key: cord-0280762-5qd7koo0
authors: Hernandez, M. M.; Banu, R.; Gonzalez-Reiche, A. S.; van de Guchte, A.; Khan, Z.; Shrestha, P.; Cao, L.; Chen, F.; Shi, H.; Hanna, A.; Alshammary, H.; Fabre, S.; Amoako, A.; Obla, A.; Alburquerque, B.; Helena Patino, L.; Ramirez, J. D.; Sebra, R.; Gitman, M. R.; Nowak, M. D.; Cordon-Cardo, C.; Schutzbank, T. E.; Simon, V.; van Bakel, H.; Sordillo, E. M.; Paniz Mondolfi, A. E.
title: Robust clinical detection of SARS-CoV-2 variants by RT-PCR/MALDI-TOF multi-target approach
date: 2021-09-16
journal: nan
DOI: 10.1101/2021.09.09.21263348
sha: a354639dc38c4b61b49ea609af74e57f682cd025
doc_id: 280762
cord_uid: 5qd7koo0

The COVID-19 pandemic sparked rapid development of SARS-CoV-2 diagnostics. However, emerging variants pose the risk for target dropout and false-negative results secondary to primer/probe binding site (PBS) mismatches. The Agena MassARRAY SARS-CoV-2 Panel combines RT-PCR and MALDI-TOF mass-spectrometry to probe for five targets across N and ORF1ab genes, which provides a robust platform to accommodate PBS mismatches in divergent viruses. Herein, we utilize a deidentified dataset of 1,262 SARS-CoV-2-positive specimens from Mount Sinai Health System (New York City) from December 2020 through April 2021 to evaluate target results and corresponding sequencing data. Overall, the level of PBS mismatches was greater in specimens with target dropout. Of specimens with N3 target dropout, 57% harbored an A28095T substitution that is highly-specific for the alpha (B.1.1.7) variant of concern. These data highlight the benefit of redundancy in target design and the potential for target performance to illuminate the dynamics of circulating SARS-CoV-2 variants.

In addition to the quantity of viral nucleic acids in a clinical specimen, the diagnostic and 58 analytic capabilities of NAATs depend on the complementarity of primers and probes to viral genome 59 sequences to reliably amplify targets of interest. As a result, binding of primers and probes can be 60 impacted by progressive accumulation of changes in the viral genomes at primer binding sites (PBSs). 61

Indeed, mismatches in PBSs -particularly the 2-3 nucleotides at the 3' end of the oligonucleotide -can 62 result in reduced binding and subsequent failure to amplify (termed "dropout" in diagnostic NAAT 63 assays) 5-8 . In fact, SARS-CoV-2 has diversified over the past 18 months, and mutations in the N, S, and 64 or more diagnostic targets 21, 22 . We recently reported the analytic performance of the Agena 73 MassARRAY ® SARS-CoV-2 Panel which combines RT-PCR and matrix-assisted laser 74 desorption/ionization time-of-flight (MALDI-TOF) technologies to detect SARS-CoV-2 23 . The Agena 75

MassARRAY ® platform probes for five distinct targets in the ORF1ab and N viral genes 24 , providing a 76 robust platform for diagnosis of SARS-CoV-2 in clinical specimens despite the emergence of virus 77 strains that have accumulated mutations that can interfere with some diagnostic targets. We evaluated 78 the pattern of target detections for SARS-CoV-2-positive specimens collected at the Mount Sinai Health 79 System (MSHS) to interrogate the impact of viral genetic variation on this diagnostic platform. 80

To do this, we compared detection of Agena diagnostic targets and genomic sequence data for 81 SARS-CoV-2-positive specimens that were deidentified and banked as part of our Pathogen 82

Surveillance Program (MSHS PSP) at the Icahn School of Medicine at Mount Sinai (ISMMS), which 83 has been previously described 25 . Complete viral genomes underwent phylogenetic analyses to 84 characterize emergent evolutionary lineages among the SARS-CoV-2-positive specimens at MSHS 85 (manuscript in preparation). For this analysis, we utilized a dataset comprised of 1,262 viral genomes 86 recovered from deidentified clinical specimens collected from patients seeking care at the Mount Sinai 87

