key: cord-0778270-jv1g0a85
authors: Erster, Oran; Mendelson, Ella; Levy, Virginia; Kabat, Areej; Mannasse, Batya; Asraf, Hadar; Azar, Roberto; Ali, Yaniv; Shirazi, Rachel; Bucris, Efrat; Bar-Ilan, Dana; Mor, Orna; Mandelboim, Michal; Sofer, Danit; Fleishon, Shai; Zuckerman, Neta S.
title: Rapid and High-Throughput Reverse Transcriptase Quantitative PCR (RT-qPCR) Assay for Identification and Differentiation between SARS-CoV-2 Variants B.1.1.7 and B.1.351
date: 2021-10-06
journal: Microbiol Spectr
DOI: 10.1128/spectrum.00506-21
sha: f2ab385f6f2b951f8a5e9a2c637c4e5ad7fb60ec
doc_id: 778270
cord_uid: jv1g0a85

Emerging SARS-CoV-2 (SC-2) variants with increased infectivity and vaccine resistance are of major concern. Rapid identification of such variants is important for the public health decision making and to provide valuable data for epidemiological and policy decision making. We developed a multiplex reverse transcriptase quantitative PCR (RT-qPCR) assay that can specifically identify and differentiate between the emerging B.1.1.7 and B.1.351 SC-2 variants. In a single assay, we combined four reactions—one that detects SC-2 RNA independently of the strain, one that detects the D3L mutation, which is specific to variant B.1.1.7, one that detects the 242 to 244 deletion, which is specific to variant B.1.351, and the fourth reaction, which identifies the human RNAseP gene, serving as an endogenous control for RNA extraction integrity. We show that the strain-specific reactions target mutations that are strongly associated with the target variants and not with other major known variants. The assay’s specificity was tested against a panel of respiratory pathogens (n = 16), showing high specificity toward SC-2 RNA. The assay’s sensitivity was assessed using both in vitro transcribed RNA and clinical samples and was determined to be between 20 and 40 viral RNA copies per reaction. The assay performance was corroborated with Sanger and whole-genome sequencing, showing complete agreement with the sequencing results. The new assay is currently implemented in the routine diagnostic work at the Central Virology Laboratory, and may be used in other laboratories to facilitate the diagnosis of these major worldwide-circulating SC-2 variants. IMPORTANCE This study describes the design and utilization of a multiplex reverse transcriptase quantitative PCR (RT-qPCR) to identify SARS-COV-2 (SC2) RNA in general and, specifically, to detect whether it is of lineage B.1.1.7 or B.1.351. Implementation of this method in diagnostic and research laboratories worldwide may help the efforts to contain the COVID-19 pandemic. The method can be easily scaled up and be used in high-throughput laboratories, as well as small ones. In addition to immediate help in diagnostic efforts, this method may also help in epidemiological studies focused on the spread of emerging SC-2 lineages.

variant B.1.351 contains 18 unique mutations, compared with the original Wuhan strain, it is practically impossible to detect all of them in one quantitative PCR (qPCR) assay. Thus far, the Thermo Fisher SC-2 detection kit (CAT CCU002) has been utilized to identify suspected B.1.1.7 samples (3) . One of the reactions in this kit is directed to the viral spike gene and is negative when the template sequence contains the 69 to 70 deletion, which is one of the B.1.1.7 variant mutations. However, this deletion was also detected independently, is not unique to the B.1.1.7 variant, and can therefore often be misleading. Moreover, the absence of the S reaction in this assay may result from inhibition of this reaction and therefore may not necessarily indicate the presence of the deletion. Variant B.1.1.7 was first reported in the United Kingdom on September 2020 and by December 2020 became the dominant strain in the country (2) . Its increased infectivity led to its rapid spread, with severe consequences on public health and the global economy (4) . Likewise, from its first detection in Israel on December 23, this variant now comprises over 90% of the positive cases (O. Erster and N. Zuckerman, unpublished data).

An additional SC-2 VOC is the B.1.351 variant, which contains both B.1.351-unique mutations and mutations that are present in other notable VOC, such as the SGF deletion in the nsp6 gene and the N501Y substitution in the spike gene that is identified also in B1. 1.7 (5) . Like variant B.1.1.7, this variant was found to be more infectious than the wild-type (WT) strain. In vitro studies also show that it has increased resistance to serum of recovered and vaccinated patients, thereby posing a serious threat to the efficacy of current vaccination campaigns (6) (7) (8) .

