Combined RT-qPCR and Pyrosequencing of a SARS-CoV-2 Spike Glycoprotein Polybasic Cleavage Motif Uncovers Rare Pediatric COVID-19 Spectrum Diseases of Unusual Presentation

Background: Surveillance of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infections is essential for the global containment measures with regard to the ongoing pandemic. Diagnostic gold standard is currently reverse transcription of the (+)RNA genome and subgenomic RNAs and subsequent quantitative polymerase chain reaction (RT-qPCR) from nasopharyngeal swabs or bronchoalveolar lavages. In order to further improve the diagnostic accuracy, particularly for the reliable discrimination between negative and false-negative specimens, we propose the combination of the RT-qPCR workflow with subsequent pyrosequencing of a S-gene amplicon. This extension might add important value mainly in cases with low SARS-CoV-2 load, where RT-qPCR alone can deliver conflicting results. Results: We successfully established a combined RT-qPCR and S-gene pyrosequencing method. This method can be optionally exploited after routine diagnostics or for epidemiologic studies allowing a more reliable interpretation of conflicting RT-qPCR results. This may occur in specimens with relatively low viral loads and close to the detection limits of qPCR, practically for CT values >30. After laboratory implementation and characterization of a best practice protocol we tested the combined method in a field study on a large pediatric cohort from two German medical centers (n=769). Pyrosequencing after RT-qPCR enabled us to uncover previously unrecognized cases of pediatric COVID-19 spectrum diseases, partially exhibiting unusual and heterogeneous presentation. Moreover, it is notable that in the course of RT-qPCR/pyrosequencing method establishment when routinely confirmed SARS-CoV-2-positive specimens were used we did not observe any case of false-positive diagnosis. Conclusions: The proposed protocol allows a specific and sensitive detection of SARS-CoV-2 close to the detection limits of RT-qPCR. Combined RT-qPCR/pyrosequencing does not negatively affect preceding RT-qPCR pipeline in SARS-CoV-2 diagnostics and can be optionally applied in routine to inspect conflicting RT-qPCR results.


Keywords 73
(+)RNA, COVID-19 surveillance, pediatric SARS-CoV-2-associated diseases, epidemiology 74 75 Background 76 As of April 26 th 2020 at Helios University Hospital Wuppertal during a period of attenuation which 77 followed the first peak phase of SARS-CoV-2 transmission in Germany only 2.3% of all laboratory-78 confirmed SARS-CoV-2-positive cases were children or teenagers (age group: 0-19 yrs). Nationwide 79 as of September 8 th 2020, the contribution of this age group to all SARS-CoV-2-positive cases had 80 increased to 10.5% [1]. This development was in agreement with observations made globally, in 81 particular in the US, whose population remains among the most affected of the SARS-CoV-2 82 pandemic. Since then, a trend of weekly median age decline for persons with COVID-19-like illness 83 was reported for an observation period between May 3 rd -August 29 th 2020. In particular, a steady 84 increase of the percentage fraction of all confirmed SARS-CoV-2-positive cases for the 0-19 year olds 85 was observed in the United States starting from 7.4% in May, 10.8% in June, 14.0% in July and 86 preliminary being elevated up to 15.5% in August 2020 [2]. Reminiscent of this, at least since early 87 autumn 2020 the 7-days-incidence for all pediatric age groups increased in similar ways as observed 88 for most other age groups in Germany ( Figure 1). As evidence is accumulating that pediatric 89 presentation of SARS-CoV-2 associated diseases can possibly be more heterogeneous than adult 90 COVID-19 there is an urgent need to recognize the full spectrum of unusual pediatric SARS-CoV-2-91 borne diseases. Reliable diagnosis of SARS-CoV-2 infection will eventually contribute to effective 92 personalized treatment and optimized containment measures. To address this problem, we 93 systematically screened two pediatric cohorts from Helios University Hospital Wuppertal (North 94 'Charité protocol' amplicon targets (RdRP, E, N) [4] (Table S1). Before RT-qPCR, total RNA was 127 purified from 250 µL of nasopharyngeal swabs or bronchoalveolar lavage specimens using the 128 guanidinium isothiocyanate (GITC) extraction method. Besides, due to RT-qPCR reagent 129 manufacturer's recommendation, we tested whether the RNA purification step could be dispensable. 