key: cord-0980007-bmqc3yt1 authors: Craig, Nicky; Fletcher, Sarah L.; Daniels, Alison; Newman, Caitlin; O’Shea, Marie; Warr, Amanda; Tait-Burkard, Christine title: Direct lysis RT-qPCR of SARS-CoV-2 in cell culture supernatant allows for fast and accurate quantification of virus, opening a vast array of applications date: 2021-12-02 journal: bioRxiv DOI: 10.1101/2021.11.30.470550 sha: c94363e6bc051ec5785c94be5b4bfd153ba4c3d7 doc_id: 980007 cord_uid: bmqc3yt1 An enormous global effort is being made to study SARS-CoV-2 and develop safe and effective treatments. Studying the entire virus replication cycle of SARS-CoV-2 is essential to identify host factors and treatments to combat the infection. However, quantification of released virus often requires lengthy procedures, such as endpoint dilution assays or reinfection with engineered reporter viruses. Quantification of viral RNA in cell supernatant is faster and can be performed on clinical isolates. However, viral RNA purification is expensive in time and resources and often unsuitable for high-throughput screening. Here, we show a direct lysis RT-qPCR method allowing sensitive, accurate, fast, and cheap quantification of SARS-CoV-2 in culture supernatant. During lysis, the virus is completely inactivated, allowing further processing in low containment areas. This protocol facilitates a wide array of high- and low-throughput applications from basic quantification to studying the biology of SARS-CoV-2 and to identify novel antiviral treatments in vitro. Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is the causative agent of coronavirus 38 disease 2019 , which emerged towards the end of 2019 (Wu et al., 2020) . From (Peiris et al., 2003; Zaki et al., 2012) . 50 The global vaccination effort has eased the burden of COVID-19 slightly, but there remains an urgent 51 need for effective anti-viral treatments, especially for early administration, outpatient treatments to 52 prevent progression to severe disease, particularly in high-risk patients. Early efforts to identify 53 interventions inhibiting SARS-CoV-2 replication relied on the repurposing of existing, approved drugs 54 with known toxicity profiles rather than de novo drug development. Whilst hundreds of drugs were trialed 55 in hundreds of thousands of patients, only a panel of two drugs were given a grade A by the CORONA 56 Project database in treatment efficacy and/or research prioritization for outpatient treatments: 57 Bamlanivimab+Etesevimab and Sotrovimab (A for both) and Budesonide (A for research prioritization, 58 B/C in treatment efficacy), and Fluvoxamine (A for research prioritization) (https://cdcn.org/corona/, 59 accessed 30/11/2021). It is therefore essential to continue the effort to find new and repurposed 60 treatments against SARS-CoV-2 infection. Examining the whole replication cycle, including virus 61 release, will reveal more candidates for further investigation and narrow the search to drugs that truly 62 reduce the viral load. 63 In both diagnostic RT-qPCR and in vitro studies of SARS-CoV-2, the requirement to extract and purify 64 viral RNA (vRNA) prior to measuring virus RNA copy numbers is expensive in both time and resources. Methods for lysis and direct RT-qPCR of viral samples without RNA purification have previously been 66 developed for the study of influenza (cell lysate, (Shatzkes et al., 2014) ), Dengue virus (cell supernatant 67 (Suzuki et al., 2018) ), Zika virus (patient samples, (Li et al., 2019) ), norovirus and hepatitis A Virus 68 (foods, (Rajiuddin et al., 2020) ). In some of these, the use of expensive, commercially available lysis 69 buffers limits applicability for high-throughput, and none of the publications analyzed the impact of lysis 70 buffer on efficiency and sensitivity of the assays. For SARS-CoV-2, several groups have developed 71 direct RT-qPCR methods for detection aimed at patient swabs. The most commonly used method is 72 direct use