The emergence of these two variants, as well as other recently emerging SC-2 strains, necessitates constant improvement of the diagnostic tools used to combat the COVID-19 pandemic. In addition to performing rapid, sensitive, and specific detection of the viral RNA, diagnostic tests are now required to differentiate between circulating strains to provide valuable epidemiological data for policy decision making. The commercial assays developed recently, (http://www.kogene.co.kr/eng/sub/product/covid -19.asp) detect specific key mutations. However, recent studies showed that in spite of sharing such mutations, variants might differ in their antibody resistance capacity, prompting development of variant-specific assays (5, 9) . These findings highlight the need to determine the identity of circulating strains, not only specific mutations. By designing a multiplex PCR assay that positively detects the presence of unique mutations that are strongly associated with variants B.1.1.7 and B.1.351, rapid, economical, high-throughput screening can be performed, enabling robust and specific identification of these variants in SC-2-positive clinical samples.

In this report, we describe a differential RT-qPCR assay that detects the presence of mutations strongly associated with variants B.1.1.7 and B.1.351. We demonstrate that implementation of this novel multiplex assay allows a sensitive and specific detection of SC-2 RNA, with the advantage of variant differentiation in a single assay.

Development of differential COV19 VOC RT-qPCR. Analysis of mutations characteristic of variants B.1.1.7 and B.1.351 showed that some of them were present in both, such as the SGF deletion in nsp6 (positions 11285 to 11294 in sequence NC_045512) or the N501Y substitution in the spike protein gene (nucleotide position 23094 in sequence NC_045512). Others, such as the 69 to 70 deletion and N501Y substitution, developed independently and were detected in samples not classified as variant B.1.1.7 (see Fig. S1 in the supplemental material). On the other hand, the N gene D3L mutation is strongly associated with variant B.1.1.7 and is not associated with other currently dominant variants. Likewise, the D215G mutation and the S gene 242 to 244 deletion are strongly associated with variant B.1.351 and have not been identified in other variants thus far (Fig. S2.) . Therefore, these two regions were selected for the multiplex reaction design.

Design of the variant B.1.1.7-specific reaction. The 59p region of the variant B.1.1.7 SC-2 N gene contains a complete codon substitution, translated into D3L amino acid substitution. Additionally, there is a single "A" deletion in this region, 5 bases upstream of the N gene start codon (Fig. 1) . A selective primer was designed accordingly, to specifically detect the mutated variant. The reverse primer and the probe were based on the CDC N1 reaction (https://www.cdc.gov/coronavirus/2019-ncov/lab/ rt-pcr-panel-primer-probes.html), as detailed in Table 1 . The B.1.1.7-specific reaction was combined with an inclusive reaction based on the E-sarbeco qPCR described by Corman et al. (10) , with minor modifications (Table 1 ). This reaction was designed to detect all known CoV19 clades, thereby serving as an indicator for the presence of SC-2 RNA, regardless of the variant type. The combined assay was then tested using sequence-verified samples of the Wuhan clade SC-2 (WT SC-2) and a sequenced lineage B.1.1.7 sample (Fig. 1 ). The Wuhan clade samples were either negative for the N reaction or gave a very weak signal, approximately 10 to 15 cycles apart from the E reaction signal ( Fig. 1 and 2 ). This reaction was termed N gene D3L reaction-N D3L .

Design of the variant B.1.351-specific reaction. This variant contains two unique mutations in the spike gene, D215G and a deletion at amino acid position 242 (nucleotide positions 22281 and 22289 in sequence NC_045512). A specific reaction was developed, based on these mutations, which detects samples of lineage B.1.351, as described. The forward primer was designed to anneal to the substituted nucleotide at its 39p end. In order to increase the primer selectivity, a mismatched base was inserted 10 bp downstream to the 59p end. The reverse primer was designed to complement the region of the 242 to 244 deletion (Fig. 3 ). This reaction resulted in a negative, or very weak, signal when testing WT samples and a clear signal when testing samples previously sequenced and identified as belonging to the B.1.351 clade ( Fig. 2 and 3 ). This reaction was then termed the S 242 reaction. The virus of the first B.1.351 patient in Israel was cultured at the Israel Central Virology Laboratory (CVL), and the culture-derived RNA was used to calibrate the reaction. Using this assay, 64 samples were identified as variant B.1.351 and were confirmed by sequencing to contain the D215G and S 242 mutations (Table S2) . Generation and evaluation of the multiplex SC-2 VOC assay. The two reactions were combined with the E target reaction and a reaction targeting the RNase P gene (https:// www.cdc.gov/coronavirus/2019-ncov/lab/rt-pcr-panel-primer-probes.html), which serves as an endogenous control for RNA extraction. The multiplex reaction was then tested for sensitivity using serial dilutions of in vitro transcribed RNA containing the target sequences of all four reactions. The analytical limit of detection (LOD) was determined as the minimal number of target copies detected in all three repeats of the test. As detailed in Fig. S3 , the calculated analytical limit of detection for the three viral targets was as follows: 35 copies per reaction for the E-sarbeco reaction, 13.5 copies per reaction for the N D3L , and 15.3 copies per reaction for the S 242 reactions. The analytical LOD for the hRNAse P control reaction was 25 copies per reaction. This sensitivity was accomplished using a reaction that takes less than 70 min, thereby shortening the entire time from sample to answer.

