A simple RNA preparation method for SARS-CoV-2 detection by RT-qPCR

The technique RT-qPCR for viral RNA detection is the current worldwide strategy used for early detection of the novel coronavirus SARS-CoV-2. RNA extraction is a key pre-analytical step in RT-qPCR, often achieved using commercial kits. However, the magnitude of the COVID-19 pandemic is causing disruptions to the global supply chains used by many diagnostic laboratories to procure the commercial kits required for RNA extraction. Shortage in these essential reagents is even more acute in developing countries with no means to produce kits locally. We sought to find an alternative procedure to replace commercial kits using common reagents found in molecular biology laboratories. Here we report a method for RNA extraction that takes about 40 min to complete ten samples, and is not more laborious than current commercial RNA extraction kits. We demonstrate that this method can be used to process nasopharyngeal swab samples and yields RT-qPCR results comparable to those obtained with commercial kits. Most importantly, this procedure can be easily implemented in any molecular diagnostic laboratory. Frequent testing is crucial for individual patient management as well as for public health decision making in this pandemic. Implementation of this method could maintain crucial testing going despite commercial kit shortages.

We reviewed the published literature to search for procedures of RNA extraction that could potentially be used to replace commercial kits. Many different protocols and variations have been published over the years that optimize or simplify the RNA extraction process from various types of samples. We tested five types of procedures to identify an efficient procedure for extracting RNA from clinical samples that is compatible with downstream RT-qPCR analysis. Of the procedures evaluated, a simple method based in acid pH separation of RNA was found the most suitable. It can be carried out in approximately 40 min for ten samples, and is not more laborious than current methods using commercial kits. This procedure requires reagents and equipment that can be found in any standard molecular biology laboratory, thus avoiding supply chain issues. The resulting RNA can be used to detect SARS-CoV2 by standard RT-qPCR testing protocols with robust results comparable to those obtained using commercial RNA-extraction kits.

Results
Screening of alternative procedures for RNA extraction. For validation of the RNA extraction procedures, the RNase P target was amplified in a one-step RT-qPCR reaction, as quality control for the extraction method. As shown in Fig. 1, three of the five procedures evaluated yielded enough RNA to amplify the target gene, whereas two of them did not. The TRIzol approach was most effective, exhibiting the highest yield when amplifying the human RNase P target (Fig. 1). The BSA-based protocol also allowed for amplification of the RNase P target, albeit with a lower yield and significant variability among replicates (Fig. 1). Acid pH-based method also allowed amplification of the RNase P target, though with lower yields when compared to the TRIzol method (Fig. 1). The direct method and high-temperature method did not yield enough RNA to amplify the RNase P gene under our experimental conditions. While TRIzol appears to be the best experimental procedure in terms of yield, it is not easy to use for a diagnostics laboratory setting as it requires a chemical hood for the organic extraction step. Biosafety cabinets class II (BSL-2) necessary for operator protection are not appropriate for working with organic solvents. BSA, TRIzol, and acid pH procedures provided comparable yields, but the acid pH method was more consistent among replicates. Based on these considerations, we decided to validate the acid pH method to extract RNA from clinical samples, using High Pure Viral RNA Kit (Roche) as the gold standard.