130 Briefly, using stabilized raw specimens from SARS-CoV-2 as RT-qPCR templates, results from the 131 same specimens using purified RNA could only be matched in some cases ( Figure S1). Therefore, we 132 neglected this approach during our successive experiments. 133 Reverse transcription and subsequent amplification of cDNA were carried out using universal one-step 134

RT-qPCR reaction mastermixes. A detailed protocol including recommended reagents is provided in 135
the Methods section. Quantitative PCR programs often use an excess of cycling steps, frequently 136 leading to the incremental enrichment of unspecific byproducts during later cycles. With respect to the 137 quality of subsequent pyrosequencing, we aimed on the optimization of PCR cycle numbers for 138 singleplex, duoplex and triplex PCR approaches. We determined that 36 sequential qPCR cycles of 139 denaturation, annealing and elongation were a good compromise with respect to multiple prudential 140 reason: 1. Typically, for different amplicons we observed that the cycle threshold (CT) limit for the 141 faithful discrimination between positive and negative samples was below a CT of 35 ( Figure 3). 142 Consistently, semi-quantitative PCR and subsequent electrophoresis resulted in a specific 162 bp band 143 and no or occasionally few weak byproducts ( Figure S2A). 2. For approx. 35 or more cycles we 144 observed accelerated curve increments and thresholds crossing also for negative samples and no-145 template controls indicating ongoing amplification of PCR byproducts. Using an excess of PCR 146 cycles, multiple unspecific byproducts and DNA smear of higher molecular weight are observable in 147 confirmed positive and negative specimens ( Figure S2B). 148 Besides an excess of PCR cycles, we identified multiplex PCR approaches where more than one 149 amplicon is targeted by multiple primer sets in one PCR reaction as another possible confounding 150 factor influencing PCR quality and successive pyrosequencing. To test this hypothesis, we performed 151 triplex RT-qPCR reactions by simultaneously using RdRP, Orf E and Orf N amplicon targets. Semi-152 quantitative analyses by agarose gel electrophoresis demonstrate that numerous low and high 153 molecular weight byproducts were amplified in both cases when RNA from confirmed SARS-CoV-2-154 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted https://doi.org/10.1101https://doi.org/10. /2020 positive specimens or negative specimens was used ( Figure S3). Since we assumed that biotinylated 155 byproducts could massively impair successive pyrosequencing we decided to neglect triplex 156 approaches for the intended combinatorial use of RT-qPCR. With emphasis on the SARS-CoV-2 S-157 gene amplicon we instead aimed to compare the specificities, efficiencies and sensitivities of 158 singleplex RT-qPCR vs. duoplex RT-qPCR. In order to further characterize the performance of the 159 SARS-CoV-2 S-gene amplicon target and to select the best additional amplicon for a duoplex 160 approach, we performed comprehensive RT-qPCR tests using the S-gene, Orf E, Orf N and RdRP as 161 amplicon targets and using serially diluted specimens from the same confirmed SARS-CoV-2-positive 162 specimens for all RT-qPCR reactions. Figure 3 illustrates the performance of different qPCR assays 163 for one representative SARS-CoV-2-positive nasopharyngeal specimen. Taken together, TaqMan RT-164 qPCR for both the S-gene amplicon and the Orf E amplicon with similar sensitivity as indicated by 165 low DCT between S-gene and Orf E in singleplex reactions. Further, we deduced from a series of serial 166 template dilutions that amplification efficiency for both the S-gene target and the Orf E target is high 167 but weakens between CT=30 and CT=35. In duoplex RT-qPCR