Health System from December 1, 2020 through April 24, 2021. We identified PBS mismatches 88 associated with lineage-specific substitutions in SARS-CoV-2 variants of concern (VOC) that resulted in 89

Agena MassARRAY ® platform target dropout. 90 91 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 16, 2021. ; https://doi.org/10.1101/2021.09.09.21263348 doi: medRxiv preprint terminator nucleotide was added to the probe, using supplied extension primers and iPLEX ® Pro 114 reagents. 115

Extension products (analytes) were desalted, transferred to a SpectroCHIP ® Array (silicon chip 116 with pre-spotted matrix crystal) and loaded into the MassARRAY ® Analyzer (a MALDI-TOF mass 117 spectrometer). The analyte/matrix co-crystals were irradiated by a laser inducing desorption and 118 ionization, and positively charged molecules accelerated into a flight tube towards a detector. Separation 119 occurred by time-of-flight, which is proportional to molecular mass. After data processing, a spectral 120 fingerprint was generated for each analyte that characterizes the mass/charge ratio and relative intensity 121 of the molecules. Data acquired by the MassARRAY ® Analyzer was processed with the MassARRAY ® 122

Typer software and SARS-CoV-2 Report software. The assay detects five viral targets: three in the 123 nucleocapsid (N) gene (N1, N2, N3) and two in the ORF1ab gene (ORF1A, ORF1AB). If the IC was 124 detected, results were interpreted as positive if ≥ 2 targets were detected or negative if < 2 targets were 125 detected. If no IC and no targets were detected, the result was invalid and required repeat testing of the 126 specimen before reporting. If IC was detected and no targets were detected, the sample was interpreted 127 as negative. 128

Overall, 86,781 upper respiratory and saliva specimens underwent clinical testing in the CML at 129 SARS-CoV-2 viral RNA underwent reverse transcription, PCR amplification and next-136 generation sequencing followed by genome assembly and lineage assignment using a phylogenetic-137 based nomenclature as described by Rambaut et al. 27 Overall, of the 2,062 SARS-CoV-2-positive specimens, 1,274 (62%) had all five targets detected 188 with the remaining having one (n = 419) or more (n = 369) targets dropout. For the subset of 1,262 189 SARS-CoV-2-positive specimens sequenced in our study, all five diagnostic targets were detected in 190 943 (75%), with the remaining having one (n = 227) or more (n = 92) target dropout (Supplemental 191   Table S2 ). When we calculated the target detection rate among these SARS-CoV-2-positive specimens 192 by week, the ORF1AB target had the lowest average detection rate per week (0.87) followed by the N3 193 target (0.88) and the N2 target (0.94) (Figure 1) . Notably, the N3 detection rate declined over time with To examine the impact of mismatches on target results, we measured the number of mismatches 203 (normalized to the number of nucleotides in the PBS; see methods) in specimens with detected and 204 undetected target results (Figure 2) . Detection of each of four targets (N1, N3, ORF1A, ORF1AB) was 205 associated with perfect complementarity (0 mismatches) between the genome sequence and the 206 respective target PBSs. Specifically, > 96% of specimens with either detectable N1 or N3 targets had 207 perfect complementarity to the respective forward/reverse/probe PBS, and > 95% of specimens with 208 detectable ORF1A or ORF1AB targets had perfect complementarity to the respective 209 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 16, 2021. ; https://doi.org/10.1101/2021.09.09.21263348 doi: medRxiv preprint forward/reverse/probe PBS. The remaining specimens had -at most -only one mismatch to each of the 210 target PBSs. The exception to this was the N2 target, for which, more specimens with detectable N2 211 target had mismatches to N2 forward (43%) and N2 reverse (39%) PBSs ( Figure 2B) . Indeed, up to four 212 mismatches to the N2 forward and up to two mismatches to the N2 reverse PBSs were found in the 213 specimens for which the N2 target was detected. 214