Validation of the acid pH RNA extraction method in clinical samples. To validate the acid pH method of RNA extraction, RT-qPCR using TaqMan probes and primers recommended by the CDC were used 21 . The nucleocapsid viral proteins N1 and N2 were amplified as viral targets, and RNase P was also amplified as a control. We analyzed 50 clinical samples: 22 were positive, 11 were undetermined, and 17 were negative according to RT-qPCR recommended by CDC 21 , using RNA extracted with High Pure Viral RNA Kit (Roche). Undetermined samples are described as having a viral load around the limit of detection (LOD) of the RT-qPCR method that was reported as 10 0.5 RNA copies/µl 21 . This means that the RT-qPCR method can detect 16 RNA copies per PCR reaction. The PCR test used detects 2 targets of the virus: N1 and N2. The mean Cq value for N1 target reported for sets of dilutions that are ≥ 95% positive is around 36 21 . Therefore we analyzed the efficiency of both extraction methods in two different groups of samples: those with Cq N1 ≤ 36 and those with Cq N1 > 36. The results for the 50 samples are shown in Table 1. For samples with Cq N1 ≤ 36 there were no differences in Cq   Fig. 2A). In contrast, for samples with Cq N1 > 36, Cq values for N1 and N2 were higher for High Pure Viral RNA Kit (Roche) than those obtained with acid pH method (p = 0.026 and p = 0.022 respectively) (Fig. 2B). For samples with Cq N1 ≤ 36, Cq values for RNase P were slightly higher for acid pH method (p = 0.021), whereas for samples with Cq N1 > 36 there was no significant difference between both methods. The percentage of agreement between both methods was calculated considering samples whose report changed from positive, undetermined or negative. In total, 8 samples changed their report. The 17 negative samples were also negative using RNA extracted with the acid pH method. Out of 22 positive samples, 21 were also positive using RNA extracted with the acid pH method, whereas one sample was undetermined. Out of 11 undetermined samples analyzed, 4 were still undetermined using RNA extracted with the acid pH method. However, 3 of them were negative and 4 of them were positive. Agreement for negative samples was 100%. The percentage of agreement for samples with Cq N1 ≤ 36 was 89.5%. As expected, the percentage of agreement for samples with Cq N1 > 36 was only 57%. Importantly, the processing time and laboriousness of the acid-pH method is similar or less than that of High Pure Viral RNA Kit (Roche) method. A detailed scheme of the method is shown in Fig. 3.

Discussion
Here we tested several kit-free RNA extraction methods compatible with RT-qPCR analysis and selected one simple procedure based on RNA extraction using acid pH. We validated this method using 50 clinical samples with results comparable to those obtained with commercial kits. There are three key aspects of this method that must be pointed out. First, the acid pH-based methods that we reviewed 12,14,22 are intended for RNA extraction from tissue, cultured cells, and cell-associated virus. Therefore, the first step of these protocols is centrifugation with subsequent lysis of the cell pellet. However, we need to recover free viral particles in solution, which do not sediment after routine centrifugation at 15,000 g. For this reason we used the uncentrifuged sample directly mixed with lysis buffer, with subsequent precipitation of viral RNA in the whole mix volume. Using uncentrifuged sample is the key step for efficient RNA recovery because when centrifuged sample was used in preliminar www.nature.com/scientificreports/ tests, Cq values were much higher than those obtained with High Pure Viral RNA Kit (Roche). Second, the acid pH method uses the anionic detergent Sodium dodecyl sulfate (SDS) that can lyse cells and viral coats through disruption of noncovalent bonds in proteins causing them to lose their native conformation 12 . Third, low pH and high concentration of salt make possible the selective recovery of RNA. Within the pH range of 5.5 to 6.0, RNA degradation is minimized 22 . RNA phosphodiester bond is more stable at acidic than alkaline pH, where it is susceptible to alkaline hydrolysis at pH greater than 6 23 . Acid hydrolysis can only occur at pH lower than 2 12,24 . Moreover, DNA and RNA have different solubility at different pH, mainly due to the 2′ hydroxyl group of RNA, which increases the polarity of this nucleic acid 25,26 . Therefore, it is essential to adjust the Lysis Buffer to pH 5, as described in materials and methods It is worth mentioning that all of the samples that changed their report had Cq values that were around the cutoff value of 40. These changes occurred in both directions, meaning that some Cqs increased and some Cqs decreased. It would have been very clarifying to perform triplicated RNA extractions, in particular for undetermined samples, whose viral load is around the detection limit. Because of the above exposed information we consider the acid pH method robust and reliable. In fact, it is currently being used in our diagnostic laboratory since the 3 rd week of April 2020 for routine detection of SARS-CoV2 in clinical samples.
