Development and validation of a one-step reverse transcription loop-mediated isothermal amplification (RT-LAMP) for rapid detection of ZIKV in patient samples from Brazil

We have previously developed and validated a one-step assay based on reverse transcription loop-mediated isothermal amplification (RT-LAMP) for rapid detection of the Zika virus (ZIKV) from mosquito samples. Patient diagnosis of ZIKV is currently carried out in centralized laboratories using the reverse transcription-quantitative polymerase chain reaction (RT-qPCR), which, while the gold standard molecular method, has several drawbacks for use in remote and low-resource settings, such as high cost and the need of specialized equipment. Point-of-care (POC) diagnostic platforms have the potential to overcome these limitations, especially in low-resource countries where ZIKV is endemic. With this in mind, here we optimized and validated our RT-LAMP assay for rapid detection of ZIKV from patient samples. We found that the assay detected ZIKV from diverse sample types (serum, urine, saliva, and semen) in as little as 20 min, without RNA extraction. The RT-LAMP assay was highly specific and up to 100 times more sensitive than RT-qPCR. We then validated the assay using 100 patient serum samples collected from suspected cases of arbovirus infection in the state of Pernambuco, which was at the epicenter of the last Zika epidemic. Analysis of the results, in comparison to RT-qPCR, found that the ZIKV RT-LAMP assay provided sensitivity of 100%, specificity of 93.75%, and an overall accuracy of 95.00%. Taken together, the RT-LAMP assay provides a straightforward and inexpensive alternative for the diagnosis of ZIKV from patients and has the potential to increase diagnostic capacity in ZIKV-affected areas, particularly in low and middle-income countries.

www.nature.com/scientificreports/ achieve a low viral load (1 × 10 3 PFU/mL). Samples were infected for 1 h at 37 °C and then directly assayed by RT-LAMP assay without RNA extraction using the same volume of human biofluids. ZIKV was detected in all spiked samples, regardless of the viral dose. In semen samples, the high and low viral load gave similar results as detected by naked eye and UV irradiation, but the low viral load spike gave better amplification then the high viral load as detected by agarose gel analysis. As expected, non-template control (NTC) samples (water) and negative control (human biological sample uninfected) tested negative ( Fig. 2A-I). RT-LAMP results were compared to RT-qPCR, through which the cycle quantification (Cq) values 46 were (13.6; 13.8; 13.0; 13.1) and (24.7; 24.6; 24.3; 24.5), for high viral and low viral load in urine, serum, saliva and semen, respectively.  Table 1). RT-LAMP specifically detected ZIKV as determined by naked eye analysis, visual observation under UV light or agarose gel electrophoresis (Fig. 3). These results were confirmed by RT-qPCR, in which the Cq value for the ZIKV sample was 12.4 (Fig. 3). A common problem with highly sensitive RT-LAMP assays is cross-contamination. To prevent this, we added 1 μL of 1:10 dilution of SYBR Green I dye diluted in RNase-free water to the center of the tube lid caps before the reaction and mixed afterwards 47 . This reduces the potential for the introduction of contamination and in work performed here, no contamination was seen observed using the one step, closed tubes RT-LAMP strategy (data not shown).  www.nature.com/scientificreports/ Analytical sensitivity of ZIKV RT-LAMP assay. To evaluate the analytical sensitivity (limit of detection-LOD) of the assay, RT-LAMP was performed in human serum spiked with varying concentrations (10 5 PFU to 10 -7 PFU) of ZIKV. Spiked samples were directly assayed by RT-LAMP without RNA extraction. RT-LAMP was able to detect a broad range of virus concentration (from 10 5 to 10 −6 PFU). Then, viral RNA of same dilutions were extracted and assayed by the gold standard RT-qPCR test. The analytical sensitivity of RT-qPCR was only observed down to 10 1 PFU with a Cq value of 34.2 (Fig. 4). The experiments were independently repeated 10 times to allow probit regression analysis to accurately determine the limit of detection of RT-LAMP. The limit of detection of RT-LAMP at 95% probability was − 1.07 log 10 PFU of ZIKV with confidence interval from − 1.93 to 0.49 (Table 2 and Fig. S5), which is 100-fold more sensitive than RT-qPCR for ZIKV. Similar analytical sensitivity results were also obtained in urine, saliva and semen (data not shown).
Diagnostic performance and cost of the RT-LAMP assay for detection of ZIKV in patient samples. We next validated the RT-LAMP assay using clinical samples obtained from patients with suspected arbovirus infection. A total of 100 serum samples double blinded (20 positive and 80 negative for ZIKV by RT-qPCR) (Table S1) were used in the experiment. The Cq value in these samples ranged from 21.0 to > 40.0 and samples with Cq values of ≤ 38.0 in triplicate wells were considered positive for ZIKV by RT-qPCR. The RT-LAMP assay detected ZIKV in 25 samples, including five samples which had been determined negative by gold standard method, whereas 75 were deemed negative by ZIKV RT-LAMP (Fig. 5).
