A qualitative RT-PCR assay for the specific identification of the SARS-CoV-2 B.1.1.529 (Omicron) Variant of Concern


                  Objectives
                  The aim of this study was to develop a RT-PCR assay for the specific detection of the SARS-CoV-2 Omicron Variant of Concern (VOC) as a rapid alternative to sequencing.
               
                  Methods
                  A RT-PCR was designed in silico and then validated using characterised clinical samples containing Omicron (both BA.1 and BA.2 lineages) and the Omicron synthetic RNA genome. As negative controls, SARS-CoV-2 positive clinical samples collected in May 2020, and synthetic RNA genomes of the isolate Wuhan Hu-1 and of the Alpha (B.1.1.7), Beta (B.1.351), Gamma (P.1), Kappa (B.1.617.1), Iota (B.1.526), Epsilon (B.1.429) and Delta (B.1.617.2) SARS-CoV-2 VOC were used.
               
                  Results
                  Experiments performed using as templates the synthetic RNA genomes demonstrate the high specificity of the PCR-method for the SARS-CoV-2 Omicron. Despite the synthetic RNAs were used at high copy numbers, specific signal was mainly detected with the Omicron synthetic genome. Only a non-specific late signal was detected using the Alpha variant genome, but these results were considered negligible as Alpha VOC has been replaced by the Delta and it is not circulating anymore in the world. Using our method, we confirmed the presence of Omicron on clinical samples containing this variant but not of other SARS-CoV-2 lineages. The method is highly sensitive and can detect up to 1 cp of the Omicron virus per µl.
               
                  Conclusions
                  The method presented here, in combination with other methods in use for detection of SARS-CoV-2, can be used for an early identification of Omicron.
               


Introduction
A novel SARS-CoV-2 variant (Pango lineage B.1.1.529) was first reported to the World Health Organization (WHO) (https://www.who.int/news/item/26-11-2021-classification-of-omicron-(B.1.1.529)-sars-cov-2-variant-of-concern) from the Republic of South Africa on 24 November 2021 and was classified two days later as a VOC with the name "Omicron" (for a simplified reference) [1]. First identified in countries in the southern part of Africa, the variant has quickly appeared worldwide. Omicron has already become the dominant variant in Europe by replacing the Delta variant, and is replacing other variants in other continents. Omicron is characterised by more than 50 nucleic acid changes (including deletions). Most of these mutations reside in the S gene encoding for the spike glycoprotein [2]. The spike plays a central role in the entry of the virus into human cells through interaction with the human ACE2 receptor [3,4], and is the main antigenic component of mRNA-and viral vector-based vaccines [5][6][7][8]. Consistent evidence indicates that by acquiring these mutations, the Omicron variant infects the cells more efficiently, replicates faster and spreads at superior rate in comparison to other previously emerged variants [2,9]. These mutations may also render certain vaccines against SARS-CoV-2 less effective in preventing infection [10][11][12][13][14]. After less than one month (19 December 2021) from its first reporting, 11,615 SARS-CoV-2 genomic sequences had been deposited in GISAID [15], 10,755 of them were flagged as "complete" and with less than 5% undetermined bases. After two months from its first reporting, according to the Pango nomenclature, four Pango lineages were defined as Omicron B.1.1.529, BA.1, BA.2 and BA.3. By screening the S gene of all Omicron-flagged complete sequences, we identified unique and Omicron-specific sequences of B.1.1.529, BA.1 and BA.2 (which together include more than 99% of Omicron sequences) to be used for the development of a novel real-time polymerase chain reaction (RT-PCR) based method for the specific identification of the Omicron.