of swabs following a heating step, 30min at 65°C or increasingly shorter periods up to 95°C. 73 With or without addition of commercial buffers or detergents, a Ct difference between 4-7 cycles were 74 observed compared with extracted vRNA (Alcoba-Florez et al., 2020; Bruce et al., 2020; Fomsgaard 75 and Rosenstierne, 2020; Genoud et al., 2021; Grant et al., 2020; Hasan et al., 2020; Nique et al., 2021; 76 Pearson et al., 2021; Smyrlaki et al., 2020) . Other methods include the addition of proteinase K to patient 77 swab samples, showing 4-6 cycle differences, but no proof of virus inactivation was shown for these 78 samples (Mallmann et al., 2020; Srivatsan et al., 2021) . Commercial kits or homemade detergent-based 79 kits showed good correlation with positive clinical samples but virus inactivation and loss of Ct were not 80 determined (Castellanos-Gonzalez et al., 2021; Ladha et al., 2020; Merindol et al., 2020) . The reduction 81 in sensitivity, lack of inactivation proof, or the reliance on expensive proprietary lysis buffers makes 82 many of these methods unsuitable for quantification in in vitro-amplified viral culture supernatants. 83 Here, we show a method for direct lysis and RT-qPCR of vRNA in culture supernatant using a cheap, 84 non-commercial IGEPAL CA-630 (IGEPAL-630)-based lysis buffer which completely inactivates SARS-85 CoV-2 (>1E6 TCID50/ml reduction). The assay shows high sensitivity, detecting <0.0043 TCID50 per 86 reaction in lysate and to <1.89 copy per reaction using RNA template-spiked mock lysate. The method 87 described here can be used to accurately, rapidly and cost-effectively quantify SARS-CoV-2 production 88 in cell culture supernatant, allowing for faster workflows, saving time and resources on routine virological 89 applications as well as high-throughput screening. SARS-CoV-2 primer optimization using a 1-Step-RT-qPCR fluorescent dye system shows 92 highest efficiency and sensitivity with the CDC N3 primer. 93 1-Step-RT-qPCR using fluorescent dye detection is by far the most cost effective method of RT-94 qPCR, therefore, we focused on optimizing this for use with in vitro applications requiring quantification 95 of SARS-CoV-2 production. Primer pairs N1, N2, and N3 by the US Centers for Disease Control and 96 Prevention (CDC) (Lu et al., 2020) and the German Center for Infection Research (DZIF) against RdRp 97 and N (Corman et al., 2020), designed and widely used for the detection of SARS-CoV-2, were selected 98 for optimization in a 1-Step-RT-qPCR reaction using the Promega GoTaq system incorporating the 99 dsDNA-binding dye BYRT green on a LightCycler480. 100 Increasing concentrations of primer (symmetric and asymmetric concentrations) were tested 101 following the manufacturer's standard protocol, using target-specific, in vitro transcribed RNA templates 102 of 600-1600bp length in 10μl and 20μl final reaction volumes. An increase of the reaction volume to the 103 recommended 20µl showed no difference in efficacy or sensitivity for any of the primers (data not 104 shown), therefore data shown here illustrates the more cost-effective 10μl reaction volume. 105 All primer pairs tested showed clear improvement in sensitivity from 50 to 250nM of equal primer 106 concentration ( Figure 1A and C). Efficiencies, calculated by the LinRegPCR program (Version 11.0, 107 (Ruijter et al., 2009) ), showed a similar increase ( Figure 1C ). Detected fluorescence was highest for Analyzing the bands generated in the PCR reaction, we found that DZIF N formed larger products 116 than the expected 128bp ( Figure 1D ). Sanger sequencing was only successful on some of the products 117 but revealed circularization or multiplication of a section of the product. Increasing the annealing 118 temperature did not eliminate the occurrence of these products. 