Robust Saliva-Based RNA Extraction-Free One-Step Nucleic Acid Amplification Test for Mass SARS-CoV-2 Monitoring

Abstract

Early diagnosis with rapid detection of the virus plays a key role in preventing the spread of infection and in treating patients effectively. In order to address the need for a straightforward detection of SARS-CoV-2 infection and assessment of viral spread, we developed rapid, sensitive, extraction-free one-step reverse transcription-quantitative polymerase chain reaction (RT-qPCR) and reverse transcription loop-mediated isothermal amplification (RT-LAMP) tests for detecting SARS-CoV-2 in saliva. We analyzed over 700 matched pairs of saliva and nasopharyngeal swab (NSB) specimens from asymptomatic and symptomatic individuals. Saliva, as either an oral cavity swab or passive drool, was collected in an RNA stabilization buffer. The stabilized saliva specimens were heat-treated and directly analyzed without RNA extraction. The diagnostic sensitivity of saliva-based RT-qPCR was at least 95% in individuals with subclinical infection and outperformed RT-LAMP, which had at least 70% sensitivity when compared to NSBs analyzed with a clinical RT-qPCR test. The diagnostic sensitivity for passive drool saliva was higher than that of oral cavity swab specimens (95% and 87%, respectively). A rapid, sensitive one-step extraction-free RT-qPCR test for detecting SARS-CoV-2 in passive drool saliva is operationally simple and can be easily implemented using existing testing sites, thus allowing high-throughput, rapid, and repeated testing of large populations. Furthermore, saliva testing is adequate to detect individuals in an asymptomatic screening program and can help improve voluntary screening compliance for those individuals averse to various forms of nasal collections.

Keywords: COVID-19; LAMP; RT-qPCR; SARS-CoV-2; oral cavity swab; passive drool; pooling; saliva.

Figure 1

Diagnostic workflow for SARS-CoV-2 testing in saliva specimens. (A) Design of the study. Samples were collected from three groups of individuals: (i) individuals with the risk of being infected, (ii) COVID-19 patients admitted to a hospital and (iii) individuals without symptoms. Three saliva-sampling techniques were compared: passive drool and oral cavity swabs with nylon- and rayon-tipped applicators. NPS-paired saliva specimens were collected from asymptomatic and symptomatic individuals at the COVID-19 testing site. The specimens were analyzed for SARS-CoV-2 with RNA extraction-free one-step RT-LAMP or RT-qPCR. The NPS specimens were analyzed at the official testing laboratory National Laboratory of Health, Environment, and Food with a diagnostic RT-qPCR test, including RNA extraction. (B) Workflow of SARS-CoV-2 testing. Saliva as (i) an oral cavity swab or (ii) passive drool was collected in a tube with an RNA stabilization buffer. The specimen was heat-treated. After cooling, a one-step RT-LAMP or RT-qPCR test was conducted to detect SARS-CoV-2 in the saliva.

Figure 2

Analytical performance of saliva-based one-step RT-LAMP and RT-qPCR. (A) Effect of Syto9 fluorescent dye on RT-LAMP amplification using the N2 primer set and WarmStart Master Mix. The RNA diluted in water (315 copies per reaction) was amplified in 20 µL reaction volume (n = 11). The melting curves of the amplified target in the presence of Syto9 fluorescent dye are depicted. (B,C) LoD of RT-LAMP. (B) RNA (Twist) was amplified with N2 primer set in 20 µL reaction volume using the WarmStar Master Mix. (C) RNA (INSTAND) was amplified with N2 or E1 primer in 15 µL reaction volume using an Isothermal Master Mix. Serial dilutions of standard viral RNA were prepared in heat-treated SARS-CoV-2-negative saliva (n = 7 (B); n = 6 (C)). (D–F) LoD of RT-qPCR. (D,E) RNA (Twist) was amplified using Ultraplex Master Mix in 20 µL reaction volume as a singleplex with N1 primer–probe set (n = 20) (D) or as a multiplex with N1 and N2 primer–probe set (n = 16) (E). (F) RNA (INSTAND) was amplified using AgPath-ID Master Mix in 15 µL reaction volume as a singleplex with N1 and N2 primer–probe set (n = 16). (D–F) Serial dilutions of standard SARS-CoV-2 viral RNA were prepared in heat-treated SARS-CoV-2-negative saliva. A vertical dashed line indicates a lower LoD. A horizontal dotted line indicates the Cq cut-off value.

