Introduction
In December 2019, the world witnessed an unknown coronavirus emerging in Wuhan, China, first called 2019 novel coronavirus (2019-nCoV), and then severe acute respiratory syndrome-coronavirus 2 (SARS-CoV-2) [
1]. SARS-CoV-2 is a positive-sense single-stranded RNA virus belonging to the
Sarbecovirus subgenus. Based on the limited scale of the outbreaks of its predecessors SARS-CoV-1 in 2003 and MERS-CoV in 2012, SARS-CoV-2’s spread around the world was unfortunately underestimated. In November 2020, 52.5 million of people have been infected. The rapid increase of daily cases worldwide has made imperative the development of fast and accurate diagnostic tools of SARS-CoV-2 in order to isolate infected people quickly to reduce transmission. Accurate and fast detection also allows for a better understanding of viral transmission rate. The usual method used for
Sarbecovirus RNA detection involved reverse transcription polymerase chain reaction (rRT-PCR) on viral genes [
2]. As a coronavirus, SARS-CoV-2 contains 4 structural proteins (S spike, M membrane, E envelope and N nucleocapsid) and several non-structural proteins (such as the RNA-dependent RNA polymerase (RdRP) responsible for the synthesis of viral RNAs), all produced from cleavage of ORF1a and ORF1b [
3].
In the context of the current pandemic, our hospital laboratory receives more than a thousand samples in UTM or in RNase-free water each day for testing. Procedures that include RNA extraction are not allowing this kind of throughput with the equipment at hand. Moreover, RNA extraction kits are in limited supply due to high demands. Performing direct rRT-PCR, without any RNA extraction step, offers a quick and cost-efficient solution. However, our concern was that specimen itself or transport media such as UTM would interfere with rRT-PCR efficiency. UTM is a stable viral transport medium allowing collection and transport of viruses and bacteria that contains protective proteins and antimicrobials, which could interfere with rRT-PCR enzymes.
The first objective of this study was to address the feasibility of direct rRT-PCR and to explore the requirements of specimen dilution and PK treatment to reduce rRT-PCR interference. PK is a well-known enzyme that has several activities such as protein denaturation and nuclease inhibition and has already been shown to improve SARS-CoV-2 detection [
4‐
6].
In this study, we have optimized a method to detect SARS-CoV-2 from nasopharyngeal swabs in UTM or in RNase-free water by amplifying E, N and RdRP genes using rRT-PCR without requiring a RNA extraction step. We demonstrate that PK addition improves detection sensitivity and reduces rates of invalid results. Finally, we also show that non- extracted samples can be analysed within 3 days if stored at 4 °C without altering rRT-PCR sensitivity.
Material and methods
Reagents and sample
ONP swabs were collected either in 2 ml of UTM® (COPAN) or RNase-free water and stored at 4 °C before analysis. No patients were specifically recruited, and only samples already known as positive or negative were used in this study. Poly A (Millipore Sigma ref P9403), β-mercaptoethanol (Gibco™ ref 31350010), proteinase K solution 20 mg/ml (ThermoFischer Scientific ref 25530049).
Sample sets and experiments
All sample sets are summarised in Table
1.
1
30 known positive ONP specimens collected with flocked swabs in 3 ml UTM with all three viral genes detected. These were retested by rRT-PCR after RNA extraction as standard reference and tested by direct rRT-PCR with and without thermal lysis at the following concentrations: undiluted, 1/2, 1/5, 1/10 diluted in RNase-free water. The results obtained by direct rRT-PCR were compared to those obtained after extraction.
2
30 known positive ONP specimens collected with flocked swabs in 2 ml RNase-free water with all three viral genes detected. These were retested by rRT-PCR after RNA extraction as standard reference and tested by direct rRT-PCR with or without thermal lysis at the same dilutions mentioned above and compared to the standard reference.
3
90 known positive ONP specimens collected with flocked swabs in 2 ml RNase-free water with all three viral genes detected with a wider range of Ct values than the previous sample sets (N gene Ct values from 15 to 40).
4
60 known positive ONP specimens collected with flocked swabs in 2 ml RNase-free water with only N gene detected with low (Ct values 30–35) to very low viral loads (Ct values ≥ 36).
5
60 known negative ONP specimens collected with flocked swabs in 2 ml RNase-free water.