When compared across target result groups, the number of mismatches was significantly higher 215 in specimens with N1, N2, N3 and ORF1A target dropout (Figure 2A-D) . In addition, we found the 216 fraction of N1 probe PBS with mismatches was significantly higher in specimens with N1 target dropout 217 than in those with detectable N1 (Figure 2A) . 218

Because the position of mismatches within PBSs affect primer binding capabilities 1,6-8 , we 219 characterized the mismatch frequency by position of each primer/probe. Specifically, we measured the 220 proportion of specimen genomes with a mismatch at each independent position along the full length of 221 each target's primer/probe (Figure 3, Supplemental Figure S1 ). From 5' to 3' direction, we found that 222 among 15 specimens with N1 target dropout, 10 harbored single mismatches to the 4 th -14 th basepair 223 (bp) (SARS-CoV-2 genome positions 28714 -28704) of the 17-bp-long N1 probe PBS ( Figure 3A) . 224

Specifically, these mismatches reflected the following substitutions: G28714A (n = 1 specimens), 225 G28713A (n = 2), C28709T (n = 2), C28706T (n = 1), G28704C (n = 3), G28704T (n = 1). 226

By contrast, mismatches in the 5' end of the 22-bp-long N2 primer PBSs (forward, 1 st -3 rd bp 227 (28881 -28883); reverse, 3 rd bp (28977) and 5 th bp (28975)) were identified in sequences that yielded 228 both N2 target detection and dropout ( Figure 3B ). In 340 specimens with any one mismatch to the first 229 3 bp of the N2 forward primer, 336 (99%) harbored the concurrent substitutions G28881A, G28882A, 230 and G28883C in the N gene. Of the 72 specimen genomes with N2 target dropout, 34 (47%) had this 231 substitution trio. Although, this polymorphism was found in 304 (26%) of the 1,163 specimen genomes 232 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 16, 2021. ; https://doi.org/10.1101/2021.09.09.21263348 doi: medRxiv preprint with N2 target detection, statistically, this represents a significant association of the GGG-to-AAC 233 substitution with N2 target dropout (Fisher's exact, p = 0.0002). 234

In addition, specimens that harbor mismatches to the 5' end of the N2 reverse primer are the 235 result of the C28977T or G28975A substitutions. However, of 466 specimens that harbor either 236 substitution, only 1 had both suggesting these substitutions occur independently of one another. When 237 grouped by N2 target detection result, neither substitution was significantly associated with N2 target 238

Interestingly, we found that of the 110 specimen genomes with N3 target dropout, 63 (57%) had 240 a mismatch at the penultimate nucleotide towards the 3' end in the 20-bp-long N3 forward primer 241 ( Figure 3C ). All mismatches at this position are the result of a specific adenine-to-thymine substitution 242 in ORF8 (A28095T) of the SARS-CoV-2 genome. Of the 1,102 genomes with detected N3 target, only 243 two harbored this mismatch; overall, this represents a statistically significant association of this 244 positional mismatch with N3 target dropout (Fisher's exact, p < 0.0001). 245

We also assessed whether the association of these mismatches with target dropout is maintained 246 when the quantity of virus in the specimen is controlled. Although the Agena platform yields a 247 qualitative diagnostic result, we have demonstrated previously that the number of detected targets is 248

proportional to the quantity of virus in a given specimen 23 . When we limit our dataset only to specimens 249 for which all other (e.g., non-N3) targets are detected, the association of the A28095T substitution with 250 N3 target dropout remains statistically significant (Fisher's exact, p < 0.0001), indicating that N3 target 251 dropout due to the A28095T substitution is independent of differences in virus concentration. 252 253 Lineage-specific variation and target dropout 254 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 16, 2021. ; https://doi.org/10.1101/2021.09.09.21263348 doi: medRxiv preprint

In order to assess whether target dropout was due to lineage-specific variation, we examined the 255 phylogenetic lineages of genomes harboring distinct substitutions in our dataset. Among the 34 256 specimens with the concurrent GGG-to-AAC tri-nucleotide substitution and N2 target dropout, the 257 earliest was from December 29, 2020 (PV24926) which belonged to the B.1.1.434 lineage. This 258 polymorphism did not demonstrate bias to any one lineage in specimens that yielded N2 target dropout 259 as it was found in specimens that mapped to 15 different lineages including B.1.1.7 (alpha, n = 11), 260 B.1.1.434 (n = 6), and B.1.1 (n = 4) lineages. 261