Figure 3

RNA stabilization buffer improved one-step RT-LAMP amplification. (A) A viral RNA diluted in heat-inactivated saliva from healthy individuals was mixed in 1:1 ratio with either water, sample buffer, sample buffer with RNAsecure or Chelex, or RNA stabilization buffer; heat-treated; and RT-LAMP-amplified (n = 6). (B) The RT-LAMP reaction mix with guanidine chloride was used to amplify RNA (315 copies per reaction) (n = 12). (A,B) RNA was amplified with N2 primers in 20 µL reaction volume using a WarmStart Mix.

Figure 4

Passive drool outperformed oral cavity swabs in SARS-CoV-2 detection. (A) Oral cavity swab applicators with rayon or nylon tips. The swab tip was saturated with saliva and placed in a microcentrifuge tube with RNA stabilization buffer (100 µL). (B) Three drops of saliva (~100 µL) collected as passive drool were mixed with 100 μL of RNA stabilization buffer. (C–E) RT-LAMP and (F–H) RT-qPCR performances on saliva samples collected in parallel with NPS specimens. Saliva samples were self-collected in the RNA stabilization buffer. (C–E) Tp values of RT-LAMP are plotted against the corresponding Cq^av values of RT-qPCR for NPS. Cq^av was calculated as an average of the E1 and RdRT Cq values. A rayon (C) or nylon swab (D) or passive drool (E) saliva sample (2 µL) was added to the reaction mixture for a total volume of 20 μL, and RT-LAMP amplified using WarmStart Mix and N2 primer set. (F–H) Saliva-based RT-qPCR Cq values are plotted against the corresponding Cq^av of RT-qPCR for NPS. A rayon (F) or nylon swab (G) or passive drool (H) saliva sample (2 µL) was added to the reaction mixture for a total volume of 20 μL, and RT-qPCR amplified using Ultraplex Mix and N1 primer–probe set. Numbers above each plot indicate diagnostic specificity and sensitivity (for details, see also Supplementary Materials Table S5).

Figure 5

Diagnostic performance of extraction-free one-step RT-qPCR for SARS-CoV-2 in saliva. (A) Heat map of matched Cq values (black indicates Cq values above 43 and negative results). (B) Pooled Cq values with depicted average value as a red line. (C–F) The Cq value of single RT-qPCR and (C) Cq values of multiplex RT-qPCR (D–F) for passive drool and the corresponding Cq^av RT-qPCR of NPS were compared. (G) RT-qPCR analysis of pooled saliva samples. One microliter of selected saliva samples with Cq ranging from 15 to 40 was combined with five saliva samples negative for SARS-CoV-2 (each 1 µL) and amplified using the S-2/20 protocol. For each sample, the matched original and pooled Cq values were plotted.

Figure 6

Longitudinal analysis of SARS-CoV-2 virus in infected individuals and hospitalized patients. (A,B) Sequential saliva sampling and detection of SARS-CoV-2 infection. Two households are shown. The appearance of COVID-19 symptoms is marked with a vertical dashed line, and a red arrow indicates the day of a positive NPS-RT-qPCR test. The viral infection among individuals is schematically presented at the top. (C,D) Saliva collected from hospitalized COVID-19 patients was analyzed for the SARS-CoV-2 virus. Results grouped as positive and negative for SARS-CoV-2 in saliva are presented in relation to the time between symptoms or hospital admission and saliva tests. (A–D) Saliva samples were analyzed with saliva-based one-step RT-LAMP or RT-qPCR, and combined results were presented.