Table 1
Sample sets characteristics
n specimens | 30 | 30 | 90 | 60 | 60 |
Previous rRT-PCR | + | + | + (3 viral genes) | + (N gene only) | - |
Ct range N gene | random | random | Wide 15–40 | Low 30–35 Very low 36–40 | not detected |
Medium | UTM | RF water | RF water | RF water | RF water |
Storage | − 70 °C | − 70 °C | − 70 °C | − 70 °C | − 70 °C |
Thermal lysis | Y vs. N | Y vs. N | Y | Y | Y |
RNA extraction | Y vs. N | Y vs. N | N | N | N |
Direct rRT-PCR | Y | Y | Y | Y | Y |
Dilution in RF water | 1/1 vs. 1/2 vs. 1/5 vs. 1/10 | 1/1 vs. 1/2 vs. 1/5 vs. 1/10 | 1/2 | 1/2 | 1/2 |
Proteinase K | N | N | Y vs. N | Y vs. N | Y vs. N |
The specimens in sample sets 3, 4 and 5 (kept at − 70 °C) were retested by direct PCR, at the optimal dilution of 1/2, with thermal lysis, either with or without PK treatment prior to thermal lysis. Experiments on sample sets 3, 4 and 5 were designed to further evaluate the performance of our optimized direct rRT-PCR on a larger group of ONP specimens collected in water and to evaluate the impact of PK pre-treatment on sensitivity and reduction of invalid results.
OMEGA BIO-TEK E.Z.N.A.® Total RNA Kit I (R6834) protocol for manual extraction of viral RNAs was used. Briefly, 150 μl of nasopharyngeal swab from UTM or RNase-free water were added to 500 μl of TRK lysis buffer supplemented with carrier RNA (10 μg/ml) and β-mercaptoethanol (1 mM) into a 1.5 mL microcentrifuge tube. Tubes were vortexed for 30 s and kept at room temperature for 5–10 min. 350 μl of 100% ethanol were then added and tubes were vortexed for 30 s. Samples were transferred (including any precipitate) to a HiBind® RNA Mini Column. Columns were centrifuged at maximum speed (≥ 13,000 g) for 15 s. Columns were then washed once with 500 μl of RNA Wash Buffer I and twice with 500 μl of RNA Wash Buffer II at 10,000g for 1 min. Columns were centrifuged one last time at maximum speed (≥ 13,000g) for 2 min to remove residues. Columns were transferred into clean nuclease-free 1.5 ml tubes. 40–70 μl Nuclease-free Water was added directly on membranes into columns for 1 min and then columns were centrifuged at maximum speed (≥ 13,000g) for 2 min. Eluted RNAs were stored at -70 °C.
Reverse transcriptase-polymerase chain reaction
Non-extracted samples were diluted at 1/1, 1/2, 1/5 and 1/10 in RNase-free water in a 96-well PCR plates. RNA extracted samples were used undiluted. Plates were then either stored at 4 °C while preparing master mix or heated at 90 °C for 3 min to perform thermal lysis and cooled down at 4 °C. Allplex™ 2019-nCoV assay from Seegene Inc. were used according to the manufacturer protocol to perform rRT-PCR. Briefly, for one reaction: 5 μl of 2019-nCoV MOM (MuDT* Oligo Mix (MOM):—Amplification and detection reagent *MuDT is the brand name of Seegene’s oligo mixture), 5 μl of buffer 5 ×, 5 μl of RNase-free water, 1 μl of internal control (IC) and 2 μl of enzymes. 18 μl of master mix were distributed in each well and added with either 8 μl of sample, 8 μl of positive control or 8 μl of RNase-free water for negative control (final volume of 26 μl). Plates were then spun down at 2500 rpm for 5 s and analyzed on a CFX96 Touch Real-Time PCR from BioRad. Reverse Transcription reaction 1 cycle: 50 °C/20 min – 95 °C/15 min. PCR reaction 45 cycles: 94 °C/15 s – 58 °C/30 sec. Gene amplifications were analyzed by FAM (E gene), HEX (IC), Cal Red 610 (RdRP) and Quasar 670 (N gene) fluorophores. Results were compiled and analyzed using 2019-nCoV viewer from Seegene Inc. according to the manufacturer’s instructions [
7] (Table
2).
Table 2
Allplex™ 2019-nCoV assay interpretation
Case 1 | ± | + | + | + | 2019-nCoV detected |
Case 2 | ± | + | − | + | 2019-nCoV detected |
Case 3 | ± | + | + | − | 2019-nCoV detected |
Case 4 | ± | − | + | + | 2019-nCoV detected |
Case 5 | ± | − | − | + | 2019-nCoV detected |
Case 6 | ± | − | + | − | 2019-nCoV detected |
Case 7 | ± | + | − | − | Presumptive positive |
Case 8 | + | − | − | − | Negative |
Case 9 | − | − | − | − | Invalid |
PK treatment
PK was directly added in RNase-free water prior to dilution 1/2 at a concentration of 200 µg/ml (final concentration after dilution 1/2 is 100 µg/ml). Microwell plates were then heated in a thermal cycler at 50 °C for 15 min to perform enzyme activities and then heated at 90 °C for 3 min for inactivation.
Thermal lysis
A heat shock treatment in which the specimen is brought to 90 °C for 3 min followed by a rapid cooling step at 4 °C prior to the rRT-PCR process.