We next examined the phylogenetic lineage of genomes harboring the A28095T substitution in 262 our dataset to assess whether N3 target dropout was due to lineage-specific variation. We found that the 263 earliest specimen with this substitution was from January 8, 2021 (specimen PV25263) and belonged to 264 is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 16, 2021. ; https://doi.org/10.1101/2021.09.09.21263348 doi: medRxiv preprint any other mismatches in the N3 reverse or probe PBS. Furthermore, the two mismatches closest to the 3' 278 end of the N3 forward primer -C28087T and C28093T -were significantly associated with N3 target 279 dropout (Fisher's exact, p = 0.0407 and p < 0.0001, respectively), but the mismatch at the first position 280

was not significantly associated with N3 target dropout (Fisher's exact, p = 0.0914). 281 282 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Additionally, by using publicly available primer/probe sequences to map lineage-specific 303 substitutions, we were able to further evaluate the impact of mismatches on target results and to 304 demonstrate an association between variation in SARS-CoV-2 PBSs and target dropout. Our analysis 305 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 16, 2021. ; https://doi.org/10.1101/2021.09.09.21263348 doi: medRxiv preprint revealed that several mutations result in N1 and N3 target dropout. Interestingly, although specimens 306 from other lineages harbor mismatches in the N3 target region, we identified a distinct association 307 between the B.1.1.7-associated A28095T substitution and dropout of the N3 diagnostic target on the 308

Agena MassARRAY ® SARS-CoV-2 Panel. This finding represents the first description of a lineage-309 specific substitution that introduces a mismatch to a publicly available primer sequence and yields 310 diagnostic target dropout. This underscores the utility of publicly available sequences to further monitor 311 their diagnostic ability as SARS-CoV-2 continues to evolve and new lineages emerge. 312

The B.1.1.7 lineage (alpha) has been designated as a variant of concern by the World Health 313 In this study, we describe diagnostic target dropout that can be utilized to promptly identify 327 specimens of interest for whole genome sequencing and variant classification. Notably detection of the 328 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. performance. This warrants further study, particularly as the B.1.617.2 VOC continues to expand 343 globally 44-46 since its parent lineage was first identified in India in October 2020 47-49 . 344

An important potential limitation of our study is that target performance can be affected when 345 the quantity of viral nucleic acids in diagnostic specimens is at or near the assay limit of detection, and 346 that the limit of detection varies for different targets. We have demonstrated previously that Agena 347

MassARRAY ® target detection is proportional to quantity of viral nucleic acids 23 . Thus, detection of 348 other targets can be used as a control to evaluate the performance of an individual target, such as N3 349 target dropout in the setting of the B.1.1.7 A28095T variant. In addition, we have also identified other 350 mutations in this study that are associated with N1 and N3 target dropout. However, unlike the B.1.1.7 351 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 

. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 16, 2021. ; https://doi.org/10.1101/2021.09.09.21263348 doi: medRxiv preprint

