls were established according to the manufacturer's recommendations. The reaction was established at 65 °C for 45 min in a Thermo Block (Thermo Scientific Inc, USA). The same procedure was developed with each LAMP sets.

Limit of detection of 2019-nCoV_RdRp positive control RT-LAMP. The limit of detection (LoD)
of the RT-LAMP to detect SARS-CoV-2 virus was carried out in triplicate using the positive control 2019-nCoV_RdRp (ORF1ab) (Integrated DNA Technologies), which contains the envelope gene and a portion of the RNA-dependent RNA polymerase (RdRp) of 2019-nCoV, which has been synthesized at a concentration of 200,000 copies/µL. The predetermined copy numbers of the biochemically synthesized RNA were serially diluted ten-fold from 10 6 copies to 10 -1 copies of the target gene per reaction. RT-qPCR to detect the RdRp gene was performed using the same dilutions of the positive control.

Results
Primer evaluation-RT-LAMP. The oligonucleotides designed in this study were evaluated with the free NCBI Primer-BLAST tool. In this analysis, a set of oligonucleotides that align in the Orf1a region was also included to detect a fragment of the nps3 gene developed by Lamb et al. 16 (Table 1) Limit of detection of 2019-nCoV_RdRp positive control RT-LAMP. The limit of detection of RT-qPCR and RT-LAMP assays for the detection of the RdRp gene were compared. For the RT-qPCR, the standard curve was linear and a correlation coefficient (R = 0.99325), a slope of − 3.199 and an efficiency of 1.05 were obtained. The LoD of the RT-qPCR was one logarithmic unit higher (1000 copies/reaction) compared to the RT-LAMP (Table 2). Positive RT-LAMP reactions were seen visually as yellow, while negative reactions remained pink viewed with the naked eye ( Fig. 2A). The limit of detection was confirmed by a gel electrophoresis of the RT-LAMP products (Fig. 2B).

Limit of detection of clinical samples using RT-LAMP.
The concentration of the evaluated sample was 48.17 ng/µL. Based on the serial dilutions from 10 0 -10 -5 , the limit of detection for the RdRp, N, M, the triplex format (RdRp, M, N genes) and nsp3 genes was up to 10 -5 dilution (Fig. 3).  TGT TGT AGC TTG TCA CAC CGT TCC ACA CAT GAC CAT TTC ACT   RdRp_LoopF  TAA GCA TGT TAG GCA TGG CT   RdRp_LoopB  AGC TAA TGA GTG TGC TCA (RdRp, M, N, nsp3)     www.nature.com/scientificreports/ analysis with the RdRp gene presented a sensitivity of 98.8% and a specificity of 98.8%,the comparison with the N gene presented a sensitivity of 95.3% and a specificity of 98.8%; the comparison with the M gene obtained a sensitivity of 90.6% and a specificity of 99.6%; and the comparison with the nsp3 gene presented a sensitivity of 88.2% and a specificity of 98.4% (Fig. 4A).
The RT-qPCR based on the E gene detected a similar number of positives (n = 83) and negatives (n = 246). In the evaluation of the RT-qPCR of the E gene, in comparison with the four genes evaluated of the RT-LAMP and a mix of genes, it was determined that the amplification of the triplex format (RdRp, M, N, nsp3) presented a sensitivity of 96.4% and a specificity of 95.5%, while the comparative analysis with the RdRp gene obtained a sensitivity of 95.2% and a specificity 96.7%; the evaluation with the N gene presented a sensitivity of 94% and a specificity of 97.6%; while with the M gene obtained a sensitivity of 88% and a specificity of 98%; and the comparative analysis with the nsp3 gene presented a sensitivity of 85.5% and a specificity of 96.7% (Fig. 4B). In the comparative evaluation of triplex RT-LAMP (RdRp, M and N genes) with RT-qPCR based on the amplification of RdRp, N and GAPH genes, no false negatives were observed, reaching a specificity of 100% and a sensitivity of 91.9%, being the positive predictive value of 100% and negative predictive value of 96.6% (Fig. 4B).
In this comparative evaluation, the RT-qPCR directed to the RdRp and E genes for the identification of the presence of SARS-CoV2 virus RNA, the positive result was considered only if a Ct value ≤ 40 was detected by the RT-qPCR (RdRp and E).
From the concordance analysis of RT-qPCR results compared with all the genes of RT-LAMP, it is evidenced that there is a high concordance of 0.9685 between the RT-qPCR (RdRp) and the RT-LAMP (RdRp) ( Table 3)

