Development of an efficient one-step real-time reverse transcription polymerase chain reaction method for severe acute respiratory syndrome-coronavirus-2 detection

The general methods to detect the RNA of severe acute respiratory syndrome-coronavirus 2 (SARS-CoV-2) in clinical diagnostic testing involve reverse transcriptases and thermostable DNA polymerases. In this study, we compared the detection of SARS-CoV-2 by a one-step real-time RT-PCR method using a heat-resistant reverse transcriptase variant MM4 from Moloney murine leukemia virus, two thermostable DNA polymerase variants with reverse transcriptase activity from Thermotoga petrophila K4 and Thermococcus kodakarensis KOD1, or a wild-type DNA polymerase from Thermus thermophilus M1. The highest performance was achieved by combining MM4 with the thermostable DNA polymerase from T. thermophilus M1. These enzymes efficiently amplified specific RNA using uracil-DNA glycosylase (UNG) to remove contamination and human RNase P RNA amplification as an internal control. The standard curve was obtained from 5 to 105 copies of synthetic RNA. The one-step real-time RT-PCR method’s sensitivity and specificity were 99.44% and 100%, respectively (n = 213), compared to those of a commercially available diagnostic kit. Therefore, our method will be useful for the accurate detection and quantification of SARS-CoV-2.


Introduction
The COVID-19 pandemic, caused by the RNA virus SARS-CoV-2, has wreaked havoc on the global economy and many national healthcare systems. Countries with a large number of daily new cases need a reliable and inexpensive diagnostic system. In other countries, there is also an urgent need to stockpile kits for emergency use. Currently, nucleic acid detection and antigen detection are the two primary methods to confirm infection. Real-time quantitative PCR (RT-qPCR) is a highly sensitive and widely used method for detecting SARS-CoV-2 RNA. Fully automated platforms for RT-qPCR have been introduced in large medical institutions and clinical laboratories. However, there is still a global shortage of reagents and consumables. In addition, RT-qPCR involves complicated molecular biological techniques. Moreover, RT-qPCR produces false-positive [1] or false-negative results [2].
The conventional RT-PCR consists of two steps, the synthesis of cDNA using reverse transcriptase and the amplification of DNA using DNA polymerase. In contrast, one-step realtime RT-PCR is a continuous reaction that performs the two procedures in the same tube. Therefore, it is essential to adjust the reaction buffer conditions so that two or more enzymes remain functional in the same buffer.
In this study, we applied the widely used primers and probes, reported by the United States Centers for Disease Control and Prevention (CDC), in RT-PCR to detect the SARS-CoV-2 RNA. Previously, we reported a thermostable reverse transcriptase, MM4, harboring 4 amino acid substitutions in the original Molony murine leukemia virus reverse transcriptase [3]. In addition, we selected 2 available family A DNA polymerase, DNA polymerase from Thermus thermophilus M1 strain (M1pol Tth ) with a 5 0 -3 0 but not a 3 0 -5 0 exonuclease domain [4], and the genetically engineered L329A DNA polymerase variant (K4pol L329A ) originated from the Thermotoga petrophila K4 strain. K4pol L329A has acquired reverse transcriptase activity via mutagenesis; it has a 3 0 -5 0 but not a 5 0 -3 0 exonuclease domain [5]. RTX is xenopolymerase harboring reverse transcriptase and DNA polymerase activities due to the introduction of 17 amino acid substitutions [6] in a Family B KOD DNA polymerase from Thermococcus kodakarensis. Lastly, we used family I uracil-DNA glycosylase (UNG), cleaving the N-glycosylic bond between uracil and sugar to remove the uracil incorporated in DNA, to remove contaminations from the PCR products [7].
Currently, the clinical detection range of the RNA copy number for SARS-CoV-2 varies. The one-step real-time RT-PCR that we developed was evaluated for its detection range compared to that of a commercial SARS-CoV-2 detection kit authorized by the Ministry of Health, Labour and Welfare, Japan, as in vitro diagnostics (https://www.mhlw.go.jp/stf/newpage_ 11332.html). We aimed to detect 10-10 5 copies of RNA per test within 45 PCR cycles.
Eventually, we successfully amplified 5-10 5 copies of synthetic SARS-CoV-2 RNA and confirmed the specific amplification of the viral RNA from clinically isolated RNA samples.