Materials and methods
2.1. In silico design of the RT-PCR method SARS-CoV-2 genomic sequences deposited in GISAID, flagged as "complete" and with less than 5% undetermined bases have been downloaded and analysed with Nextclade [16] (Supplementary Material S1). Only mutations present in at least 80% of the analysed sequences were retained to define a set of commonly shared mutations. The S gene, in which most of these mutations occur, was manually inspected to look for genomic regions not longer than 250 bp and with at least 6 mutations. The National Center for Biotechnology Information (NCBI) Reference Sequence NC_045512.2 was used as reference [17]. A region (NC_045512.2:22950-23150) was identified with these features and fed into Primer3Plus [18] to design in silico primers and probes methods (program run with pre-loaded qPCR settings). A candidate method, specific for the genomic region NC_045512.2:22991-23128, was selected. We call this method OMicron METhod (OmMet) [19]. OmMet has been tested in silico by using the thermonucleotideBLAST software [20] with the following parameters: -e 20 -E 20 -l 200 on the selected set of SARS-CoV-2 genomic sequences. A set of Pango lineages [1] consensus sequences was obtained by running an inhouse developed script that parses data retrieved from the Broad Institute COVID CG [21] application programming interface (used consensus_threshold of 0.9). This set was used to check in silico the OmMet specificity. Sequence alignments were produced by using MAFFT online service [22] and alignments' representations were obtained by running the showseq tool of EMBOSS package [23].

Samples analysed
The samples used in this study are listed in Table 2. The five clinical samples assigned as Omicron-positive were from nasopharyngeal swabs from COVID-19 positive patients at the University Hospital of Geneva (HUG). Primary diagnosis of SARS-CoV-2 infection was made by RT-PCR using the Cobas 6800 SARS-CoV-2 RT-PCR assay (Roche, Basel, Switzerland). Furthermore, viral RNA was extracted using the NucliSENS easyMAG extraction kit (BioMérieux, CH) protocol followed by screening for S gene target failure (SGTF) with the TaqPath COVID19 assay (ThermoFisher). Sample with SGTF were further analysed by partial Sanger sequencing of the spike region [24]. Confirmation of the Omicron variant was done by full genome sequencing fusing Illumina NovaSeq and the sequence from 3 samples submitted to GISAID. The viral load expressed in cp/µL was measured by qPCR using the Twist Bioscience Control 48 as calibrant. Three additional clinical samples -two SARS-CoV-2 positive (Wuhan Hu-1 lineage) and one negative nasopharyngeal swabs-were kindly, provided in viral transport media by Dr Elke Wollants (University of Leuven, KUL) in March 2020. Viral RNAs was extracted from these samples using the QIAamp Viral RNA Kit (QIAGEN) according to manufacturer's instructions and used as negative controls in the RT-PCR. Additionally, clinical samples of lineages BA.1 and BA.2 were tested in the Virus Research & Development lab (Denmark), to verify whether OmMet performs efficiently with both. Samples were collected and tested SARS-CoV-2 positive from community testing facilities (Test Centre Denmark) and BioBank Denmark, which from part of the Danish national surveillance program [25]. Viral genomes from throat swabs collected in 1x PBS were isolated using the MagNa Pure 96 nucleic acid extraction system (Roche) with 100 µL elution. The identities of the lineages BA.1 and BA.2 were confirmed by Whole Genome Sequencing (WGS) generated by the Danish COVID-19 Genome Consortium (DCGC) [26] and consensus genomes were deposited in GISAID (gisaid.org). In vitro transcript synthetic RNAs covering the full viral genomes of the SARS-CoV-2 Wuhan Hu-1 lineage and indicated variants ( Table 2) were purchased from Twist Bioscience (South San Francisco, CA, USA) and diluted 10 time in Nuclease free-water supplemented with carrier RNA before use at a calculated concentration of 5 x 10 5 cp/RT-PCR reaction.

