Field-deployable, rapid diagnostic testing of saliva samples for SARS-CoV-2.

Abstract Rapid, scalable, point-of-need, COVID-19 diagnostic testing is necessary to safely re-open economies and prevent future outbreaks. We developed an assay that detects single copies of SARS-CoV-2 virus directly from saliva and swab samples in 30 min using a simple, one-step protocol that utilizes only a heat block and microcentrifuge tube prefilled with a mixture containing the necessary reagents and has a sensitivity and specificity of 97% and 100%, respectively.

modifications were necessary to increase the sensitivity of the assay in saliva by >1000-fold. We term the new assay High-Performance LAMP (HP-LAMP).
We first designed eight sets of six primers targeting regions across the full length of the SARS-CoV-2 genome ( Figure 1B). Typically, primers for LAMP are designed to target GC-rich regions of the viral RNA because GC-rich regions bind more tightly to primers. However, in SARS-CoV-2, these regions are found towards the 5′ and 3′ ends of the viral RNA. Because salivary nucleases degrade viral RNA from the 3′ and 5′ ends, we reasoned that the central portion of the virus would be better protected and, therefore, designed our primers to target that region. In the case of SARS-CoV-2, the central region is GC-poor (AT-rich), making it difficult to select primer candidates with optimal annealing temperatures when following standard parameters for primer design. Therefore, we designed the primers to permit large primer-mediated loop-structures while ensuring that the primers did not form stable secondary structures or self-dimerize. We also aligned the known SARS-CoV-2 genomic sequence with those of six other human coronaviruses (SARS-CoV, MERS-CoV, HCoV-HKU-1, HCoV-NL63, HCoV-OC43 and HCoV-229E) to ensure no cross-reactivity. To determine which primer set was most sensitive and specific to SARS-CoV-2, we tested the eight primer sets that we designed, along with previously published primer sets 9, 10 , using serial dilutions of 500 to 0.5 copies of SARS-CoV-2 RNA standard spiked into a 25 μ L RT-LAMP reaction ( Figure 1C). Primer set V5 detected 10 0 to 10 -1 level viral RNA in water, representing a 10-to 100-fold improvement in sensitivity and equivalent specificity compared with previously published primer sets. Additionally, primer set V5 targeted 3640 out of 3672 (>99%) complete SARS-CoV-2 genomes archived on the NCBI Virus database with no mismatches (as of June 9, 2020), and was, therefore, selected for use in our assay.
Even with the improved primer sets, the RT-LAMP reaction was still not sufficiently sensitive to detect fewer than 200 viral copies/μL in saliva, which is far higher than the < 2 viral copies/μL limit considered All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted June 16, 2020. . https://doi.org/10.1101/2020.06.13.20129841 doi: medRxiv preprint necessary for testing clinical samples ( Figure 1D). In order to achieve the necessary > 100-fold improvement in sensitivity while maintaining a 100% specificity, we systematically modified the RT-LAMP reaction conditions to improve performance. We found that sensitivity and specificity of the assay could be markedly improved by adding carrier DNA, carrier RNA, and RNase inhibitors, as well as by increasing the reaction volume (Supple. Figure 1A-1E). Because our assay was so sensitive, there was a risk that carry-over product from prior samples could cross contaminate a new sample and lead to falsepositive results. To solve this problem, we added dUTP to our reaction mixture in order for it to be incorporated into the HP-LAMP product. We also added Antarctic Thermolabile uracil-DNA Nglycosylase (UDG), which degrades any dUTP-containing product from prior reactions but is itself inactivated at temperatures below our running conditions, to the final HP-LAMP reaction mixture (Supple. Method).
To determine the analytical limit of detection (LoD) of HP-LAMP, 0.25-200 copies of SARS-CoV-2 RNA per μ L of saliva were tested using both RT-LAMP and HP-LAMP ( Figure 1D). While RT-LAMP was unable to detect 200 or fewer copies of viral RNA per μ L of saliva, HP-LAMP was able to consistently detect 1 copy of viral RNA per μ L of saliva (10/10). We, therefore, set the clinical LoD to double this amount (2-fold LoD), or 2 copies of viral RNA per μ L of saliva. Since viral transport medium (VTM) is less inhibitory to RT-LAMP than saliva, HP-LAMP was able to detect 10 0 to 10 -1 level viral RNA spike-in in VTM, making it a promising versatile detection method for saliva and swab samples (Supple. Figure 1F).
We then assessed the sensitivity and specificity of our reaction. We tested thirty contrived positive samples consisting of saliva from COVID-19 negative individuals with 2 copies of SARS-CoV-2 per μ L of saliva spiked-in, and thirty negative samples consisting of saliva samples from COVID-19 negative individuals without viral RNA spiked-in. Of the positive samples, 29/30 were detected and 30/30 of the negative samples remained negative for a sensitivity of 97% and a specificity of 100% ( Figure 1E). All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted June 16, 2020. . https://doi.org/10.1101/2020.06.13.20129841 doi: medRxiv preprint To determine if there was cross reactivity of our assay with other known respiratory viruses, we tested 19 known respiratory viruses and bacteria spiked-in to saliva from healthy individuals as well as a positive control consisting of 5 copies of SARS-CoV-2 spike-in per μ L of saliva. No cross reactivity with other respiratory pathogens was noted, while a positive signal was obtained from the SARS-CoV-2 spike-in sample ( Figure 1F).
We next determined the ability of our assay to accurately detect SARS-CoV-2 directly from patient saliva samples. This portion of our study was reviewed and approved by the Columbia University IRB (#AAAS9893) and all methods were carried out in accordance with relevant guidelines and regulations.
All study subjects signed informed consent prior to participating. From 04/29/2020 to 06/1/2020 we prospectively collected saliva samples at the time when patients presented to Columbia University Irving Medical Center's COVID-19 testing tent and cough/fever clinic for clinical SARS-CoV-2 testing via nasopharyngeal swab. Nasopharyngeal swab samples were tested on the RT-PCR-based Roche Cobas 6800 system following routine laboratory protocols according to the manufacturer's recommendations.  Table). In addition, we extracted RNA from the selected saliva samples and tested using RT-PCR following New York SARS-CoV-2 Real-time Reverse Transcriptase (RT)-PCR Diagnostic All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted June 16, 2020.  Table).
In summary, we developed HP-LAMP, which enables rapid detection of SARS-CoV-2 directly from saliva in 30 min using a simple one-step protocol with a LoD of 2 viral copies per μ L of saliva and a sensitivity and specificity of 97% and 100%, respectively. Performing the assay requires only a heat block and a 1.5 mL microcentrifuge tube prefilled with a mixture containing the necessary enzymes, primers, buffers and reagents to simultaneously perform reverse transcription and amplification of the SARS-CoV-2 viral RNA, while blocking the naturally occurring inhibitors and nucleases found in saliva. This simple method can be easily scaled and deployed to centralized laboratories, points-of-care and field locations where testing is greatly needed.   (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint this version posted June 16, 2020.

E.
All rights reserved. No reuse allowed without permission.
(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.