119 Despite the use of significantly increased primer concentrations for the DZIF N primer, efficiencies 120 did not improve beyond 78.24%. DZIF RdRp showed an efficiency of 89.61% ( Figure 1E ). However, by 121 far the best performing was primer pair CDC N3 at a symmetrical concentration of 350nM showing an 122 efficiency of 93.14% and a sensitivity of <1.89 template copies/ reaction ( Figure 1E ). Paired with the 123 production of primer dimer by CDC N1 and N2, and DZIF RdRp, and the multiple products of DZIF N, 124 CDC N3 at a symmetrical concentration of 350nM forward/reverse was selected for further development 125 of the method. 126 127 IGEPAL-630-based buffers show highest efficiency and sensitivity, shortly followed by 128 Triton X-100. 129 As a first step, heat lysis, as described for patient samples, was tested towards release of vRNA from 130 the capsid in virus culture supernatant, hereafter referred to as virus production medium (VPM). 131 However, vRNA release from VPM by heating for 5min at 95°C was found to be limited. The difference 132 in Ct values between vRNA extracted using a column RNA purification kit and heat lysis of an equivalent 133 corresponding volume of VPM was found to be >10 cycles, corresponding to a roughly 1,000x loss in 134 sensitivity (data not shown). Combined with the requirement for a heat block or, ideally, a PCR machine 135 to ensure correct core temperature during heat inactivation and concerns over RNA stability, heat 136 inactivation was abandoned as a broadly applicable method and focus shifted to a detergent lysis-based 137 method. 138 Initial optimization experiments were performed using VPM following heat inactivation, at 70°C for 139 10min using a PCR machine to ensure good heat transfer and correct core temperature, to allow 140 processing outside CL3. This virus inactivation method had previously been confirmed by serial dilution 141 and inoculation of cells with heat-inactivated virus (>6log10 TCID50/ml reduction). At later stages, once 142 virus inactivation by lysis buffer was confirmed, we validated that there was no difference in Ct values, 143 sensitivity, or efficiency between vRNA extracted from VPM with and without 70°C heat-inactivation 144 (data not shown). 145 All tested lysis buffers were based on a 150mM NaCl, 10mM Tris-HCl, pH 7.5 solution supplemented 146 with different lysis detergents for VPM to be lysed at a 1:1 ratio. The salt concentration in the lysis buffer 147 was based on (Shatzkes et al., 2014), who found a 150mM NaCl concentration to be most sensitive in 148 the RT-qPCR reaction. VPM, which in most cases is DMEM or RPMI, contains 108-118mM Cland 138-149 155mM Na + , thus in a 1:1 dilution with lysis buffer, NaCl concentration will remain just slightly under 150 150mM. 151 Three different detergents were tested to assess their efficacy in releasing vRNA from VPM for RT-152 qPCR: 1 or 10% Triton X-100, 0.25% IGEPAL-630, or 5% Tween-20 in a buffer supplemented with 153 10U/ml RNasin Plus RNA inhibitor. These initial concentrations were chosen either based on available 154 SARS-CoV-2 inactivation data or previous reports in lysis protocols. VPM lysis was performed for 20min 155 at room temperature. It was also assessed whether proteinase K treatment could improve vRNA release. 156 Therefore, proteinase K was added at 0.1AU/ml to each detergent lysis buffer, supplemented with 157 0.83mM final concentration EDTA to prevent heat damage to RNA during heat inactivation, mixed 1:1 158 with VPM or vRNA in nuclease-free water (NF-H2O), and incubated for 30min at 56°C prior to heat 159 inactivation for 10min at 95°C. Proteinase K is not inactivated by EDTA and has been shown to be active 160 in high detergent buffers. 