key: cord-0928367-l6igxl2k
authors: Diaz, L.; Johnson, B. E.; Jenkins, D. M.
title: Real-time optical analysis of a colorimetric LAMP assay for SARS-CoV-2 in saliva with a handheld instrument improves accuracy compared to endpoint assessment
date: 2021-01-15
journal: nan
DOI: 10.1101/2021.01.13.21249412
sha: 6bb7ba2a00db89b3421f8e6dfe47c6ac71a1c1a1
doc_id: 928367
cord_uid: l6igxl2k

Controlling the course of the COVID-19 pandemic will require widespread deployment of consistent and accurate diagnostic testing of the novel coronavirus SARS-CoV-2. Ideally, tests should detect a minimum viral load, be minimally invasive, and provide a rapid and simple readout. Current FDA-approved RT-qPCR-based standard diagnostic approaches require invasive nasopharyngeal swabs and involve laboratory-based analyses that can delay results. Recently, a loop mediated isothermal nucleic acid amplification (LAMP) test that utilizes colorimetric readout received FDA approval. This approach utilizes a pH indicator dye to detect drop in pH from nucleotide hydrolysis during nucleic acid amplification. This method has only been approved for use with RNA extracted from clinical specimens collected via nasopharyngeal swabs. In this study, we developed a quantitative LAMP-based strategy to detect SARS-CoV-2 RNA in saliva. Our detection system distinguished positive from negative sample types using a handheld instrument that monitors optical changes throughout the LAMP reaction. We used this system in a streamlined LAMP testing protocol that could be completed in less than two hours to directly detect inactivated SARS-CoV-2 in minimally processed saliva that bypassed RNA extraction, with a limit of detection (LOD) of 50 genomes/reaction. The quantitative method correctly detected virus in 100% of contrived clinical samples spiked with inactivated SARS-CoV-2 at either 1X (50 genomes/reaction) or 2X (100 genomes/reaction) of the LOD. Importantly the quantitative method was based on dynamic optical changes during the reaction so was able to correctly classify samples that were misclassified by endpoint observation of color.

Introduction 4 commercially available diagnostics with regulatory approval used for SARS-CoV-2 detection in 67 the Unites States FDA, European Union and Asia utilize isothermal technology. 11 68

The colorimetric SARS-CoV-2 LAMP diagnostic assay developed by COLOR was one 69 of the first of these isothermal assays to receive Emergency Use Authorization (EUA) by the 70 virus has been further optimized by pre-incubating viral samples at 95°C to denature the viral 88 capsid before adding samples to a LAMP reaction mix. 21-23 Sample processing protocols have not 89 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 January 15, 2021. ; https://doi.org/10.1101/2021.01.13.21249412 doi: medRxiv preprint 7 standards (pink solution, nominally assigned 0 a.u) and positive standards (yellow solution, 135 nominally assigned 60000 a.u.). 136 137

For detailed analysis of experimental reactions described in this manuscript we used the 139 raw "luminance" data from files generated automatically in the BioRanger app. Unless otherwise 140 stated, all values were shifted to start at "zero" luminance by subtracting the initial (t = 0 141 minutes) raw luminance value from every raw value. These baseline corrected luminance values 142 (L) were smoothed using a 2 nd order polynomial with a rolling average of 4 adjacent data points. 143

The maximum values of luminance derivatives (L') for each assay were identified from 144 derivative data (identified using forward difference and smoothed again with a 2 nd order 145 polynomial with 4 neighbors) using GraphPad Prism 9 (GraphPad Software, San Diego, CA, 146 USA). Table S1 ) developed by 152 New England BioLabs, 24 targeting the envelope (E1) and nucleocapsid (N2) genes of the SAR-153