RT-PCR assays
The sequences of the oligo's and TaqMan® probe are provided in Table 1. The TaqMan® probe for the OmMet assay was labelled at the 5'-end with the reporter molecule 6-carboxyfluorescein (FAM) and with the QSY-7 succinimidyl ester quencher [QSY™ from Applied Biosystems, CA) at the 3'-end. The QSY is designed for optimal sensitivity and performance in multiplex qPCR experiments, but it can be replaced by dark quencher probes, such as Black Hole Quencher (BHQ™) probes without changing probe sequences. OmMet targets a region (138 bp) located in the S gene (region NC_045512.2:22991-23128). The 3' mismatches on the forward primer and on the probe were chosen to increase the specificity of the assay. The assay targeting the E gene (E-Sarbeco) was used as unspecific control assay to determine the presence of the SARS-CoV-2 virus and estimate the viral load [27].
The reaction mixtures for the RT-PCR experiments were prepared as follows: for each 25 µL reaction, 5 µL of the extracted RNA were added to 20 µL of the reaction mix containing 10.75 µL of H 2 0 (RNAse free), 1 µL of primers and probe mix (forward and reverse primers at final concentration of 900 nM and FAM-QSY labelled probe at a final concentration of 300 nM) and 6.25 µL of MasterMix (TaqPath 1-step RT-qPCR MasterMix 4x, ThermoFisher Scientific). The PCR programme consists of uracil-DNA glycosylase incubation (2 min at 25°C), reverse transcription (15 min at 50°C), TaqMan® activation (2 min at 95°C) and 45 cycles of amplification (3 sec at 95°C followed by 30 sec at 62°C). All RT-PCR reactions were performed in a calibrated QuantStudio 7 Flex Real-Time PCR System (Applied Biosystems, BE); the raw data were analysed with the QuantStudio software (version 1.3) using an automatic threshold setting.
The OmMet assay was also performed using the same primers (Table 1) but with the Luna® Universal Probe One-Step RT-qPCR Kit and the Luna WarmStart® RT Enzyme Mix (New England Biolabs) at two different annealing temperatures (58 °C and 62 °C) to verify the robustness of the assay on BA.1 and BA.2 Omicron samples. Table 1. Primers and TaqMan® probe sequences and relative concentrations used for the E gene and OmMet simplex PCR assays. Omicron-specific nucleic acid base changes are highlighted in red.

Final concentratio n [nM]
Refere nce This paper Omt-R

In silico validation
The initial design of OmMet was carried out by inspecting 305 Omicron genomic sequences deposited between 16/11/2021 and 02/12/2021 in GISAID flagged as complete, and with less than 5% undetermined bases (Supplementary Material S1). The overview of the OmMet target amplicon (within the S gene) compared to the SARS-CoV-2 homologous sequence (Wuhan strain, NCBI Reference Sequence NC_045512.2, nucleotides: 22991-23128 ) is shown in Fig. 1 A. Eight point mutations were found in this region of the Omicron genome in comparison to the reference sequence These mutations lead to amino acid changes as found by align the corresponding fragments of the Omicron and reference Wuhan strain Spike proteins (amino acids 477-522) (Fig. 1 B). Primers and probe were designed considering these unique features present within the Omicron genome ( Fig.1 C). OmMet was tested in silico and found with perfect matching on the selected set of Omicron sequences retrieved from GISAID (Supplementary Material S1). Additionally, OmMet was tested in silico on the consensus sequences of all known Pango lineages [1]. In all the analysed sequences, differences were found in the regions of annealing of both the forward and reverse primers, and of the hybridisation oligo (Alignment in Supplementary Material S2). These in silico results underline the OmMet's ability to detect the Omicron sequences with very high specificity. In silico simulation predicts that the OmMet is highly specific for the detection of Omicron (Fig. 2). About 7% of the 11,615 analysed sequence data set deposited in GISAID and flagged as Omicron have long stretches of N in one or more of the target regions (group a), and can therefore not be detected in silico. Among the remaining ones (group b) more than 97.5% (sum of subgroups b.0 and b.1) have target regions perfectly matching with primers and probes of the OmMet method. Only 66 sequences (subgroup b.3) have not-N mismatches on these target regions, thus representing potential false negatives. Interestingly, 58 sequences (subgroup b.5) flagged as Omicron do not show the typical Omicron nucleic acid changes, and consequently they represent false positives, i.e. sequences wrongly assigned to Omicron. For each group and subgroups, the number of sequences analysed is provided, together with their corresponding percentage calculated on the total number of sequences analysed.