1% Triton X-100, 5% Tween-20, and 0.25% IGEPAL-630 detergent lysis buffers showed similar 162 sensitivities (Cts 14.75-15.01) compared to purified equivalent amounts of vRNA (Ct 13.53). The qPCR 163 reaction was clearly impaired by 10% Triton X-100, and 5% Tween-20 buffers, which consistently 164 produced a lower fluorescence. When the detergent lysis buffers were tested in combination with 165 heating with proteinase K for vRNA or in VPM lysis alone, a marked increase by 1.22-1.37Cts was 166 observed. This increase was even higher for 10% Triton X-100, where proteinase K digest lead to a loss 167 of 5.86Ct. Efficiencies as calculated by LinRegPCR were above 78% using 0.25% IGEPAL-630, 1% 168 Triton X-100, or 5% Tween-20. 10% Triton X-100, however, shows markedly lower efficiencies (59.2-169 62.89%) in VPM. All detergents used are likely to work at high efficiencies if used at lower detergent 170 concentrations upon further optimization. We decided to continue optimization of the IGEPAL-630 buffer 171 based on comparable sensitivity at lower concentration to the other detergents. (Figure 2A (Table 1) . 190 To test the effect of IGEPAL-630 on the RT-qPCR reaction, a small amount of template RNA (189 191 copies/reaction) was tested in an RT-qPCR reaction diluted 1:1 in 2.5% IGEPAL-630 buffer or in H2O, and added at 1µl/reaction. No significant decrease in sensitivity could be observed, however, Figure 2D ). 204 A small amount of RNA template was diluted 1:10 in either NF-H2O, 2.5% IGEPAL-630 lysis buffer, Neither Betaine addition nor increased primer concentrations improve direct lysis RT-220 qPCR sensitivity. 221 To try to improve direct lysis RT-qPCR efficiency in the face of suspected inhibitors introduced during 222 cell and virus culture, we tested the addition of betaine, known to improve PCR amplification by relaxing 223 secondary structures, and its effect on the direct lysis RT-qPCR reaction. 2.5% IGEPAL-630 lysis buffer 224 containing 10U/ml RNasin Plus was tested in combination with increasing amounts of betaine in the RT-225 qPCR reaction. 1M Betaine decreased the efficiency of the RT-qPCR significantly, whilst lower 226 concentrations (0.5M and 0.1M) showed no decrease in efficiency but a slight decrease in signal. 227 Overall, betaine addition showed no improvement in sensitivity. ( Figure 2F ) 228 To assess whether the primer concentrations of CDC N3 used were still appropriate for VPM lysate, 229 and whether further improvements in sensitivity and or efficacy could be made, we tested increasing In order to test the effect of increased amounts of lysate on the direct lysis RT-qPCR, the addition of 242 2 or 3µl of VPM/2.5% IGEPAL-630 lysate to a 10µl reaction, corresponding to 20 or 30% of the reaction 243 volume, was tested. However, this significantly decreased reaction efficiency and sensitivity. (Figure 244 2H) 245 Previous results showed that the VPM lysate contained RT-qPCR inhibitors. We therefore tested 246 whether diluting out the lysate in the reaction could improve accuracy of quantification and sensitivity 247 compared to equivalent amounts of vRNA. To reduce pipetting errors, VPM lysate was diluted in NF-248 H2O prior to addition to the RT-qPCR reaction. Whilst VPM lysate at 10% reaction volume showed a Ct In this study, we show a direct lysis protocol for a one-step RT-qPCR of SARS-CoV-2 in cell culture 257 supernatant, achieving Ct results within one cycle of those obtained using traditional viral RNA 258 purification, whilst inactivating SARS-CoV-2. 259 The best performing primer pair in our protocols, CDC N3, is designed to recognize all currently 260 known clade 2 and 3 viruses of the Sarbecovirus subgenus. It is therefore likely directly applicable to 261 other viruses of this subgenus and, with changes in primers, to other CoVs. Our results once again show 262 the importance of testing for the optimal concentration of PCR primers and checking for primer dimer 263 formation, particularly when using fluorescent dye incorporation RT-qPCR. We found that CDC N1, DZIF 264 RdRp, and to a lower extent DZIF N show primer dimers at relatively low primer concentrations. We 265 found that DZIF N forms multimers of its product as observed on the agarose gel evaluation and 266 confirmed by Sanger sequencing, which could affect the efficiency and results from probe-based RT-267 qPCR diagnostic assays. This is particularly interesting since 0.5% Triton X-100, which is the final concentration in a 1:1 lysis 278 buffer/VPM mix, has been shown to completely inactivate SARS-CoV-2 (PHE, 2020b). Tween-20 will 279 need to be used at higher concentrations since live virus can still be recovered from 30min treatment 280 with 0.5% final concentration Tween-20 (PHE, 2020a). 