The specificity of the assay was evaluated with a panel of respiratory pathogens, as shown in Fig. S4 and detailed in Table S3 . Some of the samples were derived from tissue culture and were not taken directly from clinical samples and were therefore negative for RNase P. However, they all tested positive for their designated pathogen RNA prior to the specificity test being run. Examination of RNA samples from 17 different pathogens showed no cross-detection, including other human coronavirus species (Fig. S4 , Table S3 ).

The multiplex SC-2 VOC assay was then tested using sequenced samples, and the results of the PCR were compared with the sequencing analysis for each sample. As detailed in Table S1 Table S1 . Likewise, 64 samples identified as containing the S 242 mutation by the multiplex assay, were confirmed to be of variant B.1.351 by the whole-genome sequencing (WGS) analysis (Table 2 ). These results suggest that the PCR assay can be used with a high degree of certainty to identify these variants ( Table 2) . For some samples, other mutations were either absent or inconclusive ("N" read in the WGS analysis), but these were less than 10% of the samples in each lineage (Table 2) . Notably, in some cases, often when samples with a high viral RNA concentration were tested, weak background signals were observed in the PCR assay (Fig. 2) . In some of these samples, a weak signal of the S 242 reaction was detected in non-B.1.351, such as WT or B.1.1.7 samples, and a weak N D3L reaction was observed in some B.1.351 and WT samples (Fig. 2) . In such cases, the E target reaction was used as a reference to evaluate the mutation-specific reaction (either S 242 or N D3L ). A C q difference of up to 3 to 4 cycles was considered a true signal, while a C q difference of 8 or more was considered a background signal. The C q difference between the E gene target signal and the background signals, as well as their fluorescence intensity and curvature, clearly indicated that the signal does not reflect the presence of the variantspecific mutations but, rather, reflects a low-affinity binding of the primers to the viral RNA, as shown in Fig. 2 . The decision matrix based on the different combinations of the reaction is detailed in Table 3 To confirm the accuracy of the multiplex reaction, complete genome sequencing of 122 clinical samples was performed, with complete agreement with the qPCR results. For variant B.1.1.7, over 1,000 samples were examined by both the new qPCR assay and by whole-genome sequencing, with a complete match. These results demonstrate that the new multiplex assay described here can be used as a rapid and reliable approach for primary classification of SC-2 B.1.1.7 and B.1.351 variants.

The emergence of new, more contagious, and potentially antigenically different SC-2 lineages poses an urgent need to adjust rapid detection methods to meet public health-related needs. To meet these needs, we developed a multiplex RT-qPCR assay that can distinguish between three SC-2 lineages. The assay is rapid (;1 h PCR assay The interpretation for each combination of results is detailed. 2, negative; 1, positive; DE-N D3L . 6, the difference in the C q values between the E reaction and the N D3L reaction is less than 6 cycles; DE-N D3L . 7, the difference in the C q values between the E reaction and the N D3L reaction is larger than 7 cycles; DE-S 242 , the difference in the C q values between the E reaction and the S 242 reaction is less than 5 cycles. time) and is suitable for high-throughput rapid screening. This is in contrast to Sanger sequencing or NGS, which are more informative but are far more expensive, take significantly more time, and cannot be scaled up easily. Since the beginning of the COVID-19 pandemic, several rapid tests detecting the presence of SC-2 RNA or proteins were implemented in wide-scale testing (11) . However, the capacity to distinguish between different lineages in less than 2 h is currently possible only using qPCR. SC-2 variant B.1.1.7 contains numerous synonymous and nonsynonymous mutations, of which the spike gene mutations 69-70del, N501Y, and P681H received most attention due to their potential effect on virus infectivity (12, 13) . For diagnostic purposes, however, the N501Y mutation is not variant-specific, as it was identified in several variants other than variant B. 1.1.7, such as B.1.351 and the P.1 variant (14) . The D3L substitution in the N gene used in our assay is specific to variant B.1.1.7 and was not reported in other major SC-2 lineages. Although this mutation can occur independently of other characteristic mutations, such as N501Y, its presence strongly suggests that the examined sample is the B.1.1.7 variant. Likewise, the reaction that identifies the variant B.1.351 targets mutations that are strongly associated with this variant-D215G and the triple deletion of amino acids 242 to 244 (Fig. S4) . The combinations of these two reactions therefore provide a reliable tool to identify each of these two variants with high confidence. The agreement between the multiplex PCR results and the WGS analysis performed on over 250 samples suggested that this assay is a reliable tool to rapidly classify suspected samples as B.1.1.7, B.1.351, or neither. It also allows us to determine the integrity of the sampling and extraction procedure using the control hRNAse P reaction.