assays targeting simultaneously the S-168 gene target and Orf E in one reaction no notable changes in sensitivity or qPCR efficiency were seen 169 ( Figure 3A). With respect to efficiency also the Orf N amplicon target performed well, but its 170 sensitivity lagged behind as indicated by a large DCT when compared to Orf E ( Figure 3B). In stark 171 contrast, the RdRP amplicon frequently performed volatile and appeared to lag behind the sensitivities 172 of all other amplicons in the majority of specimens examines -as in the illustrated case, where even 173 the curve for the undiluted sample did not cross the predefined threshold ( Figure 3B). Therefore, apart 174 from a singleplex S-gene RT-qPCR/pyrosequencing approach, we considered a duoplex (S-gene and 175 Orf E) RT-qPCR/S-gene pyrosequencing approach as a promising combination for the faithful 176 detection of active SARS-CoV-2 infections. 177 To test this hypothesis, we performed subsequent pyrosequencing of the biotinylated single-stranded 178 S-gene amplicon using the same serial dilution samples described above (Figure 4). The proposed 179 pyrosequencing approach could theoretically allow to read a SARS-CoV-2 (+)RNA-specific sequence 180 fragment of 55 nt. Using the undiluted sample from the singleplex S-gene amplicon RT-qPCR 181 reaction, we achieved an unbiased 53 nt sequence fragment, which could unambiguously be assigned 182 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted December 23, 2020. ; https://doi.org/10. 1101 to the SARS-CoV-2 reference genome. Successive serial dilutions led to modest enrichment of 183 miscalled bases in the resulting pyrogram, but still allowed the faithful assignment of the called 184 sequence to SARS-CoV-2 up to a corresponding approx. CT≈30. Serial dilutions were gradually 185 accompanied by a weakening of pyrogram-signal strength and a decreased signal-to-noise ratio. 186 Automated basecalling of the SARS-CoV-2 sequence failed frequently for samples with CT>30. 187 However, for a range between CT≈30 and CT≈35 the most adverse reason counteracting SARS-CoV-2 188 sequence recognition appeared to be the decremental signal-to-noise ratio, which led to incremental conducted priorly ( Figure 4). We noted a tendency of slightly increasing numbers of missing or excess 193 basecalls towards the 3'-end of the pyrosequenced SARS-CoV-2 fragment. As a preliminary 194 conclusion we first highlight that the proposed pyrosequencing approach does not negatively affect 195 preceding RT-qPCR pipeline in SARS-CoV-2 diagnostics. Second, it adds important value to  qPCR, where this method alone delivers conflicting results, particularly close to the detection limits 197 qPCR (CT values >30). In RT-qPCR alone even negative samples frequently exhibit curves crossing 198 the threshold within a range between CT≈30 and CT≈35 complicating mainly the reliable 199 discrimination between PCR-negatives and PCR-false-negatives. 200 In a next step, after protocol implementation we checked the sustainability of the combined RT-201 qPCR/pyrosequencing method in an epidemiological field test. We tested our combined RT-qPCR/pyrosequencing method on two pediatric cohorts from two German 206 medical centers comprising 769 children in total (n=599 [Wuppertal]; n=170 [Kassel], who had no 207 indication for routine SARS-CoV-2 testing at presentation according to WHO criteria. After initial S-208 gene amplicon detection, we used at least ORF E as backup amplicon as well as ORF N and RdRP in 209 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted December 23, 2020. ; https://doi.org/10.1101/2020.12.19.20243428 doi: medRxiv preprint some cases. Even for RT-qPCR results with relatively high CT values, successive pyrosequencing 210 could unambiguously confirm SARS-CoV-2 infections. Where possible, we investigated SARS-CoV-211 2-specific immunity. In sum, using retrospectively the combined RT-qPCR/pyrosequencing method, 212 we confirmed 6 pediatric cases of SARS-CoV-2-associated diseases among the entire cohort 213 exhibiting heterogenous presentation: 214 215 Brief case report 1. Adolescent male with sore throat (age group 13-17 years) 216 During the 2020 spring peak of incremental SARS-CoV-2 transmission at end of April an 217 adolescent male patient presented at the emergency department in slightly reduced general 218 condition with 37.5°C body temperature and reddened throat. The clinical examination was 219 otherwise completely normal. At time of presentation the patient did not fulfill the criteria for 220 routine SARS-CoV-2 testing. The specimen was incidentally tested and classified potentially 221

