SARS-CoV-2 Delta Variant N Gene Mutations Reduce Sensitivity to the TaqPath COVID-19 Multiplex Molecular Diagnostic Assay

As the SARS-CoV-2 virus evolves, mutations may result in diminished sensitivity to qRT-PCR diagnostic assays. We investigated four polymorphisms circulating in the SARS-CoV-2 Delta lineage that result in N gene target failure (NGTF) on the TaqPath COVID-19 Combo Kit. These mutations were detected from the SARS-CoV-2 genome sequences that matched with the diagnostic assay results of saliva specimens. Full length N genes from the samples displaying NGTF were cloned into plasmids and assayed using three SARS-CoV-2 qRT-PCR assays. These constructs resulted in reduced sensitivity to the TaqPath COVID-19 Combo Kit compared to the controls (mean Ct differences of 3.06, 7.70, 12.46, and 14.12), but were detected equivalently on the TaqPath COVID-19 Fast PCR Combo 2.0 or CDC 2019_nCoV_N2 assays. This work highlights the importance of genomic sequencing to monitor circulating mutations and provide guidance in improving diagnostic assays.


Introduction
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has continued to evolve through mutations in its genome throughout the coronavirus disease 2019 (COVID- 19) pandemic, leading to the emergence of variants of concern (VOCs) [1,2]. SARS-CoV-2 mutations can alter antibody neutralization, change receptor binding, reduce therapeutic efficacy and reduce the sensitivity of diagnostic assays [3,4]. Although at-home, point of care, and laboratory-based antigen tests are available for rapid detection of the SARS-CoV-2 infection, qRT-PCR remains the gold standard for diagnostic testing [5].
As SARS-CoV-2 has evolved, the qRT-PCR assays used to detect it have improved. Early in the pandemic, the U.S. Centers for Disease Control and Prevention (CDC) issued a set of N gene primers to detect the presence of SARS-CoV-2 by qRT-PCR [6]. These were designed to be performed as distinct, independent reactions for each sample. Commercial SARS-CoV-2 diagnostic assays have been designed using a multiplex of multiple gene targets. The TaqPath COVID-19 Combo Kit (Applied Biosystems, Waltham, MA USA; EUA200010) targets the ORF1ab gene, N gene, and S gene using one primer/probe set per region. All three qRT-PCR primer/probe sets are run in a single reaction well, and the amplification of each gene is measured over an independent fluorescence channel. The more recent TaqPath COVID-19 Fast PCR Combo Kit 2.0 (Applied Biosystems, Waltham, MA, USA; EUA210384) increases gene redundancy by using primer/probe sets that target three regions in the ORFf1a gene, two regions in the ORFf1b gene, and three regions in the N gene. All primer/probe sets are run in a single reaction well and all targets in each genomic region (ORF1a, ORF1b, and N gene) share a common fluorescence channel unique to that region.
In multiplexed qRT-PCR assays, there is usually a very low difference in the threshold cycle values between the two target genes. However, mutations in the virus genome arise that can lead to poor primer or probe binding at one gene target, resulting in discordant C t values between two of the gene targets, which is termed a gene target failure. Polymorphisms in the SARS-CoV-2 genome that can affect detection by commercial qRT-PCR assays have been reported for the E [7,8] and N [9] genes and other diagnostic targets [10]. For example, a 6-nucleotide deletion in the S gene found in SARS-CoV-2 Alpha variant leads to a failure of S gene target detection (S gene target failure, SGTF) in the ThermoFisher TaqPath COVID-19 Combo Kit real-time reverse transcription-PCR (qRT-PCR) assay. The Omicron variant also harbors the same SGTF deletion mutation, demonstrating that mutations can be recurring [11]. This failure profile has been used as a surrogate marker to estimate variant prevalence in populations [12,13].
Here, we report the characterization of mutations in the SARS-CoV-2 Delta variant genome that cause N gene target failure (NGTF) in the TaqPath COVID-19 Combo Kit qRT-PCR assay. This was discovered through baseline sequencing surveillance in Arizona, USA from 1 June 2021 to 21 August 2021. NGTF specimens had a greater C t difference (median 7.6; IQR: 4.0) between N gene and ORF1ab gene targets, compared to non-NGTF specimens (median C t difference 0.3; IQR: 1.0). Using next-generation sequencing, we identified a series of mutations associated with NGTF specimens. We performed molecular validation, demonstrating that plasmid constructs bearing these mutations led to NGTF. Finally, we show that these NGTF-associated mutations do not affect the revised TaqPath COVID-19 Fast PCR Combo 2.0 assays. Taken together, this highlights the importance of genomic surveillance and molecular diagnostics to stay ahead of the ongoing evolution of SARS-CoV-2.