In order to increase the range of the new assay, the SC-2 inclusive reaction targeting the E gene (10) was combined, thereby enabling detection of the viral RNA independently of the strain examined.

A few commercial kits partially address the detection of these SC-2 variants, but they target general mutations, such as N501Y and E484K in the spike protein, and not variant-specific mutations, such as the N protein D3L substitution or the spike protein 242 to 244 deletion (https://www.seegene.com/assays/rp, http://www.kogene.co.kr/ eng/sub/product/covid-19.asp).

The emergence of novel SC-2 variants with increased infectivity and increased resistance to current vaccines may significantly impair global large-scale detection and vaccination efforts (15) . Moreover, it has been shown that different variants having some identical mutations in the spike coding sequence still have different infectivity and vaccine resistance characteristics due to their different sets of additional mutations (5, 9; M. Mandelboim, unpublished data). As a result, the pressing need to improve detection methods accordingly requires constant adjustments. Such diagnostic tests should not only detect the presence of the viral RNA with high specificity and sensitivity, but also provide information on variant identity. An additional consideration is the relatively high cost of commercial kits and the need to perform complex interpretation of the results to determine the possible sample identity. The execution and analysis of the assay described here are simple and relatively inexpensive compared with current commercial kits. Implementation of molecular assays such as our multiplex qPCR assay will improve SC-2 diagnosis and contribute to the ongoing efforts to contain the COVID-19 pandemic.

Design of VOC-specific qPCRs. Analysis of SC-2 sequences and primer simulations were performed using the Geneious software package and the NCBI BLAST analysis tools (https://blast.ncbi.nlm.nih.gov/). SC-2 sequences were obtained from the GISAID initiative website (https://www.gisaid.org/) and analyzed using the Geneious software. All primer and probe sequences are detailed in Table 1 .

Processing of SC-2 clinical samples. Nasopharyngeal swab samples suspected to contain SC-2 in viral transport medium (VTM) were inactivated by heating at 70°C for 30 min or, if intended to be used for culturing, were inactivated by addition of 200 ml lysis buffer (Zymo Research, to 200 ml VTM. Total RNA extraction was performed with either the Roche MagNA Pure 96 system or the PSS MagLEAD instrument. The eluted RNA was stored at 280°C for further use or used immediately thereafter for the PCR test.

Design and synthesis of in vitro transcribed standard RNA segments. In order to establish the analytical limit of detection (LOD) and obtain standard controls for WT and mutant target sequences, genomic regions, including the E, S 242 , S RBD , N, and RNASE P, were amplified using primers that contain the T7 promoter minimal sequence (Table 4 ) with the MyTaq one-step RT-PCR kit. The resulting PCR products were transcribed in vitro to RNA using the T7 MEGAscript kit according to the manufacturer's instructions (Thermo Fisher). The in vitro transcribed RNA was purified, and its concentration was determined using a Nanodrop spectrophotometer and stored at 280°C.

RT-qPCR. RT-qPCR mix was prepared with the Meridian (formerly Bioline) SensiFast one-step mix. Initial optimization was performed by setting an annealing temperature gradient. Following optimization of the reaction conditions, the final mix concentration was determined. The reaction mix was assembled as follows: SensiFast Probe Lo-ROX one-step, 12 Each PCR run included at least one positive control of SC-2 RNA, and one negative control of H 2 O, as is the standard for any PCR test performed in the CVL.