Experimental design and statistical analyses
All sample sets were divided in 3 groups and analysed in an independent manner. Statistical significance was evaluated by Multiple t-test using GraphPad Prism 8.0 software. p value < 0.05 was considered significant: ****: p < 0.0001; ***: p < 0.001; **: p < 0.01; * p < 0.05. ΔCt means N gene were referenced to compare 2 different groups by calculating the difference between the N gene Ct values means.
Discussion
In this study, we successfully optimized a direct rRT-PCR method to detect SARS-CoV-2 RNA in ONP swabs with the Allplex™ 2019-nCoV assay from Seegene Inc.
From sample sets 1 and 2, we evaluated the impacts of specimen dilution and thermal lysis on detection rates of direct rRT-PCR compared to manual RNA extraction as a standard reference. Detection was optimal using 1/5 diluted specimens collected in UTM and ½ diluted specimens collected in RNase-free water. While 1/5 dilution of specimens collected in UTM exhibited an efficiency of 93.3% of positive samples for three viral genes, two samples were missed. These two samples were detected after RNA extraction with C
t values greater than 30, suggesting that these samples had low viral loads. A thermal lysis treatment slightly improved the efficiency up to 96.6% detection by recovering one additional sample. The last remaining sample had high C
t values following RNA extraction (E: 34.4, RdRP: 37.3, N: 38.4), indicating that viral RNA levels were too low to be detected without RNA extraction. Although Merindol and colleagues showed that UTM is a suitable media for the amplification of SARS-CoV-2 without RNA extraction [
8], we demonstrate that dilutions of samples are required to obtain reliable detection. UTM medium itself might be inhibitory for the rRT-PCR reaction, as suggested by the requirement of stronger dilutions for optimal detection on specimens stored in UTM compared to RNase-free water.
Thermal lysis improved detection rates and the number of detected genes in most conditions. This was especially notable in the case of 1/2 diluted UTM, where N gene detection rose from 93 to 97%, RdRP from 63 to 77% and E from 87 to 90%. Thermal lysis also improved detection rates of undiluted and 1/2 diluted RNase-free water from 90 to 100% and from 93 to 100%, respectively. Thermal lysis could contribute to specimen inhibitor denaturation, rather than a direct effect on viral template release, as supported by the loss of benefits at higher dilutions. Thus, optimized conditions for direct rRT-PCR include storage in RNase-free water as transport medium, 1/2 specimen dilution and thermal lysis treatment before rRT-PCR. In fact, these allowed detecting 30/30 positive specimens with minimal impact on Ct values as compared to manual RNA extraction. As a result of these findings, we switched from UTM as a transport medium to RNase-free water in our hospital.
Using sample sets 3, 4 and 5, our optimized direct rRT-PCR process was evaluated on a greater number of positive specimens with a wider range of Ct values. We also tested the impact of a PK pre-treatment on both detection rate, and invalid results rate using negative specimens.
In sample set 3, the detection rate increased marginally from 94.4% to 98.9% with addition of PK. In sample set 4, consisting exclusively of low to very low viral loads, the detection rate rose significantly from 53.3 to 76.7%. Finally, in sample set 5, 6.7% positive specimens were detected following PK addition. The increased sensitivity affected exclusively samples with Ct values > 33. Even in presence of PK, detection rates declined when Ct values were above 35, close to the detection limit of this test. Furthermore, PK treatment eliminated invalid results which represented 23.3% (28/120) results in sample sets 4 and 5 together. We propose that PK breaks down specimen-derived inhibitors, whose impact is greater when viral RNA levels are limited, affecting the IC amplification much more than the viral gene amplification.
In our experience, samples from newly infected patients often result in the detection of the three viral genes with C
t values lower than 25, suggesting a high viral titer. We have demonstrated that our protocol accurately detects specimens with low viral loads, detected at C
t values up to 36, and that detection rate declines beyond these levels. Interestingly, the majority of low viral load specimens are detected uniquely with the N gene using this assay. One possible explanation is a higher sensitivity of the N gene primers. All
coronaviruses do nested transcription and generate couples of subgenomic mRNAs coding for different structural viral proteins. While the N gene sequence is present in all subgenomic mRNAs, E and RdRP are less represented [
3]. We propose that specimens with only detected N gene of C
t values > 30 could be either non-infectious viral particles [
9] or fragmented viral genomes that are no longer infectious. Finally, the majority of specimens with very low viral loads in this study are derived from patients in late stages of disease, or in the convalescent phase that were subject to Covid-19 detection for infection control purposes as mandated by the provincial health authorities in the province of Quebec. An abundant body of literature documents that infectivity from these patients is limited [
10‐
15], and we conclude that missed positive samples using our method with C
t values above 35 have little clinical relevance.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.