years) 231
A female toddler was admitted to the hospital with altered general status and undulant fever. The 232 initial physical examination revealed tonsillitis without any cardio-respiratory affections. 233 Laboratory analyses revealed highly elevated C reactive protein (24.0 mg/dL [normal<0.5]) with 234 almost normal interleukin-6 levels at 55.2 pg/ml. Interestingly, no leukocytosis or lymphopenia 235 were diagnosed. In the clinical course, symptoms reminiscent of Kawasaki-like disease included 236 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted December 23, 2020. ; https://doi.org/10.1101/2020.12.19.20243428 doi: medRxiv preprint persistent fever, bilateral conjunctivitis, cheilitis and a maculopapular exanthema. Furthermore, 237 echocardiography exhibited enlargement of the left coronary artery and pericardial effusion 238 ( Figure S4A). Cardiac related blood parameters were within the normal ranges. Under the 239 suspicion of Multisystem Inflammatory Syndrome in Children (MIS-C) associated with COVID-240 19 [5, 6] she was given intravenous gamma globulins (2 g/kg), prednisolone (2 mg/kg) and 241 acetylsalicylic acid (50 mg/kg) at day five which resulted in rapid improvement of the girl's 242 general status. Apyrexia was achieved on day seven. At day ten of hospitalization SARS-CoV-2 243 RT-qPCR tests were negative. 244 245 Brief case report 3. Pre-school boy with high grade fever (age group 4-6 years) 246 A pre-school boy with bowel disease history presented at the paediatric emergency department 247 with a history of one day high-grade fever. Clinical examination did not reveal a specific focus. He Brief case report 4. Male secondary school child with isolated swollen neck lymph node (age group 260

10-12 years) 261
All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in days later without fever in good general condition. Nine days later she presented again with low 275 temperature up 38.4°C, general malaise and slightly increased liver enzymes. All other diagnostic 276 test at that time -including abdominal ultrasound, laboratory studies and urine analysis -were 277 normal. During the following two days she improved spontaneously and was discharged home. At 278 that time symptoms were believed to be associated with the recent EBV infection. No SARS-CoV-279 2 testing was performed since the patient did not fulfill the clinical criteria for suspected Covid-19 280 at that time. Mean CT value of repetitive RT-qPCR was relatively high (CT: 32.80) with 281 conspicuous curve shape. SARS-CoV-2 was unambiguously confirmed by pyrosequencing. In 282 October 2020 the patient was found to have IgA and IgG antibodies against SARS-CoV-2 (SARS-283 CoV-2 ELISA, EUROIMMUN). 284 285 Brief case report 6: Male toddler with generalized febrile seizure (age group 1-3 years) 286 A male toddler was admitted late in March 2020 due to a generalized febrile seizure which 287 spontaneously resolved after 4 min. The parents reported that he developed a low-grade fever up 288 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted December 23, 2020. ; https://doi.org/10. 1101 to 38.6°C several hours prior the event. He also had moderate diarrhoea with three to four pulpy 289 bowel movements per days and vomited three times on the day of admission. The parents did not 290 note any mucous or blood. On admission, clinical examination was unremarkable except for mild 291 pharyngitis, intestinal hyperperistalsis and mild dehydration. Differential blood count, liver 292 enzymes, blood gases, electrolytes, C-reactive protein as well as urine analyses (except for ketone 293 bodies) were all normal. Stool tests came back negative for Rotavirus, Norovirus, Shigella, 294 Campylobacter, Salmonella and Yersinia. The patient received intravenous fluids and was 295 discharged two days later without fever in good clinical condition. At that time the patient was 296 diagnosed with mild gastroenteritis and secondary uncomplicated generalized febrile seizure. RT-297 qPCR and pyrosequencing confirmed a SARS-CoV-2 infection. 298