CoV-2 viral genome (GenBank accession number MN908947), were tested individually and in 154 combination. For the final contrived clinical evaluation an internal control (IC) primer set that 155 amplifies human b-actin "housekeeping" gene (ACTB) was used to detect the presence of 156 inhibitors in saliva samples. Primers were synthesized commercially (Integrated DNA 157 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. (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 January 15, 2021. ; https://doi.org/10.1101/2021.01.13.21249412 doi: medRxiv preprint All real-time colorimetric assays were carried out in 0.2 mL reaction tubes (TempAssure,  180 Optical Caps, USA Scientific Inc., Orlando, FL, USA) in a modified 8-well isothermal amplifier 181 platform (BioRanger). Amplification progress was logged every 30 seconds for a maximum of 182 40 minutes. At the end of the assay the closed tube strip was briefly placed on ice and a 183 photograph was taken using a cell phone camera. The colorimetric master mix contains a phenol 184 red pH indicator used for visual detection of amplification and interpreted according to intensity and fluorescence sensitivity between channels. For the StepOnePlus™ PCR machine, 202 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. (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 January 15, 2021. ; https://doi.org/10.1101/2021.01.13.21249412 doi: medRxiv preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. GuHCl and standard colorimetric master mix and primer concentrations as previously described. 292

Assays were carried out and interpreted blind to the experimenter. The corresponding viral load 293 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. Fluorescent LAMP assays were used to evaluate optimal reaction conditions for assays 304 containing single and duplexed primer sets. The addition of 40 mM GuHCl to fluorescent LAMP 305 assays improved both the speed and sensitivity for the detection of synthetic RNA in spiked 306 water controls. Assays containing only the N2 primer set and guanidine hydrochloride performed 307 best overall in terms of sensitivity (~ 5.25 genome equivalents/reaction) and speed of detection 308 (mean tT = 19 minutes, Supplementary Fig. 1 ). The differences in the speed of detection for 309 duplexed reactions were unremarkable; E1 primers are less sensitive than the N2 primers by 310 approximately an order of magnitude in any condition. E1-primed reactions were also more 311 temperature sensitive: incubation at 68°C inhibited LAMP reactions. Colorimetric assays were 312 run ten minutes longer than fluorescent assays for a total time of 40 minutes to compensate for 313 the relatively slow color change of the phenol red indicator that has been demonstrated to take 314 forty to fifty minutes to detect single copy numbers of SARS-CoV-2 genomes in previous 315 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. Under optimized assay conditions the colorimetric luminance curve of PC standards is bi-333 phasic with two distinct exponential growth phases in the first and last half of the reaction. We 334 quantitatively evaluated two curve features, presumably associated with initial amplification and 335 subsequent color change, at time intervals where a log-linear luminance profile was observed to 336 occur in PC assays. Specifically, local maxima in rates of luminance increase (L') occur between 337 5 and 20 minutes (L' MAX1 ) and then again between 20 and 40 minutes (L' MAX2 ) (Fig. 2A ). The first 338 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 January 15, 2021. ; https://doi.org/10.1101/2021.01.13.21249412 doi: medRxiv preprint five minutes of luminance values were excluded from consideration for L' MAX1 because noise in 339 the baseline signal was previously observed to occur early in the reaction, particularly for assays 340 containing duplexed primer sets ( Supplementary Fig. 2B, 2E) . 341

