Assessment of a Smartphone-Based Loop-Mediated Isothermal Amplification Assay for Detection of SARS-CoV-2 and Influenza Viruses

Key Points Question Can loop-mediated isothermal amplification (LAMP)-based methodology coupled with smartphone detection provide an inexpensive, rapid, sensitive, and reliable platform for COVID-19 and influenza testing? Findings In this cohort study of saliva samples from 50 community-based patients, the smartphone-based LAMP assay detected SARS-CoV-2 infection and exhibited concordance with reverse transcriptase–quantitative polymerase chain reaction tests. Meaning These findings suggest that the smartphone-based LAMP assay offers an additional tool to detect COVID-19 that can be readily modified in response to novel SARS-CoV-2 variants and other pathogens with pandemic potential including influenza.

General work flow: The smaRT-LAMP reaction mix was assembled at room temperature in 50 L total reaction volume containing 25 L "sample mix" (20 L saliva specimen with RNA stabilizers), 25 L "master mix" (containing lysis reagents, primers and polymerase enzymes) in 8-tube PCR strips or 96-well PCR plates, with optically clear lid strips (BioRad). The resultant smaRT-LAMP reaction mix (50 L) was transferred to a 70 o C heatblock for lysis and to initiate reverse transcription and LAMP reactions. Amplification was monitored by the free, custom-built Bacticount app adapted from Barnes et al. 4 on Samsung Galaxy S7 or S9 phones (eFigure 1).
Specific work flow: Sample mix is comprised of 20 L saliva and 5 L RNase inhibitor/Tris-HCL buffer (2.5 l 400 mM Tris-HCl, pH 7.5 [final specimen mix concentration of 40mM Tris-HCl]), 1.25 L RNase inhibitor (40 U/μL), and 1.25 L nuclease-free water. Master mix and sample mix volumes are scaled up 10% relative to the actual volumes need to test a given sample number. Reagent volumes for 96 samples are given in eTable 2.
Order and composition of "reaction mix" and assembly (50 L).
Step Step 5. In order to reduce primer-dimer formation, primers were heated individually at 70 o C for 5 m and allowed to cool to room temperature; and added to master mix just prior to enzyme addition (primer dimers occur with no primer melting and earlier primer addition to master mix). Step 6. High concentration Bst 2.0 WarmStart DNA polymerase was added (0.64 U/μL) together with RTx WarmStart reverse transcriptase (0.6 U/μL). Step 7. Sample mix (25 L) is added to master mix (25 L) and the resultant reaction mix was transferred to a 70 o C heat block for amplification. Negative controls consisted of confirmed clinically-negative saliva from Cottage Hospital. For patient saliva testing, heat-treated aliquots were removed from -80 o C storage, thawed on ice and added to sample mix at room temperature. Specimen stability as a function of time and temperature was assessed in this study.

Bacticount application
The Bacticount mobile phone application for monitoring and analyzing the smaRT-LAMP assay was built on a Samsung Galaxy S7 and S9 phone and can be downloaded and installed free of charge from the Google Play Store or www.bacticount.com. The main screen offers a choice to "Start Bacterial or Viral Analysis" and the user is prompted to pick the sample type; i.e., blood, urine, feces, or saliva (eFigure 2). The user follows a three-step procedure: 1) Record Standard Curve for a pathogen of interest in contrived (spiked) samples (e.g., SARS-CoV-2; influenza); 2) Record Sample; and 3) Select and view results where the app displays the sample results in a binary manner as follows: "Pathogen Detected" -designated as red circle; or, "No Pathogen Found" -designated as green circle on the "Reaction Results" screen. Further, by clicking on the red circle that appears if a pathogen is detected, the app then displays the viral load in copies/mL on the "Detailed Reaction Results" screen.