281 Others found a proteinase K digest prior to heat lysis to be beneficial to detecting SARS-CoV-2 in 282 patient samples (Genoud et al., 2021; Nique et al., 2021) . In our hands, heat lysis of SARS-CoV-2 VPM 283 gave significantly worse results than using a lysis buffer, therefore that avenue was not pursued further 284 for in vitro samples. The addition of a proteinase K digest worsened the sensitivity of the RT-qPCR for 285 0.25% IGEPAL-630, 1% Triton X-100, or 5% Tween-20/VPM and /vRNA. This may be due to the 286 prolonged incubation time, the presence of EDTA, added to protect RNA stability during heat inactivation 287 of the proteinase K, impairing the RT-qPCR reaction, or heat degradation of the RNA. 288 One of the limiting factors in direct lysis RT-qPCR is that the volume of lysate, and thereby vRNA 289 copies, added to the reaction is limited. However, the direct lysis protocol was found to be sensitive 290 down to <0.0043 TCID50/reaction and showing a 1E6 dynamic range, which should be more than 291 sufficient for in vitro experiments. 292 We found that 2.5% IGEPAL-630 lysis buffer concentration, 1.25% final concentration, slightly 293 decreases the reaction efficiency but does not affect the sensitivity of the RT-qPCR. However, reaction 294 sensitivity was strongly affected by fresh cell culture media. The addition of as little as 20U/ml RNasin 295 rescued the decrease. This is in agreement with the findings of Pearson et al, who also found that adding 296 RNaseOUT RNase inhibitor significantly increased the sensitivity of their direct RT-qPCR of SARS-CoV-297 2 patient swab samples, achieving Cts only 3 cycles higher than using RNA purification and RT-qPCR 298 (Pearson et al., 2020) . This suggests that a broad range of RNase inhibitors can be used in direct lysis 299 assays. 300 With very low amounts of template RNA, 18.9/reaction, it was observed that 2.5% IGEPAL/NF-H2O 301 mix showed a decrease of around 3.69Cts without RNasin, indicating a small contamination with 302 RNases in the lysis buffer and/or small amounts in NF-H2O since the addition of RNasin rescued the 303 reduction in an 2.5% IGEPAL/fresh cell culture media mix even beyond the NF-H2O control. It shows 304 that despite careful handling, working in PCR cabinets, and decontamination of surfaces, RNase 305 contaminations can occur easily and may not show at higher concentration RNA. It is therefore further 306 recommended to add RNases to the lysis buffer. 307 Interestingly, the addition of RNase inhibitors had no impact on the 2.5% IGEPAL/VPM lysis. This is 308 either due to the much higher amount of RNA present, an estimated roughly 10,000 more, so that the 309 impact of small amounts of RNases may not be visible. Similarly, viral RNA could still be loosely 310 associated with nucleocapsid proteins, protecting its degradation by RNases. Whilst SARS-CoV-2 overall only contains around 38% of G and C nucleotides combined, the N region 312 is one of the most GC-rich areas of the virus at 47% with some stretches reaching over 60% However, the lack of effect of the addition of betaine suggests that there doesn't seem to be any PCR with higher GC contents. We did not pursue this avenue further, since dilution of lysates showed that 319 RT-qPCR inhibitors rather than a lack of RNA accessibility or virion lysis were the cause of decreased 320 sensitivity. 