N ORF10 PV29192  PV35332  PV35300  PV35319  PV35346  PV35350  PV35673  PV35695  PV35704  PV35928  PV35947  PV35948  PV35958  PV35971  PV35974  PV35990  PV35992  PV35996  PV36025  PV36037  PV36046  PV36059  PV36061  PV36073  PV36081  PV36094  PV36096  PV36101  PV36106  PV36401  PV36461  PV36434  PV36448  PV36454  PV36791  PV36648  PV36664  PV36678  PV36679  PV36687  PV36688  PV36890  PV36951  PV28451  PV29173  PV36680   PV25263  T  C  A  A  A  T  C  G  T  PV27023  T  C  A  A  A  T  C  G  T  PV27533  T  C  A  A  A  T  C  G  T  PV27553  T  C  A  A  A  T  C  G  T  PV27573  T  C  A  A  A  T  C  G  T  PV27616  G  T  C  A  A  A  T  C  G  T  PV27666  T  C  A  A  A  A  T  C  G  T  PV27684  T  C  A  A  A  T  C  G  T  PV27691  T  C  A  A  A  T  C  G  T  T  PV27709  T  C  A  A  A  T  C  T  T  PV28443  T  C  A  A  A  T  C  T  T  PV28450  T  C  A  A  A  T  C  T  T  PV28500 T T  C  G  T  T  C  A  A  A  T  C  T  T  T  C  A  A  A  T  C  T  T  A  T  T  C  A  A  A  T  C  T  T  T  C  A  A  A  T  C  T  T  G  T  C  A  A  G  A  T  C  G  T  T  C  A  A  A  T  C  C  G  T  T  T  C  A  A  A  T  C  G  T  T  C  A  A  A  T  C  G  T  T  C  A  A  A  T  C  G  T  T  C  A  A  G  A  T  C  G  T  T  C  A  A  A  T  C  G  T  C  T  C  A  A  A  T  C  G  T  T  C  A  A  A  T  C  G  T  T  C  A  A  A  T  C  T  T  T  C  A  A  G  T  A  T  C  T  T  T  C  A  A  A  T  C  T  T  T  C  A  A  A  T  C  T  T  T  C  A  A  A  T  C  G  T  T  C  A  A  A  T  C  G  T  T  C  A  A  A  T  C  G  T  T  C  A  A  C  T  A  T  C  G  T  T  C  A  A  A  T  C  G  T  T  C  A  A  A  T  C  G  T  T  C  A  A  A  T  C  G  T  T  C  A  A  A  T  C  G  T  T  C  A  A  A  T  C  G  T  T  C  A  A  A  T  C  A  G  A  T  C  A  A  A  T  C  G  T  T  C  A  A  A  T  C  G  T  PV27543  PV27631  PV27639  PV27705  PV28444  PV28459  PV28480  PV29143  PV29140  PV29172  PV29175  PV29219  PV29220  PV35292  PV35305  PV35328  PV35686  PV35706  PV35705  PV35709  PV35714  PV35722  PV35926  PV35929  PV35957  PV35962  PV35963  PV35960  PV36026  PV36032  PV36057  PV36058  PV36063  PV36070  PV36071  PV36079  PV36084  PV36087  PV36088  PV36097  PV36098  PV36402  PV36400  PV36406  PV36411  PV36426  PV36428  PV36429  PV36431  PV36458  PV36640  PV36645  PV36647  PV36651  PV36659  PV36948  PV36949  PV36950  PV36689  PV36691  PV36693  PV25639   A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C  A  T  C   T  C  A  A  T  C  A  A  T  C  A  A  T  C  A  A  T  C  A  A  T 

C A A A T C T T PV28507 G T C A A A T C G T PV28508 T C A A G A T C G T PV29158 A T C A A A T C G T PV29174 T C A A A T C G T T PV29188 T C A A A T C G T T C A A A T C G T T C A A A T C G T T C A A A T C G T T C A A A T C G T T C A A A T C T T A T C A A A T C T T T C A A A T C T T T C A A A T C T T T C A A A T C G T T C A A A T C G T T C A A A T C G T T C A A A T C G T T C A A G A T C G T A T C A A A T C G T T C A A A T C G T T C A A A T C G T A T C A A AC A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A T C A A G G T

Failure as a Marker of Variant B.1.1.7 Among SARS-CoV-2 Isolates in the Greater Toronto Area

Two-step strategy for the identification 420 of SARS-CoV-2 variant of concern 202012/01 and other variants with spike deletion H69-V70

Use of Whole Genome Sequencing Data for a First in Silico Specificity 424

Evaluation of the RT-qPCR Assays Used for SARS-CoV-2 Detection

Real-time RT-PCR in COVID-19 detection: issues affecting the results

Presence of mismatches between diagnostic PCR assays and coronavirus 428 SARS-CoV-2 genome

Testing at scale during the COVID-19 pandemic

Preliminary genomic characterisation of an emergent SARS-CoV-2 lineage in the UK defined by a 432 novel set of spike mutations

 T   T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T  T   PV27109