The diagnostic performance of this assay for detection of ZIKV was inferred by statistical analysis of several parameters using RT-qPCR assay as reference test. The RT-LAMP assay had a clinical sensitivity of 100% (95% CI 83.16% to 100.00%) and clinical specificity of 93.75% (95% CI 86.01% to 97.94%). The overall ZIKV prevalence in the samples was 20.00% (95% CI 12.67% to 29.18%). The positive predictive value of this assay was 80.00% (95% CI 63.13% to 90.33%) and negative predictive value was 100%. The overall accuracy of the ZIKV RT-LAMP assay was 95.00% (95% CI 88.72% to 98.36%) ( Table 3), highlighting the diagnostic positive features of the RT-LAMP assay for rapid detection of ZIKV in patient samples.
To confirm the identity of positive samples by RT-LAMP for ZIKV detection in human samples, we sequenced a positive sample of human serum by the Sanger method. Analyzes obtained from the sequenced amplicons and BLAST analysis demonstrated that ZIKV RT-LAMP amplicons match 100% with virus circulating in Brazil (Fig. 6), proving the specificity of the RT-LAMP for detection only ZIKV. Together, these results suggested that RT-LAMP protocol described here is highly specificity for detection of ZIKV.
We then compared the costs for consumables required RT-LAMP (Table S2) with RT-qPCR (Table S3) considering the prices (in US$) for reagent acquisition in Brazil at the time of the study. The estimated reagent cost for our RT-LAMP assay was US$ 0.32 as opposed to US$ 10.33 for RT-qPCR, which makes the RT-LAMP 33 times cheaper than the gold-standard method. www.nature.com/scientificreports/

Discussion
Given the rapid spread of ZIKV in the Americas and its association with increased incidence of congenital defects in newborns and Guillain-Barré syndrome in infected patients, reliable and fast POC assays for accessible and inexpensive ZIKV diagnosis are urgently needed 44 . In areas with past or active ZIKV circulation such as Brazil, monitoring virus activity and disease incidence is critical to allow timely interventions and avoid massive outbreaks 48 . Moreover, early identification of virus infection directly in the field after patient sample collection is a critical step to containing the spread of ZIKV 49 . Despite its emergence in Brazil in 2015, the country still   www.nature.com/scientificreports/ relies on expensive RT-qPCR testing for diagnosing ZIKV infection. The availability of testing for ZIKV has been even more challenging during the COVID-19 pandemic since most diagnostic supplies has been directed for SARS-CoV-2 detection.
The clinical diagnosis of ZIKV infection based on clinical symptoms alone is very difficult in countries where other arboviruses are endemic, and definitive confirmation is requires laboratory testing. Currently, RT-qPCR is considered the gold standard molecular method for lab-based diagnosis of ZIKV, but the technique has several limitations that preclude its wide use in remote or low-resource settings. Here, we developed a one-step closed tube RT-LAMP assay for ZIKV diagnosis that can be deployed in low-resource settings. The assay is rapid, specific, affordable, and allows straightforward detection of ZIKV in human samples, including serum, urine, saliva and semen even in the absence of RNA extraction or sample pre-treatment.
Serological assays are another alternative for laboratory ZIKV diagnosis, but lack specificity due to antibody cross-reactivity with other flaviviruses, especially DENV 7,15,50,51 . Here, our ZIKV RT-LAMP assay showed no cross-reactivity with other arboviruses including CHIKV, YFV and DENV 1-4 (Fig. 3). Additional confirmation of specificity was provided by Sanger sequencing of RT-LAMP amplicons, which confirmed 100% match with ZIKV circulating in Brazil (Fig. 6).