Saliva Specimens and Diagnostic Testing
This study involved analyses of 2625 saliva specimens submitted for testing at the ASU Biodesign Clinical Testing Laboratory (ABCTL) from 1 June 2021 and 21 August 2021. Saliva samples were independently obtained by patients in 2 mL collection vials, registered, and deposited at ABCTL drop-off locations. RNA was extracted from 250 µL of saliva specimen within 33 h of sample receipt using the KingFisher Flex (Thermo Scientific, Waltham, MA, USA), following the manufacturer's guidelines. Diagnostic testing was performed using TaqPath COVID-19 Combo Kit assay (Applied Biosystems, Waltham, MA, USA), following the manufacturer's guidelines. Samples were determined to have an N gene target failure (NGTF) or S gene target failure (SGFTF) if the C t value for that gene was 5 or more cycles higher than the ORF1ab gene C t value. Samples were also classified as target failures if a single gene failed to amplify when a positive result was observed from the other two genes.
RNA from all the ABCTL saliva samples that tested positive for SARS-CoV-2 underwent viral genome sequencing. Sample C t scores and patient metadata were obtained for the samples that generated consensus sequences. At the time of analysis, this collection included 10,134 samples, with the earliest collection date of 28 December 2020. For the sample population collected from 1 June 2021 through 21 August 2021, there were 1975 unique individuals. This study was approved by Arizona State University's Institutional Review Board (IRB).

Global Data Collection and Analysis
In order to evaluate global sequence prevalence, the entire GISAID.org database was accessed and downloaded through 21 December 2021. Sequences with incomplete collection dates or missing country location metadata were omitted from the analyses. The GISAID.org web interface was used to query the database for the nonsynonymous mutations found in sequence groups 641∆6 and A638G using the query strings "N_G214del, N_G215del" and "N_N213S, N_G214del, N_G215del", respectively). To find the sequences that contain the synonymous mutations found in the sequence groups C643T and C636T, sequences were queried with "TAATGGCTGTGATGCT" and "CAATGGCTGTGATGCC", respectively.

Plasmids
Plasmid constructs of the SARS-CoV-2 genomic regions were created by RT-PCR amplification on saliva RNA extracts using the SuperScript IV One-Step RT-PCR system (ThermoFisher, San Diego, CA, USA) and primers derived from the ARTIC V4 primer set (SARS-CoV-2_95_LEFT: GTGCGTTGTTCGTTCTATGAAGAC; SARS-CoV-2_98_RIGHT: TTTAGGCCTGAGTTGAGTCAGC). DNA bands were gel excised and used as a template for dA-tailing using DreamTAQ (ThermoFisher, San Diego, CA, USA) polymerase. The fragments were then ligated into pCRII-TOPO (ThermoFisher, San Diego, CA, USA) vectors. All cloned constructs were verified by Sanger sequencing.

Next-Generation Sequencing Analysis of Delta Variant N Gene Dropouts
To understand the evolution of circulating SARS-CoV-2 variants, we performed baseline surveillance genomic sequencing. A total of 2332 SARS-CoV-2 genomes were successfully sequenced, the majority of which were the Delta variant (B.1.617.2 and AY sublineages) (90.8%) (Figure 2A). Other variants within this study period included the Alpha
. Within this dataset, we obtained whole genome sequences for 37 NGTF specimens (33 Delta variants, 3 Gamma variants, 1 Alpha variant, 3 non-VOC lineages). We identified eight N gene mutation profiles associated with NGTF specimens ( Figure 3A). We queried our collection of 10,134 SARS-CoV-2 genomes using these mutation profiles and compared qRT-PCR Ct values associated with these mutation profiles. Four NGTF sequence groups had Ct values that significantly deviated from the following controls: C636T, T651C, and 641∆6, and A638G ( Figure 3B; p < 0.001; Mann-Whitney). These samples had ORF1ab Ct values of <32, indicating that low viral load was unlikely to be a factor in NGTF ( Figure 3C). The other four additional N gene sequence groups (G643T, C239G, GAT7CTA, G777A) did not significantly vary from the control Ct values ( Figure 3B, gray). Within this dataset, we obtained whole genome sequences for 37 NGTF specimens (33 Delta variants, 3 Gamma variants, 1 Alpha variant, 3 non-VOC lineages). We identified eight N gene mutation profiles associated with NGTF specimens ( Figure 3A). We queried our collection of 10,134 SARS-CoV-2 genomes using these mutation profiles and compared qRT-PCR C t values associated with these mutation profiles. Four NGTF sequence groups had C t values that significantly deviated from the following controls: C636T, T651C, and 641∆6, and A638G ( Figure 3B; p < 0.001; Mann-Whitney). These samples had ORF1ab C t values of <32, indicating that low viral load was unlikely to be a factor in NGTF ( Figure 3C). The other four additional N gene sequence groups (G643T, C239G, GAT7CTA, G777A) did not significantly vary from the control C t values ( Figure 3B, gray).