Establishing standard curve, unknown sample reactions, and data analysis
When running a sample reaction, the app launches a specialized viewfinder, allowing the user to center the reaction vials in the view-frame of the phone's camera, such that their intensity can be analyzed over time as adapted from 4 for up to 96 sample wells (eFigure 2). After entering the sample name, the user loads samples and presses "OK", which starts a timer to measure lost reaction time while setting up the box and aiming the camera. When the user selects "Begin Recording Amplification," the application proceeds to capture one photograph of the amplification reaction every 10 s over the entire course of the allotted reaction time. The user has three options: 1. Record Standard Curve option, the software also prompts the user to align each reference sample with a provided sample map so that the input starting concentrations of nucleic acid are known. The standard curve is determined through a linear regression fit of T t vs. log 10 [conc], which is stored as a '.pasc file' for determining the results in future tests.
2. Record Sample option, the app will record traces for each sample. The numerical sample traces and collected time-stamped photos are saved as a '.parr file' and as '.jpeg files', respectively, which may be extracted by the user to any computer.
3. Select and view results option, the app will prompt the user to choose a standard curve that has been recorded as outlined in the previous section with known standard concentrations. After data processing and analysis, the T t of unknown test samples are related to their initial concentrations via the standard curve. On its final screen, the app displays the viral load for each positive reaction well that was initially scored in a binary manner as either positive (red) or negative (green).
Data analysis. The Bacticount app enables the smartphone to serve as a stand-alone diagnostic for sensitivity (binary +/-call result) and quantitative detection of microbial titers as adapted from Barnes et al. 4 Real-time LAMP traces were automatically generated for each sample and used to calculate the threshold time (Tt) and concentration by the app on the smartphone. The Tt is linearly related to the logarithm of the input concentration and is used to determine the concentration of virus in saliva samples using standard curves. Alternatively, trace files from the phone were transferred to a personal computer (PC), where a custom MATLAB script was used to determine the Tt and calculate the resultant virus concentration using standard curves with Microsoft Excel.

Strains
Coronaviruses: The following reagents were obtained from BEI Resources, NIAID, NIH or ATCC. (NR-15241) and influenza B virus, B/Christchurch/33/2004 (Yamagata Lineage) (NR-36526) were obtained from BEI Resources. The viral genomic RNA and virus stocks were aliquoted and stored at -80 °C. Stocks were not thawed and frozen more than three times. Bacterial respiratory pathogens: S. pneumoniae D39, S. aureus USA 300, P. aeruginosa ATCC 10145 and K. pneumoniae ATCC 13883 were as previously described. 4,5 Oligonucleotide Primers Table S1 lists the oligonucleotide primer sequences used in this study. The smaRT-LAMP primer sets consist of two outer (F3 and B3), two inner (FIP and BIP), and two loop primers (F-Loop and B-Loop). Primers for smaRT-LAMP analysis target the SARS-CoV-2 nucleocapsid (N) and ORF1ab genes; influenza primers target genes encoding matrix protein (M1) and polymerase (PB1) for influenza A; and M1 and nonstructural protein (NS1) for influenza B. SARS-CoV-2 LAMP primer sets targeting ORF1ab and N genes (sets 9 and 16, respectively) were used for all analyses in this study. Alternative SARS-CoV-2 LAMP primers sets targeting ORF1ab and N genes have been validated for smaRT-LAMP (sets 15 and 3, respectively). Primers for CDC 2019-nCoV RT-qPCR analysis targets the SARS-CoV-2 N gene, 6 while the influenza SARS-CoV-2 (Flu SC2) RT-qPCR multiplex assay targets M1 for influenza A; and NS2 for influenza B. 7 New smaRT-LAMP primer sets (SARS-CoV-2, set 9; influenza A, set A5a; and influenza B, sets B5, B7) were designed as follows. A minimum of 42 genome sequences each for SARS-CoV-2, influenza A, and influenza B were downloaded from the NCBI virus genome database (SARS-CoV-2) or NCBI Influenza Virus Database (influenza A and B), and aligned using CLC sequence viewer (version 6.8.1), to identify conserved regions and consensus target sequences for primer design. Loop primers were designed for SARS-CoV-2 alternative primer set 15 adapted from Ganguli et al. 8 Influenza A primer set A6 was adapted from Poon et al. 9 by designing loop primers and using consensus sequences to redesign primers.

RT-qPCR reaction conditions
The CDC 2019-nCoV RT-qPCR 6 and Flu SC2 RT-qPCR 7 tests contains primers and probes for SARS-CoV-2; and SARS-CoV-2, influenza A and B, respectively. SARS-CoV-2 was evaluated using SARS-CoV-2, influenza A, or influenza B primer and probes (IDT) and the LUNA cell-ready probe one step RT-qPCR kit (NEB). RT-qPCR was performed using 10 μL saliva sample as per manufacturer recommendations.

SmaRT-LAMP platform hardware
SmaRT-LAMP hardware was adapted from Barnes et al. 4 Experiments were performed in low-profile 0. LEDs were powered using a single output DC power supply (UA8001A, Agilent Technologies, Santa Clara, CA). A Samsung Galaxy S7 or S9 smartphone (Samsung Electronics Co., Ltd.) was outfitted with a 520 ± 10 nm bandpass filter (Edmund Optics cat. no. 65-699) for visual detection of emitted green light. All RT-qPCR reactions were performed on a Bio-Rad CFX96 qPCR Thermocycler.