Statistical analysis
The data were analyzed by one-way ANOVA. The difference among the mean Ct values was analyzed using the Tukey-Kramer test. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and bivariate relationship were analyzed using the JMP 10 software (SAS) or Igor Pro (WaveMetrics).
One-Step real-time RT-PCR was performed by mixing MM4 with the respective DNA polymerases, M1pol Tth , K4pol L329A , and RTX to verify 10 4 and 10 3 copies of SARS-CoV-2 RNA.
We found that M1pol Tth enabled the detection of specific RNA in all the buffers, while K4pol L329A could amplify the target sequence only in buffers 1, 3, and 5. The family B DNA polymerase, RTX, lost its amplification ability with a fluorescent probe. On the other hand, M1pol Tth combined with MM4 achieved amplification in various buffers (Fig 1). While we analyzed clinical specimens, we found that M1polTth, less affected by buffer compositions, was likely the most suitable DNA polymerase for one-step real-time RT-PCR. There was variation in the Ct values of the amplification of 10 3 or 10 4 copies of RNA (Table 2) by different enzymes in multiple buffers. When M1pol Tth DNA polymerase was used, there was no significant difference in the sensitivity (Ct value) in detecting 10 4 or 10 3 copies of RNA per test.
Next, the detection limits of the method in each buffer with MM4 and M1pol Tth were compared. In buffers 1, 3, and 5, the one-step real-time RT-PCR system could detect more than 100 copies of RNA. On the other hand, the method could detect more than 10 copies of RNA in buffer 2 and more than 5 copies in buffer 4 (Fig 2 and Table 3). Each Ct value obtained in buffer 4 was smaller than those in the other buffers. Therefore, buffer 4, combined with M1pol Tth , exhibited the best performance in the tested conditions. We further optimized the reaction time of the reverse transcriptases. MM4's reaction time was set from 30 sec to 15 min at 50˚C to test the detection of 10, 10 2 , and 10 4 copies of synthetic RNA. The Ct value did not vary significantly when the initial standard RNA copies were low (10 and 100 copies/test). On the other hand, Ct values at 30 sec were significantly higher than at 5, 10, and 15 min with 10 4 RNA copies. Therefore, the optimal reverse transcription reaction time was determined to be 5 min (Table 4).
PCR contamination due to amplified DNA carryover is a problem in clinical testing. Therefore, we examined the applicability of UNG to our one-step real-time RT-PCR method. We found that the contamination of PCR-amplified DNA could be avoided by adding UNG at 0.4 U/test. On the other hand, the RNase P gene has been used as an internal standard in

PLOS ONE
SARS-CoV-2 detection using RT-PCR method   Mean Ct values ± S.D; n = 3. � p < 0.05 (Tukey-Kramer HSD).   coronavirus detection systems. Thus, we investigated whether the CDC primer and probe sets could be used for the simultaneous detection of RNase P. We found that our method could be used as an internal standard. The final, optimized conditions and cycling parameters for the one-step real-time RT-PCR were determined (Tables 5 and 6). Our one-step real-time RT-PCR kit was designated as the "Mother's and Children's, Osaka" (MoCO) kit.

Consistency with the commercial kit with clinical RNA samples
One-step real-time RT-PCR was performed with the MoCO kit and a commercial kit (SARS--CoV-2 direct PCR detection kit, Takara Bio Inc., Kusatsu, Japan) on a total of 213 RNA samples (S1 Table). Clinical RNA specimens were stored at -80˚C before use. The MoCO kit detected using the N1 and N2 of the CDC probes separately, while both probes are mixed in the Takara kit. The MoCO kit's accuracy was checked by setting the detection limits at 10, 5, and 1 copies; the sample was considered positive if amplification with the N1 or N2 probe exceeded the detection limit. With the detection limit of 10 copies, the MoCO kit's sensitivity, specificity, PPV, and NPV were 98.1, 96.5, 98.7, and 94.8%, respectively. With the detection limit of 5 copies, its sensitivity specificity, PPV, and NPV were 98.2, 97.9, 99.4, and 94.0%, respectively. With the detection limit of 1 copy, the kit's sensitivity, specificity, PPV, and NPV were 99.4, 100, 100, and 97.2%, respectively ( Table 7). The log-log plots were well fitted in a wide range (Fig 3). These results indicate that the MoCO kit is clinically applicable for detecting SARS-CoV-2 RNA. Our developed MoCO kit could reproducibly detect at least 5 copies of synthetic RNA (Fig  2), demonstrating comparable performance to that of the commercial kit (Fig 3 and Table 7). One of the critical problems in the clinical amplification of high-sensitive nucleic acids is cross-contamination. The MoCO kit can be used in combination with UNG to minimize DNA carryover. In addition, by using a thermostable reverse transcriptase, MM4, the kit could achieve cDNA synthesis efficiently at 50˚C for 5, 10, and 15 min (Table 4). Our protocol does not require pre-heating at 65˚C during cDNA synthesis. Also, our method could simultaneously detect RNase P as an internal standard; the one-tenth of the amount of primers and probes for SARS-CoV-2 was sufficient for RNase P. Our highly sensitive RT-PCR method is expected to be applied to identifying other pathogens and analyzing mutations.

Conclusions
The MoCO kit was developed to detect SARS-CoV-2 from the extracted RNA samples. Our development kit performed well for emergency research tests with a detection limit of 5 copies of RNA. It is necessary to continue improving the kit by simplifying sample preparation and the diagnosis of viral RNA mutations. Our high-sensitive RT-PCR method is expected to be applied for the molecular detection of various infectious diseases.
Supporting information S1 Table. Comparison of the SARS-CoV-2 copy numbers detected using the MoCO kit and a commercial kit (SARS-CoV-2 direct PCR detection kit, Takara). (PDF)