The luminance curves of NC assays testing N2 primers (Fig. 1) plus five NC assays from 342 other trials conducted that day testing different primers sets (Supplementary Fig. 2) were pooled 343 (n=8) and compared to PC purified RNA standards (n=13, Fig. 1C This indicates that our simple instrument is able to detect subtle changes in the optical 358 characteristics of the reaction that occur before easily observable color changes. The second 359 luminance derivative peaks (L' MAX2 ) in our instrument are much larger than L' MAX1 and are 360 typically observed between 25-35 minutes into the reaction which is more consistent with times 361 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. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The sensitivity of the colorimetric-LAMP assay was quantitatively assessed by following 397 optimized assay protocols for detection of SARS-CoV-2 in saliva using a handheld instrument 398 for luminance readout (Fig. 3) . To reduce the impact of saliva-bound nuclease activity, we 399 demonstrated that TE buffer is effective at stabilizing nucleic acids for sensitive detection (500-406 1000 viral particles per mL) by RT-qPCR. 20 We similarly demonstrated that addition of TE to 407 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 January 15, 2021. ; https://doi.org/10.1101/2021.01.13.21249412 doi: medRxiv preprint heat-inactivated samples containing intact gamma-irradiated SARS-CoV-2 improved the LOD of 408 the colorimetric assay by an order of magnitude compared to spiked samples diluted in water 409 only ( Supplementary Fig. 3) of eight non-reactive control (NRC) (Fig. 4A) . Applying this classification threshold, the 418 luminance-based LOD, defined by the lowest concentration where SARS-CoV-2 is detected in 419 all replicates, is 5 x 10 4 ge/mL corresponding to 50 ge/reaction (Fig. 4B) . Luminance 420 amplification curves for this section can be found in Supplementary Figure 4 . 421 SARS-CoV-2 was detected both visually and quantitatively below the LOD (5 x 10 3 ge/mL -422 2.5 x 10 4 ge/mL) in some replicates but at a lower rate. In some reactive samples below the LOD 423 and in reactions in which SARS-CoV-2 was not visually detected, a low-amplitude luminance (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (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 January 15, 2021. ; https://doi.org/10.1101/2021.01.13.21249412 doi: medRxiv preprint To evaluate the feasibility and performance of our quantitative colorimetric LAMP 462 method, a blinded and randomized contrived clinical trial was conducted. A total of 20 saliva 463 samples, 10 reactive controls containing inactivated SARS-CoV-2 at 1X and 2X the LOD (5 x 464 10 4 ge/mL) determined from saliva standards (Fig. 4) , and 10 non-reactive controls (0 ge/mL), 465

were tested using the standardized protocol (Fig. 3) . In addition to two replicate assays for the N- Although the test panels were run on two colorimetry-enabled BioRanger instruments 480 over two weeks and by multiple experimenters using different batches of reagent stocks, 481 performance analysis demonstrates that the method sensitivities, specificities and accuracy to 482 quantitatively detect SARS-CoV-2 did not vary substantially. All assays that resulted in an 483 endpoint color of yellow also showed luminance signal amplification rate above decision 484 threshold values where L' MAX2 was above 660 a.u./min. (Fig. 6A) . A failed color change at the 485 LOD correlated with no or low intensity luminance signals (L' MAX2 < 660). The SARS-CoV-2 N-486 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 January 15, 2021. ; https://doi.org/10.1101/2021.01.13.21249412 doi: medRxiv preprint gene was detected in all test sample replicates (10/10) for saliva samples spiked with 1 x 10 5 487 ge/mL (2X LOD). At least one replicate tested positive for each sample with 5 x 10 4 ge/mL (1X 488 LOD), though not all replicates at this concentration tested positive (6/10) (Fig. 6B) . SARS-489

CoV-2 was not detected in any of the contrived negative samples (NRC). Human ACTB gene amplification. 502

Panel 9 tested an NRC sample, and even though endpoint assessment of the color (light 503 orange) resulted in positive classification for both replicates, no signs of amplification were 504 observed in the luminance data. Using the real-time data, both classified correctly as negative 505 ( Supplementary Fig. 5, panel 9) . Similarly, one replicate of the panel 3 sample (also an NRC 506 sample) was light orange incorrectly indicating amplification based on endpoint color, though 507 the luminance data correctly showed no amplification ( Supplementary Fig. 5, panel 3) . In 508 contrast the internal control reactions on panel 9 showed evidence of faint and late amplification 509 in the luminance data but failed to meet threshold for positive classification, resulting in this 510 panel being inconclusive, even though the endpoint color of these internal controls was yellow 511 (positive). The N-gene was correctly detected in all positive control replicates (40/40) and not 512 detected in any of the negative control replicates (0/40) (Fig. 6A-B) . (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. As COVID-19 continues to surge, there is increasing need to expand points-of-care and 550 mobile testing platforms to quell the transmission of this virus. Nucleic acid amplification 551 emerged as the gold standard for detection efficiency in the early months of the pandemic 552 because this method is both highly specific and highly sensitive. Paired with an RNA isolation 553 sample-processing step, RT-PCR-based SARS-CoV-2 detection assays provided the necessary 554 data for viral transmission rates and initial contact tracing. RT-PCR-based testing, while accurate 555 and sensitive, is heavily reliant on thermocycling equipment that requires a specialized 556 laboratory environment and highly trained personnel to operate. These limitations have hindered 557 effort to scale the use of this testing platform to meet the exponential expansion in testing 558 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 January 15, 2021. ; https://doi.org/10.1101/2021.01.13.21249412 doi: medRxiv preprint demand. Antibody-based detection platforms are emerging as an alternative to RT-PCR based 559 assays to meet the rising demand for rapid, point-of-care testing. While antibody-based detection 560 methods can be adapted for at-home, point-of-care detection, or mobile testing, this methodology 561 is inherently less sensitive than nucleic acid amplification-based assays. 562