Discussion
Since the beginning of the pandemic caused by the SARS-CoV-2 virus 17 , it has been observed that there is a gap in the availability of molecular tests for the detection of SARS-CoV-2 based on real-time RT-qPCR. Timely and accurate laboratory diagnosis of suspected COVID-19 patients is critical during the COVID-19 pandemic response 18 . This study demonstrates the potential use of a rapid RT-LAMP-based method for the detection of SARS-CoV-2 with high specificity and sensitivity. RT-LAMP has been selected because it provides a rapid and reliable diagnostic alternative for the detection of infectious agents in settings with limited resources such as dengue virus, Zika, chikungunya, influenza, yellow fever among other viruses [19][20][21] .
Although ORF1ab is the confirmatory target gene with the highest specificity, it is considered less sensitive than other clinicallyapplicable targets [22][23][24] . Although diagnostic assays can be designed in the most conserved region of the viral genome, most of the RT-qPCR and RT-LAMP routinely applied target the ORF1ab, RdRp, S, E and N genes due to their high level of transcription and abundance of expression compared to other SARS-CoV-2 genes 25 , so the design of our oligonucleotides were directed towards these targets.
The high sensitivity and specificity of RT-LAMP obtained in this study is noteworthy, particularly since it is based on the comparison with the reference method RT-qPCR (RdRp and E) and the amplification of multiple genes of RT-LAMP. These findings have great importance, both clinically and epidemiologically due to the high cases of symptomatic and asymptomatic patients of COVID-19 26,27 , which explain the necessity of the development and application of a portable rapid diagnostic test with good sensitivity and specificity to identify these cases.
In this RT-LAMP optimization assays, the LoD was one logarithmic unit lower (1000 copies/reaction) than the RT-qPCR, when using the positive control 2019-nCoV_RdRp (ORF1ab) ( Table 2). This result showed RT-LAMP assays are more sensitive than RT-qPCR. Several studies using the protocol developed by Corman et al. 14 reported RT-qPCR assays could not detect most samples with Ct > 30. This group of samples is usually subject to wrong diagnostic treatment or to no treatment at all, spreading the disease to others 28 . Previous studies indicate that Ct > 30 values showed patients without infective capacity, but there is not a real consensus 29 . Additionally, RT-LAMP assays previously reported reaffirm being more sensitive than RT-qPCR, detecting a smaller number of copies, including samples with Ct values between 31-35. For example, Bharda et al. reported that the use of a set of six primers for RT-LAMP can detect 100 to 1000 copies of SARS-CoV-2 genomic RNA 30 . Similarly, Chan et al. 31 described a RT-LAMP with a LoD of 11.2 RNA copies per reaction using in vitro RNA transcripts, while Yan et al. 32 adapted the ORF1ab gene to develop a RT-LAMP with a limit of detection of 20 copies per reaction. Most of these diagnostic tests have a high level of sensitivity, specificity, and repeatability; however, these mainly lack clinical validation 33 .
In this study, we performing a RT-LAMP using five sets of oligonucleotides, the triplex format (RdRp, M, N) was the one which presented a sensitivity of 98.9% and a specificity of 97.9% when was compared to the RT-qPCR (RdRp gene and E gene), since the ORF1ab gene is very specific and the N gene is very sensitive in addition to the M gene, data very similar to the reported by Yang et al. 34 with a sensitivity of 99%. Likewise, the highly conserved RdRp gene presented a sensitivity and a specificity of 98.8%, very similar to those reported by Kitagawa et al. 35 , who obtained a sensitivity and a specificity of 100% and 97% respectively, as well as the sensitivity and specificity for the N and M genes: 95.3% and 98.8%; 90.6% and 99.6% respectively, data very similar to that reported by Jiang et al. 36 with a sensitivity and specificity of 91.4% and 99.5%, respectively, while the sensitivity and specificity for the nsp3 gene was 88.2% and 98.4% respectively, data very similar to the reported by Hu et al. 37 38 and Dao Thi et al. 39 reported sensitivities of 97.6%, and 97.5% respectively, which were comparable to results of this study. However, the study of Escalante-Maldonado et al. 40 obtained lower sensitivity (87.4%) than this study. In contrast, RT-LAMP triplex assays were reported based on different sets of primers, having advantages and disadvantages when they are compared to this study. Yamazaki et al. 41 , using primers for ORF1ab, S and ORF7a regions, obtaining a sensitivity of 82.6% and specificity of 100% and showing the poorest sensibility compared to nasal-pharyngeal samples. Another, study reported by Sherrill-Mix et al. 42 using primers E1 (E protein), As1e (ORF1a), and Penn (ORF1ab), getting a LoD greater than 100 copies/ reaction that was comparable to this study results. Juscamayta-Lopez et al. 43 included the RdRp, M and N genes for their triplex RT-LAMP with sensitivity and specificity values of 100.0% and 98.6% that were comparable to this study. However, the method needs to use a turbidimeter equipment to measure the color change during the assay which is difficult to apply in low-resource settings.
An important limitation for RT-LAMP assay is the RNA extraction and purification, which is essential during the use of nasal-pharyngeal or saliva samples without RNA purification. Some studies overcome these difficulties by applying different strategies such as Wei et al. 44 that designed a method without the need for prior RNA extraction, processing samples inside transport media in 30 min and obtaining a high sensitivity. However, the sensitivity is strongly affected by processing samples after 30 min. In this study, we performed RT-LAMP assays including samples stored to − 80 °C from a few days to several months, and the sensitivity was not affected.
In conclusion, the RT-LAMP using the triplex format (RdRp, M and N genes) was the one which obtained best numbers of sensitivity (98.9%) and specificity (97.7%), obtaining comparable results with RT-qPCR to detect SARS-CoV-2 RNA virus in nasal and pharyngeal swabs, opening the opportunity to perform this method from patients with suspected COVID-19 at primary health-care centers.