321 Looking at our combined results thus far, we found that IGEPAL-630 on its own was not reducing 322 RT-qPCR sensitivity. Fresh cell culture media decreased sensitivity but this was restored by the addition 323 of RNase inhibitors. However, 2.5% IGEPAL-630/VPM lysate was still showing sensitivities 1.52 (±0.78 324 SEM) Cts lower than extracted vRNA when added at 10% reaction volume. Testing lower percentage 325 reaction volumes shows that the sensitivity difference can be reduced to less than 1 Ct when lysate is 326 diluted out. This indicates that it is not a lack of complete virus lysis but the presence of RT-qPCR 327 inhibitors in the virus infection media that is limiting the sensitivity. 328 In conclusion, our method allows inactivation and direct RT-qPCR of SARS-CoV-2 in cell culture Table S1 IDT Invitrogen N/A All the RNA fragments produced by in vitro transcription are described in Table S1 N/A N/A LinRegPCR. RT-pPCR products were analyzed on a 2% Agarose (Invitrogen) gels using SYBR Safe DNA gel 394 stain (Invitrogen) according to the manufacturer's instructions. Larger DZIF N product fragments were 395 excised and purified using the QIAquick Gel Extraction kit (Qiagen) and analyzed by Sanger sequencing 396 using the DZIF N forward and reverse primers, respectively. To obtain absolute reaction efficiencies, 10-fold serial dilutions of template RNA were added to the RT-398 qPCR reaction (as described above). A semilog fit was performed using Graphpad Prism to determine 399 the slope to determine the RT-qPCR efficiency. LinRegPCR. n=2; curves represent the average. D) PCR products obtained were analyzed on an 588 agarose gel to assess primer-dimer formation in high and low template (indicated above gel pictures) 589 RT-qPCR reactions. E) Serial dilutions of template were run in RT-qPCR reactions to assess efficiency 590 of the PCR reactions for the best performing primer pairs. A semilog fit curve was calculated using 591 GraphPad prism to assess reaction efficiency from the slope. Samples excluded from the efficiency 592 calculation are greyed out, as they were beyond reaction sensitivity limits. n=3*2, error bars represent 593 min and max. Fast SARS-CoV-498 2 detection by RT-qPCR in preheated nasopharyngeal swab samples Efficiencies for each 601 amplification were calculated from amplification curves using LinRegPCR. For serial dilutions, a semilog 602 fit curve was calculated using GraphPad Prism to assess reaction efficiency from the slope 10% and 1% Triton X-100, and 5% Tween 20 (Tw20) 604 were assessed for their ability to release vRNA from cell supernatant and use in a VPM:lysis buffer, vRNA and VPM were lysed 1:1 and 606 incubated for 30min at 56°C in lysis buffers containing 0.1AU/ml proteinase K (PK) and 0.83mM EDTA 607 before heat inactivation for 10min at 95°C. n=2. B) Increasing concentrations of IGEPAL-630 were 608 Template RNA in NF-H2O was lysed in buffer containing 2.5% IGEPAL-630. n=2 D) A serial dilution of 610 template RNA in NF-H2O or fresh cell culture media was IGEPAL-630. n=2, error bars represent min and max. E) 18.9 RNA template copies were incubated in 612 lysis buffer containing 2.5% IGEPAL-630 and increasing amounts of RNasin Plus VPM was incubated using a low or a high amount of RNasin 614 in the lysis buffer, and compared to vRNA extracted from an equivalent amount of VPM. n=2 F) An RT-615 qPCR reaction was set up containing 10% either VPM lysate using 2.5% IGEPAL-630 or vRNA in NF Increasing symmetric 617 concentrations of CDC N3 primer were added to the VPM/lysate RT-qPCR reaction. n=2. Serial dilutions 618 of VPM in lysis buffer or template RNA in mock lysis buffer were run in an RT-qPCR assay and analyzed 619 as described above. n=3*2. H) Decreasing and increasing amounts of VPM RT-qPCR reaction, corresponding to 0.5-30% of the reaction volume (as indicated for each curve) Corresponding amounts of vRNA in NF-H2O were run in parallel. Curves represent an average of 3*2, 622 except for 20-and 30%, representing an n=2. Ct values n=3*2 show mean +/-SEM and were analyzed 623 using a 2-way ANOVA