Isothermal nucleic acid amplification merges the mobility of antibody-based assays with 563 the high sensitivity of RT-PCR as an efficient platform to expand COVID-19 testing. In this 564 study, we have built upon the colorimetric endpoint readout of LAMP-based testing and have 565 provided proof-of-design for a mobile platform that utilizes isothermal nucleic acid amplification 566 in conjunction with direct detection of SARS-Cov-2 from saliva. We have paired SARS-CoV-2 567 detection during isothermal nucleic acid amplification with a mobile, handheld device that 568 detects colorimetric changes during isothermal amplification similar to that of a microplate 569 reader used in previous studies. 12 This method is capable of detecting purified (synthetic) SARS-570

CoV-2 RNA template within an order of magnitude of single particle detection (5.25 571 genomes/reaction) therefore comparable to the theoretical maximum sensitivity of other RT-572 PCR-based approaches as well as other LAMP-based SARS-CoV-2 detection designs. 573 COLOR initially received FDA approval for merging LAMP-based testing with a 574 colorimetric endpoint readout that streamlined detection and circumvented the need for highly 575 specialized diagnostic equipment. The majority of FDA-approved LAMP-based assays have only 576 been authorized for samples that have first undergone RNA purification. In order to develop a 577 direct-detection assay from saliva samples that can vary in pH, nucleic acid amplification and pH 578 change must be measured independently. Previous studies have examined the accuracy and 579 sensitivity of endpoint colorimetric changes indirectly by conducting parallel experiments with 580 real-time fluorescent-based assays. 13,33,34 These approaches, however, cannot directly verify 581 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 January 15, 2021. ; https://doi.org/10.1101/2021.01.13.21249412 doi: medRxiv preprint whether a colorimetric readout corresponds to amplification or a spurious pH change. This 582 indirect approach has hindered progress towards developing a direct-detection assay from saliva 583 samples. Our approach, however, provides a platform to independently measure both nucleic 584 acid amplification and colorimetric endpoint readout for every sample and provide a means to 585 determine whether SARS-CoV-2 can be detected directly from saliva. 586

Using a commercial BioRanger, the time course of the initial luminance derivative peak, 587 L' MAX1 , corresponded to the time to the threshold of detection for fluorescent-based LAMP 588 assays, while the time course of the second luminance derivative peak, L' MAX2 , corresponded to 589 the times reported to be required for endpoint color change. These data indicate that our method 590 can detect subtle optical changes at the initiation of LAMP amplification, perhaps due to 591

increased scattering from small particulates of magnesium pyrophosphate, as well as more 592 intense optical changes later in the reaction corresponding to transition of pH through the pK a of 593 the phenol red indicator. The modified BioRanger, therefore, provides an ideal platform for 594 evaluating whether a LAMP-based assay can be implemented for the direct detection of SARS-595

Cov-2 from saliva. 596

Direct detection of viral particles from saliva can present a host of additional 597

complications. In addition to fluctuations in pH that can destabilize the viral capsid or cause a 598 spurious color change, saliva contains nucleases that can degrade the viral genome, reducing the 599 amount of virus that can be detected by LAMP. Previous work has shown that encapsulated 600 SARS-CoV-2 is stable in saliva 20,35,36 and isothermal nucleic acid amplification efficiency is 601 enhanced when the heat-inactivated saliva matrix is buffered with TE and Tween 20. Processing 602 samples in this way, we implemented a low-cost alternative to VTM that was completely 603 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 January 15, 2021. ; https://doi.org/10.1101/2021.01.13.21249412 doi: medRxiv preprint

Saliva is more sensitive for SARS

CoV-2 detection in COVID-19 patients than nasopharyngeal swabs. medRxiv

Saliva is a reliable, non-invasive specimen for

SARS-CoV-2 detection

Saliva specimens for detection of severe 658 acute respiratory syndrome coronavirus 2 in Kuwait: A cross-sectional study