The RNA extraction procedure with acid pH described here has many advantages over commercial kits to test for SARS-CoV-2 in the context of the current pandemic. This experimental procedure utilizes low cost reagents and equipment that can be found in standard molecular biology laboratories. The cost of extraction is a critical issue in most clinical laboratories, and the cost of our in-house method is around ten times lower than extraction kits. Moreover, DNase treatment is not necessary because SARS-CoV-2 detection is not altered in the presence of DNA. In fact, residual DNA may serve as the template for RNase P gene amplification. Because of current environmental concerns, we would also like to highlight the lower plastic contamination generated by this in-house method. Column-based extraction kits use several disposable tubes per sample, columns, bottles of buffer solutions, and plastic bags. Our in-house extraction method is by far, much more environmental friendly; it requires only two Eppendorf tubes per sample. Finally, our in-house method is comparable in hands-on time to commercial kits: it can be carried out in approximately 40 min for a set of 10 samples. However, it is important to mention that additional care must be taken in handling to avoid cross-contamination between samples.
In conclusion, the RNA extraction procedure with acid pH described here is an excellent alternative to commercial systems to test for SARS-CoV2. Our results support a new method for RNA extraction from swab samples www.nature.com/scientificreports/ that can be used to detect SARS-CoV2 by standard RT-qPCR testing protocols. This procedure can be a helpful alternative for laboratories facing supply-chain disruption and commercial kit shortages.

Materials and methods
Biological samples. Two types of biological samples were used. For preliminary evaluation of the RNA extraction methods we used saliva samples obtained from two asymptomatic volunteers. Saliva is routinely collected for the initial assessment of viral infection. Two saliva samples were obtained from each volunteer and at least three independent RNA extractions were performed from each sample, obtaining a minimum of six RNA preparations to test each experimental procedure. For validation of the RNA extraction method selected, we used nasopharyngeal swabs in Universal Transport Medium (UTM). Swabs were obtained from 50 patients that attended the outpatient service of Red Salud UC-CHRISTUS (Santiago, Chile) because of suspected coronavirus infection. Only one sample was obtained per patient: one portion of the sample was extracted using the High Pure Viral RNA Kit (Roche), and another portion of the same sample was extracted using the acid pH method. Informed consent was obtained from all participants and/or their legal guardians. Samples were processed in the Laboratory of Diagnostic Microbiology of the same institution. All methods were performed in accordance with the relevant guidelines and regulations. All procedures were approved by the Ethics Committee of the Pontificia Universidad Católica de Chile.
RNA extraction methods evaluated. The following experimental procedures were tested in this study.
Saliva samples were centrifuged before taking an aliquot of supernatant for processing as described below.
(1) TRIzol. The standard TRIzol-based method was evaluated 9,11,18 . First, 800 µL of TRIzol were added to 200 µL of sample and vortexed briefly. Then, 200 μL of chloroform were added, vortexed, and centrifuged at 12,000 g for 10 min at room temperature. The aqueous phase (600 μL) was recovered in a clean tube containing 600 μL of isopropanol. The tube was mixed by inversion and incubated at room temperature for 10 min.
The tube was then centrifuged at 12,000 g for 10 min at 4 °C, and the supernatant was discarded. The pellet was washed with 500 μL of 70% ethanol, centrifuged at 7,500 g for 5 min at 4 °C and the supernatant was discarded. The pellet was dried at room temperature for 10 min and resuspended in 25 μL of RNase-free water by incubating at 37 °C for 10 min. (3) Acid pH-based method. Under acidic pH, RNA can be separated from DNA and other molecules due to the differential polarity given by its hydroxyl groups, which maintains it in solution 12,22,25,26,29 . Based on the methods described by Heath (1999) 22 , Sambrook and Russell (2001) 12 , and Chomczynski and Sacchi (2006) 14 , 300 µL of pH 5 Lysis Buffer (69,4 mM SDS, 68 mM sodium citrate dihydrate, 132 mM anhydrous citric acid and 10 mM EDTA, then adjust the buffer to pH 5) were added to 200 µL of uncentrifuged sample and mixed by pipetting three times. Then, 150 µL of Precipitation Buffer (17 mM sodium citrate dihydrate, 33,3 mM anhydrous citric acid, and 4 M NaCl) were added and mixed by inversion 10 times. Samples were incubated on ice for 5 min and centrifuged at 15,000 g for 3 min at room temperature. 600 μL of the supernatant were transferred to a clean tube containing 600 μL of isopropanol and incubated for 10 min at room temperature. A new centrifugation step was made at 15,000 g for 5 min at room temperature. The supernatant was removed, and the pellet washed with 300 μL of cold 70% ethanol and centrifuged at 15,000 g for 3 min at room temperature. Supernatant was discarded and tubes were inverted in paper towel. The pellet was dried, leaving the tubes open for 10 min. Finally, the pellet was resuspended in 50 μL of nuclease-free water pre-warmed at 70 °C. NOTE: If the buffer is stored for later use, it precipitates at 4 °C, so it needs to be heated for 5 min at 60 °C for its use.