Molecular Validation of NGTF Mutations
To molecularly validate the NGTF mutations, we cloned the N gene from the specimens that harbored the C636T, T651C, 641∆6, and A638G mutations ( Figure 4A). As controls, we included plasmid constructs containing the Wuhan1 reference sequence and the Delta variant sequence G643T. First, we performed the 2019_nCoV_N2 qRT-PCR assays that targets a region of conserved sequences in the plasmid panel (nucleotides 891-957). As expected, serial dilutions of plasmid constructs in the panel performed similarly in the 2019_nCoV_N2 assay ( Figure 4B, left). We next performed the TaqPath COVID-19 Combo Kit assay. Both the reference and G643T constructs were equally detected. However, the NGTF mutation constructs had marked shifts in C t values ( Figure 4B, center). Construct A638G was the most impacted (mean difference 14.12 ± 0.87), followed by construct 641∆6 Viruses 2022, 14, 1316 6 of 9 (mean difference 12.46 ± 0.41), construct T651C (mean difference 7.70 ± 0.54), and construct C636T was the least impacted (mean difference 3.06 ± 0.33). Finally, we tested the TaqPath COVID-19 Fast PCR Combo Kit 2.0 that was redesigned with multiple target redundancy per gene. All plasmid constructs, including the previously associated NGTF mutations, performed equivalently without a significant shift in C t values ( Figure 4B, right). These results indicate that N gene mutations between nucleotides 636 and 651 affect the TaqPath COVID-19 Combo Kit. However, they did not impact TaqPath COVID-19 Fast PCR Combo Kit 2.0 performance.

Molecular Validation of NGTF Mutations
To molecularly validate the NGTF mutations, we cloned the N gene from the specimens that harbored the C636T, T651C, 641∆6, and A638G mutations ( Figure 4A). As controls, we included plasmid constructs containing the Wuhan1 reference sequence and the Delta variant sequence G643T. First, we performed the 2019_nCoV_N2 qRT-PCR assays that targets a region of conserved sequences in the plasmid panel (nucleotides 891-957). As expected, serial dilutions of plasmid constructs in the panel performed similarly in the 2019_nCoV_N2 assay ( Figure 4B, left). We next performed the TaqPath COVID-19 Combo Kit assay. Both the reference and G643T constructs were equally detected. However, the NGTF mutation constructs had marked shifts in Ct values ( Figure 4B, center). Construct A638G was the most impacted (mean difference 14.12 ± 0.87), followed by construct 641∆6 (mean difference 12.46 ± 0.41), construct T651C (mean difference 7.70 ± 0.54), and construct C636T was the least impacted (mean difference 3.06 ± 0.33). Finally, we tested the TaqPath

Molecular Validation of NGTF Mutations
To molecularly validate the NGTF mutations, we cloned the N gene from the specimens that harbored the C636T, T651C, 641∆6, and A638G mutations ( Figure 4A). As controls, we included plasmid constructs containing the Wuhan1 reference sequence and the Delta variant sequence G643T. First, we performed the 2019_nCoV_N2 qRT-PCR assays that targets a region of conserved sequences in the plasmid panel (nucleotides 891-957). As expected, serial dilutions of plasmid constructs in the panel performed similarly in the 2019_nCoV_N2 assay ( Figure 4B, left). We next performed the TaqPath COVID-19 Combo Kit assay. Both the reference and G643T constructs were equally detected. However, the NGTF mutation constructs had marked shifts in Ct values ( Figure 4B, center). Construct A638G was the most impacted (mean difference 14.12 ± 0.87), followed by construct 641∆6 (mean difference 12.46 ± 0.41), construct T651C (mean difference 7.70 ± 0.54), and construct