Since the emergence of ZIKV in the Western hemisphere, many LAMP assays have been developed for the diagnosis of ZIKV by research groups across the world, including one-step and two-step procedures 44 . The two-step LAMP protocol requires the addition of a reverse transcriptase (RT) enzyme together with the DNA polymerase for amplification and detection of the viral genome. In addition, the two-step protocol needed longer times, requires additional samples and reagents handling, which increases the probability pipetting errors and contamination 44,45 . In order to overcome these drawbacks of two-step protocol, we used the Bst 3.0 DNA polymerase 3.0 WarmStart due to the fact that it is an enzyme that allows the assay to be performed in a one-step protocol without RNA extraction or pre-treatment of the sample. This is possible because Bst 3.0 enzyme has reverse transcriptase and DNA polymerase activities at fixed temperature incubation. Moreover, this enzyme maintains  www.nature.com/scientificreports/ its performance even in the presence of amplification inhibitors and remains stable at room temperature compared to other prototypes of Bst enzyme (wild-type Bst DNA polymerase or Bst 2.0 DNA polymerase) 44,52 . Despite their resistance to the presence of inhibitors, we found that samples with high semen concentration seemed to be have an effect on target amplification (Fig. 2I) which may be due to the presence of inhibitors in semen which are known to affect other DNA polymerases 53 . Another potential limitation of molecular methods is to detect low viral loads in ZIKV-infected that samples usually have low titers after the acute phase of ZIKV infection. This makes it very difficult to confirm ZIKV infection in patient samples, even using the RT-qPCR 54 . Wang et al. developed a diagnostic test based on RT-LAMP for detection of ZIKV in human samples. The limit of detection of this assay was determined to be 0.02 PFU/ mL using ZIKV-spiked samples including serum, urine, and saliva 55 . Tian et al. also reported a assay based on RT-LAMP combined with AC susceptometry in a portable reaction and LOD was determined to be 1 aM (aM: 10 −18 mol per liter) using human serum spiked with synthetically oligonucleotides of ZIKV 52 . These assays were able to detect low ZIKV concentrations, but the experiments and evaluation of the assay performance were conducted only with spiked samples, which not correlate with viral load presents in clinical samples obtained from infected patients.
Moreover, several studies based on the RT-LAMP assay have reported that the analytical sensitivity (limit of detection-LOD) is lower or similar when compared to RT-qPCR 54,55 . In contrast, we developed a RT-LAMP assay for detection of ZIKV in mosquito samples from Brazil, which was able to detect up to 10 -5 PFU. This represented a 10,000-fold greater sensitivity compared with RT-qPCR 45 . In human samples, our assay was 100-fold more sensitive than RT-qPCR. In the latter work, the virus was diluted in serum whereas in the former paper the virus was spiked in mosquito lysate and then serially diluted in water before the RT-LAMP. The reasons for the decreased sensitivity of the assay in human samples compared to mosquito samples are not clear, but molecules present in serum such as IgG, hemoglobin and lactoferrin are known to directly inhibit DNA polymerases 56 . Nevertheless, the results obtained here showed that the analytical sensitivity of the RT-LAMP assay is superior than conventional RT-PCR and RT-qPCR, corroborated with results found in other studies 45,[57][58][59][60] . There are a number of reasons that explain the variation in analytical sensitivity of ZIKV LAMP assays in different studies, including differences in protocols (one and two-step protocol), primers, enzymes and research suppliers, LAMP modes of output, and type of samples 44 .
The performance of LAMP assays for ZIKV detection in patient samples has been evaluated in a few studies and compared to RT-qPCR 54,57,61,62 . In this study, we validated the ZIKV RT-LAMP assay using 100 serum specimens obtained from patients with suspected ZIKV infection cases in the state of Pernambuco, Brazil, which is at the epicenter of the last Zika epidemic. RT-LAMP assay designed in this study was shown to be 100% sensitive www.nature.com/scientificreports/ and 93.75% specific, and overall accuracy of 95.00% as compared to RT-qPCR. These results indicate that our assay has the potential could be used as a diagnostic alternative for ZIKV detection in human samples. Despite the advantages of RT-LAMP for POC diagnostics, the possibility of cross-contamination of products is considered a major challenge with this system 54,57,63-65 . The reason for this possible cross-contamination has not yet been fully elucidated, but the chance of contamination is greatest when the reaction tube caps are opened at the end of the incubation time to add dye for visualization of the result 65 . Moreover, several studies reported that SYBR, when added before incubation along with the other reagents, may inhibit LAMP reactions 66,67 . To overcome this, Antarctic Thermolabile UDG (Uracil DNA Glycosylase) or agar dye capsule have been used 58,68,69 . However, these strategies increase the cost of the reaction and make the technique more laborious. To address these concerns, we performed RT-LAMP reactions adding 1 μL of SYBR Green I to the center of the tube caps before the incubation time based on closed tube. We did not experience with false-positive results during the development and validation process. Furthermore, in order to produce reliable results and reduce false-positive results, we optimized all parameters and conditions of the ZIKV RT-LAMP assay including Mg 2+ concentration, dNTPs concentration, enzyme concentration, primers, time reaction, and temperature reaction.