SmaRT-LAMP sensitivity and specificity assays
SmaRT-LAMP can simultaneously detect SARS-CoV-2, influenza A and/or B viruses via addition of cognate primers to individual wells, which is clinically important when both SARS-CoV-2 and influenza viruses are cocirculating as the two disease syndromes are similar 10 . Sensitivity and specificity tests were performed using contrived saliva specimens as recommended under FDA Emergency Use Authorization (EUA) guidelines. 11 Sensitivity: Limit of detection (LOD) was evaluated following serial dilution of viral and genomic RNA (genome copies/mL) or viral stocks quantified by the 50% tissue culture infective dose assay [TCID 50 , the highest dilution causing a cytopathic effect in one-half of tissue culture samples]. 12 Viral pathogens quantified by TCID 50 /mL include: influenza A, 1.4 x 10 5 TCID 50 /mL; influenza B, 8.0 x 10 4 TCID 50 /mL; HCoV-NL63, 8.0 x 10 3 TCID 50 /mL; HCoV-229E, 8.0 x 10 5 TCID 50 /mL; HCoV-OC43: 4.5 x 10 4 TCID 50 /mL. LOD was determined by the largest serial dilution giving a signal in > 19/20 biological replicates.
Specificity: Cross-reactivity tests were performed with contrived saliva specimens using EUA recommended amounts of 10 5 and 10 6 copies/mL specimen for viral and bacterial pathogens, respectively; or via undiluted viral stocks quantified by TCID 50 . 11 Specificity was determined by the presence or absence of signal (binary +/-call). n= 10 biological replicates for SARS-CoV-2 variants; n = 20 biological replicates for all other pathogens.

Analysis of virus present in saliva
Spiked saliva. Known amounts of inactivated SARS-CoV-2 as well as other viral and bacterial respiratory pathogens were added to virus-negative human saliva; serial dilutions were made using saliva diluent. Samples were analyzed using both smaRT-LAMP and RT-qPCR to assess specificity and sensitivity of the assay. Patient saliva. Saliva specimens were collected at Santa Barbara Cottage Hospital of the Cottage Health system, Santa Barbara, CA. Two sub-groups of participants (symptomatic and asymptomatic) were enrolled. The symptomatic group consisted of recruited patients who tested positive for SARS-CoV-2 with symptoms; the asymptomatic patients were recruited from the same community, through negative admission screening tests for SARS-CoV-2 infection. Patient saliva specimens were collected in sterile plastic tubes and stored frozen at -20 o C on site. Upon transport to UC Santa Barbara, frozen specimens were thawed, heat-inactivated at 95 o C for 30 min, aliquoted and stored frozen at -80 o C. For processing, saliva samples were thawed on ice and the reaction mixture was assembled at room temperature. Human subjects approval was obtained from the Institutional Human Subjects Use Committee of the University of California, Santa Barbara and the Institutional Review Board of Santa Barbara Cottage Hospital. To assess potential for real-world application to detect virus in patient saliva samples, we carried out a temporal analysis of patient saliva samples stored up to a week at either 4 o C or 25 o C using smaRT-LAMP. This study *Loop primers were designed for SARS-CoV-2 primer set 15. 8 Influenza A primer set A6 was modified from ref. 9 by designing loop primers, and by using consensus sequences to redesign primers.  Nuclease-free water 1. The smaRT-LAMP reaction mix, containing "sample mix" (saliva specimen with RNA stabilizers) and "master mix" (lysis reagents, primers and polymerase enzymes), is assembled at room temperature and loaded onto a 70 o C heat block, which initiates both the reverse transcription and LAMP reactions. The mobile phone app displays the sample results in a binary manner as follows: "Pathogen Detected" -designated as red circle; or, "No Pathogen Found" -designated as green circle on the "Reaction Results" screen. Clicking on the red circle results in the app displaying the viral load in copies/mL on the "Detailed Reaction Results" screen.

eFigure 2. Workflow of the Bacticount SmaRT-LAMP Mobile Phone App.
The user is prompted to pick the sample type (saliva) followed by a three-step procedure: 1) Record Standard Curve for a pathogen of interest in contrived (spiked) samples (e.g., SARS-CoV-2; influenza); 2) Record Sample; and 3) Select and view results where the app displays the sample results in a binary manner as follows: "Pathogen Detected" -designated as red circle; or, "No Pathogen Found" -designated as green circle on the "Reaction Results" screen. By clicking on the red circle that appears if a pathogen is detected, the app displays the viral load in copies/mL on the "Detailed Reaction Results" screen. Blue arrows represent connecting app screenshots.