Numerous confounding factors influence the outcome of RT-qPCR in SARS-CoV-2 diagnostics 301
Whereas reverse transcription and subsequent quantitative PCR (RT-qPCR) are key methods for the 302 detection of SARS-CoV-2 and local as well as global pandemic surveillance, it is well known that 303 several confounding factors can lead to false-negative results. It seems clear that the time course of 304 SARS-CoV-2 load in the days after infection massively influences the predictive value of the RT-305 qPCR tests. A recent study discovered a median false-negative rate as high as 38% (CI, 18% to 65%) 306 even at the day of symptom onset, although concomitantly SARS-CoV-2 load seems to be close to its 307 highest levels. Moreover, few days before and after symptom onset the false-negative SARS-CoV-2 308 discovery rate by RT-qPCR seems to worsen dramatically [7]. Besides the time point of specimens 309 sampling post-infection, heterogeneities in specimens sampling technique, transportation, storage 310 conditions, nucleic acids purification, laboratory equipment, staff experience but also RT-qPCR 311 conditions might be among the most important factors influencing test qualities. Very early in 2020 312 the WHO started to disseminate the 'Charité protocol' for the diagnostic detection of SARS-CoV-2 by 313 RT-qPCR. Therein, a strategy for the combinatorial use of PCR primers (#name_F/#name_R) and 314 TaqMan probes (#name_P) was described with amplicon targets in the SARS-CoV-2 RNA-dependent 315 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in  (Table S1)  Therefore, we aimed to develop a confirmatory test for a stably expressed SARS-CoV-2 sgRNA 335 amplicon target that can complement RT-qPCR strategies without disturbing established and 336 automatable laboratory workflows. In particular, we intended to develop a pyrosequencing assay that 337 would allow, subsequent to RT-qPCR, the categorical confirmation of SARS-CoV-2 infections in 338 acute cases, where the clinical suspicion is high, but the SARS-CoV-2 infection cannot be ruled out by 339 RT-qPCR alone. The proposed pyrosequencing approach does not negatively affect preceding RT-340 qPCR pipelines in SARS-CoV-2 diagnostics and can therefore add important value to 341 where this method alone delivers conflicting results. Particularly, this can happen close to the 342 detection limits qPCR, practically CT values >30. Frequently, even negative samples exhibit curves 343 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted December 23, 2020. ; https://doi.org/10.1101/2020.12.19.20243428 doi: medRxiv preprint crossing the threshold within a range between CT≈30 and CT≈35 complicating the reliable 344 discrimination between PCR-negative and PCR-false-negative specimens. Theoretically, also PCR-345 false-positives results could happen, but practically we did not observe any case of false-positive 346 diagnosis within the numerous confirmed SARS-CoV-2-positive specimens from clinical routine 347 testing used during the entire phase of RT-qPCR/pyrosequencing development. Here, we have shown 348 that pyrosequencing can be a powerful complementary method of specific and sensitive SARS-CoV-2 349 case confirmation, without affecting foregoing routine RT-qPCR. But even here, lower template 350 concentrations led to the occasional occurrence of miscalled bases (missing bases: red triangle; excess 351 bases: blue triangle) and gradual convergence of signal and noise peaks. Gradually, this led to 352 increasingly misinterpreted peak heights particularly for nucleotide repeat motifs and lower 353 complexity motifs. Whereas concomitantly, automated basecalling gradually failed to separate signal 354 from noise, the SARS-CoV-2 specific sequence could be identified by manual inspection much longer.