Analytical specificity of the Omicron assay
We evaluated the specificity of the OmMet RT-PCR assay on RNA extracted from five characterized Omicron positive patients and two SARS-CoV-2 positive samples collected at the onset of the pandemic in March 2020 (containing the Wuhan Hu-1 prototype). Together with the OmMet, we also analyzed these samples with the E-Sarbeco method, which is widely used for the detection of SARS-CoV-2 [27]. The presence of Omicron was confirmed in Omicron positive samples by both OmMet and Sarbeco methods with similar Ct values ( Table 2). On the contrary, the clinical samples containing the Wuhan Hu-1 were detected only by the E-Sarbeco method, providing evidence of the high specificity of OmMet for the Omicron variant. One Omicron positive sample was serially diluted (1x, 10x, 100x, 1000x). Despite the dilutions, OmMet was still able to easily detect the virus. A nasopharyngeal sample tested positive with the E-Sarbeco assay with a very low Ct of 12.5 was negative with the Omicron assay, confirming the high specificity of the method.
The specificity of the assay was also verified by testing synthetic RNA covering the full genome of the Wuhan Hu-1, the Alpha, Beta, Delta, Gamma, Kappa, Iota, Epsilon and Omicron at a high concentration (calculated to be around 5 x 10 5 copies per PCR reaction) (as listed in Table 2). Typical amplification plots are illustrated in Fig. 2, the late parasite amplifications observed for the Alpha (and to a lesser extension for the Gamma variant) show a potential limitation of the RT-PCR, which despite the presence of 8 mismatches among the oligos and their targets, still provides a weak signal after many PCR cycles (plateau phase). However, potential false positives with the Alpha variant are not likely to occur as this "de-escalating" variant has been replaced in Europe by the Delta one, which gives a very week and late signal even when present at very high concentration in the RT-PCR assay ( Fig. 3 and Table 2). After the design of OmMet, two novel lineages of Omicron (BA.2 and BA.3) have been described. These two novel Omicron lineages are lacking the mutation A23040G in the S-region which is present in BA.1. As the OmMet was in silico designed using as a template the Omicron BA.1 sequences, the TaqMan probe developed contains this specific mutation (at position 9) (Fig. 1C and Supplementary Material S1). Therefore, there is a potential risk that OmMet would perform less efficiently with BA.2 and BA.3 that do not share this mutation. We decided to challenge OmMet with samples containing the Omicron BA.2 lineage. This part of the study was carried out at the Virus Research & Development laboratory (Statens, Denmark), also for testing the applicability of the method in another laboratory and its robustness when using different reagents and different annealing temperatures for the RT-PCR.
In particular, in addition to the recommended temperature of 62°C, the annealing temperature of 58°C was tested.
The results showed that both BA.1 and BA.2 are detectable with very similar Ct values at both annealing temperatures, although, a superior specificity against the Hu-1, the Alpha and Delta variants was achieved with the recommended annealing temperature of 62°C rather than 58°C ( Table 3).
As expected, the RT-PCR assay is performing with similar efficiency when using different MasterMixes and different concentrations of the primers. representing a limitation when assessing the analytical sensitivity. To estimate the analytical sensitivity of the method, extracted RNA from the sample VD-HUG-36220991v (Table 1) was diluted 10, 100 and 1000 times and the respective Ct values for the S-gene target recorded. For the highest dilution, corresponding to 61 copies per reaction, a Ct value of 30.8 was recorded which suggests an even lower detection limit, knowing that signal up to Ct values of approximatively 37 can be detected.
The analytical limit of detection and linearity of the signal were therefore further studied by recording the Ct values of diluted synthetic Omicron genome in the PCR reaction (Fig. 4). Concentration up to 1 cp/µL was detected by OmMet and the signal was linear between 5 to 50000 cp/reaction showing a sufficient analytical sensitivity of the method and a PCR efficiency of 97.3 %.

