Rapid And high throughput RT-qPCR assay for identification and differentiation between SARS-CoV-2 variants B.1.1.7 and B.1.351

Design of VOC-specific qPCR reactions

Analysis of SC-2 sequences and primer simulations were performed using the Geneious software package (https://www.geneious.com/) and the NCBI BLAST analysis tools (https://blast.ncbi.nlm.nih.gov/). SC-2 sequences were obtained from 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 media (VTM) were inactivated by heating at 70°C for 30 minutes, or, if intended to be used for culturing, inactivated by addition of 200µl lysis buffer (Zymo research, https://www.zymoresearch.com/products/dna-rna-lysis-buffer) to 200µl VTM.. Total RNA extraction was performed with either the Roche MagNA pure 96 system (https://lifescience.roche.com/), or the PSS MagLEAD instrument (http://www.pss.co.jp/english/). The eluted RNA was stored at -80°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, S₂₄₂, S_RBD, N and RNASE P were amplified using primers that contain the T7 promoter minimal sequence (Table 2) 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, https://www.thermofisher.com/). The In-vitro transcribed RNA was purified and its concentration was determined using Nanodrop spectrophotometer and stored at -80°C.

RT-qPCR

RT-qPCR mix was prepared with the Meridian (formerly Bioline) SensiFast one-step mix (https://www.bioline.com/). 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: 12.5µl, E Sarbeco-F1b: 400nM, E-Sarbeco-R: 400nM, E-Sarbeco Probe: 200nM, 21257VOC Fwd: 750nM, CDC N1 Rev: 750nM, CDC N1 probe: 300nM, 22201B SA Fwd: 600nM, 22238 SA Rev: 800nM, 22230 probe: 300nM, RNASE P Fwd: 300nM, RNASE P Rev: 300nM, RNASE P probe: 200nM, RT Enzyme: 0.2µl, RNAse inhibitor: 0.2µl, ddH₂O – to a final volume of 1µl.

The amplification was performed in Bio-Rad CFX96 thermal cycler (https://www.bio-rad.com) using the following conditions:

(a) 45°C for 10’, (b) 95°C for 2’:20’’, 45X[(C) 95°C for 4’’, (d) 62.2°C for 28’’]. Fluorescence was recorder at each cycle during the annealing and extension step (62.2°C for 28’’). The reaction data were analyzed using the Bio-Rad CFX Maestro software.

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 Nucleocapside (N) genes are detailed in Table 2.

Endpoint PCR reactions were performed with MyTaq One-step RT-PCR kit (Meridian, https://www.bioline.com/mytaq-mix.html) according to the manufacturer’s instructions. Resulting PCR products were analyzed using agarose gel electrophoresis and sequenced using the ABI 3500 Bioanalyzer.

Next Generation Whole genome sequencing of clinical samples

COVID-seq kit was used for library preparation as per manufacturer’s instructions (Illumina, https://www.illumina.com/). Library validation and mean fragment size was determined by TapeStation 4200 via DNA HS D1000 kit (Agilent, https://www.agilent.com/). Libraries were pooled, denatured and diluted to 10pM and sequenced on NovaSeq (Illumina).

Bioinformatic and phylogenetic analysis

Fastq files underwent quality control using FastQC (www.bioinformatics.babraham.ac.uk/projects/fastqc/) and MultiQC [10] and low-quality sequences were filtered using trimmomatic [PMID: 24695404]. Mapping to SARS-CoV-2 genome (NC_045512.2) was performed with BWA mem [PMID: 19451168]. SAMtools suite [PMID: 19505943] was used to filter unmapped reads, 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 sequencing depth lower than five. Sequences were aligned with the SARS-CoV-2 reference sequence (NC_045512.2) with MAFFT [11] and mutation analysis was done with a custom R code using Bioconductor package seqinr [citation info in: https://cran.r-project.org/web/packages/seqinr/citation.html].

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 mutual to both, like the SGF deletion in nsp6 (positions 11285-11294 in sequence NC_045512), or the N501Y substitution in the spike protein gene (nucleotide position 23,094 in sequence NC_045512). Others, such as the 69-70 deletion and N501Y substitution, developed independently and were detected in samples not classified as variant B.1.1.7 (Supplementary Figure S1). 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 242-244 deletion are strongly associated with variant B.1.351 and were not identified in other variants thus far (Supplementary Figure S2.). Therefore, these two regions were selected for the multiplex reactions design.

Design of variant B.1.1.7 specific reaction

The 5’p region of variant B.1.1.7 SC-2 N gene contains a complete codon substitution, translated into D3L a.a. substitution. Additionally, there is a single “A” deletion in this region, 5 bases upstream to the N gene start codon (Figure 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. (12), with minor modifications, as detailed in Table 1. This reaction was designed to detect all known CoV19 clades, thereby serving as an indicator for the presence of SC-2 RNA in the examined sample. The combined assay was then tested using WT and sequence-verified samples identified as belonging to lineage B.1.1.7 (Figure 1). All samples identified as belonging to clade B.1.1.7 were detected using the new multiplex assay. WT samples were either negative for the N reaction, or gave a very weak signal, approximately 10-15 cycles apart from the E reaction signal (Figure 1). This reaction was termed N gene D3L reaction.