(4) High temperature-based method. Based on the method described by Fomsgaard and Rosenstierne (2020) 30 , 50 µL of the sample were directly heated at 98 °C for 5 min and cooled at 4 °C. Then 19 µL of the sample were mixed with 1 µL of BSA (20 mg/mL) and kept on ice for immediate use or at -80 °C for later use. (5) Direct use of the samples. An aliquot taken from the original sample was directly used to perform RT-qPCR analysis 31 .
The 50 nasopharyngeal swabs used for the validation of the RNA extraction method selected, were extracted using High Pure Viral RNA Kit (Roche) according to instructions provided by the manufacturer. This RNA extraction method was considered as the gold standard for comparison purposes, and it is based in capture of RNA using columns with silica filters.
RT-qPCR analysis. For preliminary evaluation of RNA extraction procedures, we used RT-qPCR against the human RNAse P gene with primers and a Taqman probe previously described 32 . RP1-F: AGA TTT GGA CCT GCG AGC G, RP1-R: GAG CGG CTG TCT CCA CAA GT, and RP1-probe: TTC TGA CCT GAA GGC TCT GCGCG. The RNase P gene is used as an internal control because many copies of it exist in the human genome, and it is readily detectable. The source of RNase P comes from the human cells that are present in every sample used. It is assumed that if human nucleic acids were extracted to detect the human gene RNase P, viral nucleic acids were Scientific RepoRtS | (2020) 10:16608 | https://doi.org/10.1038/s41598-020-73616-w www.nature.com/scientificreports/ also successfully extracted. The RNase P target is also amplified as a quality control for the extraction method and to corroborate the absence of PCR-inhibitors in the sample. For RT-qPCR 5 μL of RNA from saliva samples, 2 μL of RNase-Free water, 1 μL of each RNase P primer, 1 μL of TaqMan RNase P probe and 10 μL of 2X TaqMan Fast Universal PCR Master Mix (Applied Biosystems), were used in a final reaction volume of 20 μL performed in a StepOnePlus Real-Time PCR System (Applied Biosystems).
For validation of the selected RNA extraction procedure, RT-qPCR using Taqman probes and primers recommended by the CDC was used 21 . Two viral targets were amplified: the nucleocapsid viral proteins N1 and N2. Primers and probe for N1 were N1-F: GAC CCC AAA ATC AGC GAA AT, N1-R: TCT GGT TAC TGC CAG TTG AAT CTG , and N1-probe: FAM-ACC CCG CAT TAC GTT TGG TGG ACC -BHQ1. Primers and probe for N2 were N2-F: TTA CAA ACA TTG GCC GCA AA, N2-R: GCG CGA CAT TCC GAA GAA , and N2-probe: FAM-ACA ATT TGC CCC CAG CGC TTCAG-BHQ1. Primers and probe for RNase P were RP2-F: AGA TTT GGA CCT GCG AGC G, RP2-R: GAG CGG CTG TCT CCA CAA GT, and RP2-probe: FAM-TTC TGA CCT GAA GGC TCT GCGCG-BHQ1. A one-step RT-qPCR reaction was performed in a StepOnePlus Real-Time PCR System (Applied Biosystems). Cutoff points for Cq values (Cycle of quantification, or Cycle Threshold) required to decide whether a result is COVID-19 positive or negative were those specified by CDC as follows. To report a positive result, both viral targets N1 and N2 must be Cq < 40. To report a negative result both viral targets must be Cq ≥ 40. If one of the viral targets is Cq < 40 and the other is Cq ≥ 40, the result must be reported as undetermined. The RNase P target must be Cq ≤ 35.

Statistical analysis.
Mean Cq values obtained through both methods for each target gene were analyzed in pairwise comparisons using a paired Student's t-test. The analysis was performed using GraphPad Prism 8 software.