Discussion
The epidemic potential of the new SARS-CoV-2 Omicron variant is of considerable concern due to its apparent higher transmissibility and the likely negative effect that many of the mutations it features might have upon natural or vaccine-induced immunity and on our ability to detect the virus. Currently, public health officials mostly rely on the so-called S gene target failure (SGTF) phenomenon to quickly identify Omicron in patient samples [24]. SGTF occurs in commercial assay designed to detect sequences from three SARS-CoV-2 genes -S, N and ORF1ab. Due to the large number of mutations in the S gene of Omicron, assays give positive signal for N and ORF1ab presence, for the new variant, but negative for S and such a test result is interpreted as from a sample containing Omicron. Although much more rapid than sequencing identification, SGTF suffers from lower specificity. For example, the S gene of many rare SARS-CoV-2 variants features a genetic sequence that is not recognised by the assays. In addition, swab samples with lower virus concentration may fail to detect the S target. Furthermore, the S gene of some actual Omicron lineages can in fact be detected by the tests, resulting in a lack of discrimination with the SGTF approach. The method we propose has a very high accuracy, typical for the sequencing approaches, while allowing much shorter turn over times needed for public health response and decision making.
SARS-CoV-2 continues to evolve with its spreading. We provide also evidence of the robustness of OmMet. The method performs equally well with different masterMixes (TaqPath 1-step RT-qPCR MasterMix and the Luna® Universal Probe One-Step RT-qPCR Kit) and it is expected to work with reagents from other providers, too. The RT-PCR assay was tested using primers and probe at 900 nM and 300 nM, and at a lower concentration of 450 nM and 214 nM, respectively, without decrease in performances (data not shown). For a high specificity towards Omicron, we recommend to use the protocol with an annealing temperature of 62 °C.
This method is simple to be settled up and implemented in any biotech/clinical/environmental laboratory that is able to carry out PCR assays , and should aid in globally monitoring the spread of the Omicron variant. The method could be combined with other RT-PCR methods in use for the detection of SARS-CoV-2 (e.g. the E-Sarbeco method [27]) to discriminate between Omicron and other SARS-CoV-2 variants. In this regard, it can also be used as part of a monitoring approach aimed to reveal emerging of novel potential VOCs. In fact, being currently Omicron the predominant variant circulating in the world, negative results obtained with OmMet, but positive to other SARS-CoV-2 methods which recognize conserved elements of the SARS-CoV-2 genome, may indicate that other SARS-CoV-2 variants with mutations in the S gene target region recognized by OmMet primers and probe are emerging.
To the best of our knowledge, when this paper was published as a preprint [19,28], OmMet was the first single one-step RT-PCR assay reported for the specific detection of Omicron. Since then, two other PCR-based assays have been developed for the detection of Omicron in clinical and environmental samples [29,30]. These methods take advantage of other unique features present in the S gene of Omicron (e.g. typical deletions or insertions) and have been described to be highly specific and sensitive for Omicron. However, according to our in silico amplification, the del69-70 based assays as described in [29,30], should not be able to detect the lineage Omicron BA.2, which is currently (end of April 2022) prevalent, as it is missing that particular deletion.
OmMet is suitable to be combined in a multiplex approach with other methods to closely track the evolution of the virus in surveillance programs. In this regard, recent results performed in our laboratory (not shown, but in preparation) and in national reference laboratories [31] show that OmMet equally performed well with RNAs extracted from wastewater samples. Alternatively, Bloemen et al. have recently described a method that combines PCR amplification of the S gene (corresponding to the region coding for amino acids 360 to 588 of the Omicron Spike) with Sanger sequencing [32] to detect VOCs . This method in comparison to the others may unequivocally ascertain the presence of VOCs (including Omicron) in given samples. However, as it uses a sequencing step, it is necessarily more laborious and time consuming in surveillance programs.
In conclusion, this study describes the design of primers and probe to identify the Omicron variant by a RT-PCR. The method has the necessary specificity and sensitivity to discriminate the Omicron variant from other SARS-CoV-2 variants.

Conclusions
The epidemic potential of the new SARS-CoV-2 Omicron variant is of considerable concern and even if whole Genome Sequencing (WGS), or at least complete or partial S gene sequencing remains the best method for characterising a specific variant, we provide here a diagnostic RT-PCR method that can be used for an early identification of Omicron. The analytical sensitivity and specificity were demonstrated in silico, on a (limited) number of clinical containing both BA.1 and BA.2, and on control samples. We feel its sharing with the scientific community is critical and invite control laboratories to use that method to confirm the clinical sensitivity and specificity of this assay on Omicron positive samples.

Declaration of Competing Interests
The authors declare no competing interests.