Global Incidence of Delta Variant N Gene Dropout Mutations
We next investigated the global prevalence of these NGTF mutations by querying the sequences deposited in the GISAID database through 1 December 2021 (6,297,034 samples). There were 3,619,255 Delta lineage samples. The B.1.617.2 parent lineage contained 171,882 samples and 3,447,373 samples belonged to Delta AY sub-lineages. The A638G motif was found in four (<0.01%) samples, all of which were Delta sub-lineages. The 641∆6 motif was found in 3986 (0.11%) Delta sub-lineage genomes ( Figure 5A) and 241 (<0.01%) non-Delta sub-lineage genomes. It was found predominantly in the B.1.617.2 sub-lineage ( Figure 5B). The C636T motif was found in 9194 (0.25%) Delta sub-lineage genomes (Figure 5A). It was found primarily in the AY.103 sub-lineage ( Figure 5B). The T651C motif was found in 401 (0.011%) Delta sub-lineage genomes ( Figure 5A). It was found primarily in the Delta AY.44 sub-lineage ( Figure 5B). .

Discussion
In this report, we identified the SARS-CoV-2 Delta lineage mutations responsible for NGTF on the TaqPath COVID-19 Combo Kit. We performed molecular validation experiments to demonstrate that they caused NGTF on the TaqPath COVID-19 Combo Kit. We also showed that the updated TaqPath COVID-19 Fast PCR Combo 2.0 assay was not affected by these mutations. These results demonstrate the importance of genome sequencing in diagnostics to account for SARS-CoV-2 evolution.
The functional consequences of the NGTF mutations are not well understood. The NGTF mutations are located within the highly disordered linker region within the N protein that is adjacent to the serine-arginine (SR) rich motif [20]. Both the 641∆6 and A638G mutations result in a two amino acid deletion of G214 and G215 at the edge of a predicted B cell epitope [21]. Mutations at position G215 have been previously implicated in viral transmissibility [22], whereas C636T and T651C are synonymous mutations.
Studies investigating target failure have been reported for the E [7,8] and N [9] genes, as well as other diagnostic targets [10]. The 641∆6 mutation has been shown to cause NGTF in the nucleic acid amplification-based Allplex SARS-CoV-2 Assay, but N protein was still detected using the FREND COVID-19 Ag rapid antigen test [23]. This study

Discussion
In this report, we identified the SARS-CoV-2 Delta lineage mutations responsible for NGTF on the TaqPath COVID-19 Combo Kit. We performed molecular validation experiments to demonstrate that they caused NGTF on the TaqPath COVID-19 Combo Kit. We also showed that the updated TaqPath COVID-19 Fast PCR Combo 2.0 assay was not affected by these mutations. These results demonstrate the importance of genome sequencing in diagnostics to account for SARS-CoV-2 evolution.
The functional consequences of the NGTF mutations are not well understood. The NGTF mutations are located within the highly disordered linker region within the N protein that is adjacent to the serine-arginine (SR) rich motif [20]. Both the 641∆6 and A638G mutations result in a two amino acid deletion of G214 and G215 at the edge of a predicted B cell epitope [21]. Mutations at position G215 have been previously implicated in viral transmissibility [22], whereas C636T and T651C are synonymous mutations.
Studies investigating target failure have been reported for the E [7,8] and N [9] genes, as well as other diagnostic targets [10]. The 641∆6 mutation has been shown to cause NGTF in the nucleic acid amplification-based Allplex SARS-CoV-2 Assay, but N protein was still detected using the FREND COVID-19 Ag rapid antigen test [23]. This study expands upon these previous studies by including molecular validation, in addition to bioinformatic analysis.
Due to the proprietary nature of the TaqPath assays, we are limited in our ability to confirm where the NGTF mutations lie in relation to the assay primers and probe. Nonetheless, our molecular validation experiments demonstrate that they interfere with the TaqPath COVID-19 Combo Kit assay performance. Constructs with the N gene polymorphisms performed equivalently to the controls on the TaqPath COVID-19 Fast PCR Combo 2.0 Kit assay. The uncertainty in the primer binding location limits confirming whether this is because fluorescence from the other two N gene targets compensates for the target failure at a single locus, or because the assay no longer targets the nucleotide region used in the original assay.
This study highlights the public health benefit of genomic sequencing in epidemiological surveillance. Pairing diagnostic C t values with genome sequences provides the ability to identify polymorphisms that may evade diagnostic assays. As SARS-CoV-2 continues to evolve, new mutations and variants are expected to arise. Knowledge of their effect on diagnostic assays is important, as novel variants of concern may arise with recurrent mutations [11].