A rapid field detection system for citrus 661 huanglongbing associated "Candidatus Liberibacter asiaticus" from the psyllid vector

Diaphorina citri Kuwayama and its implications in disease management

Non-instrumented nucleic acid amplification 665 (NINA) for rapid detection of Ralstonia solanacearum race 3 biovar 2

Development of a genome-informed loop-668 mediated isothermal amplification assay for rapid and specific detection of Xanthomonas 669 euvesicatoria

Clinical specimen-Direct LAMP: A useful 674 tool for the surveillance of blaOXA-23-Positive carbapenem-resistant acinetobacter 675 baumannii

Defining the relative performance of 680 isothermal assays that can be used for rapid and sensitive detection of foot-and-mouth 681 disease virus

Coronavirus Test Tracker: Commercially Available COVID-19 Diagnostic

Color Genomics SARS-CoV-2 LAMP Diagnostic 687

Visual detection of isothermal nucleic acid amplification 689 using pH-sensitive dyes

Id now TM product insert

Performance of Abbott ID NOW rapid SARS

CoV-2 NAAT. Block Caving -A Viable Altern

In vitro diagnostic assays for COVID-19: Recent advances and emerging 695 trends

Evaluation of specimen types and saliva 697 stabilization solutions for SARS-CoV-2 testing

Clinical assessment and validation of a rapid and 700 sensitive SARS-CoV-2 test using reverse-transcription loop-mediated isothermal 701 amplification

Facilitating Detection of SARS-CoV-2 Directly from Patient 703 Samples : Precursor Studies with RT-qPCR and Colorimetric RT-LAMP Reagents

Saliva-Based Molecular Testing for SARS

CoV-2 that Bypasses RNA Extraction

Inactivation of SARS coronavirus by means of povidone-709 iodine, physical conditions and chemical reagents

Clinical assessment and validation of a rapid 712 and sensitive SARS-CoV-2 test using reverse-transcription loop-mediated isothermal 713 amplification

SARS-CoV-2 -Escape from the NA extraction kit-shortage

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

Enhancing colorimetric loop-719 mediated isothermal amplification speed and sensitivity with guanidine chloride

Real-Time duplex applications of Loop-Mediated AMPlification ( 725 LAMP ) by assimilating probes

FRET-Based Assimilating Probe for

Sequence-Specific Real-Time Monitoring of Loop-Mediated Isothermal Amplification 729 (LAMP)

Emergency Use Authorization molecular diagnostics for the detection of SARS-CoV-2

Scaling diagnostics in 734 times of COVID-19: Colorimetric Loop-mediated Isothermal Amplification (LAMP) 735 assisted by a 3D-printed incubator for cost-effective and scalable detection of SARS-CoV-736 2

Rapid and extraction-free detection of SARS

CoV-2 from saliva with colorimetric LAMP

Temporal profiles of viral load in posterior

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

CoV-2: an observational cohort study

A rapid, highly sensitive and open-access SARS

CoV-2 detection assay for laboratory and home testing

Isothermal Amplification (RT-LAMP) Assays Targeting SARS-CoV-2

A rapid, highly sensitive and open-access SARS-CoV-750 2 detection assay for laboratory and home testing

Saliva or Nasopharyngeal Swab Specimens for 753 Detection of SARS-CoV-2

Use of Saliva for Diagnosis and Monitoring 755 the SARS-CoV-2: A General Perspective

Rapid isothermal amplification and portable 758 detection system for SARS-CoV-2

compatible with both luminance detection and endpoint colorimetric detection of isothermal 604 amplification of SARS-Cov-2 nucleic acid. 605 606 Future Implementation 607Through the assay that we developed in this study, we were able to directly detect 608 inactivated SARS-CoV-2 from saliva with an LOD of 50 genomes/reaction. This detection 609 sensitivity is similar to a previously-described LAMP-based SARS-CoV-2 direct detection assay 610 (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (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 January 15, 2021. ; https://doi.org/10.1101/2021.01.13.21249412 doi: medRxiv preprint