Sanger sequencing of suspected samples. In order to identify mutations of the variants of interest in suspected samples, several rapid sequencing reactions of the Spike gene were designed and implemented. This enabled the first identification of the B.1.1.7-related mutations 69-70Del, 144Del, N501Y, S982A, and D1118H and B.1.351-related mutations D215G, 242Del K417N, N501Y, and E484K. Primer sets for the rapid sequencing reactions of the spike and nucleocapsid (N) genes are detailed in Table 4 .

Endpoint PCRs were performed with a MyTaq one-step RT-PCR kit (Meridian) according to the manufacturer's instructions. The resulting PCR products were analyzed using agarose gel electrophoresis and sequenced using the ABI 3500 Bioanalyzer.

Next-generation whole-genome sequencing of clinical samples. A COVID-seq kit was used for library preparation as per the manufacturer's instructions (Illumina). Library validation and mean fragment size were determined using TapeStation 4200 with a DNA HS D1000 kit (Agilent). Libraries were pooled, denatured, and diluted to 10 pM and sequenced on a NovaSeq instrument (Illumina).

Bioinformatic and phylogenetic analysis. Fastq files underwent quality control using FastQC (www .bioinformatics.babraham.ac.uk/projects/fastqc/) and MultiQC (16) , and low-quality sequences were filtered using Trimmomatic (PMID: 24695404). Mapping to the SARS-CoV-2 genome (GenBank accession number NC _045512.2) was performed with the Burrows-Wheeler Aligner MEM algorithm (BWA-MEM) (PMID: 19451168). The SAMtools suite (PMID: 19505943) was used to filter unmapped reads, and sort and index bam files. Consensus fasta sequences were constructed for each sample using iVar (https://andersen-lab.github.io/ivar/ html/manualpage.html), with Ns inserted in positions with a sequencing depth lower than 5. Sequences were aligned with the SARS-CoV-2 reference sequence (GenBank accession number NC_045512.2) with MAFFT (17) , and mutation analysis was done with a custom R code using the Bioconductor package seqinr (https://cran.r-project.org/web/packages/seqinr/).

Supplemental material is available online only. SUPPLEMENTAL FILE 1, PDF file, 2.5 MB.

We thank the members of the Israel Central Virology Laboratory for their valuable help in this work. In the left column are the gene name and the amino acid position of mutations within the amplified sequence. The forward (Fwd) and reverse (Rev) primer sequences are detailed in the middle columns. The expected PCR product size is shown in the right column. All primers were designed in this study.

Estimated transmissibility and impact of SARS-CoV-2 lineage B.1.1.7 in England

Assessing transmissibility of SARS-CoV-2 lineage B.1.1.7 in England

Increased mortality in community-tested cases of SARS-CoV-2 lineage B.1.1.7

Case fatality risk of the SARS-CoV-2 variant of concern B.1.1.7 in England

Antibody resistance of SARS-CoV-2 variants B.1.351 and B.1.1.7

Emergence and rapid spread of a new severe acute respiratory syndrome-related coronavirus 2 (SARS-CoV-2) lineage with multiple spike mutations in South Africa

Multiple SARS-CoV-2 variants escape neutralization by vaccine-induced humoral immunity

Network for Genomic Surveillance in South Africa; COMMIT-KZN Team. 2021. Escape of SARS-CoV-2 501Y.V2 from neutralization by convalescent plasma

Neutralizing response against variants after SARS-CoV-2 infection and one dose of BNT162b2

Van den Bruel A, Cochrane COVID-19 Diagnostic Test Accuracy Group. 2021. Rapid, point-of-care antigen and molecularbased tests for diagnosis of SARS-CoV-2 infection

Immunoinformatic analysis of structural and epitope variations in the spike and Orf8 proteins of SARS-CoV-2/B.1.1.7

Structure-function analyses of new SARS-CoV-2 variants B.1.1.7, B.1.351 and B.1.1.28.1: clinical, diagnostic, therapeutic and public health implications

Resurgence of COVID-19 in Manaus, Brazil, despite high seroprevalence

On the road to ending the COVID-19 pandemic: Are we there yet?

MultiQC: summarize analysis results for multiple tools and samples in a single report

MAFFT multiple sequence alignment software version 7: improvements in performance and usability

The usage of clinical samples was performed in accordance with approval number 7045-20-SMC, granted by the Sheba Medical Center Committee for clinical trials involving SC-2-related samples in the Sheba Medical Center.We declare no conflicts of interests.