• Download figure
• Open in new tab

Figure 1. Design of variant B.1.1.7-specific reaction.

A specific primer was designed based on the single nucleotide insertion and the codon substitution in the 5’p of the SC-2 N gene, both circled and highlighted. The specific “UK VOC” reaction was combined with an inclusive E-sarbeco reaction (Corman et al. 2020) that detects all SC-2 variants. The presence of the B.1.1.7 RNA in the assay results in two clear amplification curves (A). A sample containing a non-B.1.1.7 variant sequence results in a faint amplification, with a Cq different larger than eight between the E and the N curves (B), or a negative signal for the N reaction (C).

Design of 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 3’p end. In order to increase the primer selectivity, a mismatched base was inserted 10 b.p. downstream to the 5’p end. The reverse primer was designed to complement the region of the 242-244 deletion (Figure 2). This reaction resulted in a negative, or a 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 (Figure 2). This reaction was then termed S₂₄₂ reaction.

• Download figure
• Open in new tab

Figure 2. Design of variant B.1.351-specific reaction.

Two primers were designed, to specifically complement the sequences containing the D215G mutation, and the 242-244 deletion. The resulting reaction gives a clear signal of the S₂₄₂ reaction with sample belonging to clade B.1.351 (A), and a negative, or a faint signal with a WT sample (B). The inclusive E-sarbeco reaction detects the presence of both the WT and clade B.1.351 samples.

Generation and evaluation of a multiplex SC-2 VOC reaction

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) that 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. As detailed in Supplementary Figure S3, the calculated analytical limit of detection for the three viral targets was as follows: 29 copies per reaction for the E-sarbeco reaction, 26 copies per reaction for the N_D3L, and 39 copies per reaction for the S₂₄₂ reactions. This sensitivity was accomplished using a reaction that takes less than 70 minutes, thereby shortening the entire time from sample to answer. Specificity test performed with a panel of respiratory pathogens showed no cross-detection of pathogens other than SC-2, except for SARS CoV-1 (SC-1), which was detected, as expected, by the E-sarbeco reaction. Nucleic acid samples of the following pathogens were tested: Chlamydia pneumonia, Rhinovirus, Influenza A, Influenza B, Swine influenza, RSV A, RSV B, Parainfluenza 3, HMPV, Adenovirus, CoV OC43, CoV 229E, CoV HKU1, CoV NL63, SARS-1 and MERS CoV.

The multiplex SC-2 VOC assay was then tested using sequence-verified samples, and successfully detected all samples in full agreement with the sequencing results, thereby confirming the accuracy of the assay. Notably, in samples with high viral RNA concentration, a weak signal of S₂₄₂ variant was observed. However, the Cq difference between the E gene target signal and the S₂₄₂ reaction clearly indicated that the signal does not reflect the presence of the D215G and 242-244 deletion mutations, but rather a low-affinity binding of the primers to the RNA.

Confirmatory Sequencing of suspected samples

The first three samples suspected as variant B.1.1.7 (numbered 5824, 6021 and 7075) were negative for the S reaction in the Thermo-Fisher SARS-COV-2 test, were subjected to further examination in the Israel Central Virology Laboratory (CVL). Focused Sanger sequencing of the S gene region of samples 5824 and 6021 showed the presence of the 69-70 deletion, but the B.1.1.7 variant-associated mutations 144del and N501Y substitution were absent. They both contained, however, a unique substitution - N439K, which is not associated with the B.1.1.7 variant (Supplementary figure S4). Sanger sequencing of the spike gene region from sample 7075, contained the following mutations, all of which characteristic of the B.1.1.7 variant: 69-70 deletion, 144 deletion, N501Y, S982A and D1118H (Supplementary figure S5). Subsequent complete genome analysis via next generation sequencing (NGS) of that sample confirmed the presence of all the defining mutations of the B.1.1.7 variant, thereby corroborating the initial results of the Sanger sequencing. This was the first known case of the B.1.1.7 variant in Israel.

Samples suspected as the B.1.351 variant were initially examined by Sanger sequencing of two regions of the spike protein: the 242-244 deletion, which is unique to this variant, and the receptor binding domain (RBD, nucleotide positions 22,493-23,218). Out of 20 suspected samples, four contained the following mutations in the spike gene: D215G, 242-244 deletion, K417N, E484K, and N501Y, all associated with the B.1.351 variant (Supplementary figure S6). Subsequent complete genome analysis via NGS confirmed the presence of the B.1.351 defining mutations.

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 complete match. These results demonstrate that the new multiplex assay described herein can be used as a rapid and reliable approach for primary classification of SC-2 B.1.1.7 and B.1.351 variants.