Here we have standardized and validated a one-step RT-LAMP assay combined with strategy based on closed tube and demonstrated that it is sensitive, specific, and practical for ZIKV detection. The simplicity and high efficiency of RT-LAMP assay to rapidly amplify DNA under isothermal conditions suggests that RT-LAMP could be a potential alternative for detecting ZIKV. Moreover, our strategy is capable of drastically reduce the cross-contamination, when on than previous methods. The cost per reaction was less than $1 USD, which is considerably cheaper than RT-qPCR ($10 USD). Our POC assay is suitable and represents a great alternative for the diagnosis of ZIKV-infected patients and suspects cases in ZIKV endemic countries, especially in remote areas.

Conclusion
The ZIKV RT-LAMP assay described here represents a potential alternative and inexpensive POC tool for the molecular diagnosis and routine screening of ZIKV-infection. The test is a simple, rapid, robust, and represents a fast molecular method for ZIKV detection in patient samples with performance equal to, or even superior to, RT-qPCR. It could also be useful in monitoring the efficacy of ZIKV control programs and to increase the diagnostic capacity of ZIKV-affected, especially for low and middle-income countries. Our POC assay have a great potential for producing rapid and reliable results to assist physicians in decision-making and can bring decentralization of health care through diagnosis in public health services.
RT-LAMP assay. RT-LAMP reactions were carried out in triplicate. First, the reaction condition was optimized to several parameters, including a range of temperatures (59- ). These primers targeted in the envelope protein of the genome and have been previously described 70 . In order to optimize visualization of positive reactions and prevent crossover contamination, 1 μL of SYBR Green I (ThermoFisher Scientific) diluted 1:10 dilution in RNase-free water (Promega) was added to the center of the tube caps before the reaction and mixing afterwards. After addition of sample, reactions were incubated at 72 °C for 40 min in a heat block, and then inactivated at 80 °C for 5 min. The reaction temperature of 72 °C was used for all experiments, except the screening for temperature. All experiments were independently replicated at least three times. To evaluate the performance of the RT-LAMP assay for POC applications, all set-up and execution of reactions were done in a conventional lab bench in an enclosed room using designated pipettes and filter tips. The capture and analysis of images occurred in different rooms to avoid contamination.
After the incubation time of the reactions, the RT-LAMP products were detected using three different methods. In the first method, the amplification products were observed by naked eye under natural light and images were captured using a conventional smartphone camera. The subsequent visual change of color orange to greenish was used to identify positive amplifications (positive sample), while a negative sample remained orange.  To evaluate the analytical sensitivity (limit of detection) of the assay for detection of ZIKV, RT-LAMP was performed using a series of tenfold dilutions of strain PE243 in human biological samples uninfected. Virus concentration in serum samples ranged from 10 5 PFU to 10 -7 PFU. After dilution, samples were directly assayed by RT-LAMP. To compare the results of our RT-LAMP with RT-qPCR, viral RNA was extracted from 140 μL of same dilutions using the QIAamp Viral Mini Kit (QIAGEN, Germany) according the manufacturer's instructions. The RNA was eluted in 60μL of elution buffer and then tested by the reference test (RT-qPCR) currently used for the diagnosis of ZIKV in human samples 7 .

Validation of RT-LAMP for detection of ZIKV in patient samples.
To evaluate the ability of the RT-LAMP assay for detection of ZIKV in clinical samples, 100 samples from patients with suspected arbovirus infection cases in the state of Pernambuco, Brazil were included. Peripheral blood was obtained from patients, who presented symptoms including fever, arthralgia, rash and neurological disorders. Serum samples were separated and stored at − 80 °C until use. The intrinsic diagnostic utility of the ZIKV RT-LAMP assay was determined using several statistical parameters compared to reference method to detect ZIKV reported by Lanciotti 7 . Sequencing of the ZIKV RT-LAMP amplicons. The genetic characterization of the RT-LAMP amplicons from some positive samples from human serum obtained patient infected with ZIKV was executed by the Sanger sequencing method as previously described 45 . RT-LAMP amplicons were directly purified using Gel Band Purification Kit (GE) and illustra GFX PCR DNA according to the manufacturer's protocol and eluted in 30 μL of water. Purified RT-LAMP fragments were directly sequenced using the primer FIP and the BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, USA) according to the manufacturer´s instructions and run on an ABI Prism 3100 Capillary Automatic DNA Analyzer. After sequencing by the Sanger method, the sequences were analyzed using the Bioedit software, v7.0.5 and submitted to NCBI BLAST database (http:// www.ncbi.nlm.nih.gov/blast /Blast .cgi) to find the most closely ZIKV strain.

Consumable reagents (price).
To analyze the costs between the reagents of both techniques, including our RT-LAMP assay and RT-qPCR for the diagnosis of ZIKV. The value per one reaction (sample test) was calculated based on Brazilian reagent prices at the time of the study, and are converted to US $) (Tables S2 and S3). Graphs