Several cases with unusual presentations of pediatric SARS-CoV-2-infections were uncovered 367
Exemplarily for the implementation of the developed experimental pipeline for an epidemiological 368 survey, we conducted a field study in search for SARS-CoV-2 infected patients without obvious 369 common SARS-CoV-2 associated symptoms, since the majority of cases pediatric SARS-CoV-2 370 infections develop only very mild disease courses [14]. In a very recent UK national cohort study on 371 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted December 23, 2020. ; https://doi.org/10.1101/2020.12.19.20243428 doi: medRxiv preprint neonatal SARS-CoV-2 infection 66 SARS-CoV-2-positive babies could be identified, who at day of 372 presentation exhibited hyperthermia, poor feeding, vomiting, coryza, other respiratory signs and 373 lethargy as the most common signs of infection or, respectively, no signs of infection at all [15]. On 374 the other hand, an unknown fraction of SARS-CoV-2 infections in childhood seem to fundamentally 375 differ from adults and can be more heterogeneous in their presentation. The most striking example is 376 Multisystem Inflammatory Syndrome in Children (MIS-C), a rare SARS-CoV-2-induced Kawasaki-377 like hyperinflammatory syndrome [5,6]. Another recent study reports that children and adults can 378 exhibit a very different antibody responses upon SARS-CoV-2-infections across the clinical spectrum 379 of associated diseases, which do not obligatorily match the adult COVID-19 spectrum [16]. 380 Necessarily, the association unusual symptoms with acute infections will contribute to our 381 understanding about the heterogeneity of SARS-CoV-2-borne diseases in general and particularly in 382 children. Using two large pediatric cohorts for a field study, we have successfully demonstrated in this 383 study that combined RT-qPCR/pyrosequencing is a reliable tool that allows the faithful confirmation 384 of less distinctive or asymptomatic cases with SARS-CoV-2 infection close to the detection limits of 385 RT-qPCR. Here, the determination of a SARS-CoV-2-specific sequence fragment can decisively help 386 to improve reliability for cases where the discrimination between negative/false-negative RT-qPCR 387 reports can be important. Whereas this might possibly be of subordinate importance for practical 388 containment measures because low viral load could be associated with low infectiousness, it can 389 contribute to recognize the still probably underestimated extent of SARS-CoV-2 prevalence and 390 associated mild or asymptomatic presentations. In particular, from a large cohort of children who did 391 not fulfil the criteria for SARS-CoV-2 testing we have retrospectively identified 6 cases of SARS-392 CoV-2 related symptoms. The identification and characterization of this cases will thus contribute to 393 the understanding of the full heterogenous picture of presentation of SARS-CoV-2 infections. 394 395

Conclusions 396
The proposed protocol allows the specific and sensitive detection of SARS-CoV-2 close to the 397 detection limits of RT-qPCR. Combined RT-qPCR/pyrosequencing does not negatively affect 398 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted https://doi.org/10.1101https://doi.org/10. /2020 preceding RT-qPCR pipeline in SARS-CoV-2 diagnostics and can be optionally applied in routine to 399 inspect conflicting RT-qPCR results. Particularly in research, the proposed workflow can valuably 400 contribute to identify and characterize the unknown fraction of cases of SARS-CoV-2 -associated 401 diseases in childhood, which can fundamentally differ from adults and can be more heterogeneous in 402 their presentation.

Storage and nucleic acids isolation 415
Specimens included nasopharyngeal swabs or bronchoalveolar lavage specimens, which underwent 416 routine COVID-19 diagnostic testing. Pediatric cohort specimens were collected using brushes from 417 the Gentra Puregene Buccal Cell Kit (100) (Qiagen, Cat. No. 158845) and then stabilized in 500 µL 418 RNAlater™ (Thermo Fisher Scientific, Cat. No. AM7021). All specimens were stocked at -80°C. 419 Total RNA was purified from 250 µL liquid specimen using 750 µL QIAzol lysis reagent (Qiagen, 420 Cat. No. 158845) upon manufacturer's recommendations. RNA quality and quantity were assessed by 421 microcapillary electrophoresis using the Small RNA kit (Agilent, Cat. No. 5067-1548) and the Agilent 422 Bioanalyzer 2100 instrument. 423 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted December 23, 2020. ; https://doi.org/10. 1101 Importantly, we determined that long term storage (to date, up to 9 months) under these conditions 424 allows unbiases RT-qPCR analyses. However, we note that repeated freeze and thaw cycles of stored 425 specimens as well as purified RNA affect sample quality and result in gradually increasing CT values. 426

RT-qPCR 428
Quantitative analyses of SARS-CoV-2 (+)RNA from human specimen was carried out combining 429 reverse transcription and qPCR in a one-step protocol using Luna Universal Probe One-Step RT-qPCR 430 Kit w/o ROX (New England Biolabs, Cat. No. E3007E) on a Corbett Rotor-Gene 6000 instrument. 431 Primers and probes are described above and listed (Table S1) RT-qPCR amplicon were loaded to each well of a PyroMark Q48 Disc and then 3 µL PyroMark Q48 448 Magnetic Beads (300) (Qiagen, Cat. No. 974203) were added to each reaction. The sequencing 449 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted December 23, 2020. ; https://doi.org/10. 1101 reaction was initiated after denaturation and subsequent magnetic bead capture of the biotinylated 450 single-stranded S-gene amplicon. The pyrosequencing software records the light signals corresponding 451 to each well in the PyroMark Q48 Disc and saves the data graphically. Basecalled pyrograms were 452 then manually inspected and analyzed for SARS-CoV-2 sequence verification. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted https://doi.org/10.1101https://doi.org/10. /2020 We thank Prof. Parviz Ahmad-Nejad (Institute of Medical Laboratory Diagnostics, Helios University 499 hospital Wuppertal) and the involved staff, in particularly Jennifer Ortelt, Anna Wagener and Petra 500 Menk, for providing specimens for method establishment and for the seamless collaboration. 501 Institute (www.RKI.de) the 7-days-incidence for pediatric infections increased very reminiscent of 558 most other age-groups. At least for the age groups between 0-14 years, this development contrasts the 559 spring situation, when these children were less affected than other age-groups. 560 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted https://doi.org/10.1101https://doi.org/10. /2020  to conduct multiple-sequence alignments using the translated protein sequences of the SARS-CoV-2 564

Figures and Figure Legends
Spike (S) glycoprotein and homologous S protein sequences from other human coronaviruses. Here, a 565 segment harboring polybasic cleavage motif (Q644 to T720 with respect to SARS-CoV-2 S protein). 566 The highlighted residues are conserved in most human coronavirus (black shaded) or are similar 567 between some human coronaviruses (grey shaded). The colored boxes are framing residues, which 568 correspond to the target position of tested oligonucleotides: forward primer (red), probe/sequencing 569 primer (blue), reverse primer (magenta). B. The aligned protein sequences were backtranslated into the 570 encoding cDNA sequences. Similarly, as described above for protein sequences, black or grey shading 571 was used to illustrate identical or similar nucleotide positions. For combined RT-qPCR and 572 pyrosequencing, we selected the following marked sequences for oligonucleotide design: 1. Forward 573 primer (red box/arrow); 2. TaqMan probe with 5'-HEX and 3'-BBQ-650 modifications (blue 574 box/line); 3. Sequencing primer without end-modification (blue box -same sequence as TaqMan 575 Probe); 4. Reverse primer with 5'-Biotin-TEG (magenta box/arrow). 576 577 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted December 23, 2020. ;

E>N>R. 583
All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted December 23, 2020. ; https://doi.org/10.1101/2020.12.19.20243428 doi: medRxiv preprint perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in  preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted December 23, 2020. ; https://doi.org/10.1101/2020.12.19.20243428 doi: medRxiv preprint Figure S1. Results of semi-quantitative PCR for the performance comparison using GITC-647 purified RNA or stabilized raw specimens as RT-qPCR substrates. Four confirmed SARS-CoV-2-648 positive specimens (p1 to p4) were used in combination with primer pairs targeting Orf E, RdRP, Orf 649 N and S-gene amplicons in singleplex qPCR reactions. From GITC-purified templates, amplicons with 650 the correct size could be amplified from all specimens. Notably, we observed lower molecular weight 651 byproducts for Orf E PCR. From stabilized raw specimens, in contrast, the Orf E amplicon was not 652 amplified from none of the four samples. With varying band intensity, amplicons with correct sizes 653 were amplified from stabilized raw samples for all cases for RdRP and S-gene targets and in some 654 cases for the Orf N target. However, for unknown reasons, band intensities appeared less stable when 655 compared with purified RNA samples. Moreover, we observed lower molecular weight byproducts in 656 some cases. 657 658 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted December 23, 2020. ; https://doi.org/10. 1101  perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted December 23, 2020. ; https://doi.org/10. 1101  perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for this this version posted December 23, 2020. ; https://doi.org/10. 1101